Digital Humanities in Biblical, Early Jewish and Early Christian Studies [1 ed.] 9789004264434, 9789004264328

Citation preview

Digital Humanities in Biblical, Early Jewish and Early Christian Studies

Scholarly Communication Series Editors

Adriaan van der Weel, Leiden University, Netherlands Ernst Thoutenhoofd, University of Groningen, Netherlands Ray Siemens, University of Victoria, Canada Editorial Board

Marco Beretta, University of Bologna, Italy Amy Friedlander, Washington, DC USA Steve Fuller, University of Warwick, UK Chuck Henry, Council on Library and Information Resources, USA Willard McCarty, King’s College London, UK / University of Western Sydney, Australia Mariya Mitova, Leiden, The Netherlands Patrik Svensson, Umeå University, Sweden Melissa Terras, University College London, UΚ John Willinsky, Stanford University, USA Paul Wouters, Leiden University, The Netherlands

VOLUME 2

The titles published in this series are listed at brill.com/sc

Digital Humanities in Biblical, Early Jewish and Early Christian Studies Edited by

Claire Clivaz Andrew Gregory David Hamidović In collaboration with

Sara Schulthess

Leiden • boston 2014

Cover illustration: Matt Katzenberger, http://katzmatt.com

This publication has been typeset in the multilingual “Brill” typeface. With over 5,100 characters covering Latin, IPA, Greek, and Cyrillic, this typeface is especially suitable for use in the humanities. For more information, please see www.brill.com/brill-typeface. ISSN 1879-9027 ISBN 978-90-04-26432-8 (hardback) ISBN 978-90-04-26443-4 (e-book) Copyright 2014 by Koninklijke Brill NV, Leiden, The Netherlands. Koninklijke Brill NV incorporates the imprints Brill, Global Oriental, Hotei Publishing, IDC Publishers and Martinus Nijhoff Publishers. All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher. Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill NV provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA. Fees are subject to change. This book is printed on acid-free paper.

contents List of Contributors ......................................................................................... vii List of Abstracts ................................................................................................ ix Preface ................................................................................................................. xvii Claire Clivaz, Andrew Gregory and David Hamidović Introduction: Digital Humanities in Biblical, Early Jewish and Early Christian Studies .............................................................................. Claire Clivaz

1

part one

digitized manuscripts The Leon Levy Dead Sea Scrolls Digital Library. The Digitization Project of the Dead Sea Scrolls .............................. Pnina Shor Dead Sea Scrolls inside Digital Humanities. A Sample ........................ David Hamidović The Electronic Scriptorium: Markup for New Testament Manuscripts .................................................................................................. H.A.G. Houghton Digital Arabic Gospels Corpus ..................................................................... Elie Dannaoui The Role of the Internet in New Testament Textual Criticism: The Example of the Arabic Manuscripts of the New Testament ............................................................................................ Sara Schulthess The Falasha Memories Project. Digitalization of the Manuscript BNF Ethiopien d’Abbadie 107 .................................................................... Charlotte Touati

11 21

31 61

71

83

vi

contents part two

digital academic research and publishing The Seventy and Their 21st-Century Heirs. The Prospects for Digital Septuagint Research ..................................................................... Juan Garcés

95

Digital Approaches to the Study of Ancient Monotheism .................. 145 Ory Amitay Internet Networks and Academic Research: The Example of New Testament Textual Criticism ......................................................... 155 Claire Clivaz New Ways of Searching with Biblindex, the Online Index of Biblical Quotations in Early Christian Literature ............................. 177 Laurence Mellerin Aspects of Polysemy in Biblical Greek. A Preliminary Study for a New Lexicographical Resource ............................................................... 191 Romina Vergari Publishing Digitally at the University Press? A Reader’s Perspective .............................................................................. 231 Andrew Gregory Does Biblical Studies Deserve to be an Open Source Discipline? .... 261 Russell Hobson Author Index ..................................................................................................... 271 Subject Index ..................................................................................................... 274

list of contributors Ory Amitay is Lecturer in the Department of General History of the University of Haifa (Israel). He is author of From Alexander to Jesus (University of California Press, 2010). Claire Clivaz is assistant Professor in New Testament and Early Christian Literature, at the Faculté de théologie et sciences des religions at the University of Lausanne (Switzerland). Her other publications include notably L’ange et la sueur de sang (Lc 22,43–44) (Peeters, 2010), and, as editor and contributor, the ebook Reading Tomorrow. From Ancient Manuscripts to Digital Era (PPUR, 2012). Elie Dannaoui is Director of the Digital Humanities Centre at the Institute of History, Archaeology and Near Eastern Studies, University of Balamand (Lebanon). Juan Garcés is Academic Coordinator of the Göttingen Centre for Digital Humanities, at the Georg-August-Universität of Göttingen (Germany). He is co-author of ‘Open Source Critical Editions: a Rationale’ (in Text Editing, Print, and the Digital World, ed. by Marilyn Deegan and Kathryn Sutherland, Ashgate Press, 2009). Andrew Gregory is Chaplain and Fellow of University College, Oxford and a member of the Faculty of Theology and Religion in the University of Oxford (United Kingdom). His other publications include The Reception of Luke and Acts in the Period before Irenaeus (Mohr Siebeck, 2003) and, as editor and contributor, New Studies in the Synoptic Problem (Peeters Press, 2011). David Hamidović is ordinary Professor in Jewish Apocryphal Literature and History of Judaism in Antiquity at the Faculté de théologie et sciences des religions at the University of Lausanne (Switzerland). He is author of Les traditions du jubilé à Qumran, Orients sémitiques (Geuthner, 2007) and editor of Aux origines des messianismes juifs, VTS 158 (Brill, 2013).

viii

list of contributors

Russell Hobson is Honorary Associate in the Department of Hebrew, Biblical and Jewish Studies, at the University of Sydney (Australia). He is author of Transforming Literature into Scripture: Texts as Cult Objects at Nineveh and Qumran (Acumen Publishing, 2012). H.A.G. Houghton is a Senior Research Fellow in the Institute for Textual Scholarship and Electronic Editing at the University of Birmingham (United Kingdom). His publications include Augustine’s Text of John. Patristic Citations and Latin Gospel Manuscripts (OUP, 2008) and the critical edition of the Vetus Latina Iohannes. Laurence Mellerin is Researcher of the Centre National de la Recherche Scientifique (CNRS) at the Institut des Sources chrétiennes in Lyon (France). She is editor (with Hugh Houghton) of Biblical Quotations in Patristic Texts, Studia Patristica LIV, vol. 2 (Papers presented at the Sixteenth International Conference on Patristic Studies held in Oxford 2011, M. Vinzent ed., 2013), including ‘Methodological Issues in Biblindex, An Online Index of Biblical Quotations in Early Christian Literature’, 11–32. Sara Schulthess is a Swiss National Fund PhD student at the Institute for the Study of Antiquity and the Institut Romand des Sciences Bibliques at the University of Lausanne (Switzerland). She is author of ‘Die arabi­schen Handschriften des Neuen Testaments in der zeitgenössischen For­schung: ein Überblick’, Early Christianity 3 (2012/4), 518–539.  Pnina Shor is Curator and Head of Dead Sea Scrolls Projects, Israel Anti­ quities Authority (Israel). Charlotte Touati is a Swiss National Fund Post-Doctoral Fellow at the Universities of Neuchâtel (Switzerland) and Hamburg (Germany). She is author of ‘The Apocalypse of the seven Heavens: from Egypt to Ireland’, in The End and Beyond: Medieval Irish Eschatology, ed. by J. Carey et al. (Oxbow Books, 2013) and co-editor of Nouvelles intrigues pseudo-­clémentines. Actes du deuxième colloque international sur la littérature apocryphe chrétienne 2006 (PIRSB 6) (Zèbre, 2008). Romina Vergari is a Post-doctoral Fellow at the Faculty of Protestant Theology of the University of Strasbourg (France). She is currently working on a monograph entitled Semantic Phenomena of Septuagint Greek. From Inherent to Selectional Polysemy.

list of abstracts The Leon Levy Dead Sea Scrolls Digital Library. The Digitization Project of the Dead Sea Scrolls By Pnina Shor The digitization project of the Dead Sea Scrolls (DSS) was initiated as part of the Israel Antiquities Authority’s (IAA) conservation efforts to preserve them for future generations. From the time of their discovery by a Bedouin shepherd in 1947, until the establishment of the IAA’s unique conservation lab dedicated solely to the conservation and preservation of the scrolls in 1991, the scrolls were heavily damaged by the ravages of time, as well as previous handling and treatment. The task of conservation and preservation of the scrolls continues to be an ongoing mission due to their extreme fragility and the need to make use of the most upto-date conservation methods known worldwide. We have been following a protocol drawn together with world experts, but since we have been practicing for about 20 years we decided it is time to reassess our work, and address unsolved issues. The idea to use spectral imaging for monitoring the wellbeing of the scrolls was first suggested a few years ago by a scientist from the Weitzman Institute. Eventually the project developed into an overall undertaking whereby we mean to monitor their physical condition on the one hand, and to expand and facilitate access to them, to scholars and the public worldwide, on the other. The digitization project will thus sustain the scroll’s conservation and preservation by creating high-quality spectral images, using the most advanced and innovative technologies available. These will support and provide active conservation assessments, non-invasive testing and monitoring tools, and a better record and documentation of the state of the scrolls. Furthermore the IAA has decided to make these highest resolution images, accompanied by data including transcriptions, translations, commentaries and bibliography available and accessible via the Internet to all, as well as to provide for scholarship by supplying interactive tools within a supportive research framework. Uploading the images and data to the Internet and allowing diverse searches will be achieved with the assistance of Google-Israel.

x

list of abstracts Dead Sea Scrolls inside Digital Humanities. A Sample By David Hamidović

The discoveries of almost 930 manuscripts near the site of Khirbet Qumran, on the western shore of the Dead Sea, contribute to a better knowledge of Ancient Judaism and Early Christianity. All serious study needs to have a well-established text-source. Thus the paleographical study of manuscripts is a necessity. The Qumran scrolls present the difficulty to be very fragmentary and many fragments have disappeared forever. Could we restore or restitute the lost text, or more precisely, how best can we evaluate the probability of restorations or restitutions? We propose a method based on software of computer aided drawing to write with the scribal hand. Thus the textual reconstructions by scholars can be evaluated. Moreover, this outlook may allow to better know the scribes of the Qumran community. Indeed, we wish to make a directory of scribal scripts to know if the manuscripts have been copied by the same scribes. Thus we wish to know the composition and evolution of the scribal milieu at Qumran. The Electronic Scriptorium: Markup for New Testament Manuscripts By H.A.G. Houghton The creation of scholarly editions has been transformed by the adoption of digital tools, affecting every stage of the process from approaching primary sources to final publication. One of the most significant is the way in which electronic media overcome some of the constraints of printed texts and permit a fuller and more flexible presentation of the data, as well as facilitating subsequent alteration and future re-use. This is exemplified by the full-text manuscript transcriptions produced by the International Greek New Testament Project (IGNTP) and the Institut für neutestament­ liche Textforschung (INTF) in their work towards the first scientific edition of the New Testament, the Editio Critica Maior. The present study offers an overview of the process of transcribing biblical manuscripts and outlines the encoding in Extensible Markup Language (XML) developed as part of the Workspace for Collaborative Editing, a joint Anglo-German project to create a suite of digital tools for the production of the Editio Critica Maior and, in time, for other textual traditions.



list of abstracts

xi

Digital Arabic Gospels Corpus By Elie Dannaoui This article is about a research project conceived by researchers from the University of Balamand in Lebanon to study the textual traditions of the Arabic versions of the four Gospels. The main objective of this project is to identify the different families of the Arabic translations preserved in the manuscripts’ tradition. To reach this goal, this research continues the previous researches in this field and tries to benefit from new technologies. This is why we decided to approach the topic from the angle of Digital Humanities in terms of strategies, tools and techniques. After a short overview about the state of research in the field of the Arabic manuscripts of the Gospels, the article presents our contribution in the international effort to study the text of the Arabic versions of the Gospels by building an Online Corpus containing the transcriptions of the Arabic manuscripts of the Gospels. The last part of the article presents the technological aspects and components of the project. The Role of the Internet in New Testament Textual Criticism: The Example of the Arabic Manuscripts of the New Testament By Sara Schulthess Over the years, the Arabic manuscripts of the New Testament have suffered from a comparative lack of interest in Western research: even today this field is barely investigated by textual criticism. This can be explained by the history of New Testament textual criticism, which focused more on finding the ‘original’ text than showing interest in the ‘secondary’ versions of the New Testament. Furthermore, the study of the Arabic manuscripts of the New Testament refers to issues that are discussed in cultural and postcultural studies and requires an interdisciplinary approach that leads us to think about identity and dialogue at the intersection of the socalled ‘religions of the Book’. However, new interest in these manuscripts is emerging: this interest is not unrelated to the present revolutions in New Testament textual criticism, made possible by the digital era and the Internet. The Internet supports and contributes to a renewed study of the Arabic manuscripts of the New Testament on several levels. Firstly, various issues that crystallize in the study of Arabic manuscripts—for instance, were there any Arab Christian writings before Islam?—are reflected on websites, often in a polemical way: the sphere of the Internet opens up a space for a new form of discourse, where New Testament manuscripts

xii

list of abstracts

raise questions for different communities and for people of different religions. Then, on another level, it should be noticed that the Internet platforms that promote collaborative work, such as the Yahoo Forum ­Textual Criticism of the Bible, give birth to a new dynamism concerning the Arabic manuscripts; this new type of collaborative work allows progress in the field, which demonstrates more broadly the importance of such research for the discipline. My article aims to show how the Internet and the digital framework encourage research around the Arabic traditions of the New Testament through ways that may be unexpected in order to analyze these new approaches. The Falasha Memories Project. Digitalization of the Manuscript BNF Ethiopien d’Abbadie 107 By Charlotte Touati The term ‘Falashas’ denotes the Jews from Ethiopia. They have developed a literature and a specific culture on the verge of disappearing. Their main texts, written in Ge’ez, are preserved in the manuscript BNF, Ethiopian d’Abbadie 107. The Memories Falashas Project aims to display online the digitized ms. 107, its transcription and the translation into English and French of its content. This later can be fully understood only through the reconstruction of its formal context. Useful information taken from the archives of Antoine d’Abbadie (1810–1897), first owner of the manuscript, and Joseph Halévy (1827–1917), a scientist mandated to investigate the Falasha community that produced the texts of ms. 107, will be directly connected with the content of ms. 107. The project follows the traditional practices of philological edition, but aims also to surpass them by providing access to a large amount of related information starting from the facsimile of the manuscript or its translation, without the artificial reconstruction of a text that probably never existed. The apparatus will be replaced by hyperlinks. The project is a reflection on critical edition, but it is also a reflection on the publication of the results of research in the field of humanities as multimedia objects, rather than paper books. The Seventy and Their 21st-Century Heirs. The Prospects for Digital Septuagint Research By Juan Garcés This essay gives an overview of some pertinent digital approaches to the research of the Septuagint (LXX), the Greek translation of the Hebrew



list of abstracts

xiii

Bible. Taking the pioneering Computer Assisted Tools for Septuagint Studies (CATSS) project as a point of departure, it aligns each of the project’s three modules to a seminal area of LXX research—in turn: lexicography, translation technique, and textual criticism. Using selected examples of the use of CATSS and comparable resources in LXX research and other relevant work, the essay takes advantage of hindsight to assess the prospects of digital LXX research in the context of said digital approaches. In doing so, the essay looks at two closely interrelated issues and asks what methodologically innovative approaches are facilitated by the refashioning of the LXX as digital data and, conversely, what kind of resources are desirable in order to facilitate approaches that cater to the advantages of digital research on the LXX. Digital Approaches to the Study of Ancient Monotheism By Ory Amitay This paper sets out a research program, with the aim of providing collaborative and methodological tools for studying ancient Monotheism. The program takes for granted that the huge amount of information, from many fields of study, puts such a task well beyond any one scholar, and in fact requires a considerable team of researchers. The technological solution suggested for such collaboration is Mediawiki software. As its proven success in supporting Wikipedia has shown, it is both a highly efficient for recording and cataloguing data, and a tried and true tool for online collaborative work. Methodologically, this project faces multiple challenges. The most immediate of these is the question of definition: what is the subject matter? My aim here is to go beyond the classic approach of first reaching a theological definition of Monotheism and then looking for its origins. Rather, it is to define the basic traits (theological, mythological, institutional, visual etc.) most prevalent in Monotheism today, and to trace their development in antiquity. For this purpose I plan to employ new theories which have not yet seen extensive use in the Humanities: memetics and network theory. Both methodologies are nicely applicable for use with Mediawiki. In the paper I give a rudimentary explanation how these two methodologies may open new possibilities for research, with hope of paving a way towards a digitally based quantitative analysis as a possible basis for historical argumentation.

xiv

list of abstracts

Internet Networks and Academic Research: The Example of New Testament Textual Criticism By Claire Clivaz As we are facing the most important turning point in the sciences in the last 300 years, the purpose of this article is to review some points about the Digital Humanities landscape and the emergence of open scientific networks. Such online collaborative tools and forums of discussion produce efficient (pertinent) knowledge when they are focused, open and when they offer the possibility for academic distinction. The second part will present some examples of the online networks used in the field of New Testament criticism, to evaluate their function according to the observations of the first part. The conclusion offers suggestions about what could be done to improve the collective online scientific production in New Testament textual criticism, sociology and new technologies. New Ways of Searching with Biblindex, the Online Index of Biblical Quotations in Early Christian Literature By Laurence Mellerin This chapter presents new ways of searching made possible by an online tool for patristic studies developed by Sources Chrétiennes in Lyon (France), Biblindex, http://www.biblindex.org. This index of the biblical references found in both Western and Eastern Christian literature currently covers the first four centuries. About 500,000 references have already been gathered. The eventual goal is to cover the whole of Late Antiquity and the early Middle Ages. IT development is currently in ­progress. Via a site using e-working public software, researchers worldwide will have the opportunity to contribute to its improvement without costly investment. This will both optimize their own research and develop the data through interdisciplinary exchanges. Two particularly innovative features of Biblindex are its spatio-temporal breadth and its multilingual corpus. By providing access in a single database to the quotations of the Bible in Jewish and Christian literature in any ancient language including those of the East, Biblindex offers a new prospect for biblical studies, paying attention to the reception of the text through the study of its different versions and subsequent interpretations.



list of abstracts

xv

Aspects of Polysemy in Biblical Greek. A Preliminary Study for a New Lexicographical Resource By Romina Vergari The paper departs from a corpus-based analysis of a portion of the Biblical Greek lexicon. Attention is focused on the nouns associated with the semantic sphere of ‘Law’ and their Hebrew equivalents, within a welldefined and homogeneous textual corpus. The analysis is underpinned by a study of word-meaning representation, carried out within a lexicological framework which takes into account the flexibility produced by the context, in a range from inherent to selectional polysemy. It assumes the syntagmatic processes of sense modulation, which contribute to determine the meaning of complex syntagmatic structures. This framework suggests new paths for scholars of Biblical Greek. The paper suggests ‘Septuagint Word-Clustering Database’ (SWCD) as a lexicographical tool for the study of the Biblical Greek lexicon. This resource will enhance the possibilities available to scholars, supplying the following new features: a) Search all the nominal heads of a given adjectival modifier; b) Search all the direct objects of a given predicate; c) Search all the occurrences of a given nominal/verbal phrase; d) Cluster the readings of an inherently polisemous word (inherent polysemy occurs when multiple interpretations of an expression are due to the semantics inherent in the expression itself); e) Cluster the readings of a word on the basis of selectional restrictions (selectional polysemy arises when a novel interpretation of an expression is due to context); f) Cluster nouns on the basis of their type in terms of logical formalism. Offering a refined representation of semantic flexibility with regard to the processes of sense modulation, SWCD is also able to show whether a regular pattern of polysemy in a given Greek noun matches its Hebrew equivalents. Publishing Digitally at the University Press? A Reader’s Perspective By Andrew Gregory This chapter discusses new developments in commercial academic publishing that use digital platforms rather than print, and traces some of the similarities and differences between these two media and the ways in which readers may use them. It considers how digital publication may give new life to the traditional monograph and other forms of printed work that flourished in the age of print, and also how its potential may

xvi

list of abstracts

be limited if publishers, readers and authors continue to approach digital platforms in a manner that is typographically conditioned. Thus it notes not only the possibilities that new technology allows, but also the continuing influence of analogue habits and expectations, and suggests that a tipping point may come when older readers and authors who grew up with print are succeeded by those who learned to read and to write as natives of a digital world. By considering not only the changing form of the academic book, but also the changing expectations of the academic reader and author, it raises the possibility that what may be most decisive for the future of academic publishing is not whether or not particular books are born digital, but whether their likely readers are. A number of digital platforms published by OUP provide specific examples that show how this leading academic publisher is embracing digital technology on a large scale, but in a way that continues to be shaped by the characteristic form of printed texts and by the expectations that those texts both reflect and engender. The essay asks what is digitally distinctive about this new approach to publishing, or to what extent it produces ‘electronic incunabula’ that retain the characteristics of an earlier form of written text, but do not make full use of the possibilities that new technology allows. Does Biblical Studies Deserve to be an Open Source Discipline? By Russell Hobson As work on the digitization of many ancient texts progresses, a critical difference between the fields of Biblical studies and Assyriology is emerging. While Biblical studies is dominated by proprietary products, developed commercially and sold under license, the vast majority of online resources for Assyriology are based on open source and open access models. As a result, biblical scholars often pay to license the electronic tools they work with while Assyriologists enjoy an increasing array of resources freely available on the web. Why is there such disparity between the tools used in these two disciplines? What are its real and potential effects on the way biblical scholars work now and in the future? This paper argues that open source models promote the development of better applications that are more ‘field-ready’ for scholarly use, and that commercially neutral, webresident software will provide the best benefit to future research.

preface Claire Clivaz, Andrew Gregory and David Hamidović This collection of essays is—as far as we know—the first attempt to illustrate under the general label ‘Digital Humanities’ a selection of what is already happening in the closely related fields of Biblical Studies, Early Jewish Studies and Early Christian Studies in their encounter with digital culture.1 Colleagues in another related field, Classical Studies, have already done something similar, and our work may be compared to theirs, both in subject matter and in approach: Digital Research in the Study of Classical Antiquity.2 We readily acknowledge that the use of digital technology is well established in the disciplines to which the contributors to this volume belong. Many scholars from different countries and different institutions have been involved both in the digitization of ancient Jewish and Christian manuscripts and in the development of digital research tools for a significant period of time. Collaborative projects including a digital component are more and more common in these fields. But our concern here is less to draw attention to the increasing number of projects that make use of digital technology than to draw attention to the emergence of the conscious recognition of something new in the way that we study ancient texts, a shift that would not be possible were it not for digital technology, and that will almost certainly affect the way in which our disciplines will develop. Thus we hope that this volume will contribute to the conscious and explicit recognition of an emerging phenomenon that is more than the sum of its parts, and will offer critical reflections on how this development transforms our fields of study.

1 For some earlier examples, see Parker, D., ‘Ancient Scribes and Modern Encodings: The Digital Codex Sinaiticus’, and Schmid, U., ‘Translating the New Testament Online’, both in Text Comparison and Digital Creativity: The Production of Presence and Meaning in Digital Text Scholarship, Scholarly Communication 1, ed. by W. van Peursen, E.D. Thoutenhoofd and A. van der Weel (Leiden: Brill, 2010), resp. 173–188 and 189–205. 2 Mahony, S., and G. Bodard, eds., Digital Research in the Study of Classical Antiquity, Digital Research in the Arts and Humanities (Farnham: Ashgate Publishing, 2010).

xviii

preface

Earlier versions of many but not all of the essays included in this book were originally presented at the Digital Humanities research group which held its inaugural sessions at the EABS/international SBL meeting in Amsterdam in July 2012.3 Others were commissioned for this volume. Abstracts for all the essays are included at the beginning of this book, so the introduction to this volume focuses not on summarizing the essays that follow, but on questions and issues that illustrate—and may perhaps contribute to—the development of Digital Humanities in the context of Biblical, Early Jewish and Early Christian studies. We are glad to express our thanks not only to all those who joined us in Amsterdam, but also to those who have contributed essays to this volume. We are particularly grateful to Sara Schulthess, a PhD student and research assistant at the University of Lausanne, who devoted much of her time to the editing and formatting of this volume.

3 At the time of writing this introduction, further meetings were scheduled to take place at the International SBL Meeting in St-Andrews in July 2013, the EABS Meeting in Leipzig in August 2013, and the annual SBL Meeting in Baltimore in November 2013.

introduction: digital humanities in biblical, early Jewish and early Christian studies Claire Clivaz The essays in this book are written not only for specialists in the related field of biblical, Jewish and Christian studies, but also for scholars across the spectrum of the humanities and social sciences who share a common interest in trying to understand changes in the ways in which we read and write in a setting where more and more information and data is being made available at increasingly faster rates. The contributors are all historians and philologists who specialize in the study of Judaism and Christianity, especially in antiquity, and in the study of their written texts, but all are actively engaged with wider questions about how digital technology may reshape, enlarge or otherwise affect not only the specific disciplines in which they are trained, but also the whole spectrum of the humanities and social sciences. Their essays are therefore written not primarily for those who want to know how other scholars in their own subject areas are already using such technology (although it certainly serves that purpose) but for those who have an interest in what such new developments might mean, and their likely impact on the way in which the practice of scholarship may develop. Although our contributors and their sources are drawn from related fields, their analysis and discussion raises questions that should be of interest for anyone with an interest in ‘digital humanities’, regardless of their specific field of study. The label ‘Digital Humanities’ emerged in oral discussions between English speaking scholars around 2001,1 and corresponds to institutional traces and DH degrees in several countries.2 It signals a new moment in Western epistemology,3 and beyond. But the meeting between the Humanities and

1 Kirschenbaum, ‘What Is Digital Humanities and What’s It Doing in English Departments?’ 2 For the masters, look for example at http://www.iath.virginia.edu/hcs/MDST.MA.html; http://www.ctsdh.luc.edu/?q=ma_digital_humanities; http://www.digitalhumanities.net/ node/19; http://humanexperience.stanford.edu/digital, last accessed 28 April 2013. 3 For a presentation of this topic, see Clivaz, ‘Common Era 2.0. Mapping the Digital Era from Antiquity and Modernity’.

2

claire clivaz

the computer is as old as the 1940s: the first attempt to create a Humanist digital tool, the Index Thomisticus,4 was the work of a Jesuit scholar, Roberto Busa. The next two generations offered some impressive starting points, as some of the contributors to this present volume observe. Juan Garcés reminds us that ‘the most prominent project making use of electronic data management for the study of the LXX is still the Computer Assisted Tools for Septuagint Studies (= CATSS) project, which was called into life through the initiative of the IOSCS in the early 1970s’.5 Pnina Shor underlines for her part that ‘Dr. Bearman had already in the mid-1990s, together with B. and K. Zuckerman, experimented spectral imaging of a few of the [Dead Sea] scrolls in attempt to aid in their deciphering’.6 The Israel Antiquities Authority (IAA) developed this experiment with such energy that it is today committed to ‘a new spectral imaging system, the MegaVision (MV) system, which combines the initial three working stations together [. . .]. This system is the most advanced imaging technology existent today’.7 In comparison, New Testament research arrived quite late in the digital field. As H.A.G. Houghton observes, ‘the first example of the re-use of electronic New Testament transcriptions was the Digital Codex Sinaiticus, an online edition of this fourth-century Greek Bible combining images from all four holding institutions, a complete transcription of the text and a translation’.8 This project started in 2005 and was inaugurated in 2010. In 2010 work ‘began on a new online environment to integrate all the tools required by the partners involved in producing the Editio Critica Maior. Scheduled for completion in late 2013, the Workspace for Collaborative Editing aims to connect each stage of the editorial process’.9 If NT research came late in the experiment of digitalizing manuscript and developing new ways to edit its text, nevertheless it is worth emphasizing

4 See http://www.corpusthomisticum.org/it/index.age, last accessed 10 March 2013. On Busa as pioneer, see McCarty, ‘What is Humanities Computing? Toward a Definition of the Field’, 8. 5 Garcés in the present volume, ‘The Seventy and Their 21st-Century Heirs. The Prospects for Digital Septuagint Research’, 98. 6 Shor in the present volume, ‘The Leon Levy Dead Sea Scrolls Digital Library. The Digitization Project of the Dead Sea Scrolls’, 15. 7 Shor in the present volume, ‘The Leon Levy Dead Sea Scrolls Digital Library. The Digitization Project of the Dead Sea Scrolls’, 16. 8 http://www.codexsinaiticus.org/, last accessed 30 April 13. Houghton in the present volume, ‘The Electronic Scriptorium: Markup for New Testament Manuscripts’, 35. 9 Houghton in the present volume, ‘The Electronic Scriptorium: Markup for New Testament Manuscripts’, 35–36.



introduction

3

that Houghton’s article shows the very high standards to which it now adheres, for example in its use of the publication system Anastasia:10 ‘This ran as a server which converted SGML data into HyperText Markup Language (HTML) for viewing in a web browser. The tags were converted into HTML elements and then rendered according to a specified scheme’.11 It may be helpful here to raise a question that is not addressed in this volume, but that may be an issue for all scholars involved in textual production: does it make sense and is it possible to imagine that every textual critic or editor will become someone with advanced skills in IT and conversant with such tools as may be required to produce and edit digitized texts? Such a scenario is probably not even imaginable, let alone useful. But, at the same time, one has to say with the same conviction that electronic encoding of texts cannot be transferred only to some highly specialized ‘DH-ers’ among the Humanists. Indeed, in a digital edition process, data and metadata are related. Sources (data) and their interpretations (metadata) are put in relationship, as was the case at the moment of their production: it is print culture that encouraged scholars in the habit of separating them from each other. As Charlotte Touati demonstrates in her article, a digital edition of an ancient text allows scholars to aggregate information that was produced at the same time in research, but is usually separated in a printed publishing process. Studying the Ethiopian Falashas, she would like digitally to edit together three corpora, which one would be usually considered as a source (data) and two as secondary literary productions (metadata): the manuscript Ethiopien d’Abbadie 107 and the writings of two modern scholars, Antoine d’Abbadie and Joseph d’Halévy. As she observes, ‘The first step’s publication is the building of the website core, on which the further two stages will depend. Thus the data from the last two corpora, that is the cross-references, will be aggregated around the nucleus (ms. 107, the primary evidence)’.12 As may be seen, that digital edition belongs to the core of the knowledge to which the scholar in the humanities aspires, and is able to deal with data and metadata, sources and secondary literature, together. We can be confident that technology and its uses will continue to develop, and that increased familiarity

10 See Robinson, Anastasia: Analytical System Tools and SGML/XML Integration Applications, Version 2.0 (Computer Program). 11  Houghton in the present volume, ‘The Electronic Scriptorium: Markup for New Testament Manuscripts’, 34. 12 Touati in the present volume, ‘The Falasha Memories Project. Digitalization of the Manuscript BNF, Ethiopien d’Abbadie 107’, 86.

4

claire clivaz

with the benefits that digital culture offers will enable scholars in the humanities to make use of the new possibilities that digital textuality will permit, regarding the promising developments of the linked data topic. These benefits include, for example, the possibility for the scholars easily to ‘draw’ missing parts of letters in manuscripts, following the ductus of various scribes, as David Hamidović explains in his article: I use the software Corel Draw. Besides the filters and clipping, the function insertion of special characters allows drawing on a layer above the picture. We can choose an originally integrated font character. This can be chosen and sized to the general size of read letters. Then we position an integrated letter on a letter on the picture. Of course, the format of the integrated letter does not precisely correspond to the letter written by the scribe. In clicking on an integrated letter, nodal points appear all around the letter and they form a Bezier curve whatever the chosen font.13

We have here a clear illustration of these Humanities ‘made with the fingers’, the digiti according to the Latin word digitus: ‘digital’ humanities.14 In such a practical work, the scholar is lead to reconsider the text as an object, as a document, made by specific people at a specific time and in a specific place, not merely the vehicle of disembodied ‘ideas’. This epistemological and technological shift could lead to impressive new scientific results, as Hamidović indicates: I am convinced that we shall manage to establish that some manuscripts have been copied by the same scribes. We shall have a more precise overview of the number of scribes who have copied the Dead Sea Scrolls between the end of the third century BCE and the middle of the first century CE, during two centuries and half. We may find scribes specialized in the writing of theonyms, the sacred name of God.15

Several other challenges that belong to the emerging digital academic culture are also illustrated in this volume, notably the mixing-up of Western and non-Western academic cultures. Digital support increasingly

13 Hamidović in the present volume, ‘Dead Sea Scrolls Inside Digital Humanities. A Sample’, 26. 14 See Darnton, The Case for Books: Past, Present, Future, XIII: ‘we find our way through the world by means of a sensory disposition that the Germans call Fingerspitzengefühl. If you were trained to guide a pen with your finger index, look at the way young people use their thumbs on mobile phones, and you will see how technology penetrates a new generation, body and soul’; quoted in Clivaz in the present volume, ‘Internet Networks and Academic Research: the Example of New Testament Textual Criticism’, 159. 15 Hamidović in the present volume, ‘Dead Sea Scrolls Inside Digital Humanities. A Sample’, 29.



introduction

5

facilitates and contributes to an academic culture that is less focused only on Western scholars and preoccupations, as we have already seen above with Charlotte Touati’s article: her digital project will allow one to ‘respect equally the ethnographic and philological dimensions of the [Falashas] issue’.16 Arabic New Testament manuscripts seem also to benefit from renewed attention, made possible by digital approaches, as shown in two essays in this volume, one by a Middle-Eastern scholar, Elie Dannaoui, the other one by a Western scholar, Sara Schulthess. Technological novelties and new methodological approaches are used to answer challenges specific to Arab Christian literature: ‘Instead of using a well-designed relational database schema, we decided to use an unstructured database. This family of databases, also known as NoSQL,17 is non-relational, distributed, open-source and horizontally scalable. Its schema-free approach will enable us to anticipate the emergence of new text formats and will allow the corpus to host different types of texts (lection, Gospel chapter, poem, book chapter, prayer . . .)’, as Dannaoui summarizes.18 Sara Schulthess demonstrates in her article how Muslim and MiddleEastern Christian sensibilities and identities are mixed-up in diverse websites offering images and discussion of early Greek or Arab Christian manuscripts: a surprising Western and non-Western academic world is emerging. As Schulthess concludes: In this context, it is easy to understand the need of promoting a tradition, which existed in many forms long before the West provided one. The will to reaffirm a culture is illustrated by the resumption of publications in recent years. This view is also shared by the Islamic website that contains the topic on John 1:1, where we find the following remark: ‘We, Arabs, Muslims or Christians, should have an interest in our heritage, as the Latins are interested in their Latin heritage, the Syriacs in their Syriac heritage, etc. We must take care of this heritage written in Arabic to give it to the world rather than they provide us our heritage’.19

If the emergence of a plural cultural academic world is potentially one of the major features of digital research and publishing, other points are also impressive. For example, collaborative work is deeply transforming 16 Touati in the present volume, ‘The Falasha Memories Project. Digitalization of the Manuscript BNF, Ethiopien d’Abbadie 107’, 89. 17 http://nosql-database.org, last accessed 30 April 2013. 18 Dannaoui in the present volume, ‘Digital Arabic Gospels Corpus’, 65. 19 http://www.hurras.org/vb/showthread.php?t=14868, last accessed 9 January 2013; quoted in Schulthess in the present volume, ‘The Role of the Internet in New Testament Textual Criticism: the Example of the Arabic Manuscripts of the New Testament’, 80.

6

claire clivaz

research habits, as Claire Clivaz’s article underlines. If mathematical problems can be solved or galaxies discovered, research on New Testament manuscripts also benefits from the involvement of scholars on open or semi-open forums of discussion. This small community of research— scholars of New Testament textual criticism—shares focus and methodologies, and could still develop its collective capacity to produce knowledge, by encouraging open forum of discussions and by innovating with new ways of certification to recognize such scholarly work. As Clivaz concludes, ‘I am convinced that NTTC is one of the most efficient test-cases for helping Humanities to enter in the new mode of scholarly production of science’.20 Openness is really an important challenge; it is impressive to read in Russel Hobson’s article that: The study of the Bible reaches across a much broader demographic than does the study of the cuneiform traditions from ancient Mesopotamia. It is surprising, then, that the use of web based resources, as seen in Assyriology, is not even more widespread in Biblical studies. [. . .] Ultimately, we purchase expensive, feature laden commercial products to do research tasks in Biblical studies that we can do in Assyriology without cost by using openly accessible online resources.21

We shall see in the future what may be the consequences of the present economic situation. In any case, digital tools are able to reconfigure categories and classifications, as Ory Amitay underlines clearly in his article on ‘monotheism’ in Antiquity: A simple example of how this can be done is offered along the path taken by the Culturomics project, using the Google Ngram viewer (GNv) on the Google Books database. As [my results show], I have searched for both ‘monotheism’ and ‘Monotheism’, as well as for ‘Abrahamic’. The search traces the frequency of the use of each word in a vast collection of English and North American literature from the last two centuries, spread along a chronological timeline.22

20 Clivaz in the present volume, ‘Internet Networks and Academic Research: the Example of New Testament Textual Criticism’, 174. 21  Hobson in the present volume, ‘Does Biblical Studies Deserve to Be an Open Source Discipline?’, 262. 22 Amitay in the present volume, ‘Digital Approaches to the Study of Ancient Monotheism’, 147.



introduction

7

The methodology is quite simple here, but allows one to demonstrate quickly and efficiently what scholars have previously been able to do only very laboriously without the assistance of digital tools. Digital tools also allow progress from a simple printed index of patristic quotations (the Biblia patristica) to Biblindex, as Laurence Mellerin explains: Two particularly innovative features of Biblindex are its spatio-temporal breadth and its multilingual corpus. By providing access in a single database to the quotations of the Bible in Jewish and Christian literature in any ancient language including those of the East, Biblindex offers a new prospect for biblical studies, paying attention to the reception of the text through the study of its different versions and subsequent interpretations.23

Humanists attached to the complexity and specificity of their works will be surely able to develop still an amount of new tools, as one can see with the project of Romina Vergari, the ‘Septuagint Word-Clustering Database’. She describes it as able to ‘bring out from the texts analyzed, structures and relations [. . . to get] a deeper understanding of word meaning, the limits of its flexibility, its creative processes, its diachronic change and cross-cultural variability’.24 We can feel here the excitement and enthusiasm that often accompanies the creation of digital tools. Of course the hope that we can use digital resources as tools with which to develop a better, clearer, and deeper understanding of human products and cultures may prove to be deceptive and misleading. After all, the Enlightenment had the same aspiration, for the worst, for the best of what was produced in Western history. The present volume has at a simpler level the purpose to be a showroom of what is at stake in digital streams in biblical studies, early Jewish and early Christian studies. The changes are not only irreversible, but probably greater than many scholars acknowledge, even if they do not consider either themselves or the peers to be digital humanists, or doing anything fundamentally different from their predecessors. As Andrew Gregory summarizes in his article:

23 Mellerin in the present volume, ‘New Ways of Searching with Biblindex, the Online Index of Biblical Quotations in Early Christian Literature’, 177. 24 Vergari in the present volume, ‘Aspects of Polysemy in Biblical Greek. A Preliminary Study for a New Lexicographical Resource’, 228–229.

8

claire clivaz We use general-purpose information technology as a matter of course, so almost all of us are digital humanists already, at least in the weak but by no means insignificant sense that our use of everyday technology entails. We are digital humanists because we read and write texts in a digital environment, even if we cannot code those texts for ourselves, even if rarely or never do we use, let alone build and develop, more specialist digital tools and resources in the course of our academic work.25

So, nolens volens, we are all ‘DH-ers’, Digital Humanists. Only time will tell what this will mean and how it will transform the ways in which our disciplines develop. References Clivaz, C., ‘Common Era 2.0. Mapping the Digital Era from Antiquity and Modernity’, in Lire Demain. Des manuscrits antiques à l’ère digitale / Reading Tomorrow. From Ancient Manuscripts to the Digital Era, ed. by C. Clivaz, J. Meizoz, F. Vallotton and J. Verheyden, with B. Bertho (Lausanne: PPUR, 2012), 23–60 (ebook). Darnton, R., The Case for Books: Past, Present, Future (PublicAffairs, 2009). Kirschenbaum, M.G., ‘What Is Digital Humanities and What’s It Doing in English Departments?’, ADE Bulletin 150 (2010), 56–57. http://mkirschenbaum.files.wordpress. com/2011/03/ade-final.pdf, last accessed 28 April 2013. McCarty, W., ‘What is Humanities Computing? Toward a Definition of the Field’ (1998), 1–9. http://www.cch.kcl.ac.uk/legacy/teaching/dtrt/class1/mccarty_humanities_computing .pdf, last accessed 10 March 2013. Robinson, P.M.W., Anastasia: Analytical System Tools and SGML/XML Integration Applications, Version 2.0, Computer Program (Scholarly Digital Editions: Nottingham, 2003). Documentation available online at http://sd-editions.com/anastasia/index.html.

25 Gregory in the present volume, ‘Publishing Digitally at the University Press? A Reader’s Perspective’, 236.

Part One

Digitized Manuscripts

the leon levy dead sea scrolls digital library. the digitization project of the dead sea scrolls Pnina Shor 1. Introduction The involvement of the Israel Antiquities Authority (IAA) in the management, publication, documentation and preservation of the Dead Sea Scrolls (DSS) began with its foundation in 1990. An advisory committee was appointed to speed-up the completion of the publication of the hundreds of DSS manuscripts long due. Besides the renowned, well-preserved scrolls from Qumran Caves 1 and 11 there are thousands of scroll fragments, mainly recovered from Qumran Cave 4, that belong to c. 930 manuscripts.

Figure 1. View of Qumran Cave 4 (Courtesy of IAA).

12

pnina shor

Figure 2. The Scrollery in the 1950s, the trestle tables with scroll fragments arranged in plates (Photographer: Najib Albina, Courtesy of IAA).

In the 1950s and early 1960s, the scholars appointed to study the scrolls concentrated their efforts on sorting the fragments and ascribing them to different manuscripts; they had little awareness, if at all, of the scrolls’ physical state. The scholars spread out the scroll fragments on long trestle tables, and when they identified two pieces or more that matched, they joined them together with adhesive tape. The joined fragments, which were thought to have belonged to the same manuscript, were then encased within more than 1200 glass plates. While compiling the plates, N. Albina, photographer of the Rockefeller Museum, documented the working stages with infrared photographs. In the mid-1960s, the British Museum sent a conservator to prepare some of the scroll plates for exhibition. The conservator remarked that the pressure of the glass plates and the adhesive tape were causing damage to the scrolls: the residues of the tape penetrated the parchment and papyrus, resulting in their disintegration.1 Just like the scholars in the 1950s, the British conservator did what she knew best, i.e., using British Museum 1 Faulkes, unpublished notes.



the leon levy dead sea scrolls digital library

13

Leather Dressing; unfortunately, this caused the scrolls further damage. In the 1970s, yet another intervention was employed by the Israel Museum, Jerusalem, using Perspex glue and rice paper, causing some additional irreversible damage. 2. Conservation and Preservation of the Scrolls While engaged in the founding of an advisory committee for the publication of the scrolls, the IAA realized that the scrolls would have to undergo physical treatment and maintenance procedures to prevent further deterioration in their state of preservation. The damaged state of the scrolls is a consequence of both ravages of time and mishandling and mistreatment during the four decades that had passed since their dis­covery in 1947. In 1991, the IAA established a unique conservation laboratory, dedicated to the conservation and preservation of the DSS. The establishment of the lab was assisted by the leading Israeli expert in paper and parchment conservation, E. Boyd-Alkalay, as well as by an international team of world experts in the field of manuscript conservation and preservation, sponsored by the Getty Conservation Institute.2 Four conservators were recruited and trained to care for these extremely challenging manuscripts, Lena Libman, Tanya Bitler, Tanya Treiger and Asia Vexler. A work protocol was drawn up, and a new era in the handling (or rather, non-handling) and managing of the scrolls began. As the environmental stability that had once ensured the scrolls’ preservation within the caves had been disturbed, it was necessary to build a climate-controlled storeroom to house the scrolls, where the humidity and temperature would be monitored. This too was accomplished with the help of the Getty Conservation Institute. The infra-red photographs, which reflect the state of the scrolls as they were found over sixty years ago, before the various interventions, are at times as invaluable as the scrolls themselves. Thus, the IAA created a designated climate controlled storage facility to preserve these so-called PAM negatives, as well. As a first course of action, a comprehensive survey was undertaken to assess the preservation state of the scroll fragments. Later, the survey was expanded also to include the eight scrolls on exhibit in the Shrine of the

2 Quandt, Stiber and Stanley Price, Conservation of the Collection of Dead Sea Scrolls in the Rockfeller Museum, Jerusalem Getty Report, A Report to the Israel Antiquities Authority.

14

pnina shor

Figure 3. Removal of adhesive tape with tweezers (Courtesy of IAA).

Figure 4. IAA conservator preparing scroll for exhibition (Photographer: Shai Halevi, Courtesy of IAA).



the leon levy dead sea scrolls digital library

15

Book, most of which subsequently underwent treatment in the IAA DSS conservation lab.3 The IAA conservators immediately engaged in treating the hundreds of fragile manuscript fragments in the best-known, worldwide conservation procedures. This includes primarily the cleaning of the scroll fragments from previous interventions of the 1950s, 1960s and 1970s, i.e., the removal of adhesive tape and other adhering materials.4 3. The Digitization Project of the Dead Sea Scrolls The digitization project of the DSS began as a conservation effort. The task of conservation and preservation of the scrolls is ongoing due to their extreme brittleness and the need to make use of the most up-to-date, stateof-the-art, conservation methods known worldwide. After twenty years of extensive work, the IAA decided to reassess the conservation work and address some unsolved issues. To this end, the IAA collaborated with the Italian Ministry of Culture and the Institute of the Pathology of the Book in Rome. Subsequently, once again an international group of experts, headed by D. Hamburg, director of Preservation Programs at National Archives, Maryland, came to Israel for a reassessment workshop.5 The IAA consulted also with Prof. S. Weiner, a scientist from the Weitzman Institute, Rehovot, who suggested the use of spectral imaging for monitoring the well-being of the scrolls. Following his advice, the IAA approached Dr. G. Bearman, formerly a leading scientist at the Jet Propulsion Laboratory, California Institute of Technology, and an expert in spectral imaging. Dr. Bearman had already in the mid-1990s, together with B. and K. Zuckerman, experimented spectral imaging of a few of the scrolls in attempt to aid in their deciphering.6 In November 2007, celebrating the 60th anniversary of the discovery of the scrolls, the IAA convened an international committee of experts from Israel and abroad to evaluate the most advanced imaging technologies available on the market, as well as the leading programs for the management of large databases. The committee recommended the initiation of 3 Boyd-Alkalay and Libman, A Condition Report of the Exhibit of the Dead Sea Scrolls at the Shrine of the Book. 4 Boyd-Alkalay and Libman, ‘The Conservation of the Dead Sea Scrolls in the Laboratories of the Israel Antiquities Authority’. 5 Hamburg et al. Dead Sea Scrolls Preservation Review. 6 Bearman and Spiro, ‘Archaeological Applications of Advanced Imaging Techniques’.

16

pnina shor

a large-scale digitization project; including the monitoring of the scrolls, the creation of high-quality colour images and advanced near infra-red images. As the publication of the scrolls was formally completed, the IAA sought to upload the digitized scroll images online, with their transcriptions, translations, commentaries and bibliography, allowing free access to public and scholars alike. In August 2008, the IAA set up a designated studio and led a pilot project, headed by Dr. G. Bearman and S. Tanner, Director of King’s Digital Consultancy Services in King’s College London. A comprehensive report7 of the results was sent for review to three independent experts in the field of advanced imaging technologies. Following the pilot, the IAA raised the funds for the imaging project. By the time the IAA was ready to commit to the project, a new spectral imaging system, the MegaVision (MV) system, which combines the initial three working stations together, was brought to our attention. This system is the most advanced imaging technology existent today. It provides high resolution, large format (39 MP) multiband images, and is the only spectral imaging system with such qualities. Covering the visible and near infrared, it also offers the most accurate colour fidelity available for any colour camera system. Using special wide-band corrected optics, the images are all automatically registered and do not require post-processing. In January 2011, the digitization project began with the high-resolution scanning of all the PAM negatives and all other negatives taken by the IAA to date. Using the highest quality of optical lenses in the scanning of the negatives and transparencies, all 5320 negatives were scanned in 800 PPI 1X1. As a result, they can be enlarged to three times their original size, without losing any information and without pixilation. At the beginning of July 2011, MegaVision Company informed the IAA that the system, which was compiled especially for the scrolls, was ready. The Santa Barbara Museum of Art in California kindly offered to host a final pilot. In the allotted space, an exact mock-up of the IAA studio was set up. For three full days, images were taken, data was analyzed and work procedures were practiced. The system was then dismantled to be finetuned for delivery. At the end of August 2011, K. Boydston, MegaVision owner and developer of the system, Dr. B. Christens-Barry, lighting technology expert, and Dr. G. Bearman, IAA consultant for imaging technologies, came over to 7 Tanner and Bearman, Digitizing the Dead Sea Scrolls, Report on the Pilot Imaging.



the leon levy dead sea scrolls digital library

17

Figure 5. The MegaVision imaging system (Photographer: Shai Halevi, Courtesy of IAA).

Israel to set up the system in the designated studio. Two new members were chosen to operate the system—Y. Medina, Jerusalem Art Graphics, and S. Halevi, photographer—and joined the team for two intensive weeks of set-up and training. A work procedure was formulated, whereby the plates of the scroll fragments are prepared for imaging by the IAA conservators—the scroll fragments are either taken out of their capsule and netting, or detached from their acid free cardboard storage. The plate is then imaged in a workstation for documentation and numbering of the fragments. The team adapted an automatic system that reads the image, gives every fragment a unique number and creates three files: an image map of the numbered fragments; an Excel file of the coordinates and pixel size of each fragment; and another Excel file for the MV system. The latter allows the integration of each fragment number with the plate’s data, which was

18

pnina shor

Figure 6. Spectral image of Psalms scroll, 11QPsalmsa—merging of color image with near infra red image (Photographers: Yair Medina and Shai Halevi, Courtesy of IAA).

drawn from the IAA State Collections database and uploaded to the MV system in preparation for the imaging. Using the highest-quality optical lens, custom-made for the project and designated LED lights, it was decided to image every fragment, both recto and verso, in 12 wavelength—7 in the visible spectrum and 5 in the near infra-red (from 445 to 924 nanometers)—producing 28 exposures, thus creating a file of 56 monochrome exposures per fragment. The system generates a 57th file of a color image that combines all visible wavelengths. The resolution of files is 1215 PPI 1:1, capturing circa four Gigabytes per fragment. Since the field of view is fixed and relatively small, many of the fragments undergo tiling and stitching, e.g., plate 978 of 11QPsalmsa, comprising 22 tiles. Given that these images are primarily produced for developing a monitoring system of the physical state of the scrolls, the system is calibrated on a daily, weekly and monthly basis. The colour targets also help the system calibrate the values of each wavelength to produce a colour image that is three-four times more accurate than any other colour image. The IAA conceived this project on collaborative terms. A three-way collaboration of joint expertise was established between the University of Eastern Piedmont (Italy), where spectral imaging to monitor and track the well-being of the scrolls is being developed;8 the Preservation Research and Testing Division of the Library of Congress, where a project concerning the aging of parchment is under way; and the Institute of Chemistry

8 Marengo et al., ‘Development of a Technique Based on Multi-Spectral Imaging for Monitoring the Conservation of Cultural Heritage Objects’.



the leon levy dead sea scrolls digital library

19

Figure 7. The Ten Commandments Deuteronomy scroll, 4QDeutn, Spectral image in near infra red—924 nanometers (Photographers: Yair Medina & Shai Halevi, Courtesy of IAA).

of The Hebrew University of Jerusalem, where a research on the physical properties of parchment and ink has begun. Making the digital images of the scrolls and their content accessible to all online, in a number of languages and formats, was made possible due to collaboration with Google R&D Center in Israel. As stated by S. Dorfman, IAA General Director, ‘we are establishing a milestone connection between progress and the past to preserve this unique heritage for future generations.’ Ultimately, our objective is to upload to the Internet all scroll images and meta-data, including transcriptions, translations and bibliography, allowing free access to anyone anywhere in the world. Toward this end, the IAA nominated an advisory committee of scholars. Dr. S. Zoref, DSS scholar, joined the IAA staff. She has worked closely with the IT team to improve and modify the data in the IAA State Collections’ database, based upon databases compiled by Prof. E. Tov,9 for integration in the website. The IAA is also collaborating with the Orion Center of the Hebrew University of Jerusalem to better their Dead Sea Scrolls Bibliography project, so as to facilitate research on the scrolls from either venue. Following the initial launch of the website by Google in December 2012 (www.deadseascrolls.org.il), the IAA will periodically update the site by loading new spectral images, and by revising the data concerning the scroll fragments to reflect ongoing research. In the future, transcriptions and translations of the texts will be added, as well as interactive tools for scholarly research. With the goal of creating the most widely used online resource; several other scientific research projects are under way.

9 Tov, Revised Lists of the Texts from the Judaean Desert.

20

pnina shor

The Dead Sea Scrolls are a universal cultural heritage. As such, the IAA is honoured with the obligation to safeguard the scrolls, preserve them for future generations and share them with the public and scholarly community worldwide. Acknowledgements A major lead gift from the Leon Levy Foundation has enabled the IAA to embark on this unique project. Substantial funding has also been provided by the Arcadia Fund. The Yad Hanadiv foundation, the project’s constant supporters, granted the initial contribution. References Bearman, G.H., and S.I. Spiro, ‘Archaeological Applications of Advanced Imaging Techniques’, The Biblical Archaeologist, 59/1 (1996), 56–66. Boyd-Alkalay, E., and L. Libman, A Condition Report of the Exhibit of the Dead Sea Scrolls at the Shrine of the Book (1996), unpublished. ——, ‘The Conservation of the Dead Sea Scrolls in the Laboratories of the Israel Antiquities Authority’, Restaurator 18 (1997), 92–101. Faulkes, V., unpublished notes (1963). Hamburg, D., L. Morenus, L. Watteeuw, R. Danzing, G. Vigliano and T. Elper, Dead Sea Scrolls Preservation Review (2013), unpublished. Marengo, E., M. Manfredi, O. Zerbinati, E. Robboti, E. Mazzucco, F. Gosetti, G. Bearman, F. France and P. Shor, ‘Development of a Technique Based on Multi-Spectral Imaging for Monitoring the Conservation of Cultural Heritage Objects’, Analytica Chimica Acta 706 (2011), 229–237. Quandt, A., L. Stiber and N. Stanley Price, Conservation of the Collection of Dead Sea Scrolls in the Rockefeller Museum, Jerusalem Getty Report, A Report to the Israel Antiquities Authority (The Getty Conservation Institute, 1993), unpublished. Tanner, S., and G. Bearman, Digitizing the Dead Sea Scrolls, Report on the Pilot Imaging (2008), unpublished. Tov, E., Revised Lists of the Texts from the Judaean Desert (Leiden: Brill, 2010).

dead sea scrolls inside digital humanities. a sample1 David Hamidović Today, it is still difficult to give a precise number of scrolls from the Qumran caves, a part of the famous Dead Sea Scrolls, because of the fragmentary state of the manuscripts. We estimate a number of around 930 manuscripts, i.e. 930 different scrolls or rather 930 remains of scrolls. There are not 930 different documents for we count, for example, between 23 and 25 uncompleted copies of Genesis or between 10 and 13 copies of the Rule of Community. The different counts come from the difficulty in identifying a manuscript from one or two fragments. As the fragments are small, it is difficult to know if the manuscript is a copy of a well-known document or a copy of another text that quotes this document. It is the case for three manuscripts of the Rule of Community. Besides the difficulty in identifying a non-biblical manuscript from one or two tiny fragments, the scholar has also the difficulty of collating the fragments. Unlike a jigsaw, all the fragments are not preserved. It is very difficult to place the fragments in respect to the others and thus to propose a setting in columns close to the setting of the original scroll. The use of high-resolution pictures allows us sometimes to arrange leather fragments on the basis of the epidermis, especially the colour of the skin. For papyrus, it is more complex because intermingled fine vertical and horizontal strips weave the papyrus. The fragments of papyrus are broken. They are very fragile and it is often impossible to manipulate them. The remains of ink are placed on vertical and horizontal bands extracted from the weaving of the papyrus. Therefore, it is very difficult to set up the reed strips in respect to one another. To my knowledge, digital pictures are not helpful at the moment because we cannot weave a virtual papyrus. A digital edition may be a major advance in another field: the critical reconstruction of texts. For the biblical scrolls of Qumran, the use of the Masoretic Text, the Samaritan Text or the retroversion into Hebrew of

1 Many sections come from my article, ‘In Quest of the Lost Text. From Electronic Edition to Digital Edition of the Dead Sea Scrolls’, in Lire demain. Des manuscrits antiques à l’ère digitale, edited by Clivaz, C. et al. (Lausanne: PPUR, 2012), 153–166. They are taken with permission of PPUR.

22

david hamidović

the Septuagint often allows us to restitute the missing text with a high degree of confidence. However, we can note that several biblical texts are difficult to link to one of the three well-known versions.2 The digital edition may precisely be an evaluation of reconstruction of the lost text for the non-biblical manuscripts. This applies both to previously unknown texts, many of which were not available before they were discovered among the scrolls at Qumran, and also to text such as the book of Jubilees and 1 Enoch. We knew both documents before the discovery of Qumran manuscripts but only in classical Ethiopic (ge‘ez). These documents were originally written in Hebrew or Aramaic and translated first into Greek then into Ethiopic in the Christian Church of Abyssinia and also among the Falashas, the Jews of Ethiopia who preserved them because they recognize them as sacred. It seemed curious to try to reconstruct missing sections of texts that were not known from other sources. Therefore, many scholars reject the reconstructed text. But the restitutions are often necessary to understand the text of a tiny fragment and thus identify it. For example, the manuscript 4Q173a is known by only one fragment:3 [. . .]‫ יע[ברו מעל‬. . .]   1 [. . .] ‫ ב[ית מכשול‬. . .]  2 [. . .]·‫ קרנ[ות המזבח י‬. . .]  3 [. . . ‫ זה[ השער לאל צדיק]ים יבאו בו‬. . .]  4 [. . . ‫ נ[קובי שמות וחרות]ים‬. . .]  5 [. . .]‫[ל ליעקוב‬. . .]  6 1   [. . .] they [will br]ing iniquity[. . .] 2  [. . . hou]se of scandal [. . .] 3  [. . . horn]s of altar [. . .] 4  [. . . This is] the gate of God, [what] the righte[ous enter by it! . . .] 5  [. . . those who have been de]signated by names and engrave[d . . .] 6  [. . .] to Jacob[. . .]

The restitutions are bracketed. The longest restitution is on line 4. Without finding securely the lost text, we can think here that the added words have a good chance to be on the manuscript. Indeed, the preserved words correspond to the formulation of another known text: Psalm 118:20. Both of the first Hebrew words preserved in 4Q173a 4 signify ‘the gate of God’. But in Psalm 118:20, the theonym is the tetragrammaton for the God

2 Tov, Scribal Practices and Approaches Reflected in the Texts Found in the Judean Desert. 3 Hamidović, ‘Le retour au Temple de Jerusalem ? (4Q173a olim 4Q173 5)’.



dead sea scrolls inside digital humanities. a sample

23

of Israel. The word is el in the fragment; it is a common way to name the God of Israel in the Dead Sea Scrolls.4 The change of theonym has no effect on the hypothesis of a quotation of Psalm 118:20 at line 4. The letters read after the two first words are the first letters of the Hebrew word for ‘the righteous’. This word is also the word read in the Masoretic Text of Psalm 118:20. Thus we can propose a quotation of this verse on line 4 with high probability; therefore we restitute the text both before and after the words that may still be read. The next question is to establish if the entire fragment quotes Psalm 118. If that is the case, the fragment would be a part of a copy of the Psalm. The complete word read on line 3 may be another proof: ‘the altar’. We read this word in Psalm 118 but not in verse 19 as expected, only in verse 27. The hypothesis of a fragment of Psalm 118 cannot be maintained. But we have restituted the word ‘horns’ in plural because we read that both last letters are the feminine plural ending. The expression ‘horn of the altar’ is in Psalm 118:27 but the fragment cannot be a copy of Psalm 118. Therefore we can easily doubt the proposed restitution of the first word. Then the palaeographer uses a concordance to find other expressions with the word ‘altar’ preceded by another word with a feminine plural ending. We find the expression ‘horns of the altar’ in Exodus 8:15; 9:9; 16:18; 1 Kings 1:50.51; 2:28 and a similar expression in Greek in the book of Revelation 9:13. A lot of words ending with a feminine plural form and followed by the word ‘altar’ are known in the Old Testament. Besides, none of the quoted passages with the expression ‘horns of the altar’ is followed by a word beginning with the letter yod like the studied fragment. If we read the letter waw which is often difficult to distinguish from the yod in the scribal writing around our era, the passages of 1 Kings 1:50 and 2:28 may correspond because we read wayyigad, ‘he announced’, ‘he informed’, but in this Hebrew word, the following letter after the waw is a yod. This cannot correspond to the traces of the second letter of this word in 4Q173a. Thus the palaeographer must look for other solutions of restoration outside the Old Testament. We summarize. The sequence composed by a word with feminine plural ending, then the word ‘altar’, then a word beginning by a waw or a yod then a second letter which cannot be a yod must be checked in the Old Testament with the concordance function. The request fails. It is the same conclusion in the Dead Sea Scrolls. Thus we must use specific concordances in paper format for other documents of Jewish apocryphal Literature or online concordances for the 4 Hamidović, Les traditions du jubilé à Qumran, 285–296.

24

david hamidović

Mishna or the Talmuds. All these requests fail. Thus the palaeographer is blocked; he cannot reasonably propose a reconstruction of line 3 in 4Q173a. We can finish the overview of the palaeographer’s work with line 5 on the same fragment. This time, there is no quotation of Psalm 118. The decipherment of the preserved word is a word in plural to say ‘the names’. The letters read before and after the word allow us to reconstruct literally ‘those who have been designated of names and graved’. Instead of the letter nun for the first word, we could restore the letter ayin to signify ‘the disappointed of’ something. The expression is in Jeremiah 17:9. We prefer to restore a nun because the expression ‘those who have been designated’ following by the word ‘names’ in plural is a locution attested in Numbers 1:17; 1 Chronicles 12:31; 16:41; 2 Chronicles 28:15; 31:19; Ezra 8:20. In context, the expression designates probably a group that benefits from the divine favour. With this meaning, we read ‘those who are designated by names’ in 1 Chronicles 16:41 for a group which praises God. According to this likely context in line 5, we have also restored an ending in masculine plural for the last word on the line. The word ‘engraved’ in plural corresponds to the plural ending of ‘those who have been designated’. The word may be appropriated to the deciphered context because we may understand that people designated by names have their named engraved as said in Exodus 32:16. For line 5 of 4Q173a, even if the read context has no direct link with the sentences of the Hebrew Bible, the read context as determined by the first words read has informed the reconstruction of the lost text. From this quick look at manuscript 4Q173a, we can observe that the degree of hypothesis is not the same for the three studied lines. For line 4, the restoration presents a high degree of probability because the read words quote Psalm 118:20 (with a change of theonym but the change is explained by the scribal practices in the scrolls of Qumran). For line 5, the context of the first read words orientates us towards precise expressions. These expressions, found independently in the Hebrew Bible, have given us a meaning: the designation of a peculiar group by God. The coherence of read expressions have allowed to establish this meaning and thus can propose restorations a minima: the first letter of the first word according to a combination with the following word and a masculine plural ending for the third word to make the word compatible with the same ending as the first word. The existence of parallels with the Hebrew Bible and the small proportion of text that is reconstructed also allow us to regard the reconstruction of line 5 with a high degree of confidence. But for line 3, the huge number of possible ways in which we might reconstruct the text



dead sea scrolls inside digital humanities. a sample

25

does not allow us to find one reconstruction that is preferable to other possible reconstructions. The probability of reading the word ‘horns’ is very low because only the feminine plural ending is preserved. The electronic edition of Qumran manuscripts allows us to use the picture with the function zoom to decipher the traces of letters on the edges of fragment. But we fail for the last trace on line 3. Too many letters may be attributed to the trace. We have also used the concordance function but it does not give a general view of early and medieval Jewish literature. Therefore, the possibility may not be excluded that an expression exists in another corpus and that we cannot attain it. Today, few palaeographers use online concordances for rabbinic Texts. The use of this literature remains very useful; therefore many Qumran manuscripts of legislative gender have been studied again these last years.5 The restitutions of texts also benefits from image editing software. The software Photoshop and Gimp are well-known. The website InscriptiFact6 of Bruce Zuckerman, Marylin Lundberg and Leta Hunt gives several pictures with high resolution but very few images of Qumran manuscripts. With the help of many filters, all these tools allow us sometimes to improve the pictures of manuscripts that were taken during the 1950s. The making of new high-resolution pictures for the project IAA Google will completely or partly render inoperative the use of filters. With the software, we may also use the function of clipping to distinguish the letters from the support. This function mainly uses the contrasts of colour to determine objects with the same colour. The holes of the support or the mushrooms gnawing the manuscript are often encompassed in the clipping. And the letters almost erased cannot be read. In the best case, if we succeed in clipping enough letters, we can establish a script chart of each letter written by the scribe. The way of writing the letters, the ductus, could be understood by the palaeographer. Thus he can propose a date of copy for the manuscript in comparing with other script charts. For the manuscripts of Qumran, the works of Frank Moore Cross are still the reference.7 The palaeographers often give a margin of error of between twenty and thirty years. I think it would be more secure to speak about a margin of half a century. If the writing of each letter evolves during the last centuries before our era, very few manuscripts of Qumran attest a parallel evolution of all the letters 5 Schiffman, The Halakha at Qumran. 6 www.inscriptifact.com, last accessed 8 May 2013. 7 Cross, ‘The Development of the Jewish Scripts’.

26

david hamidović

together. Often the manuscripts of Qumran possess letters characteristic of several palaeographical epochs. The palaeographer often deduces a date of the manuscript according to the adequacy between the majority of letters and an epoch, and the characteristic evolution of some letters. He modulates this date with a low date if all the letters belong to the preceding palaeographical epoch. In practice, the extraction of each clipping letter could be tedious and often imprecise. Moreover, the compatibility of images of extracted letters with the softwares of word processing is not easy to use because we must manage pictures of letters with word processing software. The majority of the palaeographers prefer to realize an ink drawing from a layer on photographic plate and then cutting and pasting. Thus the software of image editing are not easy to handle and the function of filters may become quickly obsolete. Therefore, I prefer to use another type of software: a software of computer aided drafting. I use the software Corel Draw. Besides the filters and clipping, the function insertion of special characters allows drawing on a layer above the picture. We can choose an originally integrated font character. This can be chosen and sized to the general size of read letters. Then we position an integrated letter on a letter on the picture. Of course, the format of the integrated letter does not precisely correspond to the letter written by the scribe. In clicking on an integrated letter, nodal points appear all around the letter and they form a Bezier curve whatever the chosen font. Just zoom on the text and move the nodal points on the traces of letters. Thus the complete letter has been drawn. After, we repeat the same process for each visible complete letter on the fragment. Sometimes, a scribe writes the same letter in two different ways, with two different ductus. This is not surprising because everybody writes a letter or a number in two different ways. There is, thus, a palaeographical observation before the drawing. We choose, of course, the complete letters to draw each time.8 Once all the letters have been drawn, we have a directory of scribal writing. Then we can try putting letters on traces of incomplete letters of the fragment in simply sliding the drawn letters on the traces of letters. Thus, we can propose restorations of incomplete letters from letters already drawn. From preserved traces on the fragment, we 8 A. Lange helps the scholars for the reconstruction of scrolls based on digitized fragments: Lange, Computer Aided Text Reconstruction and Transcription CATT Manual. He also proposes to scan individual letters in order to place them in the lacunae. But the scanned letters are not transparent; therefore they cannot be adjusted to the rest of letters on the edges of the preserved fragment.



dead sea scrolls inside digital humanities. a sample

27

can try the letters and thus establish a partially missed letter or possible letters to restore or reject impossible letters. The traces of a letter which are not in the pre-drawn letter allow us to doubt the restoration. Another interest of the computer-aided drawing is the restoration of missing words in the line. Without the manuscript’s margins, it is difficult to know the length of lines.9 As we have seen, the reconstructed texts are mainly based on parallels with readings found in other corpora or other manuscripts. The most probable restitution can be set in column. Each restitution can be set in the following of the line with the help of the directory created before. For example, we have this result for the fragment 2 of manuscript 4Q225. The fragment 2 gives an unknown version of the famous story of the sacrifice of Isaac. The restitutions based on the preserved letters are linked to other versions of the story in the book of Genesis, the book of Jubilees, the targums, Genesis Rabba or the pirqey of rabbi Eliezer. The words and preserved expressions show that the milieu of 4Q225 did not use one text source rather than another.10 Therefore, there are many possibilities for restitution. Of course, the preserved text orientates towards the phraseology of a particular version. The preserved context seems to privilege one version then another according to the passages. The interest of the computer aided drawing software is to add a new criterion to evaluate the probability of the restitution. We can check the setting up in columns with the length of restituted lines. Many cases are possible. In the writing of Qumran manuscripts, we observe, without any surprise because the support is precious, that the margins are often inferior to 2.5 cm and that the lines are completely written. Few manuscripts have a vacat. It means a space without writing. In restitution, if the content seems to present a break with the following passage, we can think of the possibility of a vacat but this becomes very doubtful if no vacat is preserved on the manuscript. We also observe that the more the scribe approaches the end of the line, the more the spaces between the words are reduced, the more the size of letters decreases to fit the text so that the scribe can write the rest of the text in the margin horizontally or vertically. When the reconstructed lines have an irregular length, we can doubt the validity of the proposed reconstructions. When the length is regular, it is an added proof towards the probability of the reconstruction. We can   9 Stegemann, ‘How to Connect Dead Sea Scroll Fragments’. H. Stegemann has set a method to count the length of the scroll from the preserved fragments. He deduces the size of margins and the setting in columns. 10 Hamidović, Les traditions du jubilé à Qumran, 280–281.

28

david hamidović

analyse the manuscript 4Q225, fragment 2 and column 1. Everyone knows the story: God commands Abraham to sacrifice his son Isaac at the top of the mountain. Here is the translation of the last lines of column 1: 10 [with Go]d and accused Abraham about Isaac. And [G]od says 11  [to Abra]ham: ‘Take your son, Isaac, [your] unique, [tha]t 12 [you, you lov]e and offer him to m[e] in holocaust on one of [hig]h mountains 13 [that I shall indicate] to you. He climb[ed and we]n[t a morning] from ‘the wells’ at the place 14 [which G]o[d had said] to him. [The third day,] Ab[raha]m rose

At line 13, we read that Abraham ‘climbed’ like in Genesis 22:3 but the verb is missing in the version of Jubilees, Jub 18:3. The author of the text follows rather the version of Genesis in this passage. If we look at the manuscript, we see that the end of the column has disappeared. To fill the small space in the middle of the line, the official editors of the manuscript, Josef Milik and James VanderKam, have proposed to restitute ‘to the morning’, the Hebrew word boqer with the prefixed preposition beth. Their proposal is based on Genesis 22:3 and Jubilees 18:3 which both contain ‘to the morning’. Both versions give this detail of the story’s chronology. But it is not possible to fit the proposed reconstruction where the manuscript is lacunose because there is not enough space, even if we remove the space after the preceding word. Unlike the drawing on a layer, we really write with the script of the scribe. This distinction is important because drawings made by hand do not replicate the scribe’s regularity of the hand of the scribe who wrote the text. The palaeographer is often tempted to increase or decrease the size of letters and thus to change the space between two words to fit his reconstruction of the line. The chosen example in 4Q225 is on one letter, which does not affect the content of the passage, but in other manuscripts, we have seen a succession of different drawings with a perfect setting in column for all of them. But a closer look at some of these renditions using tracings of letters, ductus, sizes of letters and spaces between the words shows that they are not preserved in the fragment. In this case, the drawn reconstruction gives the impression of an argument to justify the proposal of the palaeographer. How does one explain different reconstructions so perfect in column setting? With the manual method, the palaeographers make a lot of choices in the writing of restitutions. We propose with computer-aided drawing to limit these empirical choices in using strictly the directory of letters already traced by the scribe. It means a respect of ductus, a respect of the size of letters and of course, the scripts. Therefore, the computer aided drawing allows



dead sea scrolls inside digital humanities. a sample

29

more precise restitutions and thus, less questionable. If a reconstruction remains a proposition, we add a supplementary proof towards the validity of the proposed reconstruction. Finally, another outlook is opening for the study of Qumran manuscripts. Since the first decipherments, the scholars have had the surprise of discovering a great variety of scribal hands. On the 930 manuscripts, we count almost as many different scribes11 and sometimes a same manuscript presents several scribal hands.12 Also the formal writing of most of the manuscripts suggests that the scribes are professionals of writing.13 Their activity is thus to copy many manuscripts. Such a large number of skilful scribes is very surprising in Antiquity. The few manuscripts copied by the same hand probably mean that a significant part of the Qumran library has disappeared or that these scribes had several activities of copy outside the community of Qumran. The number of scribes does not correspond to any other place where a scribal activity is attested. For example, the huge Greek corpus discovered at Oxyrhynchus in Egypt has been written by a limited number of scribes.14 With the help of the software Corel Draw, we may establish a directory of traced letters for each manuscript and thus realize a real database of digital signatures of Qumran scribes. I am convinced that we shall manage to establish that some manuscripts have been copied by the same scribes.15 We shall have a more precise overview of the number of scribes who have copied the Dead Sea Scrolls between the end of the third century BCE and the middle of the first century CE, during two centuries and half. We may find scribes specialized

11 The same scribe may have copied 1QS, 1QSa and 1QSb on the same scroll around 100–75 BC. He may also have copied a text of Samuel (4Q53) and he may probably have corrected 1QIsaa (Tigchelaar, ‘In Search of the Scribe of 1QS’). Another scribe may have copied the scrolls of Genesis: 4Q6 and 4Q7. The pesharim 4Q166, 4Q167 and 4Q168 present the same hand of scribe. Also two copies of 1 Enoch, 4Q207 and 4Q214, are written by the same scribe. On the same manuscript, 4Q259 and 4Q319 may have been written by the same scribe. The manuscripts 1QHa and 4Q266 may present the same hand. Another scribe may have copied 4Q280, 5Q13 and probably 4Q390 and 5Q11. A same scribe may have copied 4Q388, 4Q388a, 4Q392 and 4Q393. Another scribe may have copied 1QpHab and 11Q20. A same hand is identifiable in 11Q12 and 11Q21. But it remains difficult to identify a scribal hand. Therefore, other manuscripts may be bought closer each other. Also the alleged comparisons may be criticized. 12 For example, 1QIsaa; 1QpHab XII 13; 1QHa XIX–XX (XI–XII); 1Q VII 1; 4Q176; 4Q216 12; 4Q393 12 iii; 4Q448; 11Q19 IV; VI–LXVII. 13 Tov, Scribal Practices and Approaches, 14. 14 Turner, ‘Scribes and Scholars of Oxyrhynchus’, 141–146. 15 Yardeni claims that more than 50 Qumran manuscripts were copied by the same scribe: Yardeni, ‘A Note on a Qumran scribe’.

30

david hamidović

in the writing of theonyms, the sacred name of God. We find this idea in the rabbinic literature later to prevent a wrong writing of divine name.16 It seems to be the case in the complete scroll of Isaiah because one different hand of scribe seems to have written some theonyms.17 By crossing the digital signature of scribes with the palaeographical dates of manuscripts, we may define the community of Qumran scribes, its variety and its evolution. This community of scribes may be a part of the members of the movement known today as the movement of Essene Jews. References Cross, Frank Moore Jr., ‘The Development of the Jewish Scripts’, in Leaves From An Epigrapher’s Notebook (Winona Lake: Eisenbrauns, 2003), 3–43. Gibert Pierre, Comment la Bible fut écrite (Paris: Fayard, 2011). Hamidović, David, Les traditions du jubilé à Qumrân (Paris: Geuthner, 2007). ——, ‘Le retour au Temple de Jerusalem ? (4Q173a olim 4Q173 5)’, Revue de Qumrân 94 (2009), 283–286. ——, ‘Le texte biblique et son interprétation. Une archéologie de l’interdépendance’ (à paraître). Lange, Armin, Computer Aided Text Reconstruction and Transcription CATT Manual (Tübingen: Mohr Siebeck, 1993). Laplanche, François, La crise de l’origine. La science catholique des Evangiles et l’histoire au XX e siècle (Paris: Albin Michel, 2006). Meier, John Paul, Un certain Juif Jésus. Les données de l’histoire, vol. I (Paris: éditions du Cerf, 2009). Schiffman, Lawrence H., The Halakha at Qumran (Leiden: Brill, 1975). Stegemann, Hartmut, ‘How to Connect Dead Sea Scroll Fragments’, Understanding the Dead Sea Scrolls, ed. by Hershel Shanks (New York: Vintage Books, 1992), 245–255. Tigchelaar, Eibert J.C., ‘In Search of the Scribe of 1QS’, Emanuel: Studies in Hebrew Bible, Septuagint and Dead Sea Scrolls in Honor of Emanuel Tov, ed. by S.M. Paul et al. (Leiden: Brill, 2003), 439–452. Tov, Emanuel, Scribal Practices and Approaches Reflected in the Texts Found in the Judean Desert (Leiden: Brill, 2004). Turner, Eric G., ‘Scribes and Scholars of Oxyrhynchus’, Ackten Des VIII Internationalen Kongressen für Papyrologie (Wien: R.M. Rohrer, 1956), 141–146. Ulrich, Eugene and Peter W. Flint, Qumran Cave 1.II The Isaiah Scrolls. Part 2, DJD 32 (Oxford: Clarendon, 2010). De Vaux, Roland, Archaeology and the Dead Sea Scrolls (London: British Academy, 1973). Yardeni, Ada, ‘A Note on a Qumran scribe’, New Seals and Inscriptions: Hebrew Idumean, and Cuneiform, ed. by M. Lubetski (Sheffield: Sheffield Phoenix Press, 2007), 287–298.

16 Cf. b. Menahot 29b–32b; y. Megillah 71b–72a; b. Shabbat 103a–105a; b. Baba Bathra 13b–14b. 17 Cf. 1QIsaa XXXIII 7 (Is 40:7); XXXV 15 (Is 42:6). See Ulrich and Flint, Qumran Cave 1.II The Isaiah Scrolls. Part 2, 63–65.

The electronic scriptorium: markup for new testament manuscripts H.A.G. Houghton1 1. A History of Transcriptions Every act of copying is a transcription, even though the end product may differ from its source. When a New Testament manuscript was copied by hand, the exemplar would often have been marked up in advance by an editor; corrections and notes added to the exemplar during its use may also have been incorporated into the new copy by the scribe.2 In the early days of printing, most textual editions were effectively transcriptions of single manuscripts, again often marked up for the guidance of the compositors, as can still be seen in certain codices.3 The new printing technology meant that the resultant text was reproduced identically on each occasion. One of the results of this was to establish a standard for subsequent scholarship. The consistency of the printed text meant that it could be used as the basis for a collation: instead of providing the full text of multiple manuscripts of the same work, textual variation could be far more economically expressed as a list of differences from a printed version. This can be seen in the alternative readings reproduced in the margins of Stephanus’ Textus Receptus and other editions of the Bible.4 While editors may have adjusted the main text to create a composite form based on the most compelling readings from all witnesses sampled, the format of subsequent editions remained relatively stable. Manuscripts of particular importance occasionally merited the print publication of a full 1 The author is Principal Investigator (UK) of the Workspace for Collaborative Editing, a project funded by the Arts and Humanities Research Council and the Deutsche Forschungsgemeinschaft. He would like to thank his colleagues Catherine Smith, David Parker, Troy Griffitts and Zeth Green for their comments on an earlier version of this paper and their contribution towards the development of the XML schema itself. 2 For an example of this, see Schmid, ‘Scribes and Variants—Sociology and Typology’. 3 The marks are still evident in Oxford, Corpus Christi College MS 26 and MS 27, used for sixteenth-century editions of Basil: see Wilson, A Descriptive Catalogue of the Greek Manuscripts of Corpus Christi College. 4 Estienne, Novum Iesu Christi D(omini) N(ostri) Testamentum ex Bibliotheca Regia (third edition). A set of IGNTP guidelines on how to make a collation of a manuscript is explained in Parker, An Introduction to the New Testament Manuscripts and their Texts, 95–100.

32

h.a.g. houghton

transcription, some of which reproduce the page layout and even the letter forms of the original.5 Other technological developments led to the production of facsimile editions, usually at even greater cost.6 Each critical edition of the New Testament text had to be created afresh, based on a collation of collations, with new typesetting on each occasion (and the potential introduction of new errors). Additional evidence and corrections to previous editions could be incorporated in each printing but, as the number of known witnesses increased, the majority of the information was usually taken on trust from one edition to the next. The scale of such an endeavour meant that no comprehensive edition was produced in the twentieth century to succeed the proliferation of New Testament texts from the second half of the nineteenth century.7 The advent of computers offered not just the possibility of storing and retrieving the huge amount of data required but a new paradigm for editing and publication. Collaboration on the Editio Critica Maior may be traced back to the adoption by both the INTF and IGNTP of the Collate software developed by Peter Robinson in the early 1990s.8 The core function of this program is the automatic generation of an apparatus of readings from separate files containing full-text transcriptions of individual witness. This removes the scope for human error in the mechanical task of collating multiple sources. At the same time, it re-focusses attention on the individual documents themselves. A by-product of the gathering of data for the new edition is that the primary sources can be presented in full. The inclusion of information about page layout, such as the extent of each line

5 The classic example is Tischendorf ’s pseudo-facsimile of Codex Sinaiticus (von Tischendorf, Bibliorum Codex Sinaiticus Petropolitanus) although mention should also be made of editions of insular biblical manuscripts which reproduce characteristic scribal features, such as Gwynn, Liber Ardmachanus: The Book of Armagh, and Hoskier, A New and Complete Edition of the Irish Latin Gospel Codex Usser. 2 or r2 otherwise known as ‘The Garland of Howth’. 6 The earliest biblical example of which I am aware is Rettig’s 1836 edition of the St Gall bilingual gospels using lithographic technology: Rettig, Antiquissimus Quatuor Evangeliorum Canonicorum Codex Sangallensis Graeco-Latinus Interlinearis Nunquam Adhuc Collatus. Photographic facsimiles became common at the turn of the twentieth century, e.g. Thompson, Facsimile of the Codex Alexandrinus (4 vols.), Lake, Codex Sinaiticus Petropolitanus: The New Testament, the Epistle of Barnabas and the Shepherd of Hermas. 7 For more on this historical period, see Parker, Textual Scholarship and the Making of the New Testament, 106–122. 8 Robinson, Collate: Interactive Collation of Large Textual Traditions, Version 2 (Computer Program). On the use of Collate see, for the INTF, Wachtel, ‘Editing the Greek New Testament on the Threshold of the Twenty-first Century’ and, for the IGNTP, Parker, ‘Electronic Religious Texts: The Gospel of John’.



the electronic scriptorium

33

or the size of individual letters, makes it possible to generate a facsimile of each witness in digital typography. Additionally, the production of the collation from electronic files means that the same transcriptions can be re-used in multiple editions, rather than starting afresh on each occasion. In essence, the task of the first digital editors of the New Testament is initially to produce a diplomatic edition of each document: the editing of the work itself is a later stage during which the disparate data from individual witnesses is brought together into a standardised form.9 The encoding of the transcriptions for Collate followed and extended the conventions developed by Robinson.10 In addition to changing an electronic base file of the work to correspond to the text found in a particular document, transcribers were able to include information about the physical and textual characteristics using a system of markup. This was divided into four categories: 1. Block markers, indicated by angled brackets: < > These are present in the base file and indicate the standard division of the work (into book, chapter and verse). They are used for orientation within the text and as the identifiers for collation. 2. Location markers, indicated by pipes: | | These describe the physical layout of the text in each manuscript, dividing it into pages, columns and lines. They are not taken into account during collation. 3. Tags, indicated by square brackets: [ ] These designate a portion of text as distinctive in some way, for example identifying section numbers written in the manuscript, capital letters or abbreviations. They may be used to mark sections which are difficult to read or have been reconstructed. They also indicate the readings of different hands, where the original text has been altered by the copyist or subsequent users. An opening tag is placed at the beginning of the relevant portion of text, and a corresponding closing tag at the end.   9 For a definition of the key terms ‘document’, ‘work’ and ‘text’ and their application to the New Testament tradition, see Parker, Textual Scholarship, 10–14 and 29. On the shift in the task of the editors, see Parker, ‘Through a Screen Darkly: Digital Texts and the New Testament’, 404. 10 A full description of the capability of the software and the conventions is provided in Robinson, Collate 2. The markup itself was based on the Oxford Concordance program produced several years earlier by Susan Hockey and Ian Marriott. A worked example of Collate encoding is given by Parker, ‘Through a Screen Darkly’, 405–407.

34

h.a.g. houghton

4. Comments, indicated by braces/curly brackets: { } These enable transcribers to make observations which are not treated as part of the text for collation. These may include glosses or page numbers, or simply provide a commentary on certain readings or the state of the manuscript. The Collate software ran on a Macintosh computer using an operating system before the introduction of the Unix-based OS X in 1999. It also relied on the ASCII character set, with only 95 printing characters. Transcriptions of Greek manuscripts therefore required a font which substituted Greek letters for the standard Roman characters. Symbol Greek was the betacode font adopted for this purpose. However, the character mapping of this font led to the substitution of the brackets indicating markup elements with other, often unusual, characters. The result was a file which, although it benefited from an economical system of markup that allowed the transcriber to focus on the biblical text, often appeared impenetrable to the human eye. Furthermore, the use of a betacode font for both text and markup and the creation of tags beyond those originally specified by the software means that, while files could be shared between specialists working in the two collaborating institutions, the encoding was not transparent for external users. The publication of the transcriptions in an electronic edition relied on the conversion of the plain text files into Standard Generalized Markup Language (SGML) with each Greek letter converted into a separate unique entity in order to enable the presentation of both Roman and Greek characters on the same page. The publication system adopted by both projects was Anastasia,11 also developed by Peter Robinson. This ran as a server which converted the SGML data into HyperText Markup Language (HTML) for viewing in a web browser. The tags were converted into HTML elements and then rendered according to a specified scheme (e.g. red characters in square brackets for supplied text, or blue characters for first-hand readings and green characters for corrections): hyperlinks were used to navigate around the edition, and images of each page could also be incorporated. The transcriptions could be viewed in two modes: ‘Page Layout’, reproducing the organisation of text on each page, and ‘Chapter

11 Robinson, Anastasia: Analytical System Tools and SGML/XML Integration Applications, Version 2.0 (Computer Program). The documentation is available online at http:// sd-editions.com/anastasia/index.html, last accessed 8 May 2013.



the electronic scriptorium

35

View’ with the text arranged by verse. The principal New Testament editions using this software were the New Testament Transcripts prototype (NT Transcripts), published by INTF in 2003, and the three online editions of different traditions of the Gospel according to John published by the IGNTP in 2007.12 Even as these editions were being produced, two significant developments paved the way for future innovations. The first was the widespread adoption of the Unicode character encoding, used by the World Wide Web Consortium (W3C) with the introduction of HTML 4.0 from 1997 and installed as standard in Macintosh OS X. With over a million possible character encodings, this meant that Greek letters (and, from 2005, Greek numerals and other special characters) could be uniquely identified regardless of font, and thus appear in the same file with symbols from other languages without the need for further differentiation. The second was the Text Encoding Initiative (TEI), established as a consortium in 2000, which employed XML as the standard encoding format with effect from its P4 set of guidelines published in 2001.13 However, while its collation engine remained the key to providing the data for editions of the New Testament, Collate did not support Unicode, could not collate files in XML (although it could convert transcriptions into this format for publication in Anastasia) and did not work natively on Mac OS X. The challenge of creating a successor, called CollateX, was undertaken by the Interedition COST action funded by the European Science Foundation from 2008 to 2012.14 In anticipation of this, the IGNTP and INTF started to make transcriptions in Unicode from 2009, using the standard Collate markup.15 The first example of the re-use of electronic New Testament transcriptions was the Digital Codex Sinaiticus, an online edition of this fourthcentury Greek Bible combining images from all four holding institutions,

12 All of these were available online at the date of writing, at http://nttranscripts.unimuenster.de/ and http://www.iohannes.com/ (comprising editions of the Greek majuscule manuscripts of John, the Vetus Latina manuscripts of John and the Byzantine text of John). Anastasia was also used for the publications of the Canterbury Tales Project (http://www .canterburytalesproject.org/) and other online and CD-ROM editions published by Scholarly Digital Editions (http://sd-editions.com/). 13 Burnard et al., TEI P4: Guidelines for Electronic Text Encoding and Interchange. 14 See http://www.interedition.eu/. 15 The guidelines for transcribers were deposited in the University of Birmingham Institutional Research Archive at http://epapers.bham.ac.uk/751/; the latest version (5) was published in October 2012 at http://epapers.bham.ac.uk/1676/.

36

h.a.g. houghton

a complete transcription of the text and a translation.16 The files of this manuscript prepared by the INTF for the NT Transcripts edition were made available to the project and subsequently enhanced by the addition of information such as Eusebian canons and marginal glosses in order to match the conventions adopted for the rest of the manuscript. They were then converted to the markup developed for the electronic edition, a customised version of XML, and published online in 2009 using a bespoke system created for the project.17 In the autumn of 2010, work began on a new online environment to integrate all the tools required by the partners involved in producing the Editio Critica Maior. Scheduled for completion in late 2013, the Workspace for Collaborative Editing aims to connect each stage of the editorial process.18 These comprise the initial transcription of manuscripts, the automated collation of witnesses and production of an initial critical apparatus and database, the addition of patristic and versional evidence, the establishment of the initial text (Ausgangstext) using the CoherenceBased Genealogical Method and the eventual publication of the Editio Critica Maior in print and electronic form. Early on in the project, the decision was made to use TEI-compatible XML as the format for encoding and storing transcriptions. A subset of the latest version of the TEI Guidelines, P5, was therefore developed by the project in order to meet the requirements for work towards the Editio Critica Maior.19 These comprised both the conversion of earlier Collate files to the necessary standard for incorporation into the new environment and the creation of new transcriptions. Despite the many advantages of XML for standardisation and storage, the verbose character of the markup makes it very inefficient for transcribers to work directly in this encoding. Instead, one of the components of the Workspace for Collaborative Editing is an online Transcription Editor 16 http://www.codexsinaiticus.org/. The history of this project is related in Parker, Codex Sinaiticus. The Story of the World’s Oldest Bible. 17 Parker, Codex Sinaiticus, includes a sample of the plain text and XML transcriptions on pp. 178–9. The Old Testament and other writings were transcribed by the project team, following the same methodology. The process of transcription is also described in Houghton, ‘The Electronic Transcription of Codex Sinaiticus’. There is documentation on the website: the full transcription can be downloaded from http://codexsinaiticus .org/en/project/transcription_download.aspx. 18 See further Parker, Textual Scholarship, 113–119 and 138–141. 19 The full version of the P5 Guidelines is Burnard and Bauman, TEI P5: Guidelines for Electronic Text Encoding and Interchange. These are available online at http://www.tei-c .org/release/doc/tei-p5-doc/en/html/index-toc.html.



the electronic scriptorium

37

used within a web-browser.20 The interface mimics the display of the XML of transcriptions already published online by the projects, providing a ‘what you see is what you get’ environment. As with the earlier electronic transcriptions, users do not start from scratch, but are able to choose from a selection of base texts with the standard divisions of the work already in place. The transcribers change this to correspond to the reading of the manuscript. The ‘hidden’ parts of the markup are supplied in dialogue boxes which appear when an additional element is selected, such as a line break or a correction. Many of these can be added through shortcut keys. One particularly notable feature is the ‘editor within an editor’ dialogue box which enables transcribers to edit the markup of the text supplied by a corrector, specifying unclear or supplied characters or other types of formatting. The transcription tool is not intended to stand alone but to be integrated into a suite of tools such as the Workspace for Collaborative Editing or the New Testament Virtual Manuscript Room (NT.VMR),21 drawing on other sources of information such as databases with bibliographical details for New Testament manuscripts and indexes of biblical content for digital images of each page. In the second half of this paper, the XML encoding adopted for the Workspace for Collaborative Editing will be described, along with observations on how it has developed over the course of the project and some problems which have been encountered. There is always a balance to be struck regarding the amount of information included in a transcription, which reflects the potential of these initial electronic files for multiple uses. For the purpose of creating a critical edition of a work, details of formatting and layout are superfluous; a ‘digital facsimile’, however, tries to match the document as closely as possible. The practice of the IGNTP has been to include some information which goes beyond the purely textual, in order both to enable the transcription to form the basis of an electronic facsimile with explanatory information for non-specialist users, and also to have the possibility of more complex searches based on specific phenomena, such as abbreviations or spelling conventions. Of course, electronic transcriptions can always be altered and improved, and it may be that other researchers with, for example, a particular interest in 20 This was developed by Thomas Burch, Martin Sievers and Gan Yu of the Kompetenzzentrum für elektronische Erschließungs- und Publikationsverfahren in den Geisteswissenschaften at the University of Trier (KoZe). 21  This is a community portal developed by the INTF for work on New Testament manuscripts: the current version, NT.VMR 2.0, is hosted at http://ntvmr.uni-muenster.de/.

38

h.a.g. houghton

punctuation or textual divisions, will enhance files in this way. Similarly, although the Workspace for Collaborative Editing itself was specifically commissioned for collaborative work on New Testament manuscripts leading to the production of a critical edition, it is hoped that, like many of the major developments in textual scholarship which originated in work on the New Testament, it may also be applied to other textual traditions. At each stage, therefore, this potential expansion of material has been kept in mind so that, with minimal adjustment, it should be possible to use the same framework in a different context. 2. An XML Encoding for Manuscript Transcriptions The XML encoding developed for the IGNTP transcriptions of New Testament manuscripts is a subset of the TEI P5 Guidelines.22 Compatibility with the TEI means that generic documentation for the markup is provided online by the TEI. Furthermore, tools developed for use with the entire range of the P5 Guidelines, such as those for parsing, visualisation and analysis may be applied directly to these transcriptions. Given the variety of potential combinations of XML elements, however, the selection of a smaller group was necessary in order to ensure a manageable standard format for developing tools within the Workspace for Collaborative Editing, displaying the transcriptions with a standard XSLT template and interchange with other encodings, such as JSON for the CollateX engine. The choice of elements was informed by the encodings used for the Digital Codex Sinaiticus and New Testament transcriptions published using Anastasia, along with the TEI P5 Guidelines.23 The initial scheme was issued on 1st December 2010; a revised version was adopted by the IGNTP in early 2011 and published online in the University of Birmingham Institutional Repository (UBIRA).24 Ongoing work has led to alterations, periodically released as revised versions: the guidelines described here are Version 1.4, published in July 2013, which may be downloaded from http:// epapers.bham.ac.uk/1727/. A schematic overview is also included in the Appendix below. 22 Burnard and Bauman, TEI P5: Guidelines for Electronic Text Encoding and Interchange. For a description of how to make an electronic transcription of a New Testament manuscript, see Parker, An Introduction, 100–106. 23 Unfortunately, it was only after the first schema had already been created that I encountered Timothy Finney, ‘Manuscript Markup’. 24 Version 1.1: http://epapers.bham.ac.uk/738/.



the electronic scriptorium

39

Unlike the Collate markup described above, all XML markup is enclosed within angled brackets: < >. The first group of letters within the brackets identify the element. This may be further qualified by attributes, whose value is expressed between quotation marks. So, for example, indicates a word element (w) with two attributes: a numerical identifier (n) of 4 and a language value (xml:lang) of English (en). Elements may either be empty, providing punctual information such as a single space or a line break, or may enclose a portion of text. In the latter case, the element must be closed after the last character to which it applies, using a forward-slash within angled brackets, such as ; for empty elements, the closing slash appears within the same brackets as the rest of the element, as in the element for page break. In certain cases, multiple elements may apply simultaneously to the same portion of text: this is known as ‘nesting’. The overall principle for the transcription of manuscripts is that what appears on the page is transcribed as text and everything else is indicated by markup. This is facilitated by editing programs which display text and markup in different colours, sometimes also differentiating between elements and attributes.25 2.1. Header The TEI header, , is an obligatory part of an XML transcription, providing details about both the electronic file and also the manuscript transcribed.26 The amount of information provided may vary from project to project: the more that can be entered from a database or series of drop-down menus the better, in order to ensure standardisation of identifiers. Most of the details of Greek New Testament manuscripts are already stored in the electronic version of the Gregory–Aland Kurzgefasste Liste available in the NT.VMR,27 while details of the originating project and encoding procedures are common to multiple transcriptions. In order to publish individual transcriptions as self-contained files, however, some information may be repeated as part of the header. The different ‘type’ attributes of the element reflect the distinctions expressed by the Documents, Works, Texts project in developing an 25 For a further introduction to XML, focussing on the hierarchy of elements, see Finney, ‘Manuscript Markup’, 276–279. 26 The header violates the principle that the text field only contains words from the manuscript: the whole element is an editorial construct, as are editorial notes (discussed below). 27 http://ntvmr.uni-muenster.de/liste.

40

h.a.g. houghton

ontology environment for the identification of electronic material relating to manuscripts.28 Document refers to the manuscript. Although the main description is free-text (e.g. ‘Codex Alexandrinus’), two attributes permit the precise identification of the document according to standard systems: the ‘key’ attribute has the five-digit number used in the electronic Liste, which also underlies the IGNTP and INTF file naming scheme, while the ‘n’ attribute gives the siglum of this witness in the Editio Critica Maior.29 Work denotes the customary abstract identification of an authorial creation (e.g. ‘The Gospel according to John’, abbreviated to ‘John’ in references). The use of collection to refer to the whole New Testament puts each work in its wider context although, of course, individual manuscripts vary in the selection of books they contain. Details of creators, funders, editions and dates are included in order to provide recognition and keep track of the publication of individual transcriptions. As noted above, the IGNTP has a policy of depositing completed files in an Institutional Repository (currently UBIRA). These are freely available, as has subsequently become a condition of research funded by UK Higher Education Funding Councils. Furthermore, they are issued with a Creative Commons licence which permits their re-use.30 Just as photographic facsimiles are of interest to a far wider range of scholarly disciplines than textual editing alone, electronic transcriptions offer the potential for fresh investigation of linguistic and scribal phenomena as well as other types of data analysis. Bibliographical information about the manuscript can be included within the element, incorporating parts of the TEI P5 Guidelines module on Manuscript Description (e.g. , , , ). Although much of this information is already present in the Liste, its inclusion in individual transcriptions assists in opening them up to discovery by search engines. The use of elements permits the identification of the manuscript in a variety of different catalogues or editions. The inclusion of references to the Leuven Database of Ancient Books is part of a reciprocal arrangement for the provision of links to these transcriptions in their catalogue.31

28 This project, funded by the UK Joint Information Services Committee (JISC) ran from 2009–10; as noted above, the differences are described in Parker, Textual Scholarship and the Making of the New Testament, 29. 29 This naming scheme is explained in Parker, An Introduction, 105–106. 30 On Creative Commons licences, see http://creativecommons.org/. 31  http://www.trismegistos.org/ldab/.



the electronic scriptorium

41

For the purpose of Collate, details of transcription practice and the history of each file were recorded in an initial status note. The former is incorporated into the section: the element is a free-text field where information on the treatment of features such as punctuation, capitalisation or rubrication may be specified. The section should also include an empty element, , specifying the procedure adopted for handling different readings.32 The file history may be detailed in the section where details of each alteration to the file are listed. A header which employs the full scope of the TEI P5 Guidelines would provide an exhaustive amount of information, from bibliographical references for secondary literature to the enumeration and identification of members of the project team and their interventions in the file. However, a compromise has been made in order to provide enough information to enable each file to stand by itself while not distracting from the task of transcription. For example, it would be good practice to include in the header a list of all scribal hands which worked on the manuscript, either as copyists or correctors. However, the header information is normally added at the start of a transcription, when the nature and composition of the copying team may be unknown; furthermore the identification of scribal hands can require specialist palaeographical expertise. As noted above, one of the benefits of an electronic transcription is the potential for its enhancement by subsequent users. So long as the header provides enough information in a valid form for the identification and re-use of the transcription by external projects, then it will have served its purpose. 2.2. Divisions of the Work and Document Scholars of the New Testament benefit from a commonly-accepted system of divisions into book, chapter and verse brought to completion by Stephanus in the sixteenth century.33 It is worth remembering that this scheme of chapters and verses is not present in earlier manuscripts, which preserve evidence for a variety of series. These latter divisions may be 32 See further 2.5. below on corrections. 33 It may be noted in passing that there is not always complete agreement between different versions or editions: for variations in the nomenclature of biblical books, see the chapter by Laurence Mellerin in the present volume; differences in the versification of John between the Nestle–Aland Greek text and the Stuttgart Vulgate are listed in Burton et al., Vetus Latina 19. Iohannes, 6. The IGNTP and INTF always follow the versification of the Nestle–Aland edition.

42

h.a.g. houghton

recorded as paratextual elements (see 2.3 below), but the modern system is used for ease of reference and to connect transcriptions with each other and the critical apparatus. In XML, the longer units of ‘book’ and ‘chapter’ are treated as elements; the use of (‘anonymous block’) for biblical verses is exemplified in §16.3 of the TEI P5 Guidelines. In order to locate each verse, a concatenation of ‘n’ attributes is created: the biblical book is identified by a two-digit code preceded by B (e.g. ‘B04’ for John), the chapter by this and the chapter number prefaced by K (e.g. ‘B04K5’ for John 5) and the verse in a similar manner (e.g. ‘B04K5V21’ for John 5:21). Although it would be possible to use an XPath query to search recursively for verse 21 within chapter 5 of book 04, in a plain text editor a single identifier at each level enables easy navigation. In an earlier version of the schema, this value was encoded as a unique attribute (‘xml:id’). However, while each biblical verse is unique, within a manuscript it may appear on more than one occasion (such as lectionaries with overlapping readings, bilingual codices, dittography of a longer passage or verses split over page breaks when a transcription is stored by page). For this reason, the ‘n’ attribute was used, which can be repeated as necessary. In bilingual manuscripts, the use of an ‘xml:lang’ attribute on each element permits the differentiation of each version.34 Within each element, each word is treated as a element and punctuation tokens as elements. The elements are numbered with an ‘n’ attribute: this is inherited from Anastasia, where it formed part of the information connecting the critical apparatus to the transcription. This attribute may also be used as a link with the presentation of the Editio Critica Maior, where each word in the base text is allotted an even number, although this would require the numbers to be added after the collation stage. This attribute could also be used to map translations back to a Greek text, as well as link to external resources such as concordances. Other information may be nested within the elements, such as abbreviations () and formatting information (), discussed below. 34 The most common languages for New Testament bilinguals are Greek, Latin and Coptic. However, the XML indication of Greek (xml:lang=“el”) refers to the modern rather than the ancient language. The BCP47 standard adopted elsewhere in the TEI Guidelines indicates that a private use code, such as xml:lang=“el-x-koine”, would be the best way of indicating New Testament Greek. However, for simplicity, we have adopted the IANA codes of ‘grc’ for ancient Greek and ‘cop’ for Coptic (see further http://www.iana.org/assignments/language-subtag-registry). Of course, Coptic should be further subdivided into the different dialects but these are already indicated in the SMR sigla for Coptic manuscripts, so the use of ‘cop’ as a blanket language indicator may suffice.



the electronic scriptorium

43

The element may be used for other material, either within or outside the elements denoting ‘book’ and ‘chapter’. Preliminary material, such as the Letter to Carpianus or canon tables, may be allocated its own at the same level as a biblical book. Within each book, prefaces and kephalaia (lists of chapter titles) may precede the first chapter. Short titles at the beginning or end of the work are also treated in their own right, as and . (In Collate, these were assigned verse number 0 in imaginary chapters at the beginning and end of each book so that they could be collated.) A longer colophon may also be treated as another , subdivided into or

elements as desired. As already described, a single XML transcription can also be used to generate a textual facsimile of each physical page of the document if information is included on the disposition and formatting of the text. In the TEI P5 Guidelines this is done by a series of empty elements which indicate page, column and line breaks: and .35 All these have the ‘n’ attribute giving an editorial numeration. (These may or may not correspond to any numerals written on the page, which would be transcribed as elements.) As there is only one instance of each of these numbered elements, they are also given an ‘xml:id’ attribute which connects them to the document, using a concatenating hierarchy similar to that mentioned earlier for divisions of the work. Thus identifies page 246, column 2, line 3 of manuscript 044. In order to distinguish between different ways of numbering, the attribute ‘type’ is included on elements, with the value ‘page’ when each side has a separate number and ‘folio’ when a page is given a single number and the sides are identified as recto or verso. Because the papyrological recto and verso may not correspond to the assigned page numbers, the direction of the fibres may also be specified with the ‘subtype’ attribute. In the case of palimpsests, information about the different systems of page numeration can be included within the element. The attribute ‘facs’ can be used to give a URL for a digital image of the page. At the level of , indentation, ekthesis and justification are encoded using the ‘rend’ attribute. The downside of this approach to encoding layout, however, is that it leaves open a problem of multiple hierarchies when producing a digital

35 In addition, the element (‘gathering break’), to indicate a manuscript’s quire structure, was introduced into version 1.8.0 of the P5 Guidelines in late 2010.

44

h.a.g. houghton

facsimile. As its name implies, the Text Encoding Initiative is primarily concerned with texts (where layout is usually secondary) rather than documents (where the emphasis is on the physical object). Thus the TEI scheme makes it easy to extract a chapter or a verse (as a division of the work), but much more difficult to extract a single page of text, since a page break is a single empty element at the beginning of a page rather than an opening and closing element surrounding a portion of text. The only way of discovering where a page, line or column ends is to go to the next break element and assume that everything between the two is part of the first element. This leads to problems, such as when a number of pages (or lines) may be missing, or at the end of a document or page where there is no closing tag. Furthermore, because page divisions rarely coincide with textual divisions, when considering a single page in isolation the opening words are usually not identified (because they form part of a block which began on the previous page), while the closing block may remain unfinished—and is therefore invalid XML. The fact that breaks are encoded as empty elements also makes it difficult to use a standard transformation scenario to extract a single page of XML to display alongside a facsimile of the manuscript, although this is less of an issue when the transcription is stored in an XML database. In part, the problem results from the mismatch between the task of transcription, proceeding page by page, image by image, and the abstract divisions of the work. When working from individual digital images (particularly if, as in the NT.VMR, more than one transcriber may be responsible for different pages of a single manuscript) the hierarchy of layout takes precedence over the system of chapters and verses, especially if the base text is automatically supplied using indexing information already collected in a database. However, when the same data is fed into a collation engine, this relies for its reference on the customary sequence of the work. One work-round is to close and reopen and elements at the extremities of each page, using the ‘part’ attribute, to ensure that each page is self-contained. This has the unwelcome result of fragmenting the text: in order to collate it, the various parts of the have to be identified (using the ‘n’ attribute) and re-assembled. It also places an additional burden on the transcriber by making them responsible for ensuring that all the relevant elements are closed and then reopened at every page break, although in the online interface this can be partially automated. This is the solution adopted in these guidelines. Other alternatives have been suggested. The minimal level is represented by Finney’s use of the empty element to record each



the electronic scriptorium

45

division of the work in place of containers.36 This has the advantage of not privileging one hierarchy over another, but would require a second stage of processing to extract a particular section by page or biblical text. A maximal level is found in the XML developed by Peter Robinson for the Digital Codex Sinaiticus, which not only departs from the TEI P5 Guidelines in turning break elements into containers (meaning that text was contained between opening and closing page, column and line elements), but also by subdividing the breaks into pairs of ‘start’ and ‘end’ subtypes. This is indicated in the ‘id’ attribute, with the details of the corresponding other member included as a ‘corres’ attribute.37 The advantage of this is that marginal material can be transcribed at exactly the point at which it appears in the text: if it is in the top margin of the page, then it would be enclosed by the opening and closing elements of the for the start of the page; if it is at the end of a line, it would be added between the indications of the for the end of that line. As a check, an innovative element was also added inside the break elements, confirming that the text was located in the margin. The disadvantage is a substantial increase in the size and complexity of the file, much of it unnecessary: as marginal material only occurs within a small proportion of break elements, most of the co-ordinated ‘start’ and ‘end’ breaks are simply recorded as empty elements. Even though the output XML was generated automatically, it often had to be manipulated by hand. Furthermore, the departure from the TEI Guidelines (including the creation of a element) meant that it was considered unsuitable for the IGNTP transcriptions. 2.3. Paratextual Elements and Formatting Electronic transcriptions offer the possibility of comparing far more details found in manuscripts than simply the biblical text, so long as these are also included in the transcription. Indeed, the goal of an ‘electronic facsimile’ is to represent all the information on each page. Features such as page numbers, quire signatures, running titles and other elements of the mise-en-page can all be described as elements. The following explanation is given in §11.6 of the TEI P5 Guidelines: ‘Although the name derives from the term forme work, used in description of early printed

36 Finney, ‘Manuscript Markup’, 272, 279–280. 37 This procedure is sometimes known as ‘Trojan Markup’: see further DeRose, ‘Trojan Markup and Other XML Milestone-Tagging Techniques’, available at http://www.mulberry tech.com/overlap/index.html.

46

h.a.g. houghton

documents (the ‘forme’ being the block used to hold movable type), the fw element may be used for such features of any document, written or printed’.38 In addition to the three types mentioned above, lection titles and chapter titles are also treated as elements. When an element includes a value, especially if this is in Greek or Latin numerals, the equivalent value is supplied as an ‘n’ attribute. Certain paratextual elements are numerals: chapter numbers and the two parts of the Eusebian apparatus (Ammonian Sections and Eusebian Canons) are encoded as elements, again with the value given as an ‘n’ attribute. The ability to collate or compare series of section numbers, especially those which are unusual, can offer significant information about the relationship of manuscripts.39 The decoration of the text is recorded by the element. Its principal deployment is to indicate letters written in rubrics or other coloured ink, horizontal lines above or below the text (except where these are a substitute for a missing letter) and oversize letters. The convention developed in Collate was to describe the size of large capital letters as a multiple of the height of a standard line: this provides the numerical value for the ‘height’ attribute. When texts written in minuscule include capital letters within the vertical span of a single line, these are simply transcribed as capitals and not otherwise indicated (although care is sometimes needed to identify unusual features of certain scripts). The recording of decorative lines often has to be approached with caution, in order to distinguish whether they were added by the first hand (as in the case of nomina sacra or titles) or a later user: if the latter are transcribed, it may be appropriate to treat them as corrections (or, better, alterations). 2.4. Lacunae, Spaces, Supplied and Damaged Text Many New Testament manuscripts, especially the early papyri, have only been partially preserved. It is possible to give an extremely detailed description of the nature and cause of damage or gaps in XML, but this has been restricted in the IGNTP scheme in order to accommodate earlier practice in the plain-text transcriptions made for Collate. Lacunae,

38 http://www.tei-c.org/release/doc/tei-p5-doc/en/html/PH.html#PHSK. 39 See, for example, McGurk, ‘The Disposition of Numbers in Latin Eusebian Canon Tables’; Popovic, ‘Du nouveau sur les É vangiles de Split’, especially 290–291; O’Loughlin, ‘The Biblical Text of the Book of Deer (C.U.L. Ii.6.32): Evidence for the Remains of a Division System from its Manuscript Ancestry’.



the electronic scriptorium

47

when the writing material is no longer present and no reconstruction is offered, are expressed by an empty element. The approximate size of the gap is given by the combination of two attributes, ‘unit’ (with values ‘page’, ‘line’ or ‘char[acter]’) and ‘extent’. The benefit of ‘extent’, as defined in the TEI P5 Guidelines, is that it can be used to give a range (e.g. ‘3–4’) rather than just a single number. In an earlier version of the guidelines, a ‘ghost page’ was added to fragmentary manuscripts beginning in the middle of a verse, with the opening and elements before the of the surviving text in order to avoid giving the impression that the lacuna occurs on a page which is completely extant. The adoption of the ‘part’ attribute means that this is no longer necessary and only extant text need be transcribed.40 In addition, a second type of element, with the values ‘verse’ and ‘chapter’ for the ‘unit’ attribute, may be added to indicate a textual lacuna. This means that even though the extant page may be physically complete, the incompleteness of the text (with regard to the standard form) can still be recorded, This is also helpful for aligning witnesses in an automatic collation. In the case of shorter lacunae, the transcriber may choose to supply a reconstruction of the missing text. In this case, the words are enclosed within a element, which replaces the . Where the parchment is still extant but the text is no longer readable, the same procedure may be followed but with the attribute ‘reason=“illegible” ’ rather than ‘reason=“lacuna” ’. The source of the reconstruction may also be specified: values for this attribute may include ‘transcriber’, a standard text such as ‘NA28’ or ‘Maj’, or the editio princeps. Where the text is partially extant but the reading is unclear, the relevant characters may be given inside an element, with the reason again specified as an attribute.41 There is also an XML element which can be used to enclose sections of damaged text: this is not adopted within the IGNTP schema, however, as the use of and gives a more precise indication of exactly which characters may be read. Detailed guidance on the use and combination of these various elements is given in §11.5.2 of the TEI P5 Guidelines. It should be noted that where a copyist has deliberately left a 40 Unfortunately, the XML elements and , which could be employed in transcriptions to indicate the extremities of missing portions, are intended for use in a critical apparatus and can only appear within or elements. 41 There is the scope to give an indication of degrees of certainty as an XML attribute: see further Finney, ‘Manuscript Markup’, 270–272. The approach to unclear and supplied text in transcriptions for the Editio Critica Maior reflects the requirements of printed editions following the Leiden conventions.

48

h.a.g. houghton

blank space, this should be transcribed using the element , the dimensions specified in the same way as for a . 2.5. Corrections and Alternative Readings The practice of the IGNTP and INTF, following procedures established for Collate, is to encode variant readings within the flow of the text at the point to which they refer.42 It is also a convention that variation units must consist of complete words, in order to assist with the automatic generation of an apparatus. In XML, each variation unit is enclosed within an element, while the respective readings are identified as elements. The attributes ‘type’ and ‘hand’ identify the nature of the readings and the person responsible. The first element in the sequence is normally , followed by all the words in that unit.43 As many corrections to that section of text may follow as are required, usually in chronological order (so far as this can be ascertained) and with each allocated a separate ‘hand’ attribute. The entire variation unit must be included within the element even if some words are left unchanged by a particular corrector. The numbering of the ‘n’ attributes of the elements in the first continues the sequence of the verse, and is then repeated in each subsequent . If a single word is omitted, the relevant element is missed out; if an entire is blank (i.e. deletion of the entire text or a first-hand omission) this is indicated by the entity &om; in order to assist with collation and display.44 The combination of and elements may also be used to transcribe commentary readings or alternative readings, such as those found in the margins of Family 1 minuscule manuscripts. In this case, the reading in the body of the manuscript is identified as ‘type=“orig” ’ and the alternative is ‘type=“alt”’ or ‘type=“comm” ’ as appropriate, with details of the hand responsible. Recording the position of these readings follows the process detailed for marginal material below: all the words within

42 As noted in 2.1. above, this is indicated in the TEI header by . 43 It would also have been possible to designate the first reading as the lemma, the element rather than . However, because the earliest reading of a manuscript may not be entirely legible, or may have been immediately changed by the first hand, the hierarchy implicit in the use of lemma could be misleading. 44 For example, omissions in Codex Sinaiticus and Nestle–Aland are indicated by the ⸆ symbol.



the electronic scriptorium

49

the should be enclosed within . The ability to combine all the different types of encoding is a particular strength of XML: changes in punctuation, the form of abbreviations, the decoration of the text and even the position of readings (if a text added in the margin was subsequently written above the line by a later hand) are as easy to describe as alterations to the text itself. 2.6. Abbreviations and Non-Standard Characters As already observed, details of abbreviations and character forms go beyond the information required in transcriptions for the purpose of creating a critical edition. Nonetheless, two types of abbreviation found in most New Testament manuscripts are regularly recorded in transcriptions for the Editio Critica Maior. These are nomina sacra and numerals, which are also usually indicated by an overline. The element is added after the opening and followed by the formatting for overline . The three nested elements are closed in reverse order. The type attribute identifies whether the abbreviation is ‘nomSac’ or ‘num’. Other abbreviations may be silently expanded or have the omitted letters supplied within an element. The latter allows the display to toggle between the visible characters, matching the manuscript, and the full form for reading purposes. Sometimes more than one character may be visible in an abbreviation, such as τλ, the standard form for τελος, which could be rendered τελος. However, as other non-letter like symbols are used for abbreviations, it may be better to restrict the ‘visible characters’ to those which are normal size, as in τελος. Non-standard characters sometimes feature in a manuscript, often involving abbreviations. Transcribers should, as far as possible, supply the appropriate character in Unicode even if this is not supported by all fonts: examples include ϛ and ϙ for the numerals 6 and 90.45 When a character is not represented in Unicode, there is the option of using a different glyph (or combination of glyphs) of similar appearance, such as the use of a macron in place of a ‘nu-superline’, although this is not transparent and should be mapped in a one-to-one relationship for consistency.46

45 The code points for these characters are, respectively, U+03DB and U+03D9. 46 An alternative would be to expand the nu-superline into ν, but this does not of itself give the precise form which the abbreviation has taken. Conversely, the kaicompendium does have its own glyph in Unicode, ϗ (U+03D7), but IGNTP practice is to

50

h.a.g. houghton

A better alternative is to use entities to represent non-Unicode characters. So, for example, the single character abbreviations for Latin autem and enim in Insular script are transcribed as &autem; and &enim; rather than trying to reproduce them by combinations such as hr or ++. Similarly, while most punctuation characters are included in Unicode, even in Greek, non-standard forms (or those which conflict with XML encoding, such as > or ;) can be treated in this way. Ligatures may also be expressed as entities. However, it is necessary to declare all entities which occur in an XML document, as well as their intended representation, above the TEI header in order for the file to parse successfully. The treatment of bilingual manuscripts was mentioned in 2.2 above, with an ‘xml:lang’ attribute on each element. On other occasions when two languages are found in the same manuscript (such as untransliterated Greek words in a Latin codex), these may be enclosed in a element, again specifying the language as an attribute. The use of Unicode, however, renders this less of a problem than formerly. 2.7. Editorial Notes The descriptiveness of XML markup means that most features of a transcription can be encoded without the need for further comment. A , however, is one of the most flexible TEI elements and allows the editor and transcriber to supply a free-text description. In the IGNTP scheme, an editorial note should be added at the point to which it refers and given a unique ‘xml:id’ attribute. A distinction is made between notes intended for publication in an online edition and internal queries raised locally by transcribers, using the ‘type’ attribute. The element constitutes an exception to the principle that the only text which does not occur within markup brackets is that which is physically extant on the page. Despite its flexibility, however, it seems advisable to restrict to editorial material rather than use it to convey complex written material, such as paratextual elements. A further use for the element is in the editorial identification of lectionary passages or the biblical references for initial chapter or canon titles. In addition, it may be used in conjunction with the empty element to indicate a change of copyist. expand this as και. In the Digital Codex Sinaiticus, almost all the unusual marginal punctuation characters were reproduced by combinations of other glyphs (e.g. Quire 59 fol. 5r or Quire 82 fol. 2v).



the electronic scriptorium

51

2.8. Marginal Material and Text Positioning The shortcomings of the TEI P5 Guidelines regarding the encoding of the physical layout of text in a document were mentioned above in 2.2. The description of text in margins is a case in point for New Testament manuscripts. A small selection of elements (including and ) has the attribute ‘place’, where the location of the element may be supplied. However, this is not available for section numbers (when encoded with ) or variant readings (apart from additions). Instead, the solution has been to adopt the same approach for all marginal material, which is to designate it as an arbitrary segment, , which may contain (and be contained by) the majority of XML elements.47 The use of a ‘subtype’ attribute permits the further specification of the precise location (e.g. ‘lineleft’, ‘pagetop’ etc.). In addition, the ‘n’ attribute links it to the xml:id of the relevant page, column or line in that witness (e.g. n=“@ P7rC1L45-33”). This should enable it to be displayed at the appropriate place on the page as well as being processed at the relevant point in the flow of the text.48 According to the same logic, the most suitable place to encode chapter or section numbers is at the beginning of the verse to which they refer rather than the beginning of the line; the downside of this is that the transcriber does not handle them in sequence, unless they are already present in the base text. The element can also be used to record the position of other displaced text on a page, for example corrections added above the line or the final characters of a word added in blank space after the line below. 3. Conclusion This paper has attempted to give an account of the development of encoding formats for electronic transcriptions used by the INTF and IGNTP and explain some of the reasoning behind the decisions made with regard to the current scheme of XML encoding. It is hoped that this—as well as the other fruits of the Workspace for Collaborative Editing—will be of interest and assistance both to those who use these transcriptions and those 47 The absence of , however, could cause a problem, although the hierarchy of work outlined above expects all text within a to be further located within or elements. 48 In the Digital Codex Sinaiticus, a element was additionally inserted at the relevant place in the margin, linked to the xml:id of the corresponding element.

52

h.a.g. houghton

working on other textual traditions who are faced with similar problems. The aim has been to describe the approach from a philological rather than a technical standpoint: full documentation for each user group is available when consulting the TEI P5 Guidelines themselves, to which the reader is referred. Keeping within the broad guidelines of the TEI brings a welcome uniformity to electronic transcriptions but it is rarely possible to have a ‘one size fits all’ approach, hence the subset adopted for the Editio Critica Maior of the New Testament. As noted, both the transcriptions and guidelines are made available through UBIRA, where future updates may also be found. The switch to the digital medium has affected not just the format of a transcription, but also the process of transcribing. Starting from an electronic base text, which will usually agree around 90% with the text of the manuscript under consideration, means that transcribers focus on the differences between the manuscript and the base rather than simply trying to copy the manuscript’s text word for word. One advantage of this is the reduction of the potential for harmonisation to a ‘mental text’ when transcribing, although there is the concomitant danger of failing to notice a difference and leaving the base text unchanged. Secondly, encoding a manuscript to permit its presentation as a ‘textual facsimile’ means that all formatting information must be explicitly encoded. In a handwritten copy of a text, column and line breaks (and other aspects of the miseen-page) can be reproduced almost without thinking: the task of including this information means that the transcriber inherits the duties of a compositor as well as a copyist. Nonetheless, this attention to format as well as text, to the physical characteristics of a manuscript as well as its contents, is in keeping with the recent emphasis in New Testament textual criticism on manuscripts as documents as well as tradents of the text. The flexibility of presentation offered by XML encoding means that not only can transcriptions serve as ‘textual facsimiles’ alongside images of a manuscript, but that it will become ever easier to search for and compare physical information (as well as orthographic data and other scribal habits) in the study of New Testament manuscripts. Finally, the potential of electronic transcriptions for revision, adjustment and re-use means that, with the entry of the New Testament into the digital sphere, a new chapter will have to be written on the transmission of the text. As with the transcriptions themselves and the manuscripts before them, some information will be explicitly included by the transcriber (in the form of editorial notes or comments in the header), while other trends may only become apparent over time. In this way, perhaps,



the electronic scriptorium

53

the electronic scriptorium is no different from its predecessors. Even so, while attempts at standardisation often founder in the face of the mass of material, different practices, and varieties of use for the end product, certain innovations have become accepted. Most of these are in the realm of formatting, such as chapter and verse numbers, book titles, or the Eusebian Apparatus. It remains to be seen what lasting effects may be produced by the transition to electronic transcriptions. References Burnard, L., and S. Bauman, eds., TEI P5: Guidelines for Electronic Text Encoding and Interchange (Charlottesville VA: Text Encoding Initiative Consortium, 2007). http://www .tei-c.org/release/doc/tei-p5-doc/en/html/index-toc.html. Burnard, L., S. Bauman and S. DeRose, eds., TEI P4: Guidelines for Electronic Text Encoding and Interchange (Charlottesville VA: Virginia UP, 2001). Burton, P.H., H.A.G. Houghton, R.F. MacLachlan and D.C. Parker, eds., Vetus Latina 19. Iohannes (Freiburg-im-Breisgau: Herder, 2011–). DeRose, S., ‘Trojan Markup and Other XML Milestone-Tagging Techniques’, in International Workshop on Markup of Overlapping Structures 2007. http://www.mulberrytech .com/overlap/index.html. Estienne, R. (Stephanus), Novum Iesu Christi D(omini) N(ostri) Testamentum ex Bibliotheca Regia (third edition) (Paris: Stephanus, 1550). Finney, T.J., ‘Manuscript Markup’, in The Freer Biblical Manuscripts: Fresh Studies of An American Treasure Trove. (SBLTCS 6), ed. L.W. Hurtado (Atlanta: Society of Biblical Literature, 2006), 263–287. Gwynn, E.J., Liber Ardmachanus: The Book of Armagh (Dublin: Hodges, Figgis and Co., 1913). Hoskier, H.C., A New and Complete Edition of the Irish Latin Gospel Codex Usser. 2 or r2 otherwise known as ‘The Garland of Howth’ (London: Quaritch, 1919). Houghton, H.A.G., ‘IGNTP Guidelines for XML transcriptions of New Testament manuscripts’ (2013). http://epapers.bham.ac.uk/1727/. ——, ‘The Electronic Transcription of Codex Sinaiticus’, in Синайский кодекс и памятники древней христианской письменности: традиции и инновации в современных исследованиях [Codex Sinaiticus and Old Manuscripts of Early Christian Writing: Traditions and Innovations in Modern Research], ed. by E. Krushelnitskaya and Z. L. Levshina (St Petersburg: National Library of Russia, 2012). Lake, K., Codex Sinaiticus Petropolitanus: The New Testament, the Epistle of Barnabas and the Shepherd of Hermas (Oxford: Clarendon, 1911). McGurk, P., ‘The Disposition of Numbers in Latin Eusebian Canon Tables’, in Philologia Sacra I. Altes und Neues Testament, ed. by R. Gryson (Freiburg-im-Breisgau: Herder, 1993), 242–58, repr. in Gospel Books and Early Latin Manuscripts, P. McGurk (Aldershot: Ashgate, 1998). O’Loughlin, T., ‘The Biblical Text of the Book of Deer (C.U.L. Ii.6.32): Evidence for the Remains of a Division System from its Manuscript Ancestry’, in Studies on The Book of Deer, ed. by K. Forsyth (Dublin: Four Courts Press, 2008), 3–31. Parker, D.C., ‘Through a Screen Darkly: Digital Texts and the New Testament’, Journal for the Study of the New Testament 25 (2003), 395–411. ——, ‘Electronic Religious Texts: The Gospel of John’, in Electronic Textual Editing, ed. by L. Burnard, K. O’Brian O’Keefe and J. Unsworth (New York: MLA, 2006), 202–205.

54

h.a.g. houghton

——, An Introduction to the New Testament Manuscripts and their Texts (Oxford: Oxford UP, 2008). ——, Codex Sinaiticus. The Story of the World’s Oldest Bible (London: British Library, 2010). ——, Textual Scholarship and the Making of the New Testament (Oxford: Oxford UP, 2012). Popovic, V., ‘Du nouveau sur les É vangiles de Split’, Bulletin de la société nationale des antiquaires de France 1990 (1992), 275–93. Rettig, H.C.M., Antiquissimus Quatuor Evangeliorum Canonicorum Codex Sangallensis Graeco-Latinus Interlinearis Nunquam Adhuc Collatus. Ad Similitudinem Ipsius Libri Manu Scripti Accuratissime Delineandum Et Lapidibus Exprimendum (Zurich: Schulthess, 1836). Robinson, P.M.W., Collate: Interactive Collation of Large Textual Traditions, Version 2, Computer Program distributed by the Oxford University Centre for Humanities Computing (Oxford, 1994). ——, Collate 2. Manual (Oxford: The Computers and Variant Texts Project, 1994). ——, Anastasia: Analytical System Tools and SGML/XML Integration Applications, Version 2.0, Computer Program (Scholarly Digital Editions: Nottingham, 2003). Documentation available online at http://sd-editions.com/anastasia/index.html. Schmid, U.B., ‘Scribes and Variants—Sociology and Typology’, in Textual Variation: Theological and Social Tendencies (T&S 3.6), ed. by H.A.G. Houghton and D.C. Parker (Gorgias: Piscataway NJ, 2008), 16–23. Thompson, E.M., Facsimile of the Codex Alexandrinus, (4 vols.) (London: British Museum, 1879–1883). von Tischendorf, C., Bibliorum Codex Sinaiticus Petropolitanus (Leipzig: Giesecke & Devrient, 1862). Wachtel, K., ‘Editing the Greek New Testament on the Threshold of the Twenty-first Century’, Literary and Linguistic Computing 15 (2000), 43–50. Wilson, N.G., A Descriptive Catalogue of the Greek Manuscripts of Corpus Christi College, Oxford (Cambridge: D.S. Brewer, 2011).



the electronic scriptorium

55

Appendix: Schematic Overview of the xml Structure for New Testament Transcriptions (based on Version 1.4, July 2013)

                                    

                  All the above location elements may be followed by:                  subtypes: pagetop pagebottom pageleft pageright          coltop colbottom colleft colright lineleft lineright            Optional        

56

h.a.g. houghton

        Contents of may include:            The elements may contain or be contained within:        (rend="colour", rend="cap", height="")        (reason="")        (source="" reason="")                   The elements may be contained within:       . . . .     The element may contain:            The element must contain type="" and may contain id=""       note-types: editorial local canonRef    (optional quire element)     



the electronic scriptorium

57

Illustration: Sample xml of a Page of Transcription of a New Testament Manuscript (Codex Bezae: Gregory–Aland 05, Cambridge, University Library Nn. 2.41, fol. 114v. The transcription may be seen in context alongside the image online in the Cambridge University Digital Library http://cudl.lib.cam.ac.uk/view/MS-NN00002-00041/198.)

IIIIκαταιωαννην της σαμαρητιδος

οτιιηςπλειοναςμαθηταςποιει καιβαπτιζειηϊωαννης·καιτοιγεαυτοςιης ουκεβαπτιζεναλλοιμαθηταιαυτου

λβαφηκεντην ϊουδαιανγηντελος καιαπηλθεν παλινειςτην γαλιλαιαν

58

h.a.g. houghton

ανναγνοσμαλγ εδειδεαυτον διερχεσθαιδιατηςσαμαριας ερχεταιουνειςπολιντηςσαμαριας λεγομενηνσυχαρ·πλησιοντουχωριου ουεδωκενϊακωβ·ϊωσηφτωϋιωαυτου ηνδεεκειπηγητουϊακωβ: οουνιηςκεκοπιακως εκτηςοδοιποριαςεκαθεζετοουτως επιτηπηγηωραηνωςεκτη ερχεταιγυνηεκτηςσαμαριαςαντλησαιϋδωρ λεγειαυτηοιηςδοςμοιπειν οιγαρμαθηταιαυτουαπεληλυθισαν



the electronic scriptorium

59

ειςτηνπολιν·ϊνατροφαςαγορασωσιν λεγειουναυτωηγυνηησαμαριτις συϊουδαιοςωνπωςπαρεμου πειναιτειςγυναικοςσαμαριτιδος απεκριθηοιηςκαιειπεναυτη ειηδειςτηνδωρεαντουθυ καιτιςεστινολεγωνσοιδοςμοιπειν συνανητησαςαυτον καιεδωκενανσοιϋδωρζων λεγειαυτωηγυνηκεουδεαντλημαεχεις καιτο φρεαρεστιβαθυ· ποθενεχειςϋδωρζων

μησυμειζωνει

60

h.a.g. houghton

τουπρςημωνϊακωβ οςεδωκενημειντοφρεαρ καιαυτοςεξαυτουεπιεν καιοιϋιοιαυτουκαιταθρεμματα απεκριθηιηςκαιειπεναυτηπαςοπεινων εκτουϋδατοςτουτου·διψησειπαλιν

digital Arabic gospels corpus Elie Dannaoui 1. Project Summary The history of the Arabic text of the Gospels remains one of the most unstudied topics among the ancient translations of the New Testament.1 Historically, many attempts were made to study this text starting with critical editions of the Bible in the nineteenth century, then continuing in the work of Orientalists and lately in the work of scholars interested in the Arab Christian heritage. Although these efforts have not yet covered every aspect of this issue, they have succeeded in drawing attention to the richness of this tradition and to its potential contribution to Biblical studies and to research related to Church history, linguistics, Muslim-Christian relations and other topics. The main obstacle in studying the history of the Arabic text of the Gospels is its heterogeneous aspect. This is due to many elements: the wide spectrum of translation dates, the different Vorlagen of the translations and their linguistic origins, the plurality in terms of compilations and forms (continuous text, lectionary . . .), etc. Why were the Ancient Arabic Translations of the Gospels excluded from the search for the Urtext of the four Gospels? This remains among the top unanswered questions until now. The traditional answers to this type of questions are based mainly on the following hypothesis: the Arabic tradition is late and is a translation from Syriac and Latin and not directly from Greek. We think that the reason for keeping the Arabic text in the shadow has two aspects: first, the Arabic was considered as a recent translation compared to Syriac and Latin. Second, the Arabic text was preserved mainly in lectionaries, not in continuous texts of the Gospels; accordingly, some scholars considered lectionaries to contain a weak text compared to the continuous or canonical version of the four Gospels mainly because lectionaries are church compilation of the original text. 1 For an up-to-date overview about the state of research in the field of the Arabic Manuscripts of the New Testament see: Schulthess, ‘Die arabischen Handschriften des Neuen Testaments in der zeitgenössischen Forschung: ein Überblick’.

62

elie dannaoui

The following chronological survey shows the controversial nature of this issue. In the nineteenth century, Baron Hermann von Soden2 discounted lectionaries as a possible source of information relevant to continuous Gospels books and did not even choose to classify them. Caspar René Gregory3 was one of the first scholars to attempt to establish the importance of the lectionary text in the quest for the Urtext of the four Gospels. In the twentieth century, many scholars started to pay more attention to lectionaries as authentic witness of the Gospels. Ernest Cadman Colwell4 highlighted the importance of the lectionary text in the framework of the Chicago Lectionary project. He indicated the possible existence of a lectionary text distinguishable from the regular Gospel text and disputed the earlier and contemporary notion that the lectionary was unworthy of serious studies because it was a bad text. Lately, Christopher Robert Jordan5 defended the hypothesis saying that the text of the lectionaries derives from the continuous text tradition of the eighth to the eleventh centuries. The Center for Study and Preservation of the Majority Text6 considers that the textual tradition is found and preserved in the Greek New Testament and lectionary text of the Orthodox Church. The complex7 nature of the text requires a comprehensive approach that tackles the Arabic text of the Gospels from different angles. According to this approach, our project deals with the whole tradition of the Arabic translations of the Gospels as one dynamic entity made up of inter-related components and witnessing to the life of Christian communities. Accordingly, the project aims to build an online digital corpus of the Gospels in Arabic. Over time, this corpus will include the transcribed texts, citations and allusions of the Arabic translations of the Gospels. This online database will include both explicit and implicit verses of the Gospels with different layers of metadata (textual, paleographical, codicological, linguistic, etc.). In addition to this digital corpus, the project will provide a set of tools to enable and facilitate the study of the text. 2 Von Soden, Die Schriften des neuen Testaments, in ihrer ältesten erreichbaren Textgestalt / hergestellt auf Grund ihrer Textgeschichte. 3 Gregory, Textkritik des Neuen Testaments. 4 Colwell, Studies in Methodology in Textual Criticism of the New Testament. 5 Jordan, The Textual Tradition of the Gospel of John in Greek Gospel Lectionaries from the Middle Byzantine Period (8th–11th century). 6 http://www.cspmt.org, last accessed 8 May 2013. 7 The same verse may derive from different heterogeneous formats. For example: a regular verse from a chapter in a Gospel or a verse from a lection or a patristic citation or an allusion in an apologetic text.



digital arabic gospels corpus

63

2. Objectives The project has two main objectives: – Contributing to the scholarly quest for the first translation of the four Gospels into Arabic. When and where was the earliest Arabic translation of the Gospels compiled? There were have been some debates over the dating of this version, possibly because this question has been tackled from different angles. Historical accounts confirm the existence of an Arabic Christian community before the rise of Islam, but cannot confirm that Arabic was the scriptural and liturgical language that they used.8 – Identifying the different families of the Arabic translations preserved in the manuscripts’ tradition. Many attempts have been made to classify the textual traditions of the Arabic Gospels. They include the work of the following scholars: Ignazio Guidi (1888),9 Samir K. Samir (1992),10 Jean Valentin (2003)11 and Hikmat Kashouh (2012).12 These important studies highlight the complexity of this issue, but they were made on the basis of only a few test texts selected from a large number of manuscripts. Their classifications were based mainly on verbal agreement. Our project intends to continue these efforts by extending the study to include a larger number of texts and manuscripts and by using new automated methods.

  8 Kashouh, The Arabic Versions of the Gospels: The Manuscripts and their Families, 34–35.   9 Guidi, Le traduzioni degli Evangelii in arabo e in etiopico. 10 Samir, La version arabe des évangiles d’al-As’ad ibn al-‘Assāl. 11  Valentin, ‘Les évangéliaires arabes de la bibliothèque du monastère Ste.-Catherine (Mont Sinaï): essai de classification d’après l’étude d’un chapitre (Matth. 28): traducteurs, réviseurs, types textuels’. 12 Kashouh, The Arabic Versions of the Gospels.

64

elie dannaoui 3. Methodology

To achieve these objectives, the project will proceed in the following way: – Collecting digital copies of all known Arabic Gospels texts. To achieve this task, we are in the process of consolidating accounts from manuscript catalogues that are held in different countries. We use the monumental work of Graf on Arabic Christian Literature as our basis and update it with new findings from catalogues, manuscripts examination and complementary literature published since the work of Graf. Thus far we have completed a detailed list of Arabic Gospels manuscripts conserved in Lebanon, Syria and Germany.13 – Building a digital corpus for the Gospels in Arabic containing the transcriptions of the identified texts. – Defining types and techniques of analysis that will be performed on the content of the corpus. – Designing, developing and implementing appropriate tools for textual analysis. – Conceiving a formal evaluation grid that may be used to define the identity of a text and to build a taxonomy of the Arabic translations. 4.  Building the Corpus The Arabic text of the Gospels is present in two formats: – Explicit: This format includes the verses of the Gospels transmitted to us in the manuscript tradition. Those verses are present in the different compilations of the four Gospels (continuous text, lectionaries) and as citations in other writings (writings of Church Fathers, liturgical texts, Arabic and Islamic literature, etc.). – Implicit: This format includes allusions to Gospel verses or contents in different types of writings (Arabic poetry, apologetic literature . . .). These allusions are witnesses of a certain version or translation of the Gospels’ translation and may contribute to identifying this tradition if they were formally presented and integrated in the corpus.

13 This part was developed in 2012 in collaboration with the German ‘Orient Institut’ in Beirut.



digital arabic gospels corpus

65

Both explicit and implicit texts of the Gospels verses will be included in the corpus according to a formal taxonomy. This approach will allow the corpus, when queried, to return all the occurrences of a specific verse regardless of its format or type. For example, if one asks the corpus about Matthew 1:12, the platform will return the text of this verse as found in the entire Arabic manuscript tradition of the Gospels, in all the lections, in patristic citations and allusions, and in liturgical prayers or in any other sources. For example, if a pre-Islamic poem refers indirectly to Matthew 1:12, its text will also be included in the results. The modularity of the system will allow the scholar to define his or her own way of displaying, grouping and sorting the results so as to suit best the requirements of the research. The transcription will include different layers of metadata formalized in a set of authority lists. (An authority list contains an array of values belonging to a specific domain. The metadata values are selected from those lists. In this way we guarantee a high level of consistency and we avoid having the same aspect of the text described in different ways.) This will guarantee a higher level of data consistency and integrity. 5. The Technology 5.1. The Database Taking into consideration the diverse nature of these sources, the main challenge was to conceive a database capable of handling and consolidating these different types of content. Traditionally, relational databases are used in similar projects and the main issue for decision makers is mainly to select the appropriate relational database to use. In our project, we decided to go in the opposite direction. Instead of using a well-designed relational database schema, we decided to use an unstructured database. This family of databases, also known as NoSQL,14 is non-relational, distributed, open-source and horizontally scalable. Its schema-free approach will enable us to anticipate the emergence of new text formats and will allow the corpus to host different types of texts (lection, Gospel chapter, poem, book chapter, prayer . . .). The main features of this approach are as follows:

14 http://nosql-database.org, last accessed 8 May 2013.

66

elie dannaoui

5.1.1. Document Oriented Data has a flexible schema. The corpus does not enforce document structure. This means that transcribed texts in the corpus do not need to have the same set of fields or structure, and common fields in the documents contained in the corpus may hold different types of data. For example, a lection containing a reading from the Gospel according to John only needs to contain fields that are relevant to the fourth Gospel (evangelist, chapter, verses and a set of optional fields). These fields will allow the database to place this lection alongside patristic citations or allusions that contain the same features (which will allow the database to group together all the different witnesses to the same text together, regardless of their origin). Schema flexibility means that we can model the corpus documents so that they can closely resemble and reflect application-level objects. This data model reflects how data will grow and change over time, and the kinds of queries our application will perform. 5.1.2. User-Centered One of the most important advantages of this technological approach is the shift from developer-centered to user-centered applications. In Relational Database Management Systems, the load is put on the back-end operations (analysis, design, development, programming, etc.). The end user (the researcher in our case) is a consumer of the system and any modification on the level of data structure requires an intervention at the back-end, and affects the front-end on several levels. Our approach gives the user the possibility to interact with the platform not only on the client side, but also allows the scholar for example to define a new document type, to populate it with appropriate existing fields, or even to add new fields. The integrity of the data is guaranteed by a set of validation schemes. 5.1.3. Full Index Support An index is a data structure that allows the user quickly to locate documents in the corpus based on the values stored in certain specified fields. The platform supports indexes on any field or sub-field contained in documents within the corpus. This allows the scholar to locate word occurrences in all the transcribed texts of the corpus.



digital arabic gospels corpus

67

5.2. The Document: Transcription Text Encoding Initiative (TEI)15 is the standard for transcriptions and annotations. TEI is based on XML technology. In our project, we decided to use JSON16 (JavaScript Object Notation) instead of XML for the following reasons: – JSON is a lightweight data format easy for human and machines to read, write and parse. – XML processing is slower and requires the programmer to have certain level of technical knowledge. – JSON can easily be enriched with new data. – JSON does not require nested attributes to build the document’s tree. A reference system is used for this purpose, which simplifies the document’s structure.17 – JSON handles both diplomatic (document content) and critical (content and layout) transcriptions in a smart way. Content and layout get separated immediately with a referencing system. The data stored in the database can be retrieved as diplomatic or critical edition depending on the scholar’s decision. 6. Corpus Processing As we mentioned previously, the corpus will allow different types of content analysis: 6.1. Structural The critical transcription will populate the corpus with both text and metadata related to the format, the inner structure, the versification and the reading boundaries. The analysis of these elements (and many other elements) will help the researcher to identify different aspects of convergence or divergence between textual traditions. For example, the internal structure of the lectionary in terms of number of lections, verses included in a lection, and sequence of lections will help the scholar to build a grid

15 http://www.tei-c.org/index.xml, last accessed 8 May 2013. 16 http://www.json.org, last accessed 8 May 2013. 17 http://json-schema.org, last accessed 8 May 2013.

68

elie dannaoui

system that may be used to identify the styles of the different compilations, the date of a writing, its original community, geographical identity and many other properties. The dynamic structure of the corpus will make these types of analysis possible through a user-friendly, interactive and visual interface. 6.2. Linguistic Several attempts were made to study the linguistic features of the Arabic translations of the Gospels in order to identify the different textual traditions. The majority of these projects use verbal agreement between texts to define their identities. In some cases, other techniques were used on a reduced scale in analyzing selected texts. The limitations of these attempts reside in the fact that the study used only a small number of readings selected from only a few manuscripts and that the selection was not formalized or automated. To fill these gaps, our project offers automated linguistic corpus processing features. All transcribed texts are subject to a morphosyntactic annotation. Lexical, grammatical and inflectional properties (tense, grammatical mood, grammatical voice, aspect, person, number, gender and case) are associated with the annotated text. These linguistic properties allow the system to perform complex searches based on abstract representations of a specific word, sentence, paragraph, syntax and occurrence. In order to formalize all possible verbal tokens, we defined a taxonomy of inflectional classes for Arabic verbs. This taxonomy allows the system to encode simultaneously in the lexical representation three variations: inflectional, morphophonemic and orthographic. These features will enable the scholar to study the following variations: 6.2.1. Variations in Dialect It is common in Arabic manuscripts of the Gospels to have different forms for the same inflected word. For example, the same lemma or canonical form of a word may have different forms of broken plural. The current type of analysis allows the scholar to perform a search based on the dialectic pattern of a broken plural identified in a specific text. In this case, the search operation will return all the lemmas or lemmata affected by this pattern. A more advanced search may return the geographical or historical distribution of the lemmas affected by a specific inflectional pattern.



digital arabic gospels corpus

69

6.2.2. Orthographic Variations Orthographic variations may serve in revealing the identity of the scribe and may also contribute in locating a text on a timeline. The use of diacritics in Arabic Gospel manuscripts remains unstudied. The formalization and abstraction of this use may help the scholar to define sets of standards that enable the text’s grouping according to orthographic variations. 6.2.3. Grammatical Variations A language changes over time and varies according to place and social setting. In the case of Arabic, we can observe grammatical variation like differences in the structure of words, phrases or sentences by comparing the same translated Gospel text taken from different manuscripts. One of the most common differences between these manuscripts is the way in which tenses are formed. Our corpus processor provides an appropriate tool capable of formalizing all the instances of the same lemma and drawing the variation graph of its tenses on a timeline. This type of analysis assists the researcher in building a complex system in order to classify the textual traditions. For example, if a Greek verb requires an accusative for its object, the corpus processor can identify the case of all the objects in the Arabic translation of the same verb wherever it is found in the transcribed manuscripts. The results can be grouped and sorted by case. This process can be done in both horizontal (in the same manuscripts) and vertical (in all the manuscripts) ways. 6.2.4. Semantic Variations Many aspects of lexico-semantic variation are linked to a sociolinguistic framework. This interplay between society, culture, religion and language is very strong in Arabic manuscripts of the Gospels. In this perspective the corpus allows the scholars to define semantic arrays through a process of abstraction. 6.2.5. Syntax Variations The syntax reflects the principles and processes by which sentences are constructed in particular languages. In the case of Arabic Gospels, the syntax underlying the text is one of the strongest characteristics of the linguistic origin of the manuscript. Being a sacred text, the translators did their best to make their Arabic translation very close to the original (Greek, Syriac, Latin and potentially other oriental languages). As a result of this effort, they compromised the Arabic syntax and imported the original one, but now using Arabic words. The annotated corpus allows the

70

elie dannaoui

scholar to define a specific syntax as a pattern and to find all the Arabic instances in the transcribed manuscripts that follow it. For each verse, the researcher can extract its syntax and compare it to that of the same verse in other languages (this requires that the same verse in other languages is annotated in the corpus). This tool may help in identifying formally the Vorlagen of the translations. 7. Conclusion The major added value of this corpus would be its capacity to give precise answers both to text critics and to scholars of Arab Christian studies in response to still unanswered questions about the importance of the different forms of Gospels text (mainly lectionary) and the real value of the Arabic tradition. We believe that a systematic and formal approach will guarantee a higher level of precision in measuring the evaluating the nature and impact of Arabic tradition on the text of the Gospels. References Colwell, E.C., Studies in Methodology in Textual Criticism of the New Testament (Leiden: Brill, 1969). Gregory, C.R., Textkritik des Neuen Testaments, Leipzig, 1900. Guidi, I., Le Traduzioni degli Evangelii in arabo e in etiopico, Atti della reale accademia dei Lincei, 4 (Rome: Tipografia della r. accademia dei Lincei, 1888). Jordan, C.R., The Textual Tradition of the Gospel of John in Greek Gospel Lectionaries from the Middle Byzantine Period (8th–11th century) (Ph.D. thesis, Birmingham, 2009). Kashouh, H., The Arabic Versions of the Gospels: The Manuscripts and their Families, Arbeiten zur neutestamentlichen Textforschung 42 (Berlin/Boston: De Gruyter, 2012). Samir, S.K., ‘La version arabe des évangiles d’al-As’ad ibn al-‘Assāl’, Parole de l’Orient 19 (1992), 156–190. Schulthess, S., ‘Die arabischen Handschriften des Neuen Testaments in der zeitgenössischen Forschung: ein Überblick’, Early Christianity, Volume 3, Number 4 (December 2012), 518–539. Valentin, J., ‘Les évangéliaires arabes de la bibliothèque du monastère Ste.-Catherine (Mont Sinaï): essai de classification d’après l’étude d’un chapitre (Matth. 28): traducteurs, réviseurs, types textuels’, Le Muséon 116 (2003), 415–477. Von Soden, H., Die Schriften des neuen Testaments, in ihrer ältesten erreichbaren Textgestalt / hergestellt auf Grund ihrer Textgeschichte, 4 volumes (Berlin: Glaue, 1902–1910).

the role of the internet in new testament textual criticism: the example of the Arabic manuscripts of the new testament Sara Schulthess 1. Introduction The aim of this article is not only to discuss the role that Internet can play in relation to the study of the Arabic manuscripts of the New Testament, but also to explore some related issues raised by that role. At first, it must be said that the Arabic manuscripts of the New Testament have suffered from a comparative neglect in New Testament textual criticism. This neglect has two main explanations: firstly, for a long time, New Testament textual criticism focused mainly or exclusively on finding the ‘original’ Greek text and indeed showed little interest in the ‘secondary’ versions of the New Testament. In this context, the Arabic manuscripts were discarded as useless, as is shown in Ewert’s comment: ‘Since the Arabic versions are so late, they are not useful as witnesses to the original text of New Testament.’1 Secondly, the study of the Arabic manuscripts of the New Testament belongs to the general context of what we call today ‘orientalism’. This set of attitudes and prejudices was problematized by the scholar Edward Said in the 1970s and is discussed today in the field of cultural and postcolonial studies. The idea of the superiority of Western culture was also predominant in universities in the nineteenth and twentieth centuries and especially touched the Arabic language and Islamic traditions, with consequences for the study of Arabic Christianity and the Arabic versions of the Bible. The fact is that since the extensive study by Georg Graf Die Geschichte der christlichen arabischen Literatur,2 published at the Vatican in 1944, nothing really important has been published until recent years.3 But this period of 1 Ewert, A General Introduction to the Bible: from Ancient Tablets to Modern Translations, p. 171. 2 Graf, Geschichte der christlichen arabischen Literatur. 3 For a more complete state of the research, see Kashouh, The Arabic Versions of the Gospels. The Manuscripts and their Families, 9–33, or my article Schulthess, ‘Die arabischen Handschriften des Neuen Testaments in der zeitgenössischen Forschung: ein Überblick’.

72

sara schulthess

disinterest seems now to be over. Indeed, in recent years, there is a large number of publications in this field, marking a resurgence of interest in the Arabic manuscripts of the New Testament. I mention first the thesis of Hikmat Kashouh, The Arabic Versions of the Gospels4 which is a huge work that classifies more than 200 manuscripts. Samir Arbache has already published L’Évangile arabe selon saint Luc. Texte du VIIIe siècle, copié en 897. Édition et traduction;5 the author has been working on this manuscript, the Sin. Ar. 72, for twenty years6 but it is the first time that a part of the manuscript is edited. Sydney Griffith also published on this topic, with a volume The Scriptures of the People of the Book in the Language of the Qur’an.7 These are only a few examples. So we may ask: Why the resurgence? What does it mean? Does this have something to do with the Internet? Can we find some answers to such questions on the Internet? 2. Digitization and Democratization In general, we may affirm that New Testament textual criticism has been boosted by the rise of the digital era and the Internet. Claire Clivaz explains: Viewed in the second part of the 20th century as a subsidiary task, textual criticism is today one of the most rapidly expanding fields in New Testament studies, thanks notably to the ‘explosion’ of new manuscripts discovered or published online. David Parker has already drawn attention to the significance of computers and the new tools they provide for the present ‘dramatic change’ in textual criticism and the editing of the New Testament, but the extent of this change is still currently underestimated.8

It is clear that the rediscovery of the Arabic versions participates in this renaissance of the textual criticism, a renaissance with several consequences. Digitizing the manuscripts and putting them online, but also enjoying an easier access to basic works of textual criticism in general,

4 Kashouh, The Arabic Versions of the Gospels. 5 Arbache, L’Évangile arabe selon saint Luc. Texte du VIIIe siècle, copié en 897. Édition et traduction. 6 See his PhD thesis: Arbache, Une version arabe des évangiles: Langue, texte et lexique (PhD diss.). 7 Griffith, The Bible in Arabic: The Scriptures of the ‘People of the Book’ in the Language of Islam. 8 Clivaz, ‘Homer and the New Testament as “Multitexts” in the Digital Era?’, 3.

the role of the internet in new testament textual criticism 73 allows for a ‘democratization’ of the discipline: anyone interested can educate oneself and consult the manuscripts. This development can give a new dynamism to the research, especially in the field of the Arabic versions that has been discarded for too long. But it requires prudence. Indeed, we can note the phenomenon of ‘pseudoscholarship’, a term used by Ulrich Schmid: ‘The Internet is full of pseudoscholarship that exhibits images equipped with uneducated interpretations as if knowledge is easily culled from a quick glance at an image taken out of context.’9 For example, on the website http://scholarly-faith.blogspot.com, an Egyptian tour guide, who is fond of biblical textual criticism, offers a topic10 on an ancient Arabic manuscript of the New Testament.11 On this webpage, he presents some divergent readings, including the omission of the famous episode of the adulterous woman in the Gospel of John (John 7:53–8:11). In a comment, a reader is surprised at this lack, wondering if the ancient Arabs could not understand the story. It is indeed not explained in the topic that the Pericope Adulterae follows a complicated textual history and is missing in some textual traditions. This is an ambivalent point: Of course, it is interesting to learn that the manuscript does not contain the Pericope Adulterae, because this gives us important clues about the translation that the manuscript contains. But we also see that Internet users can misunderstand the significance of this information, when it is given without further explanation. 3. The Collaborative Work: An Advantage for the Study of the Arabic Manuscripts of the New Testament The Internet also allows for the emergence of collaborative work, like the best-known example of the successful encyclopedia Wikipedia. This ‘wikiculture’ has much to contribute to the field of New Testament textual criticism. A good example is the Yahoo forum Textual Criticism of the

  9 Schmid, ‘Thoughts on a Digital Edition of the New Testament’, 302. 10 http://scholarly-faith.blogspot.com/2009/06/blog-post_03.html, last accessed 9 January 2013. 11  The author does not give the name of the manuscript, but the description as ‘the oldest Arabic manuscript of the four Gospels’ seems to refer to the Sin. Ar. 72. However, in his description of the manuscript, Kashouh does not mention the absence of John 7:53–8:11, see Kashouh, The Arabic Gospels of the Gospels, 87.

74

sara schulthess

Bible12 moderated by Wieland Willker where very interesting topics and new findings may be found.13 This kind of collaborative work encouraged by the Internet is an important step also for the study of the Arabic manuscripts of the New Testament. Indeed, the study of the Arabic manuscripts of the New Testament has always been a crossroad field, between ‘East’ and ‘West’, between biblical studies and Oriental studies. In the history of research, Biblical scholars seemed eager to delegate the work to the Orientalists. Important studies, such as those of Ignazio Guidi,14 Georg Graf, Anton Baumstark,15 etc., were works of Orientalists. We can also feel a certain reserve on the part of the Alands, when they write in their reference book: ‘But unfortunately the arabists of today are hardly concerning themselves with the transmission of the New Testament, although there are many interesting problems here [. . .]’.16 This quotation illustrates some distance between researchers of different fields, a distance that the Internet can help to reduce, facilitating contacts and collaborations between scholars with different and complementary competences. As mentioned before, the Yahoo Forum Textual Criticism is a significant place for discussions and exchanges, containing many interesting topics on the Arabic manuscripts of the New Testament.17 An example of such collaborative work is found in this topic, where a user wonders about an Arabic writing in the Codex Sinaiticus: ‘In Codex Sinaiticus, Revelation 7:12–9:5 (folio # 129a), at the bottom of the page there is some writing placed directly under the first three columns. It looks to me like it might be Arabic? Does anyone know what language this is? And, if so, what is the translation as well as the history behind this strange editorial insertion?’18 The scholar Jean Valentin, author of the article Les évangéliaires 12 http://groups.yahoo.com/group/textualcriticism, last accessed 9 January 2013. 13 See the article in the present volume: Clivaz, ‘Internet Networks and Academic Research: the Example of New Testament Textual Criticism’. 14 Guidi, Le Traduzioni degli Evangelii in arabo e in etiopico. 15 Baumstark, ‘Arabische Übersetzung eines altsyrischen Evangelientextes und die Sure 21:105 zitierte Psalmenübersetzung’; ‘Das Problem eines vorislamischen christlichkirchlichen Schrifttums in arabischer Sprache’; ‘Eine altarabische Evangelienübersetzung aus dem Christlich-Palästinensischen’; ‘Die sonntägliche Evangelienlesung im vorbyzantinischen Jerusalem’. 16 Aland, The Text of the New Testament: An Introduction to the Critical Editions and to the Theory and Practice of Modern Textual Criticism, 214. 17 See also the forum http://evangelicaltextualcriticism.blogspot.ch, last accessed 9 January 2013. 18 http://groups.yahoo.com/group/textualcriticism/messages/3097, last accessed 9 January 2013.

the role of the internet in new testament textual criticism 75 arabes de la bibliothèque du Monastère Ste-Catherine,19 with the help of other users, offers a translation of the Arabic comment.20 It follows an interesting discussion on the possible origin of this comment and its dating. Although these questions remain unanswered so far, such a discussion can only nourish the aspiration to study the Arabic annotations on the Codex Sinaiticus more closely. An example of collaborative work specific to the Arabic manuscripts of the New Testament is the project started by a team of the University of Balamand in Lebanon: the elaboration of a platform for the study of the manuscripts of the New Testament, which is still in progress. Such a tool will encourage all interested parties to share information, with the possibility to complete the database progressively.21 4. The Point of View of Some Islamic Websites Related to these digital revolutions, to textual criticism and also to Arabic manuscripts of the New Testament, we can note this interesting phenomenon: the case of Islamic websites interested in textual criticism of the New Testament. These websites provide reference material and links to online manuscripts and encourage the readers to learn about textual criticism of the New Testament. The purpose is most apologetic and polemical: it is expected to show the inconsistencies of the Bible or its transmission. For example, on the website http://www.sheekh-3arb.net/, a user seeks to question the divinity of Christ by showing that mentions of divine attributes were simply substitutions or later additions. He gives twenty examples of such ‘substitutions’ or ‘additions’, reviewing different editions and then some papyri and codices, with the help of digitalized pictures.22 In the first example, he presents the case of John 6:69 ‘And we believe and are sure that thou art that Christ, the Son of the living God’ (KJV), showing that the christological attribute is missing in the 19 Valentin, ‘Les Evangéliaires arabes de la bibliothèque du Monastère Ste-Catherine (Mont Sinai): Essai de classification d’après l’étude d’un chapitre (Matth. 28). Traducteurs, réviseurs, types textuels’. 20 http://groups.yahoo.com/group/textualcriticism/messages/3101, last accessed 9 January 2013. 21 See the article of Elie Dannaoui in the present volume, ‘Digital Arabic Gospels Corpus’. 22 http://www.sheekh-3arb.net/vb/showthread.php?t=2127, last accessed 9 January 2013. On this website, see my article: Schulthess, ‘Les manuscrits du Nouveau Testament, le monde arabe et le digital. L’émergence d’un discours hybride’.

76

sara schulthess

figure 1. codex Vaticanus (John 6.69) on the website Sheek-3arb.23

oldest manuscripts, as in the codex Vaticanus (image below); a fact which implies a later human alteration. this way of proceeding evokes a long tradition of islamic polemic, linked with the concept of the falsification of the scriptures, taḥrîf. the issue of taḥrîf is already in the Quran,24 but only as the idea of a falsification of meaning, taḥrîf al-ma‘na, and that is to say, misinterpretation of the text.25 the concept of textual falsification, taḥrîf al-lafẓ, comes much later in the islamic tradition: ‘it is only beginning with the time of ibn Ḥazm in the eleventh century that the taḥrîf argument became something of a starting point in the islâmic polemical discourse [. . .].’26 this idea that christians do not have the original text of the bible has become an essential part of the anti-christian polemic in the nineteenth century especially with the progress of biblical textual criticism in western universities. as christine schirrmacher notes: 23 screenshot of http://www.sheekh-3arb.net/vb/showthread.php?t=2127, last accessed 9 January 2013. 24 the verb ḥarrafa occurs in sura Al-Baqara (2) 75, sura An-Nisâ’ (4) 46, sura Al-Mâ’ida (5) 13 and 41. 25 accad, ‘corruption and/or misinterpretation of the bible, the story of the islâmic usage of tahrîf ’. 26 accad, ‘corruption and/or misinterpretation of the bible’, 86.

the role of the internet in new testament textual criticism 77 [Muslim apologists] feel confirmed in the traditional Muslim view that the Bible is corrupted just as the Qur’ân states. Muslim apologists have known this for centuries already, but now European theologians have confirmed it themselves through scientific studies in history, geology or archeology.27

It is interesting to see that Arabic manuscripts of the New Testament are also used in some controversies on Islamic websites. In a topic of the website www.alta3b-wordpress.com, classified in the taḥrîf section, the author deals with the following question: should the end of John 1:1 be translated by ‘and the word was God’ (allâh) or ‘and the word was a god’ (ilâh)? The question connects two polemical issues: First, the questioning of the divinity of Jesus Christ and the debate against the trinity, both classical themes of the anti-Christian debate; second, the falsification or misinterpretation of the Scriptures, as seen previously. A first page deals with the Greek text, browsing through biblical references and showing some pictures of manuscripts.28 A second page presents some Arabic manuscripts of the New Testament.29 The author uses an article of Hikmat Kashouh, who showed, among other things, that several Arabic translations found in ancient manuscripts have at the end of the verse the name ilâh and not allâh. Kashouh explained in the article that the scribes who translated John 1:1 followed different approaches, among them: The literal-approach scribes who translated θεός by ilâh ([a] God) and ὁ θεός by allâh ([the] God). This differentiation may possibly have caused a misrepresentation belief that the Christianity firmly holds especially in an Islamic milieu. The earlier versions of the Gospels seem to prefer this translation.30

But the author does not just copy Kashouh’s study: He completes it by adding the images of the manuscripts, not present in the article, which means he had to get the images and to extract the desired verse.

27 Schirrmacher, ‘The Influence of German Biblical Criticism on Muslim Apologetics in the 19th Century’. 28 http://alta3b.wordpress.com/2009/04/06/translation_jn-1-1/, last accessed 9 January 2013. 29 http://alta3b.wordpress.com/2011/11/03/jn1-1_arabic-mss/, last accessed 9 January 2013. 30 Kashouh, ‘The Arabic Versions of the Gospels. A Case Study of John 1,1 and 1,18’.

78

sara schulthess

figure 2. sin. ar. 75 (John 1:1) on the website Alta3b.31

on another website, we can find the same topic, completed by other manuscripts and classed into the families of manuscripts according to Kashouh.32 we can often observe such appropriation: Globally, there is little sense, on these websites, that a given point has an individual author. here, there is a kind of dissemination that defies the notion of a fixed authorship that the modernity knows. at any rate these websites are not an insignificant phenomenon. in fact, we can see that the website of the Center for the Study of the New Testament Manuscripts, which put online many digitized manuscripts of the new testament, offers an index of the codex Vaticanus,33 which was made by the previously cited islamic website http://www.sheekh-3arb.net.34

31  screenshot of http://alta3b.wordpress.com/2011/11/03/jn1-1_arabic-mss/, last accessed 9 January 2013. 32 http://www.hurras.org/vb/showthread.php?t=14868, last accessed 9 January 2013. 33 http://images.csntm.org/manuscripts/Ga_03/Vaticanus-scripture-index.pdf, last accessed 9 January 2013. 34 http://www.sheekh-3arb.net/vb/showthread.php?p=6346, last accessed 9 January 2013.

the role of the internet in new testament textual criticism 79

35

Figure 3. Title page of the Vaticanus Index.35

The CSNTM is an independent Christian Center led by professors of New Testament, wishing to provide the largest number of manuscripts for researchers and preserve them by digitalizing. The Salafist website Sheek_3arb’s mission is to study the New Testament for polemical and apologetic purposes. This raises questions about the status of each of these productions and their meeting produced what we can call a hybrid scholarly discourse.36 Such ‘hybridization’ is also found on the Yahoo Forum, where users of such Islamic websites exchange freely with more traditional searchers in textual criticism. For instance in this topic, a user offers a comparison between Wieland Willker’s commentary on the Gospel of Matthew37 and the Sinai Ar. 72.38 The Internet address of the document goes back to the

35 Screenshot of http://images.csntm.org/Manuscripts/GA_03/Vaticanus-ScriptureIndex.pdf, last accessed 9 January 2013. 36 Term developed by Clivaz, ‘Homer and the New Testament as “Multitexts” in the Digital Age’. 37 http://www-user.uni-bremen.de/~wie/TCG/index.html, last accessed 9 January 2013. It is interesting to notice that Wieland Willker is originally a chemist at the University of Bremen. This illustrates the changes of borders in the discipline. 38 http://groups.yahoo.com/group/textualcriticism/messages/6127, last accessed 9 January 2013.

80

sara schulthess

website Sheekh-3arb and the author presents himself as co-editor on an Arabic Textual Criticism Journal,39 a website connected to website Sheekh3arb as well. 5. Conclusion This hybridization can make us confused, but we should not misjudge too quickly. In his work Orientalism, Edward Said describes the multiplicity of the forms of domination of the West over the East; one is the idea that only the West can produce a meaningful discourse on the Orient, which cannot represent itself: The exteriority of the representation is always governed by some version of the truism that if the Orient could represent itself, it would; since it cannot, the representation does the job, for the West, and faute de mieux, for the poor Orient.40

This thinking is obvious in the next example: in 1865, a new translation of the Bible in Arabic was published on the initiative of the American Bible Society. This translation of the Bible, so-called Smith-Van Dyke, became very popular among Arab-speaking people. The committee describes so the aim of this project: ‘To give the Word of God to forty millions of perishing sinners [. . .]: in short, to give them a Christian literature, or that germinating commencement of one [. . .].’41 In this context, it is easy to understand the need of promoting a tradition, which existed in many forms long before the West provided one. The will to reaffirm a culture is illustrated by the resumption of publications in recent years. This view is also shared by the Islamic website that contains the topic on John 1:1, where we find the following remark: We, Arabs, Muslims or Christians, should have an interest in our heritage, as the Latins are interested in their Latin heritage, the Syriacs in their Syriac heritage, etc. We must take care of this heritage written in Arabic to give it to the world rather than they provide us our heritage.42

Perhaps, the rediscovery of a rich tradition that has been underestimated for too long comes to us from where we may expect it at least, but it is an 39 http://tcjournal.sheekh-3arb.net, last accessed 9 January 2013. 40 Said, Orientalism, 21. 41  Jessup, Fifty-Three Years in Syria, 68–69. 42 http://www.hurras.org/vb/showthread.php?t=14868. Last accessed 09.01.13. My traduction.

the role of the internet in new testament textual criticism 81 encouraging sign for a recognition of the need to study the Arabic manuscripts of the New Testament. References Accad, M., ‘Corruption and/or Misinterpretation of the Bible, the Story of the Islâmic Usage of Tahrîf ’, NEST Theological Review 24/2 (2003), 67–97. Aland, K., and B. Aland, The Text of the New Testament: An Introduction to the Critical Editions and to the Theory and Practice of Modern Textual Criticism (Grand Rapids: W.B. Eerdmans, 1987). Arbache, S., L’Évangile arabe selon saint Luc. Texte du VIIIe siècle, copié en 897. Édition et traduction (Bruxelles: Safran, 2012). ——, Une version arabe des évangiles: Langue, texte et lexique (PhD diss., Université Michel de Montaigne Bordeaux III, 1994). Baumstark, A., ‘Die sonntägliche Evangelienlesung im vorbyzantinischen Jerusalem’, Byzantinische Zeitschrift 30 (1930), 350–359. ——, ‘Das Problem eines vorislamischen christlich-kirchlichen Schrifttums in arabischer Sprache’, Islamica 4 (1931), 574–575. ——, ‘Eine altarabische Evangelienübersetzung aus dem Christlich-Palästinensischen’, Zeitschrift für Semitistik und verwandte Gebiete 8 (1932), 201–209. ——, ‘Arabische Übersetzung eines altsyrischen Evangelientextes und die Sure 21:105 zitier­te Psalmenübersetzung’, Oriens Christianus 9 (1934), 164–188. Clivaz, C., ‘Homer and the New Testament as “Multitexts“ in the Digital Era?’, Scholarly and Research Communication 3(3): 030126 (2012), 15 pp. Ewert, D., A General Introduction to the Bible: from Ancient Tablets to Modern Translations (Grand Rapids: Zondervan, 1990). Graf, G., Geschichte der christlichen arabischen Literatur, 5 volumes (Rome: Biblioteca Apostolica Vaticana, 1944). Griffith, S.H., The Bible in Arabic: The Scriptures of the ‘People of the Book’ in the Language of Islam (Princeton: Princeton University Press, 2013). Guidi, I., Le Traduzioni degli Evangelii in arabo e in etiopico, Atti della reale accademia dei Lincei, Vol. 4, ser. 4 (Rome: Tipografia della r. accademia dei Lincei, 1888). Jessup, H.H., Fifty-Three Years in Syria, vol. 1 (New York, Chicago et al.: Fleming H. Revell Company, 1910). Kashouh, H., ‘The Arabic Versions of the Gospels. A Case Study of John 1,1 and 1,18’, in The Bible in Arab Christianity, ed. by D.R. Thomas (Leiden: Brill, 2007), 9–36. ——, The Arabic Versions of the Gospels. The Manuscripts and their Families (Berlin/Boston: De Gruyter, 2012). Said, E.W., Orientalism (London: Penguin Books, 20033). Schirrmacher, C., The Influence of German Biblical Criticism on Muslim Apologetics in the 19th Century (1997). http://www.contra-mundum.org/schirrmacher/rationalism. html, last accessed 9 January 2013. Schmid, U., ‘Thoughts on a Digital Edition of the New Testament’, in Lire Demain. Des manuscrits antiques à l’ère digitale / Reading Tomorrow. From Ancient Manuscripts to the Digital Era, ed. by C. Clivaz, J. Meizoz, F. Vallotton and J. Verheyden, with B. Bertho (Lausanne: PPUR, 2012) 299–306 (ebook). Schulthess, S., ‘Les manuscrits du Nouveau Testament, le monde arabe et le digital. L’émergence d’un discours hybride’, in Lire Demain. Des manuscrits antiques à l’ère digitale / Reading Tomorrow. From Ancient Manuscripts to the Digital Era, ed. by C. Clivaz, J. Meizoz, F. Vallotton and J. Verheyden, with B. Bertho (Lausanne: PPUR, 2012), 333–344 (ebook).

82

sara schulthess

——, ‘Die arabischen Handschriften des Neuen Testaments in der zeitgenössischen Forschung: ein Überblick’, Early Christianity, Volume 3, Number 4 (December 2012), 518–539. Valentin, J., ‘Les Evangéliaires arabes de la bibliothèque du Monastère Ste-Catherine (Mont Sinai): Essai de classification d’après l’étude d’un chapitre (Matth. 28). Traducteurs, réviseurs, types textuels’, Le Muséon 116 (2003), 415–477.

the falasha memories project.1 digitalization of the manuscript bnf, ethiopien d’abbadie 107 Charlotte Touati 1. The State of the Falasha Studies The term ‘Falasha’ is used to refer to Ethiopian Jews.2 Their origins are unclear, both because of a lack of documentation, and also because Falashas form a kind of mythic tribe at the heart of numerous legends.3 The most ancient sources are narratives from medieval chroniclers and travellers, whose accounts date from the time when the Falashas ruled their own kingdom between Tana Lake and Simien Mountains, North Ethiopia. Their political independence could indicate that they had been settled there for a long time, but this is only speculation. Falashas are frequently presented as the descendants of Solomon and the Queen of Sheba or the lost tribe of Dan. No scientific evidence can support these assumptions.4 Nevertheless, the Chief Rabbi Ovadia Joseph recognized the Jewishness of the Falashas5 in 1972, on the argument they were the tribe of Dan. His decision was nothing but symbolic, since it prompted the Israeli government to grant them the right to return in 1975.6

1  The Falasha Memories project was formalized on the occasion of the call for application published by Infoclio.ch, the Swiss Portal for Historians related to the Swiss Academy of Humanities and Social Sciences, to encourage the publication of historical studies on the Internet. Our project met with success and was awarded funding. 2 The Ge’ez word ‘Falasha’ designates an exile or an expropriated peasant since a fifteenth century decree, cf. Kaplan, ‘Betä Esra’el’. The Falashas prefer to call themselves ‘Beta Israel’, nevertheless, custom and scientific literature have retained ‘Falasha’. See for example the classical books: Aescoly, Sefer ha-Falashim, and Aescoly, Recueil de textes falachas. So, we will use this term without any form of qualification or disparagement. 3 Kaplan, ‘The invention of Ethiopian Jews: three models’; Salamon, The Hyena People: Ethiopian Jews in Christian Ethiopia; Abbink, Mytho-légendes et histoire: l’énigme de l’ethnogenèse des Beta Esra’el; Abbink, ‘A socio-cultural analysis of the Beta Esra’el as an ‘infamous group’ in traditional Ethiopia’. 4 Cf. Kaplan, ‘Betä Esra’el’, 552–553. 5 Cf. Kaplan, ‘The Beta Israel and the rabbinate: law, ritual and politics’; Rapoport, Les Falashas d’Ethiopie. Une communauté en perdition, 19; 61. 6 Cf. Rapoport, Les Falashas d’Ethiopie, 29; 201–215; Kessler, The Falashas. The forgotten Jews of Ethiopia, 147–169.

84

charlotte touati

Years were spent in political considerations to postpone their repatriation, but in 1984, as Ethiopia was ravaged by one of its most severe famines,7 Israeli government and army suddenly activated the famous ‘Moses Operation’, allowing the departure of the Falasha through airlifts. Today almost all of the Falashas are settled in the State of Israel, where they are progressively assimilated, losing their genuine characteristic culture.8 Falashas’ Ge’ez memory, the language in which their sacred texts are written, gradually fades and will inevitably disappear if even the scientific community ignores it. 2. Falasha Memories 2.1. Description of the Project ‘Falasha Memories’ refer to the writings copied in the manuscript from the Bibliothèque Nationale de France, Ethiopien d’Abbadie 107, which contains all the sacred texts of the Falashas. We propose to develop a website9 presenting the facsimile of manuscript 107, a Ge’ez transcription, and two translations into modern languages, French and English (Amharic and Hebrew are not excluded in a further version of the website). The results will be published in open access, with a dynamic interface intended to switch between the different versions (facsimile, transcription and translations) thanks to tags, as well as a lexical search engine (English, French, Ge’ez), the first of the kind developed on the Internet to date. The transcription of the ms. Ethiopien d’Abbadie 107 will comply with the standards of modern editing10 with apparatus and philological

7 Cf. Rapoport, Les Falashas d’Ethiopie, 1­26. 8 Cf. N’Diaye, Les Falachas, Nègres errants du People juif. Enquête historique; Anteby, Les Juifs éthiopiens en Israël: les paradoxes du paradis; Kaplan and Salamon, Ethiopian Immigrants in Israel: Experience and Prospects; Kaplan and Rosen, ‘Ethiopian Jews in Israël’; Parfitt and Trevisani Semi, Ethiopian Jews in Ethiopia and Israel; Friedmann and Santamaria, Les Enfants de la Reine de Saba. Les Juifs d’Ethiopie (Falachas) histoire, exode, integration, 127–321; Friedmann and Santamaria, ‘Identité et changement: le cas des Falasha entre l’assimilation en Ethiopie et l’intégration en Israël’; Ashkenazy and Weingrod, Ethiopian Jews and Israel; Abbink, The Falashas in Ethiopia and Israel: The Problem of Ethnic Assimilation.   9 In collaboration with the Digital Humanities Lab at the École Polytechnique Fédérale de Lausanne. 10 According to the model exposed in Haile, Catalogue of the Ethiopic Manuscript Imaging Project, vol. 1: Codices 1–105, Magic Scrolls 1–134.    



the falasha memories project

85

c­ ommentary, but in the form of hypertext links rather than footnotes in order to allow the reader to access them, whatever version he/she is reading (transcription or translations). This multidimensional conception of one single text cannot be simply displayed on a sheet of paper. We intend to try to do something new, not just to post a traditional book on the Internet (pdf or another format). We plan to edit texts into context, that is to say to trap them in a straight web of allusions, references and confluence points. Regarding the ms. Ethiopien d’Abbadie 107, the background is as follows. Antoine d’Abbadie d’Arrast, a French geographer and a linguist, toured Ethiopia from 1838 to 1849. After more than a decade of exploration, he came back to France with an impressive collection of Ethiopian manuscripts, which constitutes today the ‘fonds d’Abbadie’ of the Bibliothèque Nationale de France. The books are currently stored in the Department of Oriental Manuscripts, Richelieu site of the BNF. As did most scholars of his time, Antoine d’Abbadie subtracted artifacts, namely manuscripts, from their original milieu. We intend to restore part of this context through archival documents. Indeed, Antoine d’Abbadie has meticulously recorded the details of his travels in his literary work,11 but his unpublished papers, preserved in the family property, Abbadia Castle (Hendaye, Atlantic Pyrenees, France), will produce even more valuable information. With hypertext links, it will be possible to indicate the place where Antoine d’Abbadie found each text, its use in liturgy, the oral development about the figures mentioned in it (all this data forms the corpus of cross-references). Pointing out names and key concepts in a Falasha writing (either in Ge’ez, in English or in French) will allow readers to obtain cultural, literary or theological information, as well as to find the other occurrences of the same word in the rest of the ms Ethiopien d’Abbadie 107. The return to France of Antoine d’Abbadie and his testimony to the existence of a Jewish population in Ethiopia12 inspired curiosity among the scientific community, but equally, what is more important to us, aroused the interest of the emerging Alliance Israélite Universelle (1860). The founders of this institution intended to identify, record and, when necessary, defend the Jews anywhere in the world. Hearing about the 11  d’Abbadie, Lettre de M. Antoine d’Abbadie à Lord Clifford sur l’Abyssinie; d’Abbadie, L’Abyssinie et le roi Théodore; d’Abbadie., ‘Sur les Abyssins’. 12 Cf. Aescoly, ‘Notices sur les Falacha ou Juifs d’Abyssinie, d’après le journal de voyage d’Antoine d’Abbadie’.

86

charlotte touati

Falashas, the authorities of the Alliance Israélite Universelle quite logically mandated Joseph Halévy, a renowned linguist, to go to Ethiopia in the fall of 1867. Joseph Halévy was profoundly convinced that the Falashas were true Jews as he explained in his conclusive report to the Alliance Israélite Universelle.13 They lost the knowledge of Hebrew because of their isolation and the high degree of illiteracy among them. He even took advantage of it. According to Halévy, their naïve and unsophisticated conception of religion is an evidence of purity, which is rather a romantic argument, but he also reported that they are totally ignorant of rabbinical Judaism, which would mean that they separated from the main Judaism as early as the second century. Halévy assumed that their cult was very close to ancient Hebrew’s. Thus, Joseph Halévy’s purpose was totally different to d’Abbadie’s since the first one was only interested in the Falashas, their religion and their culture. Halévy’s reports and books are precious because they are not only nearly contemporaneous, but also complementary to the ms. Ethiopien d’Abbadie 107. They will provide our second source of cross-references. 2.2. Method The progress of the project is planned in three steps, corresponding to the three corpora: – The ms. Ethiopien d’Abbadie 107 – The writings of Antoine d’Abbadie – Joseph Halévy’s report and books Each step consists of a heuristic stage followed by the publication of the result. The first step’s publication is the building of the website core, on which the further two stages will depend. Thus the data from the last two corpora, that is the cross-references, will be aggregated around the nucleus (ms. 107, the primary evidence). The choice of ms. Ethiopien d’Abbadie 107 is justified. Indeed Antoine d’Abbadie encountered difficulties in acquiring the Falasha manuscripts, because of their scarcity, but above all because their owners did not want to give or sell them. So, Antoine d’Abbadie asked to copy the manuscripts 13 Cf. Halévy, J., ‘Rapport de mission’; Nantet and Ochs, A la découverte des Falasha: le voyage de Joseph Halévy en Abyssinie (1867).



the falasha memories project

87

that he could see, failing to bring them back to France. The ms. Ethiopien d’Abbadie 107 is the result of this work, as can be confirmed by Antoine d’Abbadie himself in the catalogue of his own collection of Ethiopian manuscripts: N° 107.—ጸሎታተ፡ፈላሲያን Zalotata Falasyan “Prières des Falaša (Juifs)” 21 sur 14; deux colonnes; treize cahiers détachés ou 119 feuillets, dont 4 blancs. Je fis copier dans ce recueil toutes les prières des Falaša qu’il ne me fut pas possible d’acheter, afin d’y chercher plus tard quelque preuve intrinsèque de l’origine de ce peuple, plus singulier encore par sa foi que par sa langue.14

Even though it is not very old, this manuscript is nevertheless particularly important. There are two reasons why this is so. First, because Antoine d’Abbadie had access to copies which are lost today. Second, because the ms. Ethiopien d’Abbadie 107 is already a kind of edition. As he was very learned in Ge’ez, Antoine d’Abbadie selected models to copy not only for their content, but also for their linguistic or philological qualities. Here is another characteristic of the ms. Ethiopien d’Abbadie 107: the sleeping partner took indeed an active part in the copy, rereading and correcting it with his own hands.15 The ms. 107 is thus partly the work of Antoine d’Abbadie. The unique position of this manuscript allows us to use it as the core of the Falasha Memory project, around which ethnological, liturgical and literary data revolve, as well as biographical anecdotes from the travels of Antoine d’Abbadie and Joseph Halévy. Information that has a geographical location will be situated on a clickable and illustrated map. Finally, as the ms. Ethiopien d’Abbadie 107 is responsible for Halévy’s expedition and for the acknowledgement of the Falashas by the Alliance Israélite Universelle, it played an important part in the history of Judaism. Comparison with the few (fewer than twenty) other codices containing Falasha writings, which moreover did not undergo the editing work such 14 d’Abbadie, Catalogue raisonné des manuscrits éthiopiens appartenant à Antoine d’Abbadie, 119. 15 See figure 1 (below). On the folio 11r (right page), middle of the first column, several additions or corrections by Antoine d’Abbadie are readable. His pencil is narrow and the writing light by comparison with the added letters by the scribe on the first line of folio 10v (left page).

88

charlotte touati

Figure 1. BNF Ethiopien d’Abbadie 107, folios 10v–11r (Courtesy of the Bibliothèque Nationale de France).

as the one lead by Antoine d’Abbadie on his copy, reveals that the ms. Ethiopien d’Abbadie 107 is indeed an anthology, since it offers all the texts considered as Falasha in other manuscripts, while the opposite is not true. In simple terms, the other manuscripts are always shorter than the ms. Ethiopien d’Abbadie 107 and do not contain supplementary texts.16 The content of the ms. Ethiopien d’Abbadie 107 may be summarized as follows: – Hymns, fol. 1r–5v – Incantations, fol. 6r–10r – Excerpt from the Book of Isaiah, fol. 10v – Book of Gorgoryos, fol. 11v–20r – Testament of Abraham, fol. 20v–26v – Incantations, fol. 27r–v – Hymns, fol. 27v–32r 16 Cf. Aescoly, Recueil de textes falachas, 11–13, and the references given by Kaplan, ‘ “Falasha” Religion: Ancient Judaism or Evolving Ethiopian Tradition?’, 49–65.



the falasha memories project

89

– Arde’et and Acts of Moses, fol. 33v–48v – Hymns, fol. 48v–50v – Liturgical prayers of ‘Astari’ fol. 50v–75v – Tĕʾĕzâza Sanbat, 75v–119v To date, there is no synthetic study on the Falashas that respects equally the ethnographic and philological dimensions of the issue. We propose to remedy this shortcoming by establishing a database containing: – The facsimile of the ms. Ethiopien d’Abbadie 107, containing Falasha sacred texts – The transcription in Ge’ez script of all the writings from the ms. Ethiopien d’Abbadie 107 – The translation into English and French of these texts – The online publishing of previously unedited pieces concerning Falashas from d’Abbadie’s archives – The online publishing of extracts concerning Falashas from the written paper publications (catalogue, memoirs, letters, reports and essays) of Antoine d’Abbadie and Joseph Halévy – Links to documents concerning the Falashas, already posted by the Alliance Israélite Universelle – Reconstruction of the respective journey of Antoine d’Abbadie and Joseph Halévy on interactive maps, linked to photographs and extracts from the ms Ethiopien d’Abbadie 107 All this data will not only be brought together within a single site, but also networked to allow, in addition to the linear reading of each document, a personal and transverse reading. 2.3. Technical Features of the Website – After a homepage presenting the objectives of the Falasha Memory project, a visual journey through the different sources (corpus of texts and cross-references) in the form of a clickable timeline will present all the documents at the reader’s disposal, in order to make the history of the corpus available. – Each source document will be referenced according to archival standards, indicating precisely what is known about its history. This information will be exportable to international standards for archival description.

90

charlotte touati

– The website dedicated to the project will include a document viewer with both the digital copy of the manuscript (in picture mode), transcription (Ge’ez alphabet via CSS) and translations (Latin alphabet). – A system for matching line by line will be set up, so as to allow the reader to switch between the three interfaces (manuscript images, transcription and translation) and to compare the different versions. – Cross-references and hypertext links will be added to each document. – Places and people/characters encountered in the corpus will be tagged to form a separate thematic index. – Each text, commentary or reference will be indexed to enable a full text search in the source corpus (transcription and translation), the crossreferences as well as in the notes. The research will be carried out in Latin characters or Ge’ez alphabet thanks to a special keyboard. – All documents on the website will be published in open-source. 3. Contribution of the Project The Falasha Memory project is built around two main points of interest: the culture of the Falashas and the history of sciences in the nineteenth century. Moreover it sheds light on very early stages of the Judaeo­Christian literature. The publication of such texts not only shed light on Jewish apocryphal literature, it is also valuable to inform about Christian literature since some of these writings are available in two versions, respectively Jewish and Christian. Henceforth, they will be freely at the disposal of scholars. By supplying scientific editions of primary sources, the Falasha Memories’ website is also conceived to leverage peer pressure. We intend to draw attention to areas of Ethiopian, Jewish and European history, through a modern interface, as well as to contribute to the long history of literary materials. The Falasha Memory project is also a reflection on the formation process of bibliophilic collections in scholarly circles since the nineteenth century. On the level of diffusion, opting for the electronic format enables us to disclose publicly and all over the world texts formerly subtracted from their local rooting for private use. Indeed, it is possible to say that the ms. Ethiopien d’Abbadie 107 has been partially formatted by the expectations of its first reader, namely Antoine d’Abbadie, who intended to work as a historian. The Falasha Memory website will thus also illustrate the impossibility of objectivity



the falasha memories project

91

in historiography and may serve as a pedagogical example showing how 1) the conception of a document 2) its use 3) its conservation and 4) its accessibility are prone to vary under different circumstances. As many other European collectors, Antoine d’Abbadie subtracted artifacts from their genuine context. It is impossible to recover this context, but thanks to worldwide web, objects take another dimension. They can be connected to various other objects, information and people. Falashas’ sacred texts, the archive documentation about them and multimedia data extracted from it will be accessible to the Beta Israel people now settled in Israel. Falasha Memory website is intended as a work in progress, but under control. Which means that new data, new links will constantly be added, but carefully selected for their scientific value and their pertinence to the original project. References Sources d’Abbadie, A., Lettre de M. Antoine d’Abbadie à Lord Clifford sur l’Abyssinie (Lyon: Pélagaud et Cie, 1853). ——, L’Abyssinie et le roi Théodore (Paris: Douniol, 1868). ——, ‘Sur les Abyssins’, Bulletins de la Société d’anthropologie de Paris 4 (1869), 164–165. Aescoly, A.Z., Sefer ha-Falashim (Jerusalem: R. Mas, 1943). ——, Recueil de textes falachas (Travaux et mémoires de l’Institut d’ethnologie 55) (Paris: Institut d’ethnologie, 1951). Halévy, J., ‘Rapport de mission’, Bulletin de l’Alliance Israélite universelle (01.01.1868), 58–60. Studies d’Abbadie, A., Catalogue raisonné des manuscrits éthiopiens appartenant à Antoine d’Abbadie (Paris: Imprimerie impériale, 1859). Abbink, G.J., The Falashas in Ethiopia and Israel: The Problem of Ethnic Assimilation (Sociaal Antropologische Cahiers 15) (Nijmegen: ICSA, 1984). ——, ‘A socio-cultural analysis of the Beta Esra’el as an “infamous group” in traditional Ethiopia’, Sociologus 37 (1987), 140–154. ——, Mytho-légendes et histoire: l’énigme de l’ethnogenèse des Beta Esra’el (Bruxelles: CEDAF, 1991). Aescoly, A.Z., ‘Notices sur les Falacha ou Juifs d’Abyssinie, d’après le journal de voyage d’Antoine d’Abbadie’, Cahiers d’études africaines 2 (1961), 84–147. Anteby, L., Les Juifs éthiopiens en Israël: les paradoxes du paradis (Cahiers du Centre de recherche français de Jérusalem. Série ‘hommes et societés’ 8) (Paris: CNRS Éditions, 2004). Ashkenazy, M. and A. Weingrod, Ethiopian Jews and Israel (New-Brunswick: Israel Social Science Research, 1987). Friedmann, D. and U. Santamaria, ‘Identité et changement: le cas des Falasha entre l’assimilation en Ethiopie et l’intégration en Israël’, Archives européennes de sociologie 30 (1989), 90–119.

92

charlotte touati

——, Les Enfants de la Reine de Saba. Les Juifs d’Ethiopie (Falachas) histoire, exode, integration (Paris: Métalié, 1994). Haile, G. et al., Catalogue of the Ethiopic Manuscript Imaging Project, vol. 1: Codices 1–105, Magic Scrolls 1–134 (Ethiopic Manuscripts, Texts, and Studies Series 1) (Cambridge: James Clarke & Co, 2009). Kaplan, S., ‘ “Falasha” Religion: Ancient Judaism or Evolving Ethiopian Tradition?’, The Jewish Quarterly Review 79 (1988), 49–65. ——, ‘The Beta Israel and the rabbinate: law, ritual and politics’, Social Science Information 27 (1988), 357–370. ——, ‘The invention of Ethiopian Jews: three models’, Cahiers d’études africaines 33 (1993), 645–658. ——, ‘Betä Esra’el’, in Encyclopaedia Aethiopica I, ed. by S. Uhlig (Wiesbaden: Harrassowitz Verlag, 2003), 552–559. Kaplan, S. and C. Rosen, ‘Ethiopian Jews in Israël’, American Jewish Yearbook 94 (1994), 59–109. Kaplan, S. and H. Salamon, Ethiopian Immigrants in Israel: Experience and Prospects (London: Institute for Jewish Policy Research, 1998). Kessler, D. The Falashas. The Forgotten Jews of Ethiopia (New York: Holmes & Meier, 1982). Nantet, B. and E. Ochs, A la découverte des Falasha: le voyage de Joseph Halévy en Abyssinie (1867) (Paris: Payot & Rivages, 1998). N’Diaye, T., Les Falachas, nègres errants du people juif. Enquête historique (Paris: Gallimard, 2004). Parfitt, T. and E. Trevisani Semi, eds., Ethiopian Jews in Ethiopia and Israel (Richmond, Curzon Press, 1998). Rapoport, L., Les Falashas d’Ethiopie. Une communauté en perdition (Paris: Lattès, 1983). Salamon, H., The Hyena People: Ethiopian Jews in Christian Ethiopia (Contraversions 13) (Berkeley, Los Angeles: University of California Press, 1999).

Part Two

Digital Academic Research and Publishing

the seventy and their 21st-century heirs. the prospects for digital septuagint research Juan Garcés The Septuagint1 (= LXX) is in many ways a remarkable collection of texts. It represents the first known attempt to translate the Hebrew Bible into an Indo-European language, namely Hellenistic Greek. As such, it also represents an invaluable source for understanding pertinent linguistic, translational, text-critical, socio-cultural, and philosophical-theological issues that led to its creation and reception. Spanning in its inception from the first translations in the early- to mid-third century BCE to the later revisions in the second century CE, it gives scholars an insight not only into the development of the Greek language, but also into the influence of a Semitic language on its vocabulary and possibly even its syntax. Furthermore, being one of the rare cases where both the ancient translated text and source text are extant, at least in approximations, it also offers a rich source of insight into translation techniques and philosophies, and helps in establishing the possibility of a clearer understanding of other Greek texts, generally deemed to be translations from Semitic originals. Last, but not least, is its reflection of the culture and ideology of diaspora communities in the Eastern Mediterranean metropoles, which led to the emergence and shaping of two important religious groupings—Judaism and Christianity. The Septuagint is unique in having been both canonised and de-canonised by these two world religions. It should therefore hardly come as a surprise that the academic study of the Septuagint experienced an impressive growth in interest and research output in the twentieth century. Last century has not only seen the important founding of the International Organization for Septuagint and Cognate

1 The term, going back to the legend of the seventy(-two) translators of the Pentateuch as preserved in the pseudepigraphical Letter of Aristeas to Philocrates, is potentially ambiguous (see, e.g., Dines, The Septuagint, 1–3). In the context of this essay, it is understood here in a comprehensive, wider sense encompassing the writings of the canon collected by Alfred Rahlfs in his Septuaginta, including its complex transmission history and relation to cognate texts.

96

juan garcés

Studies (= IOSCS) as the institutional home for LXX scholars,2 with its own peer-reviewed periodical,3 regular congresses, monograph series,4 but also a series of seminal translation projects of the LXX into English,5 French,6 German,7 Spanish8 and other languages, as well as commentaries,9 which has further invigorated an already active research community. My vantage point, from which I would like to engage LXX Studies, is Digital Humanities (= DH), another fast-growing academic area, ‘born of the encounter between traditional Humanities and computational methods’10 and populated by a highly inter-disciplinary community of practice, of which I partake. DH is very much a product of the contemporary digitisation of society at large and, increasingly, academic research, in particular. The present networked information age not only exerts an ultimately inescapable influence on all aspects of the production of knowledge— workflows, conceptualisations, methods, etc.—it also calls for a critical reflexion on the limits, challenges and possibilities of digital research. In dealing with Humanistic subject matters, traditional Humanities are understood and constructively engaged as a product of its medialisation,

  2 The IOSCS (http://ccat.sas.upenn.edu/ioscs/, last accessed 8 May 2013) was founded at the organisational meeting on 19 December 1968 during the annual meeting of the Society of Biblical Literature.   3 The Journal for Septuagint and Cognate Studies (formerly the Bulletin of the International Organization for Septuagint and Cognate Studies).   4 The Society of Biblical Literature’s Septuagint and Cognate Studies series (http:// www.sbl-site.org/publications/Books_SepandCog.aspx, last accessed 8 May 2013).   5 Pietersma, ed., A new English translation of the Septuagint and the other Greek translations traditionally included under that title. See http://ccat.sas.upenn.edu/nets/edition/ for an electronic edition.   6 La Bible d’Alexandrie, multiple volumes (Paris: Les Éditions du Cerf, 1986ff). See http://www.editionsducerf.fr/html/index/collection.asp?n_col_cerf=239&id_theme=1&id_ cat=255 for a list of available titles and http://septante.editionsducerf.fr/ for an introductory text by Marguerite Harl, one of the editors. Last accessed 8 May 2013.   7 Karrer and Kraus, eds., Septuaginta Deutsch: das griechische Alte Testament in deutscher Übersetzung.   8 Fernández-Marcos and Spottorno Díaz-Caro, co-ord., La biblia griega: Septuaginta, 5 volumes.   9 In English alone, two commentary series are underway: the Society of Biblical Literature Commentary on the Septuagint with Rob Hiebert and Ben Wright as editors-in-chief (http://ccat.sas.upenn.edu/ioscs/commentary/, last accessed 8 May 2013) and Septuagint Commentary series, edited by Stanley E. Porter, Richard S. Hess, John Jarick and published by Brill (http://www.brill.com/publications/septuagint-commentary-series, last accessed 8 May 2013). 10 Burdick, Drucker, Lunenfeld, Presner and Schnapp, Digital_Humanities [sic], 3. Digital_Humanities, which is available Open Access as a downloadable pdf file from MIT Press at http://mitpress.mit.edu/sites/default/files/titles/content/9780262018470_Open_Access_ Edition.pdf is a suitable point-of-entry to DH. Last accessed 8 May 2013.



the seventy and their 21st-century heirs

97

and its approaches are supplemented (and sometimes challenged) in view of the realities of the new dispensation. Historians of the beginnings of Christian philology should readily appreciate the influence of medium and technology on the construction of knowledge about the LXX. In their beautifully argued book Christianity and the Transformation of the Book,11 Anthony Grafton and Megan Williams make their case for the transformative force of third-century CE book production and its effects leading to ‘new ways of organizing scholarly inquiry and collaboration’.12 Much like contemporary LXX researchers, ‘the scholars of Christian Caesarea lived in a time of seismic cultural change, a time when one regime of book production and storage supplanted another, and when the nature and practices of Christian scholarship were being redefined’.13 In the case of the Caesarean scholars, the change was due to the fact that the ‘basic physical form of the book was in a state of flux’.14 In terms of material, papyrus was being replaced by parchment, but more importantly, in terms of form, the scroll was yielding to the technically more sophisticated and versatile codex. The ‘codification’ of writing ‘threw into question existing assumptions regarding the natural relation between the book as material object and as unit of meaning’.15 Origen’s Hexapla, described by Grafton and Williams as ‘an elaborate tool for textual criticism of the Hebrew Scriptures . . . written in six parallel columns laid out across each opening of a series of massive codices’,16 a pivotal tool for early LXX research, illustrates their argument persuasively: The epochal importance of the work lies above all in its arrangement. The Hexapla was perhaps the first book—as opposed to official documents— ever to display information in tabular form: in columns intended to be read across rather than down the page.17

Origen’s path-breaking innovation in terms of technology and layout and its epistemological consequences, however, shouldn’t be seen too hastily as the work of an isolated genius. The paradigm shift was facilitated and

11  Grafton and Williams, Christianity and the Transformation of the Book: Origen, Eusebius, and the Library of Caesarea. 12 Grafton and Williams, Christianity and the Transformation of the Book, xi. 13 Grafton and Williams, Christianity and the Transformation of the Book, 6. 14 Grafton and Williams, Christianity and the Transformation of the Book, 10. 15 Grafton and Williams, Christianity and the Transformation of the Book, 12. 16 Grafton and Williams, Christianity and the Transformation of the Book, 17. 17 Grafton and Williams, Christianity and the Transformation of the Book, 17.

98

juan garcés

also relied on transformed workflows, collaborations, and infrastructures, ranging from the production of the parchment leafs, ink, and bindings to the careful planning and execution of the first copies, to the custodial work of librarians and subsequent publications, and, lest we forget, the generous funding from patrons. My modern point of departure for this essay occurs in the aftermath of further cultural-historical transformations—the invention of the Internet and the PC. For the LXX, of course, was one of the few ancient texts to receive early concerted applications of computer-assisted approaches. The most prominent project making use of electronic data management for the study of the LXX is still the Computer Assisted Tools for Septuagint Studies (= CATSS) project, which was called into life through the initiative of the IOSCS in the early 1970s.18 Located at the University of Pennsylvania’s emerging Center of Computer Analysis of Texts (= CCAT) and directed by Robert Kraft and Emanuel Tov, this project sought the use of the available computing resources towards three goals: 1. the morphological analysis of the Greek text, resulting in a tagged electronic text, 2. the comparison of the texts of the Hebrew Bible and its Greek translation, resulting in an electronic parallel aligned Hebrew and Greek text, and 3. the recording of published critical variants, resulting in an electronic Greek text, with encoded variants. The three modules are now freely available—upon the signing of a user agreement to/declaration of fair use—at the CCAT’s gopher19 and form the basis in some way or another of most, if not all, current electronic LXX text resources, research projects and studies making use of computing technology. Following a brief description of this pioneering project, this essay will align each of the modules to a seminal area of LXX research—in turn: lexicography, translation technique, and textual criticism. Using selected examples of the use of CATSS and comparable resources in LXX research and other relevant work, I will take advantage of hindsight to assess the 18 Tov, ed., A Classified Bibliography of Lexical and Grammatical Studies on the Language of the Septuagint and Its Revisions. 19 http://ccat.sas.upenn.edu/gopher/text/religion/biblical/. They are also linked from the IOSCS website http://ccat.sas.upenn.edu/ioscs/#resources. Last accessed 8 May 2013.



the seventy and their 21st-century heirs

99

prospects of digital LXX research in the context of DH. In doing so, I will be looking at two closely interrelated issues and ask myself what methodologically innovative approaches are facilitated by the refashioning of the LXX as digital data and, conversely, what kind of resources are desirable in order to facilitate approaches that cater to the advantages of digital research on the LXX. 1. CATSS’s Three Modules As mentioned above, CATSS pursued a threefold goal. This threefold goal resulted in three modules, each aimed at one of the specific goals. These are to be briefly described here: 1.1. The Greek Morphological Analysis Module (Morph Module) The files available at CCAT’s gopher are simple text documents featuring the text of Alfred Rahlfs’ 1935 hand edition of the Septuagint.20 The text is presented vertically with one word in each line, interrupted by an empty line and a line with the verse number at the beginning of each verse. The Greek characters are transliterated into ‘beta code’, where Latin characters represent Greek characters, diacritics and punctuation. In the most basic tabular format, these words are aligned to morphological codes— ‘type’ codes of up to three characters, identifying the part of speech (the main categories being: noun [‘N’], adjective [‘A’], pronoun [‘R’], conjunction [‘C’], particle [‘X’], interjection [‘I’], indeclinable number [‘M’], preposition [‘P’], adverb [‘D’], verb [‘V’], with various subcategories), followed by ‘parse’ codes of up to six characters. Finally, each line also gives the ‘dictionary entry’ for each word. The simplicity and consistency of the well-documented format makes it relatively straight-forward to transform these files into subsequent standards, such as Unicode for the Greek text and (TEI) XML for the tagging. 1.2. The Parallel Hebrew // Greek Text Module (Hebrew // Greek Module) The second module of the CATSS resources juxtaposes semantic equivalent words or phrases on each line. It basically offers two tab-delimited 20 Rahlf, Septuaginta, id est Vetus Testamentum Graece iuxta LXX interpretes, 2 volumes. This text edition was subsequently published and re-printed in one volume. In 2006, The German Bible Society published a text revised by Robert Hanhart.

100

juan garcés Gen 1:1 E)N P E)N A)RXH=| N1 DSF A)RXH/ E)POI/HSEN VAI AAI3S POIE/W O( RA NSM O( QEO\S N2 NSM QEO/S TO\N RA ASM O( OU)RANO\N N2 ASM OU)RANO/S KAI\ C KAI/ TH\N RA ASF O( GH=N N1 ASF GH=

Figure 1. Genesis 1:1 in CATSS’s Morph module, section taken from http:// ccat.sas.upenn.edu/gopher/text/religion/biblical/lxxmorph/01.Gen.1.mlxx, last accessed 8 May 2013.

columns: one featuring the Masoretic Text, according to the ‘MichiganClaremont BHS consonantal text’, in beta code, written from left to right; the other column featuring the aforementioned Rahlfs text (without morphological tagging). This module is, at first glance, the most straightforward of the three modules. A closer look, however, discloses a number of challenges. To start with, the Greek translation was not made from the Masoretic Text but from a no longer extant earlier Hebrew text. In those cases where the Hebrew source text, of which the Septuagint is translation, clearly differed from the Masoretic Text, the first column is split into two sub-columns, the second sub-column featuring a reconstructed (‘retroverted’) Hebrew text, as well as notes on the translation technique. This module was further developed by Emanuel Tov and his team at the Hebrew University in Jerusalem. 1.3. The Greek Textual Variants Module (Variants Module) The third module is, out of the three, the least developed one and has undergone several formatting changes in subsequent files. The approach taken for the initial format of this module was to take the ‘Göttingen’ text of the Septuagint (see below 2.3)—where it exists, otherwise the aforementioned Rahlfs hand edition was used—and add ‘all significant textual variations in the Greek tradition’21 by aligning the apparatus variorum of 21 Kraft, ‘CATSS: Treatment of the Greek textual variants’.



the seventy and their 21st-century heirs

101

Gen 1:1 B/R)$YT E)N A)RXH=| BR) E)POI/HSEN )LHYM O( QEO\S )T H/$MYM TO\N OU)RANO\N W/)T H/)RC KAI\ TH\N GH=N Figure 2. Genesis 1:1 in CATSS’s Hebrew // Greek module, section taken from http://ccat.sas.upenn.edu/gopher/text/religion/biblical/parallel/01.Genesis.par, last accessed 8 May 2013.

the major critical editions. This was done by offering the eclectic text, as in the other two modules, as a vertical text and integrating variant readings where appropriate. In other words, instead of following the standard print format of keeping text and apparatus in separate text blocks, the variant readings are interspersed between the main text units. While this approach necessitated ‘a great deal of reorganization and innovation in formatting has been required’, the advantage of this format is that it ‘allows for efficient searching of the file, analysis of textual groupings, recreation of the main text or of the consecutive text of particular witnesses and/ or families, etc.’ Again, the simple format, expressed in fixed-width characters, is CATSS’s strength. Each verse is preceded by a ‘locator’. ‘Main’ text (the ‘lemma’) starts in ‘column 1’ of each line. Where this lemma has variant readings, this is indicated by a square close bracket (‘]’), followed, if relevant, by the notation of witnesses that do not feature the lemma, prefixed by a close angle bracket (‘>’). The variant readings start in ‘column 2’ of a subsequent line. CATSS’s three goals and modules can intuitively be aligned to three approaches, which have proven particularly fruitful in Septuagint studies and are the focus of the remainder of this essay. 2. Lexicography and Lexical Semantics Robert Kraft, co-director of CATSS and one of the key people in the development of the project, published in 2004 an ‘autobiographical retrospective on [his] involvement with computers and textual studies’22 as his 22 Kraft, ‘How I met the computer, and how it changed my life’, SBL Forum, n.p. [cited April 2004]. Online: http://sbl-site.org/Article.aspx?ArticleID=246, last accessed 8 May 2013.

102

juan garcés #01 01 *E)N A)RXH=| E)POI/HSEN] : EPLASEN 664 : EKTISEN Ios I 27 (2nd) 0( QEO\S TO\N OU)RANO\N KAI\ TH\N GH=N

Figure 3. Genesis 1:1 in CATSS’s Variants module, section taken from a screenshot of http://ccat.sas.upenn.edu/gopher/text/religion/biblical/lxxvar/1Pentateuch/ 01Gen-beta.htm, last accessed 8 May 2013.

contribution to the Society of Biblical Literature’s SBL Forum. In this brief piece, he writes about the catalyst for the creation of the CATSS resources: one of the first aims of the newly-formed IOSCS, ‘in 1970 or soon thereafter’, was to produce ‘a new lexicon of ‘Septuagint’ Greek’ and to ‘creat[e] concordance type tools to facilitate the lexical work’.23 That the IOSCS would identify the creation of a Septuagint dictionary as one of its first priorities is hardly surprising. The most recent dictionary truly focusing on the Septuagint when CATSS started, Schleusner’s five-volumes opus Novus thesaurus philologico-criticus,24 was already very outdated in terms of its methodology and material, not least due to the discovery and publication of many Hellenistic Greek texts since its publication.25 The need 23 Kraft, ‘How I met the computer’. See Cameron Boyd-Taylor, ‘The Evidentiary Value of Septuagintal Usage: Alice’s Response to Humpty Dumpty’, 80, for how ‘the self-identity of the society [i.e. IOSCS] as a new and distinct formation within the larger scholarly community was very much bound up with this lexicographical impulse’. 24 Schleusner, Novus thesaurus philologico-criticus sive lexicon in LXX et reliquos interpretes graecos ac scriptores apocryphos veteris testamenti, 5 volumes. The out-ofcopyright work has been digitised and is now available for online view and download from the Bavarian State Library at the permalink http://opacplus.bsb-muenchen.de/ search?oclcno=39771930, last accessed 8 May 2013. 25 A key figure to integrate epigraphic and papyrological finds into Greek Bible studies was Adolf Deissmann. See his Bible studies: contributions chiefly from papyri and inscriptions to the history of the language, the literature, and the religion of Hellenistic Judaism



the seventy and their 21st-century heirs

103

for a new lexicon had been expressed repeatedly as far back as the beginning of the twentieth century. Since then, however, Septuagint lexicography has developed into a particularly fertile academic sub-field.26 Research on the LXX lexis has been shaped by the dialectics between two ostensibly antithetical paradigms, each departing from a different yet plausible insight. The first approach departs from the observation that the LXX is, generally speaking, a translation of the Hebrew (and Aramaic) Bible and should therefore be treated in juxtaposition with and, if necessary, linguistic subservience to the Hebrew Bible. The second approach departs from the equally cogent observation that Septuagint Greek represents, nevertheless, Hellenistic Greek or Koine and that it, therefore, should be understood first and foremost in its own right, without necessary recourse to its Hebrew source text. The first approach conceptualises the text as produced by the multilingual (Hebrew/Aramaic and Greek) translator(s), the second understands it as received by a contemporary Hellenistic reader without knowledge of Hebrew or Aramaic.27 The choice of paradigm has evidently methodological and conceptual consequences for LXX lexical semantics and lexicography. According to esteemed Greek lexicographer John Lee, the two approaches ‘to determine the meaning of a word in the LXX are fundamentally different’, ‘are not reconcilable and are bound to lead to different results when the conditions require them.’28 Fortunately, researchers subscribing to any one of the two approaches have now recourse to a dictionary based on their favoured paradigm: A Greek-English Lexicon of the Septuagint by Johan Lust, Erik and primitive Christianity; Deissmann, New light on the New Testament from records of the Graeco-Roman period; and Deissmann, Light from the ancient east: the New Testament illustrated by recently discovered texts of the Graeco-Roman world. A useful resource is still the now outdated lexicon by Moulton and Milligan, Vocabulary of the Greek Testament. This tradition has been kept up by the still ongoing New documents illustrating early Christianity volumes, published 1976–1987 and the Papyrologische Kommentare zum Neuen Testament commentary series (Göttingen: Vandenhoeck & Ruprecht). For a thorough use of this material in LXX lexicography, see Lee, A Lexical Study of the Septuagint Version of the Pentateuch. The cumulative result from this approach is the insight that the LXX is mirroring general Hellenistic Greek characteristics and developments, albeit not from a literary ‘register’, rather than idiosyncrasies resulting from the translator’s creative semantic innovations. 26 I refer the interested reader to the relevant sections of the bibliographies up to 1993: Brock, Firtsch and Jellicoe, comp., A classified bibliography of the Septuagint; Tov, ed., A Classified Bibliography of Lexical and Grammatical Studies on the Language of the Septuagint and Its Revisions; Dogniez, Bibliography of the Septuagint/Bibliographie de la Septante (1970–1993). For newer material see below. 27 See, e.g., Pietersma, ‘Beyond Literalism: Interlinearity Revisited’, 3–4. 28 Lee, ‘Review of T. Muraoka, A Greek-English Lexicon of the Septuagint (2009)’, 121.

104

juan garcés

Eynikel and Katrin Hauspie29 (= LEH) for the LXX Greek as translation Greek, and Takamitsu Muraoka’s equally titled A Greek-English Lexicon of the ­Septuagint30 (= MSL) for LXX as Hellenistic Greek proper.31 Each of the aforementioned dictionaries lays out its approach unambiguously in their respective introductions. LEH concludes: ‘Septuagint Greek’ is first of all translation Greek [italics JG]. Any lexicon of the LXX should, therefore, refer to the Semitic original, at least in those cases where the deviations between a Greek word and its Semitic equivalent can be explained at the level of the morphemes, but also when the Greek words as such are incomprehensible because they are transliterations or because they adopted the meaning of the underlying Hebrew or Aramaic.32

Whereas MSL states: ‘we regard the language of the Septuagint to be a genuine representative of the contemporary Greek, that is to say, the Greek of the Hellenistic and Early Roman periods’.33 It follows that information referring to its Hebrew/Aramaic source text, present in a previous instalment, ‘is not integral to the LXX lexicography as such, it has been deleted from this edition.’34 Both dictionaries share a fundamentally traditional format: headwords are alphabetically organised and listed with relevant morphological information for lemmata; a list of senses is given as translation glosses (LEH throughout; MLS occasionally) or definitions (MLS mostly) with example occurences (up to 5 in LEH and MLS ‘err[ing] on the generous side’);35 and a list of selected literature is featured.36 The differences between the lexica are as much due to differently nuanced approaches to lexicography

29 Lust, Eynikel and Hauspie, A Greek-English Lexicon of the Septuagint. 30 Muraoka, A Greek-English Lexicon of the Septuagint. See also his previous instalment A Greek-English Lexicon of the Septuagint: chiefly of the Pentateuch and the Twelve Prophets. 31  A third reference work covering the LXX vocabulary from a more inclusive vantage point, entitled Historical and Theological Lexicon of the Septuagint, edited by Eberhard Bons and Jan Joosten is forthcoming from Mohr Siebeck in Tübingen. See Bons and Joosten, eds., Septuagint Vocabulary: Pre-history, Usage, Reception, for the papers held at two preparatory colloquia and Eberhard Bons and Anna Passoni Dell’Acqua, A sample article: ‘κτίζω—κτίσις—κτίσμα—κτίστης’, in Septuagint Vocabulary, 173–187, for a sample entry. 32 Lust, Eynikel and Hauspie, A Greek-English Lexicon of the Septuagint, xxiv. 33 Muraoka, A Greek-English Lexicon of the Septuagint, ix. 34 Muraoka, A Greek-English Lexicon of the Septuagint, xv. 35 Muraoka, A Greek-English Lexicon of the Septuagint, xi. 36 I am merely providing an overview. For reviews of these dictionaries see, for example, see Lee’s review mention in footnote 21 and Taylor, ‘Review of Lust et al., A Greek-English Lexicon of the Septuagint’, pts. 1–2. http://www.reltech.org/TC/v03/Lust-etal1998rev.html, last accessed 8 May 2013.



the seventy and their 21st-century heirs

105

in general as they are to a subscription to either the interlinear or Hellenistic Greek paradigm to LXX lexicography. To name but two examples: (1) LEH provides very useful statistics about the occurrence of words in parts of the LXX, grouped into 5 groups for the Torah, the Early Prophets (including 1 & 2 Chronicles), the Later Prophets, the Writings (excluding 1 & 2 Chronicles), the ‘deuterocanonical’ books (excluding the additions to Daniel and Esther), as well as the total number—MSL does not offer such statistics; (2) unlike LEH, MSL provides crucial syntagmatic (e.g. subject and object in the broad sense of ‘deep structure’) and paradigmatic (e.g. lists of words or word groups semantically associated) lexicographic information. Does the LXX researcher not need both sets of information? And why only offer one set and not the other? LEH are explicit in their association with and use of the CATSS resources in creating their lexicon and in offering the morphological information from that very electronic resource.37 Considering that this is the case and LEH’s explicit advocacy of the interlinear paradigm, it is surprising how little use is being made of the Hebrew // Greek module. LEH does not seem to be interested in breaking down which Hebrew words are translated by a given lemma and how often and consistently a Hebrew-Greek word-pair can be established as equivalents within the source and target texts, thus potentially indicating patterns such as stereotypical ­renderings.38 The main CATSS module used seems to be, to all appearances, the Morph module and this is most probably related to the fact that this is a tool familiar to traditional lexicographic approaches, which ordinarily use a concordance approach to analyse lexical semantics. Having such a concordance in electronic format clearly makes it easier to manipulate the text and extract straight-forward statistics. One would have expected, however, that a lexicon departing from the bilingual nature of its Greek corpus would have made more substantial use of the CATSS Hebrew // Greek module in order to analyse patterns indicating cross-lingual transference. The design of a LXX lexicon based on an electronic corpus raises two questions: (1) What innovative methods are opened up for LXX lexicography? (2) What innovative information design and presentation are

37 Lust, Eynikel and Hauspie, A Greek-English Lexicon of the Septuagint, xii–xiii. 38 Ironically, it is Muraoka who has produced the more useful resource for this approach in A Greek ~ Hebrew/Aramaic two-way index to the Septuagint.

106

juan garcés

facilitated by not only using an electronic corpus but also creating an electronic lexicon? 2.1. Corpus-Based Lexicography and the LXX Corpus linguistics39 is one of the fastest growing approaches in linguistics and is particularly relevant for LXX lexicography. One key assumption made in corpus linguistics is that language can be best understood by analysing (often, by testing hypotheses) a finite amount of data (a ‘corpus’) under the conditions that this data (usually texts) has been produced in a ‘natural’ communication context (i.e. not specifically for the purpose of linguistic analysis), is representative of the linguistic variety to be analysed, and represents linguistic features of interest in a balanced (i.e. reflecting their proportional occurrence in reality) way. The difference between a corpus and a collection of texts, however, does not consist only of the quality of the data, it has a fundamental methodological implication: The . . . whole point of making something a corpus rather than a collection of texts is in order to observe things which cannot be directly observed because they are too far apart, they are too frequent or infrequent, or they are only observable after some kind of numerical or statistical process. The essence of the corpus as against the text is that you do not observe it directly; instead you use tools of indirect observation, like query languages, concordances, collocators, parsers and aligners.40

The central feature that can be inferred from a corpus is the distribution of a certain linguistic phenomenon, mainly its occurrence or nonoccurrence and its co-occurrence together with other phenomena, and, thus, quantitative methods and statistical modeling become crucial to this approach.41 It also follows that corpora should, ideally, be large: ‘The main virtue of being large in a corpus is that the underlying regularities have a better chance of showing through the superficial variations, and there’s a lot of variation in the surface realization of linguistic units in a corpus.’42 Can the LXX then be considered a corpus, albeit a small one? In terms of the LXX as finite data representing natural language, the answer has to be affirmative. Being a representative of Hellenistic Greek (with Semitic 39 For a comprehensive overview, see Lüdeling and Kytö, eds., Corpus Linguistics: An International Handbook, 2 volumes. For a quick introduction in the form of a fictitious dialogue, see Gries, ‘What is Corpus Linguistics?’. 40 Sinclair, ‘Current Issues in Corpus Linguistics’, 189. 41  Gries, ‘What is Corpus Linguistics?’, 2–4. 42 Sinclair, ‘Current Issues in Corpus Linguistics’, 189.



the seventy and their 21st-century heirs

107

interference), it is part of an ‘epigraphic’ language, where ‘the language is defined by its extant materials (most of which have been discovered and edited)’43 and these texts are unlikely to grow significantly. Can we therefore apply corpus-linguistic methods to the LXX? In a way, we cannot but use such approaches and they have been a (limited) part of Biblical philology for some time. For example, one of the key tools of corpus linguistics, the concordance, was first created in 1230 CE by a group of Dominican scholars led by Hugo de Saint-Chair who put together the Concordantiae Sacrorum Bibliorum or Concordantiae S. Jacobi on the Latin Vulgate. Methodologically, however, the philologia sacra has a lot of distance to make up in relation to the philologia profana, represented by corpus linguistics. The size of the LXX corpus is indeed a limiting factor: There is no virtue in being small. Small is not beautiful; it is simply a limitation. If within the dimensions of a small corpus, using corpus techniques, you can get results that you wish to get, then your methodology is above reproach—but the results will be extremely limited, and also the range of features that you can observe.44

LXX lexicography should therefore take full advantage of the analytical approaches to corpus-based lexicography while, at the same time, be aware that a number of assumptions concerning representativeness and balance have to be made and, given the relatively small size of the corpus, caution is demanded in view of the statistical conclusions drawn from corpus linguistic approaches to LXX lexicography. A useful example for corpus-based lexicography applied to the Greek Bible is O’Donnell’s collocational analysis of ἐγείρω and ἀνίστημι in the New Testament.45 Departing from Sinclair’s well-known triad of corpus— concordance—collocation,46 O’Donnell exemplifies the application of collocational analysis, i.e. which words regularly appear in the textual neighbourhood of a word in question. The analysis of collocational patterns can be undertaken on a simple, untagged electronic text. In the case of a highly inflected language as Hellenistic Greek, however, the analysis will benefit considerably from additional markup in terms of lexical and grammatical information.47 The immediate advantage of collocational analysis over a traditional lexicographic approach is apparent in 43 Brook O’Donnell, Corpus Linguistics and the Greek of the New Testament, 330. 44 Sinclair, ‘Current Issues in Corpus Linguistics’, 189. 45 Brook O’Donnell, Corpus Linguistics and the Greek of the New Testament, 314–396. 46 Sinclair, Corpus, Concordance, Collocation. 47 Brook O’Donnell, Corpus Linguistics and the Greek of the New Testament, 336, 341.

108

juan garcés

the ­additional data that becomes available for disambiguating different senses of the same lemma, for it shows that ‘where a word has a number of different senses, each sense is accompanied by a unique syntactical pattern’.48 The verbs ἐγείρω and ἀνίστημι in the New Testament are generally understood as synonyms. They are, however—possibly because of, rather than despite, their theological importance for a theological understanding of the resurrection from the dead—often translated inconsistently and, more importantly, analysed incompletely in terms of their base data. In progressive steps, O’Donnell shows how collocational analysis of these two lemmata might help in clarifying some of the lexico-grammatical (and theological) issues. A look at the collocates in a non-lemmatised or inflected New Testament corpus already shows an interesting pattern: ‘While νεκρός [‘dead’ (usually of a person); JG] appears to be a significant collocate of both verbs, it is more commonly found with ἐγείρω and in more varied uses than with ἀνίστημι.’49 The same exercise on a lemmatised text confirms and further refines this observation, since it further abstracts from the text by grouping differently inflected but identical words and eliminates orthographic differences. Νεκρός does indeed appear to a differently degree in the context (keyword analysed with four words on each side) of ἐγείρω (with 45 times the first most frequent collocate) and ἀνίστημι (12 times the tenth most frequent collocate). Also interesting are the observations that can be made when looking for collocates with a high semantic value for the keyword. In the case of keywords that are verbs, these tend to be nouns: ‘[T]he proper nouns occurring in the close co-text are likely to be of particular interest because they may help to identify common actors (the ‘doer’) and patients (the recipients) of the processes communicated by the verb forms.’50 For example: while Χριστός (Christ) is the 12th most frequent collocate of ἐγείρω, it is only the 30th most frequent collocate of ἀνίστημι; Πέτρος is the 13th most frequent collocate of ἀνίστημι but does not collocate with ἐγείρω in the New Testament at all. Similar analyses could be undertaken on the CATSS Morph module with, I expect, considerable benefit. O’Donnell shows however two further steps that would enhance the benefits gained from collocational analysis: one uses the result of further lexico-semantic classification, the

48 Brook O’Donnell, Corpus Linguistics and the Greek of the New Testament, 339. 49 Brook O’Donnell, Corpus Linguistics and the Greek of the New Testament, 342. 50 Brook O’Donnell, Corpus linguistics and the Greek of the New Testament, 345.



the seventy and their 21st-century heirs

109

other necessitates further syntactic markup of the corpus. Louw and Nida51 attempted to describe the underlying cognitive framework of New Testament Greek by classifying its vocabulary into semantic fields (‘domains’) and subfields, instead of the traditional lexicographic arrangement by morphology (alphabetically) or by grammatical function (part-of-speech). By further grouping collocates according to Louw/Nida domain classifications, O’Donnell demonstrates how a further step in abstraction from the untagged corpus adds another level in differentiating larger ­semantic patterns. The similarity of ἐγείρω and ἀνίστημι are underscored by their sharing of the first five most frequent domain collocates, albeit in a slightly different order.52 Also, both verbs share 24 out of the 28 most frequent domain collocates. However, the previous observation concerning the difference in relation to the collocate Χριστός (Christ) is further confirmed: ‘Collocates of ἐγείρω, not among the 28 most frequent of ἀνίστημι, are . . . Supernatural Beings and Powers (19th in list), . . . Affirmation and Negation (25th in list), . . . Hold a View, Believe, Trust (28th), and . . . Groups and Classes of Persons (27th).’53 O’Donnell’s last and most refined collocational analysis makes use of a grammatically annotated New Testament corpus.54 This approach shows that ‘both the lexical and grammatical choice of the keyword’55 are relevant when researching its semantics: ‘Just as two similar words, thought to be synonymous, may be found in the company of a very different list of nouns, adjectives and other words, the words may also exhibit very different grammatical patterns.’56 Such an analysis allows for ‘a more linguistically precise collocational analysis’.57 Engaging in a—for most Biblical scholars at least—relatively new methodology with its unfamiliar terminology and statistical models can be daunting. It can also, notwithstanding, present an opportunity to rethink fundamental assumptions. One such assumption might involve the basic unit of meaning, particularly concerning the relationship between form and meaning. In a highly inflected language such as Hellenistic Greek, 51  Louw and Nida, eds., Greek-English Lexicon of the New Testament Based on Semantic Domains, 2 volumes. 52 Brook O’Donnell, Corpus linguistics and the Greek of the New Testament, 349. 53 Brook O’Donnell, Corpus linguistics and the Greek of the New Testament, 350. 54 See Brook O’Donnell, Corpus linguistics and the Greek of the New Testament, 102–167 for his approach to annotation. 55 Brook O’Donnell, Corpus linguistics and the Greek of the New Testament, 357. 56 Brook O’Donnell, Corpus linguistics and the Greek of the New Testament, 357. 57 Brook O’Donnell, Corpus linguistics and the Greek of the New Testament, 368.

110

juan garcés

annotating a corpus with morpho-syntactic information and the lemma (i.e. the traditional dictionary entry form) for each word, as was the case with the CATSS morph module, makes a lot of sense. To infer automatically that these lemmata, even after having been disambiguated into possible homonyms, are the natural units of meaning, might be by-andlarge correct. It might also prove to be somewhat problematic. One relevant insight from corpus linguistics is the correlation between form and meaning, i.e. that ‘each meaning can be associated with a distinctive formal patterning’,58 to the degree that ‘[t]here is ultimately no distinction between form and meaning’:59 ‘There is a good case for arguing that each distinct form is potentially a unique lexical unit, and that forms should be conflated into lemmas when their environments show a certain amount and type of similarity.’60 If differences in meanings correlate with differences in formal patterns, these patterns can and should be described and their frequencies can and should be assessed as to their significance. The study of lexis is forced to move beyond the atomic word into the realm of multiple-word constructions and syntactic structures as part of its lexicographic analysis. In corpus linguistics, collocation focuses on ‘co-occurrence preferences and restrictions pertaining to specific lexical items’.61 Corpus-based lexicography has been ‘typically based on a more-or-less systematic interpretation of patterns emerging from a manual inspection of (i) a KWIC [i.e. keyword in context] concordance display providing the node word in its context and/or (ii) the node word’s collocates, i.e. frequent words within a user-specified span around the node word’.62 Combining the analysis of collocation with the analysis of construction,63 Stefanowitsch and Gries suggest a method they label ‘collostructional analysis’: ‘Collostructional

58 Sinclair, Corpus, Concordance, Collocation, 6. 59 Sinclair, Corpus, Concordance, Collocation, 7. 60 Sinclair, Corpus, Concordance, Collocation, 8. 61  Stefanowitsch and Gries, ‘Collostructions: Investigating the Interaction of Words and Constructions’. 62 Stefanowitsch and Gries, ‘Collostructions’, 214. In this respect, MSL represents a solid corpus linguistic approach: ‘In order to determine the meaning of a word, one needs to read it at least in the whole sentence of which it forms a part. It is further axiomatic to see it in paradigmatic relationships, namely by taking into account synonymic, antonymic, or some other semantically associated word or words with which it occurs’, x. 63 Construction is here understood as ‘any linguistic expression, no matter how concrete or abstract, that is directly associated with a particular meaning or function, and whose form or meaning cannot be compositionally derived’ (Stefanowitsch and Gries, ‘Collostructions’, 212).



the seventy and their 21st-century heirs

111

analysis always starts with a particular construction and investigates which lexemes are strongly attracted or repelled by a particular slot in the construction (i.e. occur more frequently or less frequently than expected)’.64 One of the main advantages of this method is that it provides an objective approach to understanding the meaning of words within particular grammatical constructions. Such methodological innovations will not be without their impact on the decisions as to whether what information should be included in a lexicon, especially if that lexicon is not restricted by the confines of the printed page, in other words: especially if that lexicon is an electronic one. 2.2. A Hellenistic Greek Lexicon as Digital Resource In dealing with the question of why and how to design an electronic lexicon for Hellenistic Greek, an even more fundamental question has to be asked at the outset: What kind of resource is a dictionary—what is or should be expected from it and to what use is it supposed to be put? John Lee, who has produced a history of New Testament lexicography65 and a useful overview of Ancient Greek lexicography,66 offers a refreshingly honest answer, while pointing out an important issue: Lexicons are regarded by their users as authoritative, and they put their trust in them. Lexicons are reference books presenting a compressed, seemingly final statement of fact, with an almost legal weight. The mere fact that something is printed in a book gives it authority, as far as most people are concerned. . . . Yet this trust is misplaced. The concise, seemingly authoritative statement of meaning can, and often does, conceal many sins—indecision, compromise, imperfect knowledge, guesswork, and, above all, dependence on predecessors.67

Lee has a number of legitimate misgivings with Hellenistic Greek lexica, in general, and Greek Bible lexica, in particular. Speaking more specifically of New Testament lexicography, Lee states: New Testament lexicography has failed to deliver the results one might expect from [the] long-sustained attention [it has experienced over the centuries]. Instead of a commodity that provides accurately described meanings and a reliable summation of the relevant data, we have haphazard coverage

64 Stefanowitsch and Gries, ‘Collostructions’, 214. 65 Lee, A History of New Testament Lexicography. 66 Lee, ‘The Present State of Lexicography of Ancient Greek’. 67 Lee, ‘The Present State’, 66.

112

juan garcés of the latter and a considerably flawed treatment of the former. The reasons for this outcome [are]: undue reliance on predecessors, an unsatisfactory method of indicating meaning, interference from translations, and inadequate means of gathering evidence and opinion.68

His conclusions are particularly intriguing in view of our posed question, since, according to Lee, the format of a resource that takes cognissance of and learns from the past and current shortcomings could only be the electronic format.69 The main reason for this is the possibility for collecting and organising the necessary data for such a desired resource that is offered by the electronic medium. The task is simply too large for the limitations of a print resource and too large to be accomplished by a specific finite project. Scope is an important aspect here. According to Lee, the LXX lexis can only be adequately understood and represented in a resource if the scope is extended in two directions, synchronically and diachronically. First, the scope needs to be extended in the direction of contemporary non-LXX usage of words, especially but not exclusively by including so-called documentary texts. Secondly, what is really needed is a comprehensive lexical database covering all of post-Classical Greek.70 In other words, Lee is proposing a radical overhaul of the lexicographical workflow—a critical evaluation of previous work and an extensive collaboration is called for—as well as a radical extension of the corpus to be considered. The envisioned electronic lexicon would therefore have to be an open collaboration that covers all of Hellenistic Greek. What kind of information would this lexicon provide and how would it do so? It would, of course, provide ‘an accurate description of the meaning of each Greek word’, functioning as ‘an ongoing, long-term means of assembling and assessing all the data relevant to the task’.71 Lee has been persuasive in arguing that word senses should be given as descriptions, not merely as translation glosses, and that the different word senses of a single lemma should be clearly differentiated. But more important for the user is the information that allows her to understand the formal conditions under which a specific word or phrase should be understood as denoting one sense over against alternatives. Traditional lexica offer little help in assessing which of the ‘meanings’ of the word listed is the one 68 Lee, A History of New Testament Lexicography, 177. 69 Lee, ‘A Lexicographical Database for Greek: Can It Be Far Off ? The Case of amphodon’. 70 See, e.g. Lee, ‘The Present State’, 72–74. 71  Lee, ‘A Lexicographical Database for Greek’, 215.



the seventy and their 21st-century heirs

113

that should be assumed for a specific use of the word: ‘The interpreter or translator is given no indication of either the relative frequency of each option or the grammatical context (co-text) in which it might occur. Instead the interpreter is left to ‘try out’ the different possibilities and see which one fits best’.72 Such information would be key for such a lexicon. It would therefore be crucial to not only offer word senses and up-to-date bibliographies, but also the syntagmatic and pragmatic information given in MSU, alongside all statistically significant linguistic patterns (including the aforementioned collostructional analysis) with which the word in question is associated. There is, however, still the important issue, mentioned at the beginning of this section, of how the LXX’s language relates to Hellenistic Greek in general and the related question of whether, unlike the two available LXX lexica, an electronic lexicographical resource, such as the one proposed here, could cater for researchers subscribing to both the interlinear and the compositional paradigms, or, as would be my preferrence: whether it could cater for somebody who would like to be able to engage in both production-centered and (first) reception-centered linguistic analysis. Albert Pietersma, a scholar associated with the interlinear paradigm, stresses the role of the ‘paradigm’ as metaphor and heuristic tool to understand the LXX’s language, rather than a historical explanation of the LXX’s origin— ‘a metaphor, a working hypothesis made into a paradigm’.73 Despite the sometimes hard rhetorics employed in the dialogue between researchers from each camp, there is a surprisingly large common ground between the two groups. There would certainly be agreement as to the foundational historical resources to be interpreted following either paradigm: ‘What we do have . . . is (1) the Greek translation itself, (2) by and large the source text from which this translation derives, and (3) compositional literature in Greek from the historical period in question.’74 Even the basic assumption is shared: ‘the LXX is “normal” [Hellenistic; JG] Greek until proven otherwise’.75 The point where the two paradigms part ways concerns the moment when different competing word senses do not seem to be easily disambiguated or even do not seem to have a ‘normal’ Hellenistic Greek sense at all, in other words: when LXX Greek ceases to be ‘normal’ 72 O’Donnell, Corpus linguistics and the Greek of the New Testament, 321. 73 Pietersma, ‘Beyond Literalism’, 17. 74 Pietersma, ‘Beyond Literalism’, 14. 75 Pietersma, ‘Response to Muraoka’, http://ccat.sas.upenn.edu/nets/discussion/­ pietersma-re-muraoka.pdf, 12, last accessed 8 May 2013.

114

juan garcés

Greek because of interference and seems in need of the Hebrew source text, in order to make sense. This, of course, begs the question: When do LXX words depart from ‘normal’ Hellenistic Greek use? In order to answer this question we, and the aforementioned electronic lexicographical resource, need a representative and balanced Hellenistic Greek corpus,76 which we do not have at the moment.77 Only then can we assess how the use of a specific word in the LXX breaks with its use in Hellenistic Greek and only makes sense on the background of the Hebrew text, which leads us to our next topic. 3. Translation Technique LXX research displays a particular fascination with the mind of the translator. This is hardly surprising, if one considers the monumental importance of this particular translation project, which began in the early- to mid-third century BCE and did not cease until the later revisions in the second century CE, and the creative achievement represented in the LXX texts. The temptation to give in to this fascination is further increased by the fact that both the Hebrew source and the Greek target texts for this translation can be recovered reasonably well for analysis. The intertextual field between source and translation text has generated one of the most prominent foci of Septuagint studies, called ‘translation technique’ and defined by one of its most distinguished practitioner as simply the analysis of the ‘relationship between the text of the translation and its Vorlage’.78 It can be compared in many ways to the redaction-critical approach to the Synoptic gospels, which discovered the evangelists as creative editors of the gospel tradition and used the intertextual relationship between the reconstructed underlying tradition and final edition as productive metaphor. In the case of the Septuagint translators, however, creativity was expressed in the more subtle choices made in transposing 76 See Brook O’Donnell’s thoughts on how to go about creating such a corpus in Corpus linguistics and the Greek of the New Testament, 103–137. 77 There is, of course, the Thesaurus Lingua Graece (http://www.tlg.uci.edu/, last accessed 8 May 2013) but it should be seen as rather an archive than a corpus (see Brook O’Donnell, Corpus Linguistics and the Greek of the New Testament, 47 and passim). It is also prohibitively expensive and is published with a very restrictive license. 78 Aejmalaeus, ‘What We Talk about When We Talk about Translation Technique’, in On the Trail of the Septuagint Translators: Collected Essays, 203. See Tov, ‘The Nature and Study of the Translation Technique of the LXX in the Past and Present’ for a classic survey on this approach.



the seventy and their 21st-century heirs

115

a text from one language system into another, while negotiating the rules of both languages. One of the parameters applied to the assessment of a particular translator’s work is the level of literal faithfulness to the source text that has shaped the rendition in Greek. The standard approach along these lines was simply to situate the translator’s technique within a continuum framed by literalness, on the one hand, and freedom, on the other, as its extremes. This approach was criticised for the lack of conceptual clarity underpinning such an evaluation. A more nuanced breakdown of the criteria for identifying literalness was needed. It is largely due to the work of James Barr79 and Emanuel Tov80 that we now have such a set of widely accepted criteria. Fortunately, at least four of these criteria are quantifiable and offer, thus, the possibility of a computational approach. These quantifiable criteria are: 1. the degree of adherence of the Septuagint translation to the word-order of the Hebrew source text (word-order equivalence); 2. the translation of Hebrew words with individual Greek equivalents for each constituent of the Hebrew word (segmentation); 3. the general translation of each element of the Hebrew source text by one equivalent element in the Greek translation (quantitative representation); and 4. the consistency with which one Hebrew word was translated by the same Greek word (stereotyping). Tov and Wright were among the first to use simple computing approaches to query the CATSS resources in terms of translation technique.81 They argued that, while granting that the Greek Rahlfs text and the Hebrew Masoretic text juxtaposed in CATSS’s Hebrew // Greek module were not the source and target texts of the historical translation project, statistical analysis of that module (using the Oxford Concordance Program) would not exceed 1–2% in margin of error.82 Taking five criteria and counting their frequency in 30 sample Septuagint texts in both absolute and

79 Barr, The Typology of Literalism in Ancient Biblical Translations. 80 Cf. Tov, The Text-Critical Use of the Septuagint in Biblical Research, 20–24. 81  Tov and Wright, ‘Computer-Assisted Study of the Criteria for Assessing the Literalness of Translation Units in the LXX’. Cf. Wright, No Small Difference: Sirach’s Relationship to Its Hebrew Parent Text. 82 Tov and Wright, ‘Computer-Assisted Study’, 156.

116

juan garcés

­relative numbers, allows Tov and Wright to categorise these texts into four categories of literalness (‘literal’, ‘relatively literal’, ‘relatively free’, ‘free’).83 It, furthermore, allowed them to draw conclusions in terms of the consistency with which the translations of certain texts deal with certain linguistic phenomena. This computational analysis, furthermore, corroborates similar, seemingly non-computational analyses by the Finnish school represented by Ilmari Soisalon-Soininen, Raija Sollamo and Anneli Aejmalaeus.84 The main achievement of this study was to move away from an intuitive approach to the analysis of translation technique and to do so, moreover, in a computerised method. Despite Tov and Wright’s straight-forward approach and clearly laidout results, not all Septuagint scholars were persuaded by the computerassisted approach to ‘measuring’ translational ‘consistency’. Anneli Aejmalaeus, one of the foremost experts in the LXX’s translation techniques, expressed her critique of this approach with the thesis: ‘Translation technique is not something that can be measured with statistical methods.’85 She further polemicised against the aforementioned statistical approach: The question is: what do such abstract percentages of literalness actually measure? Do they reliably describe the work of translators? The percentages for the various books can naturally be compared to each other, but since there are no means of knowing how great a difference, for instance, one percent is, the result of the comparison is meagre. Indeed, there is no ideal percentage by which the performance of the translators could be measured. One hundred percent—total literalness—is surely not the proper point of comparison, although one sometimes gets the impression that this is the viewpoint of some writers.86

Karen Jobes, an accomplished Septuagint scholar with a MSc in Computer Science, responded to Aejmelaeus’ critique with further methodological explanation and refinement of the statistical approach. 83 Tov and Wright, ‘Computer-Assisted Study’, 156–157, 184. 84 Wade, Consistency of Translation Techniques in the Tabernacle Accounts of Exodus in the Old Greek, 10, names Tov/Wright and the Finnish school as two approaches attempting to apply ‘more objective means of analysis’. See footnote 37 on the same page and footnote 15 on page 111 for bibliographic references on the Finnish school and a summary description of its methodology in the later page. 85 Aejmalaeus, ‘What We Talk about When We Talk about Translation Technique’, 208. The thesis is somewhat puzzling, since her own approach itself (and that of her Finnish colleagues) makes fruitful use of quantitative approaches. 86 Aejmalaeus, ‘Introduction: on the trail of the Septuagint translators’, in On the Trail of the Septuagint Translators, xiii–xiv.



the seventy and their 21st-century heirs

117

While ­acknowledging the apparent counter-intuitiveness of a statistical approach to Septuagint texts, seemingly opening up ‘an unbridgeable chasm between the quantitative precision implied by the use of statistics, and the relatively inexact and qualitative methodologies of historical and linguistic research’,87 she nevertheless makes the case for ‘descriptive statistics that allow us to organize and describe a very large quantity of syntactical data in a form that is concise and convenient, and that facilitates ease of discussion about and interpretation of the syntactic data’.88 Her basic line of argumentation is that, if textual elements, which are relevant for better understanding translation technique, can be quantified, they can also be statistically described. Therefore, statistical analysis of Septuagint texts should be accepted as one research tool alongside traditional approaches.89 Jobes’ proposal is to ‘allow [. . .] the Greek texts themselves to define the poles of what is meant by ‘free’ and ‘literal’ for each syntactical criterion examined’.90 She elaborates this point: Normalizing the relative frequencies of the syntax in a given text to the average frequency of occurrence within the larger corpus addresses the problem that, language being what it is, there is no ideal standard to which the syntax of a text can be compared. The texts themselves collectively must define the norm against which any given Greek text is compared.91

Jobes further proposes to do this by generating a ‘syntactic profile’ [sic] of a specific set of syntactic features in a text in question in comparison with the same set of syntactic features in both texts known to have been composed in Hellenistic Greek, on the one hand, and texts known to have been translated from a Semitic source, on the other hand.92 Practically, this ‘syntactic profile’ takes the form of a graphical representation in an axis, in which a series of syntactical criteria,93 traditionally seen as a sign 87 Jobes, ‘Quantitative Methods for Exploring the Relationship Between Books of the Septuagint’, 89. See Jobes, ‘A Comparative Syntactic Analysis of the Greek Versions of Daniel: a Test Case for New Methodology’, for an earlier study, the critique by Tim McLay in ‘Syntactic Profiles and the Characteristics of Revision: a Response to Karen Jobes’, as well as her response: Jobes, ‘Karen Jobes Responds to Tim McLay’. 88 Jobes, ‘Quantitative Methods’, 89. 89 Jobes, ‘Quantitative Methods’, 89. 90 Jobes, ‘Quantitative Methods’, 75. 91  Jobes, ‘Quantitative Methods’, 78. 92 Jobes, ‘Quantitative Methods’, 77–78. 93 The criteria were first identified by Raymond A. Martin in his seminal, but largely ignored study Syntactic evidence of Semitic Sources in Greek Documents. Cf. Merle Rife, ‘The Mechanics of Translation Greek’, for a precursor and Davila, ‘(How) Can We Tell If a Greek

118

juan garcés

of a Semitic source text translated into Greek, are placed within their relative frequency in composition Greek (–1) and translation Greek (+1). The graph immediately shows whether, according to the selected criteria, the text in question fits into the tendency of either composition or translation Greek. An additional application of descriptive statistics to the Septuagint presented by Jobes is the comparison of a specific feature in one Septuagint book with the same feature in other books. One such test is variance, another is the Tukey-Kramer multiple comparison test. Taking the translation of the Hebrew causal preposition ‫ כי‬by either ὅτι or γάρ in the Pentateuch in terms of variance, for example, ‘indicates with a confidence level of 95 per cent that the translation . . . was not handled consistently across the five books’.94 A Tukey-Kramer multiple comparison test in relation to the same phenomenon in the Pentateuch, however, shows that Exodus is ‘clearly different’95 from the others. As in the case of the work done by Tov and Wright, this specific analysis undertaken by Jobes seems to corroborate the research by Raija Sollamo. In his classical survey of translation technique research on the LXX, Tov stresses the impressionistic nature of the studies, which began in the nineteenth century, and the lack of meticulous and more comprehensive comparative analyses on this topic.96 The quantitative computations undertaken by Tov, Wright, and Jobes represent noticeable steps forward, away from that situation. They should be considered a success in terms of illustrating the usefulness of CATSS as a resource and as a catalyst for innovative methods in LXX Studies. It is, however, fair to say that computer-assisted quantitative approaches to the study of translation technique have not really caught on in LXX Studies. There may be several reasons for this. One reason is a misunderstanding of the role of quantitative approaches in the arsenal of analytical methods for LXX scholars. Quantitative methods should not be understood as a replacement of existing and well-established methods. When Wade states: ‘these studies will never replace the detailed analyses

Apocryphon or Pseudepigraphon Has Been Translated from Hebrew or Aramaic?’, for an interesting overview on linguistic interference as evidence of translation from a Semitic source text. 94 Jobes, ‘Quantitative Methods’, 91: ‘Analysis of variance is used to classify observations into groups on the basis of a single property.’ 95 Jobes, ‘Quantitative Methods’, 92. 96 Tov, ‘The Nature and Study of the Translation Technique’. Cf. Wade, Consistency of Translation Techniques, 112.



the seventy and their 21st-century heirs

119

of grammatical structures in their contexts’,97 she seems to imply a claim that has, to my knowledge, never been made, namely that quantitative methods have arrived to replace instead of supplement. Another reason for the lack of uptake is the perception of the limitations both in terms of the available digital data and the analytical value of the methods. This is how I understand Wade, when she advocates the necessary coexistence of qualitative and quantitative approaches: Computer-assisted studies that include context will have to wait for the next generation of computer software and tagged texts. Until then, those who are interested in the context, must struggle with the data and find means other than computer-produced statistics to help the reader grasp the nature of the translation techniques used for grammatical structures in the [LXX] Scriptures.98

Last, but not least, is the general lack of appreciation of what precisely the computed numbers signify and what statistical models might be appropriate for assessing the quantitative analyses of the LXX corpus. Biblical scholars are rarely educated in statistical modelling and no theological curriculum incorporates, to my knowledge, the necessary modules that would capacitate LXX scholars to devise and apply quantitative approaches. Concerning the analytical yield of quantitative methods for translation technique and the desideratum of introduction-level overviews of the essential quantitative methods in question, a relatively new sub-field is establishing itself as ‘corpus-based translation studies’ (= CBTS) and beginning to produce interesting studies.99 As CBTS is developing more sophisticated ways of measuring text features and appropriate tests to assess their significance, LXX scholars might well profit from extending their cross-disciplinary interlocutions with this new research area. CATTS and the studies by Tov, Wright, and Jobes have shown that opening up the sub-discipline in that direction has not only the potential to supplement the intuition-based close readings entrenched in the academic field, but also the potential to transform the discourse in dialogue with wider inter-disciplinary insights. In order for that to happen, further critical reflection and studies are necessary. And, as Wade indicates, better and more ­accessible data, which leads us to the issue of establishing a suitable critical text. 97 Wade, Consistency of Translation Techniques, 112. 98 Wade, Consistency of Translation Techniques, 112. 99 Oakes and Ji, eds., Quantitative Methods in Corpus-Based Translation Studies: A Practical Guide to Descriptive Translation Research.

120

juan garcés 4. Textual Criticism 4.1. The Göttinger Septuaginta and Its Production Workflow

In 1907, Alfred Rahlfs, together with Rudolf Smend and others, instigated the creation of the Septuaginta-Unternehmen der Göttinger und Berliner Akademien der Wissenschaften (= GSU). Its stated aim was to reconstruct, according to the historical-philological approach established in the nineteenth century by Paul Anton de Lagarde (and others), the ‘original’ text of the Septuagint, the Göttinger Septuaginta-Ausgabe (= GöSA). With the optimism characterising researchers in search of project funding, Smend supported Rahlfs’ forecast that the project could be concluded within 30 years.100 Over a century, two world wars, and 23 (expensive) volumes later, the project is still ongoing.101 Rahlfs managed to publish a preliminary Handausgabe (German for ‘hand edition’) of the Septuagint just before his death in 1935. This edition is, in waiting for the finalised Göttingen edition, and certainly for the books not yet covered in the published volumes, still for many practical purposes the de-facto standard edition, represented by the main CATSS text corpus and having the additional advantage of now being out of copyright in some countries.102 The definitive and benchmark-setting editio critica maior,103 however, remains the partly completed GöSA. The reason for this lies in the resources invested in and the meticulous methodology and workflow characterising the work of the GöSU. The basic raw material for LXX textual criticism exists in the form of the ca 2000 manuscripts featuring LXX texts,104 supplemented by those ­featuring early quotations and other relevant versions. How do the

100 Rudolf Smend, ‘Der geistige Vater des Septuaginta-Unternehmens’, 337. 101  Bernhard Neuschäfer, the current director of the Göttinger Septuaginta-Ausgabe, remarks that this is about two thirds of the editio maior of the LXX. ‘Die Göttinger Septuaginta-Ausgabe—Standortbestimmung eines editorischen Jahrhundertprojekts: Internationale Fachtagung, Göttingen, 28.-30. April 2008’, 241. 102 See footnote 1 above. Alfred Rahlfs died on 8 April 1935 in Göttingen. 103 This is how Bernhard Neuschäfer, rightly, qualifies the Göttinger SeptuagintaAusgabe: ‘für die nächsten hundert Jahre die maßgebliche und maßstäbliche “editio maior” der Septuaginta’ in ‘Die Göttinger Septuaginta-Ausgabe’, 245. 104 For a description of these manuscripts, covering up to the eighth century CE, see the magisterial Verzeichnis der griechischen Handschriften des Alten Testaments, vol. 1: Die Überlieferung bis zum VIII. Jahrhundert, edited by Alfred Rahlfs and revised by Detlef Fraenkel. For the description of manuscripts copied after the eighth century the somewhat outdated work by Rahlfs is still indispensable: Rahlfs, Verzeichnis der griechischen Handschriften des Alten Testaments, für das Septuaginta-Unternehmen zusammengestellt, available at archive. org at http://archive.org/details/mitteilungendess00akaduoft, last accessed 8 May 2013.



the seventy and their 21st-century heirs

121

researchers of the Göttingen Septuaginta-Unternehmen—for it is a collaboration of editors, researchers of the Septuaginta-Kommission proper, and research assistants—proceed with this material on their way to the critical edition of a particular LXX book? In his useful report, Frank Austermann describes this labour as unfolding in three distinct ‘phases’.105 The first phase is the preparatory phase undertaken mainly by research assistants (‘von studentischen und wissenschaftlichen Hilskräften’)106 who put together the material on the basis of which the editors proper will create the critical edition. The main product of this phase is the collation booklet (Kollationsheft). This booklet features the synthetic text of the Septuagint—something coming close to the textus receptus, though in principle it could be any text representing a reasonably widely-attested LXX text—as the lemma text, written vertically, word by word, separated if necessary by free lines to accommodate multiple variant readings, in one column. Juxtaposed to these words, a second column lists all variant readings in manuscripts up to the fifteenth century, referred to by their Rahlfs sigla, next to the relevant word in the lemma text.107 All readings differing from the lemma text are recorded, even orthographic variations or so-called nonsensical readings. This phase is completed with a thorough proofreading by two experienced researchers to ensure the quality of the recordings in the collation and gain a first impression of the text, as well as the manuscripts and their relation und groupings. Having finalized the preparatory work, the second phase introduces the critical work proper, undertaken by the editor. But before the editor starts with the reconstruction of the transmission history of the text, s/he adds relevant material from ancient versions, where extant—particularly Coptic, Ethiopic, Latin, Syriac, Armenian, to some extent also Arabic and Georgian—as well as quotations from the Greek and Latin Fathers and material from the Hexaplaric tradition, catena, patristic literature and the Syrohexapla. Crucial to the text-critical work of the editors is a critical assessment of the translation technique of the book in question. This is then followed by an analysis of the transmission history. Based on these two last steps, the collated text is meticulously stripped of the many ­secondary layers comprising both mistakes made in the process of copying and deliberate (recensional or revisional) changes to the text. 105 Austermann, ‘Die Septuaginta edieren: Zur Arbeit an der kritischen Ausgabe einer antiken Bibelübersetzung’. 106 Austermann, ‘Die Septuaginta edieren’, 20. 107 See http://adw-goe.de/uploads/pics/aufgaben_KH_Foto.jpg for a depiction of an opening of such a collation booklet for 4 Reigns = 2 Kings 3:4. Last accessed 8 May 2013.

122

juan garcés

This presupposes the grouping of the manuscripts according to their textual closeness and a thorough knowledge of the various recensions and their ‘typical’ elements. This phase is, again, concluded with a scrupulous quality-check. The final phase is devoted to the separate type-setting and proof­reading of the introduction, critical lemma text, as well as the two-fold apparatus criticus. The final outcome is indeed ambitious: A critical edition of an ancient translation has been made available that comprises all sources, which witness to the oldest reconstructable text, as well as all the variants differing from the original text. Thus, the recipient of the text edition is put in the position to profit from the research of the editor and, at the same time, to develop his/her own text-critical thoughts in view of the complete material. The work on the most influencial ancient translation aims, in the end, at providing the critical text and, at the same time, at making the editor’s decisions accesible to verification and criticism by the way it has been edited.108

Similar to LXX lexicography, though with the hindsight of an issue now largely settled, LXX textual criticism has been influenced by two competing theories. For there has not always been a consensus that there was an Ur-Septuagint, a single original text, to start with. Paul Kahle, for example, maintained that there were not one but several Greek translations of the Hebrew Bible that were produced out of liturgical necessity. According to him, occasional attempts to unify the existing translations over time notwithstanding, the authoritative version came only at the end of a longer process.109 Kahle did, in other words, reverse the assumed Lagardian development of the Septuagint. Such a reading of the manuscript evidence would render the attempt to discover or reconstruct a primitive text, which functioned as an archetype for the subsequent textual history, futile. Over the years, however, the Lagardian approach has established itself as the scholarly consensus by making a convincing case for being the theory that better explains the extant manuscript tradition, not least since the discovery of the LXX texts in the Judean Desert. Though it should be noted that the same finds also demonstrated that the text of the Hebrew text itself was still in a state of fluidity. Even though Lagarde by-and-large won the day, there are a number of peculiarities of the textual transmission of the Septuagint texts that

108 Austermann, ‘Die Septuaginta edieren’, 23, my translation. 109 See Kahle, The Cairo Geniza.



the seventy and their 21st-century heirs

123

complicate the Göttingen project. Even assuming the existence of a single original LXX text of the individual books, the development of its text is still thought to have been ongoing over four centuries. Not all LXX books were translated at the same time.110 Even after a first edition of all LXX books (not necessarily circulating together) had been produced, the LXX underwent a series of revisions and recensions. Three types of revisions usually distinguished:111 1. Before the second century CE, there was a tendency to eliminate ‘Semitisms’ and replace them with expressions found in Koine literary documents. 2. After the second century CE, the Atticist movement, which represented a return to more Classic ‘Attic’ Greek, influence further revision of the Septuagint. 3. Most importantly—and probably throughout its developmental ­process—the Septuagint was assimilated again and again to the text of the Hebrew Bible, which was also developing. These revisions make the Septuagint a particularly complex textual tradition of the Bible and the task of reconstructing an Urtext almost impossible. Taking cognisance of this situation has led the GSU to modify its aim.112 In Fernández-Marcos’s words: In fact nowadays the aim of the Göttingen editors has become more modest, merely trying to reach in their critical restoration a stage of languages as near as possible to the original; moreover, in some books even this modest goal cannot be maintained and the editor has to content himself with the restoration of two or three textual forms, as he is unable to go back to an earlier stage of the tradition.113

110 Some later books of the LXX, as understood in this essay, such as Wisdom, 2–4 Maccabees and Judith, were of course composed in Greek. 111  Fernández-Marcos, The Septuagint in Context: Introduction to the Greek Version of the Bible, 191–203. 112 Neuschäfer (‘Die Göttinger Septuaginta-Ausgabe’, 243) speaks of the reconstruction of the ‘oldest attainable text’ (‘den “ältest erreichbaren Text” ’, quoting Detlev Fraenkel). He explains: ‘Gemeint ist damit jener Textzustand, der auf Grund einer detaillierten Kenntnis der Übersetzungstechnik sowie einer Ausscheidung der bereits im 2. Jahrhundert v.Chr. belegten jüdischen und vom 3. bis ins 6. Jahrhundert nachweisbaren christlichen Bearbeitungsspuren (der sog. ‘Rezensionselemente’) als ‘vorrezensionell’ eingestuft werden kann.’ 113 Fernández-Marcos, Scribes and Translators: Septuagint an Old Latin in the Book of Kings, 24.

124

juan garcés

For some Septuagint books (Daniel, Tobit, Judges, Esther) and parts of other books (Habakkuk, Kingdoms), for example, the Göttingen edition publishes a double text, without resolving them to one single archetype. 4.2. Towards Modeling a Critical LXX Edition In view of this situation, we have to ask ourselves a question similar to the one we asked of LXX lexica: what kind of resource is a critical edition and for what uses is it supposed to cater? First and foremost, a critical edition is expected to provide a ‘standard’ reference for academic research, a sort of ‘default’ text that is referred to, without having to further legitimise the critical decisions that led to its reconstruction. As such, textual criticism provides the textual foundation for any further critical (historical, philological, literary etc.) analysis, or, to use eighteenth-century parlance: the critical edition provides the fruits of ‘lower criticism’, on which the tasks of ‘higher criticism’—criticism proper—can build.114 Without further elaborating on this somewhat infelicitous separation of concerns, two aspects have to be carefully balanced when creating a critical edition. Bernhard Neuschäfer summarises the ‘daring feat’ of the Göttingen editors as ‘to do justice to the textual tradition without denying the justified demands of the users’.115 In relation to the text tradition, this is encapsulated in the ‘duty of documenting the history of the transmission of the text’116 and delivered via the elaborate apparati critici—for the GöSA has two, one for the substantial variants from manuscript texts, cognate versions and quotations, in other words: ‘the textual evidence proper’;117 another for the Hexaplaric variants. The result is a layout masterpiece of the print era that manages to condense an immense amount of complex data on the available space offered by a printed page. There is, however, another side to this achievement concerning the ­aforementioned ‘justified

114 Though see Colwell, ‘Biblical Criticism: Lower and Higher’, for a more nuanced explanation of the terms. 115 The ‘Kunststück’ consists of ‘dass seine [i.e. the editor’s] Edition den spezifischen Erfordernissen der Textüberlieferung gerecht zu werden vermag, ohne sich den berechtigten Forderungen der Benutzer zu verschließen’—Neuschäfer, ‘Die Göttinger Septuaginta-Ausgabe’, 245. 116 Neuschäfer, ‘Die Göttinger Septuaginta-Ausgabe’, 245: ‘Textgeschichtliche Dokumentationspflicht’. 117 Kielsmeier-Jones, ‘How to Read and Understand the Göttingen Septuagint: A Short Primer, part 2 (Apparatus)’, https://abramkj.wordpress.com/2013/01/11/how-to-read-andunder­stand-the-gottingen-septuagint-a-short-primer-part-2-apparatus/, last accessed 8 May 2013.



the seventy and their 21st-century heirs

125

demands of the users’. In the words of a recent, very readable introduction to using the GöSA on a blog: [T]he Göttingen Septuagint is not for the faint of heart, or for the reader who is unwilling to put some serious work in to understanding the layout and import of the edition and its critical apparatuses. A challenge to using Göttingen is the paucity of material available about the project, even in books about the Septuagint. An additional challenge is that the critical apparatuses contain Greek, abbreviated Greek, and abbreviated Latin. The introductions to each volume are in German, though below I cite from English translations of the introductions to the volumes of the Pentateuch.118

If the GöSA were an online publication, it would surely be recommended for a serious usability overhaul. Usability, as defined by the International Organization for Standardization, is ‘[t]he extent to which a product [here: the GöSA] can be used by specified users [here: biblical researchers with interest in the Greek Bible] to achieve specified goals [more on that below] with effectiveness, efficiency, and satisfaction in a specified context of use.’119 But this merely begs a further question: what is the goal of the researcher interested in the text of the LXX and its transmission history? Or asked differently: what is the primary text that the researcher wishes to establish and to which the multitude of variants in the apparatus might witness? In a print edition the answer to this question determines to a great degree the layout and format of the critical text, since a preferred text—usually eclectically put together from text witnesses (manuscripts) according to an assessment as to their originality—will have to be chosen for the lemma, and relevant variants will have to be organised in one apparatus (or more, as is the case in the GöSA). The Göttingen edition, faithful to its intellectual forefather Lagarde, prints as its lemma text the oldest attainable text of the LXX. This layout, however, is certainly not the only way of presenting the outcome of textual scholarship on the Septuagint. By contrast, the editio maior of the so-called Cambridge Septuagint (aka the ‘Larger Cambridge Septuagint’),120

118  Kielsmeier-Jones, ‘How to Read and Understand the Göttingen Septuagint: A Short Primer, part 1’, https://abramkj.wordpress.com/2012/11/04/how-to-read-and-understandthe-gottingen-septuagint-a-short-primer-part-1/, last accessed 8 May 2013. This is an issue that doesn’t escape seasoned philologist from neightbouring disciplines either, as Neuschäfer, ‘Die Göttinger Septuaginta-Ausgabe’, documents. 119  ISO 9241-11: Guidance on Usability (1998). http://www.iso.org/iso/catalogue_detail .htm?csnumber=16883, last accessed 8 May 2013. 120 Brooke and McLean, eds., The Old Testament in Greek According to the Text of Codex Vaticanus, 9 volumes.

126

juan garcés

for example, follows a ‘best-text’ approach offering the text of a manuscript coming closest to the notional source text from which all other witnesses depart—in this case Codex Vaticanus—with variant readings in the apparatus. Nor is an ecclectic, reconstructed text or a diplomatic best-text the only lemma text in LXX editions, for some recensions are important and intriguing in themselves. The so-called ‘Antiochene’ text (also called ‘Antiochian’ or ‘Lucianic’), for example, is of great interest for the history of the Greek Bible, due to its obvious influence during a crucial period of Christian theology. This text-type was identified in a small number of manuscripts and was used by both fourth-century Christian writers, like John Chrysostom and Theodoret of Cyrrhus, and in pre-Christian papyri as well as in the Jewish historian Josephus. The Antiochene text is, therefore, not only of interest for the influence it might have had during certain periods, but also for its prime importance as a witness to an early Hebrew text. So much so that an edition for the historical Septuagint books, for which this text can be reconstructed, was edited to critical acclaim.121 Taking these three editions, we have, at least, three types of Septuagint texts, around which a critical edition has been organised and which might be the analytical goal of a researcher: 1. an archetype text, reconstructed eclectically from a host of critically assessed variant readings; 2. a certain recension, which has been reconstructed according to a mixture of identifying typical features of the recension in question and the selection of a group of manuscript representing said recension; 3. the text of a specific manuscript, which has been chosen for certain qualities. While a print edition has to choose between any of these three possibilities, an electronic edition based on a soundly modeled database, can offer any combination of these interactively. A number of emerging online critical editions serve to illustrate this potential, among them The Online Critical Pseudepigrapha122 and the cluster of impressive resources being made available by the Institut für neutestamentliche Textforschung in 121  Fernández-Marcos and Busto Saiz, eds., El texto antioqueno de la Biblia Griega, 3 volumes. Cf. the accompanying and useful: Fernández-Marcos, Spottorno Díaz-Caro and Cañas Reíllo, eds., Índice griego-hebreo del texto antioqueno en los libros históricos, 2 volumes. 122 http://ocp.tyndale.ca/, last accessed 8 May 2013.



the seventy and their 21st-century heirs

127

Münster in conjunction with the Institute for Textual Scholarship and Electronic Editing in Birmingham: among them the New Testament Transcripts Prototype123 and the Virtual Manuscript Room.124 These projects offer the user the possibility of choosing the text according to a critical edition (OCP sometimes offers several editions) or according to a certain manuscript. The apparatus reorganises itself accordingly. No longer is the researcher obliged to make do with whatever practical and methodological decisions that were taken in conceiving a critical edition, but she is now included in the process of creating her own basic virtual research environment, be that an Urtext, a recension, or the text of a specific witness. One further aspect of the critical edition has already been mentioned but deserves some further elaboration. Austermann states that the edition aims at ‘making the editor’s decisions accessible to verification and criticism’.125 That the complex assumptions and critical decisions going into the volumes of GöSA are not self-evidently clear and accessible from the information provided and the way it is laid out in the individual volumes alone, is evident. It would be, in all honesty, too much to ask from a print edition! This is why John Wevers’ authoritative edition of the Greek version of Genesis,126 for example, is also accompanied by a volume explaining his understanding of its text history,127 notes on the text, explaining its language and relation to its Hebrew Vorlage,128 and, recently, a ‘user manual’ to the edition.129 This is not a trivial issue, for it concerns the ‘critical’130 in critical edition in terms of the accountability of the editor, on the one hand, and the possibility of the user to assess any critical decision made. Concerning this point, print publications are already stretched to their limits set by the ‘real estate’ in the form of a printed page. Electronic publications are, in this respect at least, potentially limitless in the amount of hyperlinked information that can be offered. The relevant materials that would be electronically interlinked comprise 123 http://nttranscripts.uni-muenster.de/, last accessed 8 May 2013. 124 http://ntvmr.uni-muenster.de/, last accessed 8 May 2013. 125 Austermann, ‘Die Septuaginta edieren’, 23, my translation. 126 Wevers, ed., Septuaginta: Vetus Testamentum Graecum: Auctoritate Academiae Scientiarum Gottingensis editum, vol. I: Genesis. 127 Wevers, Text History of the Greek Genesis. 128 Wevers, Notes on the Greek Text of Genesis. 129 Schäfer, Benutzerhandbuch zur Göttinger Septuaginta, vol. 1: Die Edition des Pentateuch von John William Wevers, though also repeating material contained in Wevers’s Text history of the Greek Genesis. 130 I have formulated my thoughts on this issue, together with Gabriel Bodard, in ‘Open Source Critical Editions: A Rationale’, especially 88–92.

128

juan garcés

the complex represented by the manuscripts themselves—images, transcriptions, metadata—and any critically reconstructed texts—oldest attainable form, revisions, recensions—not to speak of the host of ‘cognate’ texts—‘daughter’ versions and, lest we forget, their own source texts, which, in turn, should be interlinked to their own manuscripts and textual reconstructions. The point made here is simple: the necessarily high standards set by the GöSA can only be realised in an electronic edition. The radical transparency the edition is supposed to display in relation to the documentary evidence and the critical assessment of that evidence, not least the critical decisions that led to any sort of reconstructed text, is made possible in a resource that consists of networked data. This networked data, furthermore, cannot be created by any individual research endeavour, however well-resourced and however long in running time. This vision can only be realised in a large collaboration of related projects and scholars (more on that below), a critical ecosystem that creates digital representations according to the highest critical standards and makes them openly available so it can be fully interconnected by the user. Only in this way, can the raw materials be ‘opened up’ (in the sense of the German erschlossen) and this is, in my understanding, what DH is all about. But once the material is available digitally, new methodological approaches open up along the way. 4.3. Computer-Assisted Methods and LXX Textual Criticism The advantages of digital collations based on digital transcriptions of manuscript texts are not only presentational, they can and should also influence the methodology for analysing the compared texts. In 1986, John Abercrombie published his attempt at demonstrating ‘how a new generation of [LXX] manuscript studies can be conducted with the assistance of the computer’.131 Taking one of the first books for which CATSS’s Variants module had been finalised, Abercrombie automatically collated various manuscript families against each other and compared this collation with the Masoretic Text of the book of Ruth.132 The outcome of his research corroborated Rahlfs’ conclusions about these same manuscript families, but added methodological rigour to the analysis. Unfortunately, Septuagint 131  Abercrombie, ‘A Computer-Assisted Study of a Textual Family in the Book of Ruth’, here 95. 132 Abercrombie, ‘A Computer-Assisted Study’, 96.



the seventy and their 21st-century heirs

129

scholars showed little interest in taking up this approach. (Abercrombie, very unfortunately, died not long after his article was published.) In fairness, this has been, largely, due to the lack of a comprehensive resource, the aforementioned CATSS module being far from completed at present. Despite the lack of open accessibility of the digitised text of the Göttingen edition133 and the slow-footed progress of their publications, interest in the application of computer-assisted analyses for text-critical research on the LXX, however, seems to be re-emerging at the beginning at the beginning of the twenty-first century. A sign of this can be seen in Hiebert and Dykstra’s work on 4 Macabbees.134 Robert Hiebert is currently editing this LXX text—unlike most other LXX texts, not a translation, but a text originally composed in Greek—for the GöSA.135 His approach departs somewhat from other GöSA editors in that he planned to use computing skills as part of his editorial work, from the beginning.136 This meant that he first digitally transcribed the data collected in the Göttingen collation booklet into a table, which was subsequently migrated into a relational database, in order to not only better to manage the data but also to query and to assess previous groupings of manuscript texts undertaken by Hans-Josef Klauck for his German critical edition:137 ‘What a computer database enables one to do easily . . . is to retrieve, sort and classify all readings for a particular manuscript

133 Both the Accordance (http://www.accordancebible.com/, last accessed 8 May 2013) and the Logos Bible Software (http://www.logos.com/, last accessed 8 May 2013) systems offer digitised versions of some (Accordance) or all (Logos) the published GSA volumes. While offering the advantages of being digitized—searchability, (limited) text re-use etc.—the editions seem to function more as surrogates of the print edition. The apparati of the GöSA, for example, are not deconstructed and presented in a more user- and researchfriendly way, but merely as they are laid out in print. 134 Hiebert and Dykstra, ‘Septuagint Textual Criticism and the Computer: 4 Maccabees As a Test Case’. Cf. Hiebert and Dykstra, ‘What Does the Computer Have to Do With Textual Criticism? Innovative Technology for the Management and Analysis of Collation Data and the Grouping of Manuscripts’, and Hiebert, ‘Establishing the Textual History of Greek 4 Maccabees’, for more on the results coming from this work. 135 It is listed in his bibliography as ‘[i]n preparation’ (http://acts.twu.ca/faculty/ rob-hiebert.html, last accessed 8 May 2013): Hiebert, ed., Septuaginta: Vetus Testamentum Graecum Auctoritate Academiae Scientiarum Gottingensis editum, Vol. 9.4: Maccabaeorum liber IV. 136 Hiebert and Dykstra, ‘Septuagint Textual Criticism and the Computer’, 170. It should be seen as both a sign of Hiebert’s foresight and as a sign of the (medialised) times that he considered the ‘application of computer technology to research in the humanities’, 170, in other words: Digital Humanities, as important enough that he budgeted necessary resources in his grant application accordingly. 137 Klauck, 4. Makkabäerbuch.

130

juan garcés

or group. This inevitably speeds up the completion of such work and greatly reduces the possibility of introducing errors in the process.’138 After eliminating a number of ‘orthographic and morphological readings’,139 i.e. readings that represent clear misspellings or non-standard spellings without linguistic significance, Hiebert and Dykstra established that there was no clear evidential basis for a number of groups Klauck had posited, that one distinction was ‘textually artificial’, and that some further sub-groups could be identified.140 It is encouraging to read that Hiebert and Drykstra are currently extending the use of the data to include the analysis of ‘fluctuation in manuscript affiliation’,141 i.e. to detect ‘block mixture’, where different parts of manuscript texts go back to different strands of its history of transmission. Additional uses will comprise a reconstruction of the text of any given manuscript featuring 4 Macabbees and, in the end, the lemma text and apparatus for the GöSA. Hiebert and Drykstra are a clear example of how computer-assisted resources comparable to CATSS can be put to use in LXX textual criticism. Their work illustrates the need, first, to create a properly modeled textual base with the data collated from the text tradition. This data base can not only be used to re-create the ultimately necessary wording of the individual manuscripts and the critical edition itself but, more importantly, open up the possibility of applying quantitative approaches to manuscript grouping. One question that I had in reading about their research was: given that Hiebert and Drykstra use the percentages of agreement/disagreement in variation units between manuscript texts as a way to group them, how do they discriminate between significant and insignificant percentages? This was one of the questions put by another research project almost a decade before. Acknowledging the key role of manuscript grouping in the critical methodology of LXX textual criticism, Boyd-Taylor, Austin, and Feuerverger focused on critically ‘exami[ning] the assumptions underlying the classification of manuscripts in the Göttingen edition of Psalms by Alfred Rahlfs’.142 Rahlfs organised the manuscripts he collated for his ­preliminary 138 Hiebert and Dykstra, ‘Septuagint Textual Criticism and the Computer’, 172. 139 Hiebert and Dykstra, ‘Septuagint Textual Criticism and the Computer’, 173. 140 Hiebert and Dykstra, ‘Septuagint Textual Criticism and the Computer’, 179. 141  Hiebert and Dykstra, ‘Septuagint Textual Criticism and the Computer’, 181. 142 Boyd-Taylor, Austin and Feuerverger, ‘The Assessment of Manuscript Affiliation within a Probabilistic Framework: A Study of Alfred Rahlfs’s Core Manuscript Groupings for the Greek Psalter’, 99. Rahlfs published his preliminary edition of the Greek Psalter as Psalmi cum Odis, and his preparatory studies as Septuaginta-Studien, vol. 2: Der Text des



the seventy and their 21st-century heirs

131

edition of the Greek Psalter in three old text-forms, labelled the Upper Egyptian, Lower Egyptian and Western texts. Using sound statistical modelling, Boyd-Taylor, Austin and Feuerverger asked themselves just what evidentiary value manuscript readings have when they are shared with other manuscripts. For the important principle of multiple independent attestation can only be applied after it has been established how significant it is when manuscripts agree or disagree on a specific reading, particularly in view of their overall pattern of dis/agreement on readings over against other manuscripts. This latter aspect was unfortunately not dealt with adequately by Rahlfs: ‘Rahlfs’s approach leaves him no means of determining a priori just what a significant affiliation would consist of. His groupings are, therefore, determined in an entirely post hoc manner’.143 The underlying reason for this ‘methodological infelicity’, it is argued, is that ‘Rahlfs has not taken into account the role of chance in the transmission history of his witnesses. He can offer no criteria for analysing the significance of the observed patterns’.144 Put differently: ‘When two manuscripts, over against other witnesses, agree on a reading, the text critic wants to know what the likelihood is that this agreement is independent of the respective copying histories of those manuscripts and therefore representative of the original text’.145 Boyd-Taylor, Austin, and Feuerverger therefore suggest that researchers should start out by only considering ‘those readings for which there is more than one variant and for which both manuscripts depart from the [hypothesized] OG’,146 for only those readings provide potential evidence for affiliation. Within such a list of readings, they proceeded to consider manuscripts in pairs and to assess whether their agreement/disagreement is a product of chance and whether any significance can be ascertained from such a pattern in terms of dependence or independence. In other words, an approach is proposed that allows moving beyond simple alignment to significant affiliation by framing the question within a ­probabilistic framework. Following the well-established statistical method of ‘maximum likelihood estimation’, the authors can now assess the ­significance of patterns of conjunction and disjunction involving any two manuscripts. Septuaginta-Psalters. Boyd-Taylor, Austin, and Feuerverger are influenced by their teacher’s dissatisfaction with some of Rahlfs’s methodological infelicities expressed in Albert Pietersma, ‘The Present State of the Critical Text of the Greek Psalter’. 143 Boyd-Taylor, Austin and Feuerverger, ‘The Assessment of Manuscript Affiliation’, 105. 144 Boyd-Taylor, Austin and Feuerverger, ‘The Assessment of Manuscript Affiliation’, 105. 145 Boyd-Taylor, Austin and Feuerverger, ‘The Assessment of Manuscript Affiliation’, 110. 146 Boyd-Taylor, Austin and Feuerverger, ‘The Assessment of Manuscript Affiliation’, 111.

132

juan garcés

Thus, they conclude that ‘there is insufficient evidence to conclude that A [i.e. the text of Codex Alexandrinus] and B [i.e. the text of Codex ­Vaticanus] are not independent of each other’, while the text of Codex Vaticanus and, on the one hand, the Bohairic versions and, on the other hand, though less conclusively, are significantly dependent, while ‘there is insufficient evidence to claim either that B [Codex Vaticanus] and R [a Greek-Latin Psalter at the Biblioteca Capitolare in Verona] or B [Codex Vaticanus] and Sa [the Sahidic versions]’.147 This has, of course, consequences for the presentation of the evidence in a critical edition: ‘A revised apparatus for Rahlfs’s Psalter would thus be warranted in using the notations B-S and B-Bo to indicate the fact that each pair counts as only one witness when they agree.’148 Yet, concerning R, Sa and A, the textual critic needs to revise the evidentiary value assigned to their dis/agreement: ‘When these witnesses read in support of B, all else being equal, it seems best to treat the readings as multiply attested’.149 These two studies unveil some of the research possibilities for LXX textual criticism provided by two conditions: the availability of digital textual data and computational processing according to sound statistical methods. What each study represents is nothing else but an ‘experiment’ with textual data in order to establish or assess a hypothesis. Given that the data is potentially available and the statistical processes are based on machineactionable algorithms, there is much to be said in favour of making the ‘workflows’ underpinning these ‘experiments’ available in a way that other researchers can re-create the experiment und thus critically evaluate the hypothesis.150 Hiebert and Dykstra establish the need to have the digital data at hand, while Boyd-Taylor, Austin and Feuerverger evidence the need for LXX textual critics to engage in interdisciplinary dialogue in order to learn from and collaborate with colleagues who are used to the application of quantitative methods. Quantitative methods, properly understood as an additional methodological approach to the LXX’s complex text tradition, could thereby introduce workflows and discourses from the sciences

147 Boyd-Taylor, Austin and Feuerverger, ‘The Assessment of Manuscript Affiliation’, 121–122. 148 Boyd-Taylor, Austin and Feuerverger, ‘The Assessment of Manuscript Affiliation’, 122. 149 Boyd-Taylor, Austin and Feuerverger, ‘The Assessment of Manuscript Affiliation’, 122. 150 One could imagine this happening via a social website like myExperiment (http:// www.myexperiment.org/, last accessed 8 May 2013), which allows the sharing of Research Data—in our case, the collated texts—and Scientific Workflows—in our case, the algorithms to analyse the dis/agreements in variation units between manuscript texts.



the seventy and their 21st-century heirs

133

into philology. The point here is not to ­transform philology into science at the expense of long-established approaches. The point is rather to engage philology into a cross-disciplinary dialogue, which forces it to open up and critically re-assess its assumptions and conclusions. Such a cross-disciplinary dialogue has been taking place under the name of stemmatology,151 particularly in its exposure to the application of mathematical models and quantitative approaches developed by evolutionary biologists. This cross-disciplinary borrowing is based on the structural similarity between the transmission of a text in the process of copying from one manuscript to another, on the one hand, and the reproduction of organisms, on the other: both pass on sequence data in the process of transmission/procreation—a sequence of glyphs, words, sentences, on the one hand, and a sequence of genetic information in the form of protein strings or DNA, on the other. Both of these processes can be represented in a graph—a stemma of manuscripts, on the one hand, and a phylogenetic tree, on the other. Having recognised this close parallel, the Canterbury Tales Project (http://www.canterburytalesproject .org/), for example, set out to work together with a group of Cambridge biochemists.152 The collaboration applied phylogenetic techniques to 850 lines of 58 surviving fifteenth-century manuscripts of the ‘The Wife of Bath’s Prologue’, particularly the method of ‘split decomposition’. The outcome of this analysis is interesting: the ancestor of the whole tradition, Chaucer’s own copy, was not a finished or fair copy, but a working draft containing (for example) Chaucer’s own notes of passages to be deleted or added, and alternative drafts of sections. In time, this may lead editors to produce a radically different text of The Canterbury Tales.153

Promising research similar to that undertaken in Cambridge and Birmingham, is currently underway in Leuven154 and Helsinki155 and ­yielding

151  The term can be understood in this context as roughly synonymous to, though methodologically broader than, textual criticism. See, e.g., van Reenen and van Mulken, Studies in Stemmatology and van Reenen, den Hollander and van Mulken, Studies in Stemmatology II. 152 Barbrook, Howe, Blake and Robinson, ‘The Phylogeny of the Canterbury Tales’. 153 Barbrook, Howe, Blake and Robinson, ‘The Phylogeny of the Canterbury Tales’, 839. 154 The ‘Tree of texts’ project (http://treeoftexts.arts.kuleuven.be/, last accessed 8 May 2013). 155 These were disseminated internationally via the ‘Studia Stemmatologica’ workshops (http://cosco.hiit.fi/stemmatologica/, last accessed 8 May 2013).

134

juan garcés

results on a variety of heterogeneous texts, proving the transferable nature of the methods.156 5. Conclusions More than 17 centuries ago, Origen and the Christian scholars at Caesarea took advantage of the technological advances in book production to refashion scholarship and collaboration concerning the study of the Bible. More than 4 decades ago Robert Kraft and Emanuel Tov, along with a number of LXX scholars, attempted to achieve something comparable, perhaps not yet contemplating nor comprehending the full potential effects. The fruits of their effort, still available online and—corrected,157 modified and enriched—via a number of packages, have to be seen as pioneering, bold, and visionary. They instigated some important computer-assisted LXX research, a small selection of which I was able to consider in this essay. But it is also fair to say that digital LXX research has not yet fully caught on in the LXX community. This is surely due to a number of factors, not least the need for more comprehensive corpora, methodological openness and willingness to adapt critically to the possibilities afforded by the changes in technology and knowledge cultures. With the arrival of DH and its visibly successful establishment as an energising and creative academic discourse, it is evidently time to reflect upon the prospects of LXX research in the spirit of CATSS. What would LXX studies need to do in order to experience a fullyfledged DH breakthrough? In view of the aforementioned areas of digital and digitised lexicography, translation studies and textual criticism, I would like to propose a vision concerning three closely-linked areas: electronic resources, quantitative methods, and a change in research culture.

156 For an overview of phylogenetic methods, see Felsenstein, Inferring Phylogenies. The application of these approaches to textual criticism is, of course, not without its criticisms. See Howe, Connolly and Windram, ‘Responding to Criticisms of Phylogenetic Methods in Stemmatology’. 157 CATSS, as any other pioneering resources, was, of course, not without mistakes. See, for example, Appendix 2 in Evans, Verbal Syntax in the Greek Pentateuch: Natural Greek Usage and Hebrew Interference, 270–280, who offers a commendable study linking digital resources and cutting-edge linguistic theory.



the seventy and their 21st-century heirs

135

5.1. Resources Digital scholarship needs resources in the tradition of CATSS. This is not only true for the mere fact that CATSS is electronic and therefore machinereadable, but for the fact that CATSS is (relatively) openly available. By this I mean that the resources for digital scholarship need to be available via a license that allows the freest-possible use and re-use, at least for research purposes.158 This is a difficult issue for publishers and Bible software vendors to come to terms with, since it entails some re-thinking in terms of business models and sustainability.159 Digital research resources need to be independent from any specific research environment—be they software packages, tools or websites—and not have any stipulated restriction as to the analytical processes to which they may or may not be submitted. This means that, even more important than accessing the data via welldesigned websites—important as this may be—, the raw data needs to be openly available for scholarly use via some standard transfer protocol. Only when this availability is given can we think further about how to network the openly available resources. For the LXX, this network would comprise digital surrogates of all relevant Greek manuscripts (ranging from ‘standard’ high-resolution images to elaborate digital reconstructions, always along the relevant metadata), their transcriptions and annotation according to a variety of concerns (editorial, codicological, paleographical, linguistic, etc.), eclectic reconstructions of a variety of textual ‘states’ (the oldest wording, but also the different revisions and recensions, etc.) and their respective annotations. This set of networked resources should, in turn, become linked to cognate sets like their equivalent in the realm of the Hebrew Bible, daughter translations, and larger Hellenistic Greek texts of any kind. In order to navigate such a rich network, the research community will have to agree some set of persistent and unambiguous identifiers that allows the referencing and use of any particular element of the network, including different set of annotation on the same basic text. Such a network of resources can only be realised along with a cultural change in the research community—more on that below. It will also only 158 I am thinking, of course, about license agreements like the Creative Commons licenses (http://creativecommons.org/, last accessed 8 May 2013)—the more liberal, the better. 159 I am not proclaiming a naïvely free, as in ‘free beer’, availability. There are some serious challenges that follow the proposal I (without being in any way original) am putting forward. But dealing with these challenges has to follow the main need of an Open Content approach to the critical raw materials of our trade.

136

juan garcés

be fruitful if it comes along with a research ecosystem that allows for the critical analysis of these resources—in any possible combination of their individual elements—according to innovative digital approaches. 5.2. Methods Having the aforementioned rich network of resources openly available at the disposal of the LXX scholar would be a great achievement in itself. It would mean that these resources could be readily consulted from any access point to the World Wide Web. Being able to display such information on a screen would considerably accelerate the potential for research. We could speak of digitised research in terms of the electronic screen (and, only as a derivative, the print-out) as the new preferred interface to access research data. However, settling for this scenario—as attractive as it is!—would mean to fall short of a far richer one, namely a scenario which includes computing as an intrinsic, analytical part of our research workflow, in other words: digital research. Computers are very good tools for capturing, managing and manipulating data. Take an electronic corpus, for example, and it will be easy to find an already-existing tool for writing some code to create a KWIC concordance file. The quality of qualitative research is greatly enhanced in that way. But, to somewhat exaggerate and simplify, the digital research methods promising truly innovative insights are mostly quantitative—be they in assessing the distribution of linguistic patterns in a corpus, as in corpus linguistics, or correlating and comparing said linguistic patterns across parallel corpora, as in corpus-based quantitative translation studies, or calculating the distance between manuscript copies of the same text, as in the application of phylogeny to stemmatology. I hope to have shown that there have been already promising studies in that area—some based on the CATSS modules, others on similar resources—and that there are a number of methods promising to bear interesting fruits in digital LXX studies. Quantitative methods have a series of advantages over against traditional approaches based on expert intuition. They offer additional descriptive power and precision when analysing linguistic and textual phenomena in the extant LXX and cognate texts. They also introduce empirical and therefore more objective methods in dealing with culturally and ideologically contested texts. Given that the corpora are to be openly accessible and quantitative processing are based on algorithms that are to be computed, these methods are furthermore replicable by anybody with access



the seventy and their 21st-century heirs

137

to resources and algorithms, as well as the right expertise. Results and core elements of a statistical argumentation are therefore opened up to a more transparent critical assessment. It should be emphasised here that I am not arguing for these quantitative methods to replace established traditional methods. They are rather to supplement and, at any rate, possibly challenge traditional approaches and their results. 5.3. Culture Embracing the potential offered by the technological transformation of the information society means also to embrace a change in research culture along the lines of DH. We are only slowly finding out what this really means, but a few corner pillars are already becoming clear. The above ecosystem of networked resources and related tools and scripts cannot but be realised as a comprehensive collaboration of the research community. This is not to use a misguided dualism of the anecdotal solitary scholar of the print era in contrast to the fully networked digital scholar. It is rather to say that existing ways of collaborating among scholars need to be extended and transformed in order to succeed in the new digital environment. It already begins with the posture towards the resources mentioned above: the overall attitude is one of sharing and exchange at the earliest convenient moment. To mention but one example: In the print era, the GSU assembles its material in collation booklets and makes these available to the editor in order for her to create the critical edition of a specific LXX book. By contrast, in the proposed vision, the transcription and collation would be made available electronically, as soon possible, in order to be shared with the community at large. The loss of losing editorial control would be greatly surpassed by the community’s gain of a foundational resource and the ensuing research generated by that publication. The example of favouring the ‘publication’ of foundational resource over final research output points to another aspect of the research culture likely to be affected by the digital turn: a new kind of transparency in terms of contributions to resources and research workflows. In the print era one key mode of collaboratively working towards further knowledge was ‘to stand on the shoulders of others’, namely to use their publications and, after a critical assessment, build upon them or build something altogether new, as the case may be. In the digital era, the exchange is more immediate, or more precisely: at much shorter intervals and interactive. A ‘mistake’ in the transcription of a manuscript text, for example, could

138

juan garcés

easily be corrected and that correction immediately be published, while both original transcription and ‘correction’ may be preserved.160 Contributions—in our case: original transcription and ‘correction’—can be precisely measured. This ‘modularisation’ of the research workflow also means that we ought to appreciate the accumulation of such smaller research contributions when assessing researchers for jobs. In addition to these contributions, there is also a whole host of newer research activities that should equally be acknowledged: modelling of research data and research problems, coding of analytical algorithms, development of research tools and virtual research environments, to name but a few. Last, but not least: digital LXX scholarship will have to entail opening up to and engaging in conversations with those scholars from other disciplines who are further along in the application of computing methods to a research subject. Inter-disciplinary collaboration is often proclaimed yet seldom accomplished. It entails a great deal of humility and respect and the readiness to find common conceptual and terminological ground. It also often implies a steep learning curve, in our case, not least in the area of mathematics and statistical modelling. If other interdisciplinary research in DH is anything to go by, this work will be worth the effort. There is a fair deal of scholarly disagreement as to what the original translators of the Hebrew Pentateuch intended to achieve when they translated the five books into Hellenistic Greek in third century BC Alexandria. One immediate achievement was to open up these important texts to an audience far greater than it would have otherwise been. As we left the twentieth century and entered the twenty-first century with its new technological transformations and possibilities, the Septuagint is experiencing a further reinvention that promises to open it up to even newer usage. I doubt whether the translators dared to dream what would happen with their translations, but I hope that in pursuing the digital approaches presented in this essay, we prove ourselves to be heirs who are worthy of both the work of ‘The Seventy’ and of the twenty-first century.

160 Lest the reader thinks that the need for correction is a downside of premature electronic publication, I shall only mention the need for Berichtigungslisten for edited papyrus texts. See Bodard and Garcés, ‘Open source critical editions’, 93–94. My papyrology teacher once told me: ‘The editio princeps is always incorrect!’.



the seventy and their 21st-century heirs

139

References Abercrombie, J.R., ‘A Computer-Assisted Study of a Textual Family in the Book of Ruth’, Textus 13 (1986), 95–110. Aejmalaeus, A., On the Trail of the Septuagint Translators: Collected Essays, rev. and exp. ed., Contributions to Biblical Exegesis & Theology 50 (Leuven/Paris/Dudley: Peeters, 2007). Austermann, F., ‘Die Septuaginta edieren: Zur Arbeit an der kritischen Ausgabe einer antiken Bibelübersetzung’, Editio 15 (2001), 17–24. Barbrook, A., N. Blake, C. Howe and P. Robinson, ‘The Phylogeny of the Canterbury Tales’, Nature 394 (1998), 839. Barr, J., The Typology of Literalism in Ancient Biblical Translations, Nachrichten der Akademie der Wissenschaften in Göttingen, I. Philologisch-historische Klasse 11, Mitteilungen des Septuaginta-Unternehmens (MSU) 1 (Göttingen: Vandenhoeck & Ruprecht, 1979). Bodard, G., and J. Garcés, ‘Open Source Critical Editions: A Rationale’, in Text Editing, Print, and the Digital World, ed. by M. Deegan and K. Sutherland (Farnham/Burlington: Ashgate Press, 2009), 83–98. Bons, E., and J. Joosten, eds., Septuagint Vocabulary: Pre-history, Usage, Reception, Society of Biblical Literature Septuagint and Cognate Studies 58 (Atlanta: Society of Biblical Literature, 2011). Boyd-Taylor, C., ‘The Evidentiary Value of Septuagintal Usage: Alice’s Response to Humpty Dumpty’, Bulletin of the International Organization for Septuagint and Cognate Studies 34 (2001), 47–80. Boyd-Taylor, C., P.C. Austin and A. Feuerverger, ‘The Assessment of Manuscript Affiliation within a Probabilistic Framework: A Study of Alfred Rahlfs’s Core Manuscript Groupings for the Greek Psalter’, in The Old Greek Psalter: studies in honour of Albert Pietersma, ed. by R.J.V. Hiebert, C.E. Cox and P.J. Gentry, Journal for the Study of the Old Testament Supplement Series 332 (Sheffield: Sheffield Academic Press, 2001), 98–124. Brock, S.P., C.T. Firtsch and S. Jellicoe, comp., A Classified Bibliography of the Septuagint, Arbeiten zur Literatur und Geschichte des hellenistischen Judentums 6 (Leiden: E. J. Brill, 1973). Brook O’Donnell, M., Corpus Linguistics and the Greek of the New Testament, New Testament Monographs 6 (Sheffield: Sheffield Phoenix Press, 2005). Brooke, A.E., and N. McLean, eds., The Old Testament in Greek According to the Text of Codex Vaticanus, 9 volumes (Cambridge: Cambridge University Press, 1906–1940). Burdick, A., J. Drucker, P. Lunenfeld, T. Presner and J. Schnapp, Digital_Humanities (Cambridge, MA and London: The MIT Press, 2012). http://mitpress.mit.edu/sites/default/ files/titles/content/9780262018470_Open_Access_Edition.pdf, last accessed 8 May 2013. Colwell, E.C., ‘Biblical Criticism: Lower and Higher’, Journal of Biblical literature 67 (1948), 1–12. Davila, J.R., Journal for the Study of the Pseudepigrapha 15:1 (September 2005), 3–61. Deissmann, A., Bible Studies: Contributions Chiefly from Papyri and Inscriptions to the History of the Language, the Literature, and the Religion of Hellenistic Judaism and Primitive Christianity, trans. A. Gireve (Edinburgh: T&T Clarke, 1900). ——, New light on the New Testament from Records of the Graeco-Roman Period, trans. L.R.M. Strachan (Edinburgh: T&T Clark, 1907). ——, Light from the Ancient East: the New Testament Illustrated by Recently Discovered Texts of the Graeco-Roman World, trans. L.R.M. Strachan (New York/London: Hodder and Stoughton, 1910). Dines, J., The Septuagint (London/New York: T&T Clark, 2004). Dogniez, C., ed., Bibliography of the Septuagint/Bibliographie de la Septante (1970–1993), Supplement to Vetus Testamentum 60 (Leiden: E. J. Brill, 1995). Evans, T.V., Verbal Syntax in the Greek Pentateuch: Natural Greek Usage and Hebrew Interference (Oxford: Oxford University Press, 2001).

140

juan garcés

Felsenstein, J., Inferring Phylogenies (Sunderland: Sinauer Associates, 2004). Fernández-Marcos, N., Scribes and Translators: Septuagint and Old Latin in the Book of Kings, Supplements to Vetus Testamentum 54 (Leiden/New York/Köln: Brill, 1994). ——, The Septuagint in Context: Introduction to the Greek Version of the Bible, trans. W.G.E. Watson (Leiden: Brill, 2000). Fernández-Marcos, N., and J.R. Busto Saiz, eds., El texto antioqueno de la Biblia Griega, 3 volumes, Textos y estudios ‘Cardenal Cisneros’ 50 (1–2 Samuel); 53 (1–2 Reyes); 60 (1–2 Crónicas) (Madrid: Instituto de Filología, C.S.I.C, Departamento de Filología Bíblica y de Oriente Antiguo, 1989–1996). Fernández-Marcos, N., and M.V. Spottorno Díaz-Caro, co-ord., La biblia griega: Septuaginta, 5 volumes (Salamanca: Ediciones Sígueme, 2008ff). Fernández-Marcos, N., M.V. Spottorno Díaz-Caro, and J.M. Cañas Reíllo, eds., Índice griegohebreo del texto antioqueno en los libros históricos, 2 volumes, Textos y estudios ‘Cardenal Cisneros’ 75 (Madrid: Instituto de Filología, C.S.I.C, Departamento de Filología Bíblica y de Oriente Antiguo, 2005). Grafton, A., and M. Williams, Christianity and the Transformation of the Book: Origen, Eusebius, and the Library of Caesarea (Cambridge/London: The Belknap Press of Harvard University Press, 2006). Gries, S.Th., ‘What is Corpus Linguistics?’, Language and Linguistics Compass 3 (2009), 1–17. Hiebert, R.J.V., ‘Establishing the Textual History of Greek 4 Maccabees’, in Die Septuaginta—Texte, Theologien, Einflüsse, ed. by W. Kraus and M. Karrer, Wissenschaftliche Untersuchungen zum Neuen Testament 252 (Tübingen: Mohr Siebeck, 2010), 117–139. ——, ed., Septuaginta: Vetus Testamentum Graecum Auctoritate Academiae Scientiarum Gottingensis editum, Vol. 9.4: Maccabaeorum liber IV (Göttingen: Vandenhoeck & Ruprecht, forthcoming). Hiebert, R.J.V., and N.N. Dykstra, ‘Septuagint Textual Criticism and the Computer: 4 Maccabees As a Test Case’, in XIII Congress of the International Organization for Septuagint and Cognate Studies, Ljubljana, 2007, ed. by M.K.H. Peters, Society of Biblical Literature Septuagint and Cognate Studies 55 (Atlanta: Society of Biblical Literature, 2008), 167–181. ——, ‘What Does the Computer Have to Do With Textual Criticism? Innovative Technology for the Management and Analysis of Collation Data and the Grouping of Manuscripts’, in Die Göttinger Septuaginta: ein editorisches Jahrhundertprojekt, ed. by R.G. Kratz, Abhandlungen der Akademie der Wissenschaften zu Göttingen (Berlin: de Gruyter, forthcoming). Howe, C.J., R. Connolly and H.F. Windram, ‘Responding to criticisms of phylogenetic methods in stemmatology’, SEL Studies in English Literature 1500–1900 5:1 (Winter 2012), 51–67. Jobes, K.H., ‘Karen Jobes Responds to Tim McLay’, in Bulletin of the International Organization for Septuagint and Cognate Studies 30 (1997) 32–35. ——, ‘A Comparative Syntactic Analysis of the Greek Versions of Daniel: a Test Case for New Methodology’, Bulletin of the International Organization for Septuagint and Cognate Studies 28 (1995), 19–41. ——, ‘Quantitative Methods for Exploring the Relationship Between Books of the Septuagint’, The Bible As Book: the Transmission of the Greek Text, ed. by S. McKendrick and O.A. O’Sullivan (London: The British Library, 2003), 73–95. Kahle, P.E., The Cairo Geniza, The Schweich Lectures of the British Academy (London: Oxford University Press, 1959). Karrer, M., and W. Kraus, eds., Septuaginta Deutsch: das griechische Alte Testament in deutscher Übersetzung, (Stuttgart: Deutsche Bibelgesellschaft, 2009). Kielsmeier-Jones, A., ‘How to Read and Understand the Göttingen Septuagint: A Short Primer, part 1’. https://abramkj.wordpress.com/2012/11/04/how-to-read-and-understandthe-gottingen-septuagint-a-short-primer-part-1/, last accessed 8 May 2013.



the seventy and their 21st-century heirs

141

——, ‘How to Read and Understand the Göttingen Septuagint: A Short Primer, part 2 (Apparatus)’. https://abramkj.wordpress.com/2013/01/11/how-to-read-and-understandthe-gottingen-septuagint-a-short-primer-part-2-apparatus/, last accessed 8 May 2013. Klauck, H.-J., 4. Makkabäerbuch, Jüdische Schriften aus hellenistisch-römischen Zeit 3/6 (Gütersloh: Gerd Mohn, 1989). Kraft, R.A., ‘CATSS: Treatment of the Greek textual variants’ (2004). http://ccat.sas.upenn .edu/gopher/text/religion/biblical/lxxvar/0–Explan.htm, last accessed 8 May 2013. Lee, J.A.L., A Lexical Study of the Septuagint Version of the Pentateuch, Society of Biblical Literature Septuagint and Cognate Studies 14 (Chico/CA: Scholars Press, 1983). ——, A History of New Testament Lexicography, Studies in Biblical Greek 8 (New York et al.: Peter Lang, 2003). ——, ‘The Present State of Lexicography of Ancient Greek’, in Biblical Greek Language and Lexicography: Essays in Honor of Frederick W. Danker, ed. by B.A. Taylor, J.A.L. Lee, P.R. Burton and R.E. Whitaker (Grand Rapids/Cambridge: William B. Eerdmans Publishing Company, 2004), 66–74. ——, ‘A Lexicographical Database for Greek: Can It Be Far Off? The Case of amphodon’, in Die Septuaginta—Texte, Kontexte, Lebenswelten: Internationale Fachtagung veranstaltet von Septuaginta Deutsch (LXX.D), Wuppertal 20.-23. Juli 2006, ed. by M. Karrer and W. Kraus, coll. M. Reiser, Wissenschaftliche Untersuchungen zum Neuen Testament I 129 (Tübingen: Mohr Siebeck, 2008), 214–222. ——, ‘Review of T. Muraoka, A Greek-English Lexicon of the Septuagint (2009)’, Bulletin of the International Organization for Septuagint and Cognate Studies 43 (2010), 115–125. Louw, J.P., and E.A. Nida, eds., Greek-English Lexicon of the New Testament Based on Semantic Domains, 2 volumes (New York: United Bible Societies, 1988). Lüdeling, A., and M. Kytö, eds., Corpus Linguistics: An International Handbook, 2 volumes, Handbücher zur Sprach- und Kommunikationswissenschaft 29 (Berlin/New York: Walter de Gruyter, 2008, 2009). Lust, J., E. Eynikel and K. Hauspie, A Greek-English Lexicon of the Septuagint, revised edition (Stuttgart: Deutsche Bibelgesellschaft, 2003. First published in two volumes in 1992 and 1996). Martin, R.A., Syntactic evidence of Semitic Sources in Greek Documents, Society of Biblical Literature Septuagint and Cognate Studies 3 (Cambridge: Scholars Press, 1974; repr. Eugene: Wipf & Stock Publishers, 2004). McLay, T., ‘Syntactic Profiles and the Characteristics of Revision: a Response to Karen Jobes’, Bulletin of the International Organization for Septuagint and Cognate Studies 29 (1996), 15–21. Merle Rife, J., ‘The Mechanics of Translation Greek’, Journal of Biblical Literature 52:4 (December 1933), 244–252. Moulton, J.H., and G. Milligan, Vocabulary of the Greek Testament (London: Hodder and Stoughton, 1930). Muraoka, T., A Greek-English Lexicon of the Septuagint: chiefly of the Pentateuch and the Twelve Prophets (Louvain: Peeters, 2002). ——, A Greek-English Lexicon of the Septuagint (Louvain: Peeters, 2009). ——, A Greek ~ Hebrew/Aramaic two-way index to the Septuagint (Louvain/Paris/Walpole: Peeters, 2010). Neuschäfer, B., ‘Die Göttinger Septuaginta-Ausgabe—Standortbestimmung eines editorischen Jahrhundertprojekts: Internationale Fachtagung, Göttingen, 28.–30. April 2008’, Editio 22 (2008), 241–245. Oakes, M.P., and M. Ji, eds., Quantitative Methods in Corpus-Based Translation Studies: A Practical Guide to Descriptive Translation Research, Studies in Corpus Linguistics 51 (Amsterdam/Philadelphia: John Benjamins Publishing Company, 2012). Pietersma, A., ‘The Present State of the Critical Text of the Greek Psalter’, in Der Septuaginta-Psalter und seine Tochterübersetzungen, ed. by A. Aejmelaeus and U. Quast,

142

juan garcés

Abhandlungen der Akademie der Wissenschaften zu Göttingen; Mitteilungen des Septuaginta-Unternehmens 24 (Göttingen: Vandenhoeck & Ruprecht, 2000), 12–32. ——, ed., A New English Translation of the Septuagint and the Other Greek Translations Traditionally Included Under That Title (New York/Oxford: Oxford University Press, 2007). ——, ‘Beyond Literalism: Interlinearity Revisited’, in Translation is required: the Septuagint in retrospect and prospect, ed. by R.J.V. Hiebert, Society of Biblical Literature Septuagint and cognate studies 56 (Atlanta: Society of Biblical Literature, 2010), 3–21. Rahlfs, A., Verzeichnis der griechischen Handschriften des Alten Testaments, für das Septuaginta-Unternehmen zusammengestellt, Mitteilungen des Septuaginta-Unternehmens der Königlichen Gesellschaft der Wissenschaften zu Göttingen 2 (Berlin: Wedimannsche Buchhandlung, 1914). ——, Psalmi cum Odis, Septuaginta: Societatis Scientiarum Gottingensis 10 (Göttingen: Vandenhoeck & Ruprecht, 1931). ——, Septuaginta, id est Vetus Testamentum Graece iuxta LXX interpretes, 2 volumes (Stuttgart: Württenbergische Bibelgesellschaft, 1935). ——, Der Text des Septuaginta-Psalters, Septuaginta-Studien, vol. 2 (Göttingen: Vandenhoeck & Ruprecht, 1907, 19652). Rahlfs, A., ed., and D. Fraenkel, rev., Verzeichnis der griechischen Handschriften des Alten Testaments, vol. 1: Die Überlieferung bis zum VIII. Jahrhundert, Septuaginta: vetus testamentum graecum supplementum 1 (Göttingen: Vandenhoeck & Ruprecht, 2004). Schäfer, C., Benutzerhandbuch zur Göttinger Septuaginta, vol. 1: Die Edition des Pentateuch von John William Wevers (Göttingen: Vandenhoeck & Ruprecht, 2012). Schleusner, J.F., Novus thesaurus philologico-criticus sive lexicon in LXX et reliquos interpretes graecos ac scriptores apocryphos veteris testamenti, 5 volumes (Leipzig: Weidmann, 1820–21. Repr. Turnhout: Brepols, 1995). Sinclair, J., ‘Current Issues in Corpus Linguistics’, in Trust the Text: Language, Corpus and Discourse, Sinclair, J., ed. with R. Carter (London/New York: Routledge, 2004), 185–193, 199. Smend, R., ‘Der geistige Vater des Septuaginta-Unternehmens’, in Studien zur Septuaginta—Robert Hanhart zu Ehren aus Anlaß seines 65. Geburtstages, ed. by D. Fraenkel, U. Quast, and J.W. Wevers, Anhandlungen der Akademie der Wissenschaften in Göttingen: Mitteilungen des Septuaginta-Unternehmens (MSU) 20 (Göttingen: Vandenhoeck & Ruprecht, 1990), 332–344. Stefanowitsch, A., and S.Th. Gries, ‘Collostructions: Investigating the Interaction of Words and Constructions’, International journal of corpus linguistics 8:2 (2003), 209–243. Taylor, B.A., ‘Review of Lust et al., A Greek-English Lexicon of the Septuagint’, TC: A Journal of Biblical Textual Criticism 3 (1998). http://www.reltech.org/TC/v03/Lust-etal1998rev .html, last accessed 8 May 2013. Tov, E., ed., A Classified Bibliography of Lexical and Grammatical Studies on the Language of the Septuagint and Its Revisions (Jerusalem: Academon, 19833). ——, ‘The Nature and Study of the Translation Technique of the LXX in the Past and Present’, in VI Congress of the International Organization for Septuagint and Cognate Studies, Jerusalem, 1986, ed. by C.E. Cox, Society of Biblical Literature Septuagint and Cognate Studies 23 (Atlanta: Scholars Press, 1987), 337–359. ——, The Text-Critical Use of the Septuagint in Biblical Research, Jerusalem Biblical Studies 8 (Jerusalem: Simor Ltd., 1997). Tov, E., and B.G. Wright, ‘Computer-Assisted Study of the Criteria for Assessing the Literalness of Translation Units in the LXX’, Textus 12 (1985), 149–187. van Reenen, P. and M. van Mulken, eds., assist. J. Dyke, Studies in Stemmatology (Amsterdam/Philadelphia: John Benjamins Publishing Company, 1996). van Reenen, P., A. den Hollander and M. van Mulken, assist. A. Roeleveld, Studies in Stemmatology II (Amsterdam/Philadelphia: John Benjamins Publishing Company, 2004). Wade, M.L., Consistency of Translation Techniques in the Tabernacle Accounts of Exodus in the Old Greek, Society of Biblical Literature Septuagint and Cognate Studies 49 (Atlanta: Society of Biblical Literature, 2003).



the seventy and their 21st-century heirs

143

Wevers, J.W., ed., Septuaginta: Vetus Testamentum Graecum: Auctoritate Academiae Scientiarum Gottingensis editum, vol. I: Genesis (Göttingen: Vandenhoeck & Ruprecht, 1974). ——, Text History of the Greek Genesis, Abhandlungen der Akademie der Wissenschaften in Göttingen: Pholologisch-historische Klasse, dritte Folge 81; Mitteilungen des Septuaginta-Unternehmens 11 (Göttingen: Vandenhoeck & Ruprecht, 1974). ——, Notes on the Greek Text of Genesis, Septuagint and Cognate Studies 35 (Atlanta: Scholars Press, 1993). Wright, B.G., No Small Difference: Sirach’s Relationship to Its Hebrew Parent Text, Society of Biblical Literature Septuagint and Cognate Studies 26 (Atlanta: Scholars Press, 1989).

digital approaches to the study of ancient monotheism Ory Amitay Monotheism—that ill-defined complex of so-called Abrahamic Religions—is one of the strongest, most enduring and all-pervasive ideological phenomena in human history. Remarkably, unlike its main three constituent units (Judaism, Christianity, Islam), a comprehensive history of Monotheism as a whole has not yet been undertaken by historical research.1 Scholarly discussion of Monotheism is usually limited to the field of Biblical Studies, and is directed at the theology and origins of this phenomenon rather than at its history.2 The purpose of this paper is to set out a research program, which aims to use various digital, conceptual and computational approaches in order to prepare the ground for writing a history of Monotheism in Antiquity. The most immediate obstacle to pursuing this ambitious goal is presented by the overwhelming amount of data, and of the wide variety of scholarly proficiencies, required to tackle this vast subject: Cuneiform, Egyptology, Biblical scholarship, Jewish history from Hellenism to Islam, Rabbinics, New Testament and Church Fathers, Classics, Neo-Platonism and Gnostics, Iranology, Arab studies, to say nothing of the art history, archaeology, epigraphy and numismatics of most of the ancient world. Such a task is hopelessly beyond the capabilities of any one scholar. It can only be undertaken by a large, well-coordinated team. Which is where Wiki comes in. In a little over a decade since the onset of Wikipedia, MediaWiki software has proven its worth as a platform for mass collaboration in presenting and organizing information. Its greatest

1  Important and illuminating, if partial, exceptions to the rule are: Albright, From the Stone Age to Christianity: Monotheism and the Historical Process; Armstrong, A History of God: the 4000-Year Quest of Judaism, Christianity, and Islam; Assman, Moses the Egyptian: the Memory of Egypt in Western Monotheism; Assman, Of God and Gods: Egypt, Israel, and the Rise of Monotheism. None of these, however, attempt to encompass the entire subject. 2 This was the status quo of current scholarship as expressed in the special threesession section titled ‘The Concept of Monotheism: Should it Have a Future in Biblical Studies’ in the 2011 London conference of the Society of Biblical Literature. For useful surveys of biblical scholarship on the origins and theology of Monotheism see MacDonald, Deuteronomy and the Meaning of ‘Monotheism’, 5–58; Assman, Of God and Gods.

146

ory amitay

achievement is doubtless its success in harnessing a very large number of editors worldwide to participate in such a large-scale joint project. This success inspires confidence in the suitability of MediaWiki for the suggested project, both for the purpose of organizing the acquired data and for accommodating the necessary teamwork. The fact that MediaWiki is an open sources software, and the immense popularity of Wikipedia, at least render it likely that the software will enjoy continued technical support, and that when the technology finally becomes obsolete, the possibility of transition to a newer platform will be readily available. Yet another cardinal advantage of MediaWiki is provided by the nature of hypertext. On the most basic level, links give us the facility of moving around from one context to another, through defined words or expressions. Beside its general usefulness, however, hypertextuality also challenges the inherent linear nature of traditional historical writing. By organizing the information side by side, rather than in concatenation, the research team (henceforth ‘we’) will be able more easily to keep prejudices in check, and to defer judgment to a time when a considerable mass of data had been assembled. This kind of mental flexibility is an important prerequisite in a project that is both extremely complex and emotionally charged. The need to keep an open mind is nowhere as evident as regarding the very name of our subject matter. Is ‘monotheism’ even the right word for our purpose? One disturbing fact is that the term certainly does not go back all the way to Antiquity, but is rather a neologism by the Cambridge Platonist Henry More.3 May we use the term so anachronistically?4 And how are we to avoid giving too much weight in the discussion to theological matters, ‘monotheism’ being a theological definition? A more acceptable term might be ‘Abrahamic Religions’ (used for example in the title of a recently established Chair at Oxford). On the other hand, Abraham figures much more prominently in Judaism and Islam than in Christianity, where stories about him are not as often told, and the quintessential physical manifestation of his covenant with God, circumcision, is not

3 In his in 1660 treatise An Explanation of the Grand Mystery of Godliness he wrote: ‘To make the World God is to make no God at all; and therefore this kind of Monotheisme of the Heathen is as rank Atheism as their Polytheisme was proved to be before’ (More, An Explanation, 62; cited in MacDonald Deuteronomy, 206). 4 A reply to the argument form anachronism, which has nothing to do with DH but is nonetheless worthy of mention, is that we do have references to a veritably ancient term: πολυθεΐα, used by Philo, De Mutatione Nominum 205, Procopius, Arcana Historia 19 and De Aedificiis 6.2. If there is good evidence for a parallel to our ‘polytheism’ in antiquity, we may infer ‘monotheism’ from it.



digital approaches to the study of ancient monotheism 147 0,0002%

monotheism

Monotheism

Abrahamic

0,00016% 0,00012% 0,00008% 0,00004% 0,00% 1800 1810 1820 1830 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Figure 1. Google Ngram viewer.

generally observed. Furthermore, the name ‘Abrahamic religions’ (in the plural) puts the emphasis on the constituent units rather than on the whole, as opposed to ‘Monotheism’. A good many more arguments can easily be made for and against each term, but none is likely to convince completely. Can a digital approach allow us to address the question from a new angle? A simple example of how this can be done is offered along the path taken by the Culturomics project, using the Google Ngram viewer (GNv) on the Google Books database.5 As the table above shows, I have searched for both ‘monotheism’ and ‘Monotheism’, as well as for ‘Abrahamic’. The search traces the frequency of the use of each word in a vast collection of English and North American literature from the last two centuries, spread along a chronological timeline (smoothing: 2). An immediate observation is that monotheism has a clear quantitative advantage over everything and anything Abrahamic. In terms of recognition, at least, the monotheistic trend enjoys a solid lead.6 But there is much more to learn from this simple search. One point is that ‘Abrahamic’ has seen a steady rise in the last decades, increasingly so most recently, and may have a future after all. In fact, it is now at its most frequent since the 1850s. On the other hand, since GNv is case sensitive, the ‘score’ for monotheism is actually an aggregate of the red and blue lines, way above the yellow line of ‘Abrahamic’. Interesting questions arise also from the chronological angle: what happened in the early 1840s, which gave birth to this explosion of writing about Monotheism, lasting about half a century? The answer may well lie in the question, or rather in the books which talked about Monotheism: an automatically generated ­reading

5 Michel et al., ‘Quantitative Analysis of Culture Using Millions of Digitized Books’. 6 All the more so, given that ‘Abrahamic’ may also relate to ‘covenant’ or other nouns, which lie outside the definition we are looking for here.

148

ory amitay

list, defined by the frequency of a single word during a given period, and already linked to Google Books, where many of the titles can be easily found. With early nineteenth century books happily in the public domain, this kind of research is well within the realm of possibility even without access to a superb library. Even more questions arise from what is (at the user end) a quirk of GNv—its case sensitivity. In what is perhaps the main attention grabber in this graph we see a spike in the use of both ‘monotheism’ and ‘Monotheism’ around 1870, and yet a higher one two decades later (the highest ever). However, in the 1890s there is a sharp rise for lower-case monotheism but only a moderate rise for the uppercase Monotheism (which does not even duplicate the spike of 1870). How are we to explain these data? Is there a discernible semantic difference between the different spellings? Do authors use either spelling consistently, or can an author use both? And to what end? Answering any of the questions above in earnest (let alone deciding on the monotheism/Abrahamic question) is well beyond the scope of this paper, or of this rudimentary search. To be sure, a serious study will have to include many more searches, and use any raw data that may be supplied by Google. An important lesson from this short experiment is, to my mind, that by attempting to answer a particular question we find ourselves facing a great many new ones. Even more importantly, these new questions arise not from our own prejudices, but rather from objective data, presented graphically. This, I suggest, ought to be taken into consideration from this point on: the more we expand our databases and hone our methods of searching, the more will patterns reveal themselves, allowing us a deeper insight into the inner mechanisms of our subject of study. This brings us back to our own database on the wikisite, and to the basic questions of definition and organization. In answer to these questions I propose to use two innovative methodologies, respectively: Memetics and Network Theory (NT). Memetics is the business of tracking the replication and mutation of memes. A meme (rhymes with cream) is to culture what a gene is to evolutionary biology, making up our ‘cultural DNA’. A meme can be an idea, a name or number, a form of behavior or an organizing principle, which can be replicated from one brain to another by whatever means. Memetic theory was first formulated by Dawkins, and elaborated significantly by Blackmore.7 At the heart of the theory lies

7 Dawkins, The Selfish Gene, chap. 11; Blackmore, The Meme Machine. For attempts to use Memetics in Humanities research see for example Middleton, ‘Mycenaeans, Greeks,



digital approaches to the study of ancient monotheism 149

the assumption that memes have their own evolutionary process, which runs alongside genetic evolution. Memetics is thus a holistic approach, to which no form of the human experience is foreign. The use of memetic language is therefore particularly useful in putting very different memes on equal footing, providing a powerful descriptive and interpretative tool for interdisciplinary research. It also allows us to measure our data, thus tentatively opening the road towards quantitative argumentation. The memetic approach is easily applicable to the MediaWiki format. Simply put, each meme is assigned its own page on the Wiki site. Each meme page is tagged as such. At the head of each meme page there is a longer definition of the meme, much like the leading definition in a Wikipedia article. This longer definition is fluid, setting out and explaining the exact boundaries of each meme, as they develop along with the research project.8 Below the (automatically generated) contents table each meme page records all of its appearances in all the sources covered by the database. Each source, in turn, also has a page of its own, sourcetagged, linking to the pages of all the memes it contains. Naturally, finer distinctions will have to be made. Each kind of meme (visual, ideal, narrative, onomastic, numeric, behavioral, organizational etc.) ought to have its own tag. Sources, too, will have to be tagged for more detail, including date, language, religious affiliation, including an internal tagging system according to the accepted inner division of the text (e.g.—book, chapter and verse). Consider, for example, Luke 3. The source page for this chapter is to be tagged: New Testament, Gospel, Luke, Luke_3, Greek, Christianity, 1 century CE. From this page links will go out to meme pages including (but hardly limited to): baptism, Holy Ghost, dove, serpent, forgiveness of sin, Seth, Adam, as well as to another source page—Isaiah 40 (Old Testament, Prophets, Isaiah, Isaiah_40, Hebrew, Judaism). In itself, the picture presented by this breakdown is interesting, but of limited value. However,

Archaeology and Myth: Identity and the Uses of Evidence in the Archaeology of Late Bronze Age Greece’; Campbell, Lucretius on Creation and Evolution: a Commentary on De Rerum Natura, Book Five, Lines 772–1104; Varela, ‘Vortex to Virus, Myth to Meme: The Literary Evolution of Nihilism and Chaos in Modernism and Postmodernism’; Amitay, From Alexander to Jesus; Kneis, The Emancipation of the Soul: Memes of Destiny in American Mythological Television. 8 It seems probable at this stage of theoretical speculation that with time sets of submemes will form. The question of memes and memeplexes (a group of memes in alliance and cooperation), that is of memetic hierarchy, may be expected to require deep consideration, especially in the preliminary stages of this research project.

150

ory amitay

when put into memetic context in a hypertextual database with a critical mass of pages and links, it may be subjected to empirical modes of analysis. This kind of analysis will be based on Network Theory, as formulated, practiced and popularized by Barabási.9 The fascinating find of Barabási and his team is that when the all the nodes and links are mapped, according to the number of links going in and coming out of each node, various kinds of scale-free networks, from the Internet through genetic configurations to ecological systems, show a repeating pattern: a few nodes have a disproportionately large number of links, and an increasingly large number of nodes have an increasingly small number of links. Graphically displayed, we get a small number of nodes (memes and sources, in our database) at the top, then a significant drop and a long tail. The tagging system outlined above (which will surely require expansion, possibly even reconfiguration as research advances) will enable both a search of the entire database, and more specific searches, defined through the tags. In effect, within the confines of our database network, each tag can operate as the ontology directory is designed to do for the Semantic Web, providing not only the content of the page but also various definitions (tags) regarding its nature and circumstances.10 Going back to more traditional approaches, the research team will pay special attention to those nodes (either memes or texts) that came at the top of the list with the largest number of links. If the principle developed by Barabási et al. proves true in our case as well, the result of this kind of network analysis ought to point to us what are the ‘key players’ in the monotheistic game. This kind of information may help us deal with one of the thorniest methodological questions involved in the study of ancient (and modern) Monotheism: how can we write a history of the monotheistic phenomenon itself, rather than a composite history of the various members of the Abrahamic club? Which principles, characters, myths and practices are specifically monotheistic? I suggest that the answers are to be found in the key nodes. The memes (and texts) that are linked to the largest number of other memes and texts, that is, the most prevalent, replicated and preserved memes, are the main characters of our story. By 9 Barabási, Linked: the New Science of Networks. 10 This solution is not offered in regard to the Semantic web in general. The ambitious goals set by Berners-Lee in his 2001 challenge still await a comprehensive solution. However, due to the less ambitious goal of the project suggested here, and especially due to the fact that the database is to be put together and tagged from day one, the principle presented here may hopefully be usable and successful.  



digital approaches to the study of ancient monotheism 151

taking this methodological approach we may (a) neutralize our prejudices as much as possible, (b) be better situated to appreciate Monotheism as a whole rather than as a sum of its constituent units, and (c) enjoy the random nature of digital searches, which often yield unexpected results and lead to new questions which we do not yet even know that we need to ask. The suggested methodology, calling for a definition and tracking of memes, followed by NT analysis, need not be limited to the wiki database envisioned here. On the contrary, throughout the research process, and especially in its earlier stages, when the database has not yet reached a critical mass of data, it may prove fruitful to use the same technique on other databases, as numerous and varied as possible. Only in this way can we construct, at a more advanced stage, a truly scale-free network, which would put any computational analyses on a more solid basis. Let us therefore envision a number of ‘fantasy searches’, disregarding the obvious difficulties presented by copyright laws and by computer programming needs.11 Returning to Google Books, we have already seen how a simple search using GNv has both given us an (admittedly partial) answer to the question how to define our topic, opened a number of new questions, and provided us with a distinct bibliography. But there is much more that can be done. The way proposed in this paper to circumvent the theological preoccupation of the lion’s share of existing scholarship on Monotheism is to work our way backwards. Today, we seem certain enough in deciding which religions may be classified as monotheistic or Abrahamic, and which are not. This has less to do with any theological feeling, and more with an intuitive sense of sameness. In other words: which memes form the most connected nodes in the monotheistic/Abrahamic network? Let us imagine the following search: to begin, a set of memes is defined, prejudicially, as essentially monotheistic. This set is defined as ‘mono1’. Using an enhanced search engine, the proportional frequency (tf-idf) of all memes in mono1 is checked in all titles in Google Books.12 The data

11  Both issues will, of course, require a great deal of attention when this research program is actually put into action. However, such difficulties are technical in nature, and immaterial to the theoretical aspects of the program, which are at the focus of this paper. 12 For my own GNv search I used English, 1800–present, which is currently the largest and fullest collection on offer. In order to keep the Anglophone prejudices in check, parallel searches ought to be conducted in other languages as well, to serve as a linguistic-based control group.

152

ory amitay

are presented as a table. Vertically, we get an immensely long list of titles in alphabetical order, while horizontally we get the frequency value for each mono1 meme in the same text. Organized in this fashion, the huge size of the table will likely leave us more than a little helpless. However, if organized by frequency of particular memes, it may reveal highly interesting results. Which are the books where Abraham is most frequently mentioned? Which other memes in mono1 are also frequently mentioned in the same books? How will the numbers be affected if we take ‘Lincoln’ out of the equation? More sophisticated searches can be devised, to produce better focused and more manageable results. For example, after some initial tinkering with the kind of search suggested above, it may be possible to decide on an average normalized frequency value as a ‘gatekeeper’, limiting further searches only to those books which contain a certain number of mono1 memes more frequently than gatekeeper-level. This kind of limited search is likely to yield a much shorter list of titles, one which will be easier to survey and to use for further searches. It may also be possible to extend the list of titles or to cut it even shorter, lowering or raising gatekeeper value according to need. It may be expected that when enough searches had been made and documented, memetic patterns will start to appear, with possible important implications for our subject matter. One result which may be hypothesized is a manifestation of strongly interconnected meme groups, or memeplexes. For example, it will not be a surprise that works which most frequently contain the word ‘Crucifixion’ will also be more likely to contain ‘Christianity’ and ‘baptism’ than ‘mezuzah’ and ‘Allah’. It is not as easy to guess regarding all the books with a high frequency value for Christianity, Crucifixion and baptism, which animal will be most frequently mentioned: a dove, a serpent or a donkey; which typological numbers are more likely to appear; or which biblical verses are most likely to be cited. Furthermore, will the numbers change or remain much the same, if we make a further distinction between Crucifixion and crucifixion? What other words appear significantly more frequently in these books than in Google Books in general? Are any of them worthy of being added to mono1, to create an expanded set ‘mono2’? A final point to be addressed concerns the power of crowd sourcing and crowd wisdom. It is, I believe, in the best interest of the suggested research program to conduct as much as possible of its business in complete transparency. Furthermore, it seems to me imperative to complement the program with a conscious effort to reach a wide audience on the web, and to explore ways to involve this audience in the process of



digital approaches to the study of ancient monotheism 153

research. Such an effort can benefit the program in a number of ways. First, in providing a sounding board to the project as it progresses. A community of second-tier researchers and a third-tier general reading audience may provide insights, correct prejudices, and even assist with the vast amount of reading required. Second, this peripheral community may serve as yet another control group (of sorts) to the main channel of investigation. Given that the logic of the program is based on the assumption that Monotheism is intuitively recognizable today, it seems worthwhile to get a response to this, and many other subsidiary arguments, from a large number of interested readers and part-time participants. Thirdly, such a community can help in the general dissemination of information gathered through this program, whether by incorporating it in Wikipedia articles or in blogs, social networks, etc. Such online activity may not only benefit the program directly, as suggested above, but also advance it through popularizing its contents and attracting public and media attention. In turn, it will also provide a natural avenue for giving back to the community—a desired benefit of Digital Humanities in general. To conclude, the program suggested in this paper makes use of various DH approaches. On the simplest level, a digital solution (MediaWiki) is sought in order to address an immediate technical difficulty: the need for a tried and true tool for collaborative research. Overcoming obstacles of time and space, the digital approach is a practical sine qua non for this kind of project. On a more philosophical level, wiki-style hypertextuality is a strong facilitator of mental flexibility. The nature of MediaWiki is admirably congruent with memetic theory, and easily applicable to NT theory. Our aim is to use both these methodologies in our analysis of the information, both in our own database and in other databases too. By doing this, we hope to pave a way towards a digitally based quantitative analysis of material usually limited to more traditional methodological approaches within the Humanities. References Albright, William Foxwell, From the Stone Age to Christianity: Monotheism and the Historical Process (Baltimore: Johns Hopkins Press, 1957). Amitay, Ory, From Alexander to Jesus (Berkeley: University of California Press, 2010). Armstrong, Karen, A History of God: the 4000-Year Quest of Judaism, Christianity, and Islam (New York: A.A. Knopf, 1993). Assmann, Jan, Moses the Egyptian: the Memory of Egypt in Western Monotheism (Cambridge, Mass: Harvard University Press, 1998). ——, Of God and Gods: Egypt, Israel, and the Rise of Monotheism (Madison, Wisc: University of Wisconsin Press, 2008).

154

ory amitay

Barabási, Albert-László, Linked: the New Science of Networks (Cambridge, Mass: Perseus, 2002). Berners-Lee, Tim, James Hendler and Ora Lassila, ‘The Semantic Web’, Scientific American Magazine (2001). Retrieved 20.01.2013. Blackmore, Susan J., The Meme Machine (Oxford: University Press, 1999). Campbell, Gordon Lindsay, Lucretius on Creation and Evolution: a Commentary on De Rerum Natura, Book Five, Lines 772–1104 (Oxford: Oxford University Press, 2003). Dawkins, Richard, The Selfish Gene (Oxford: University Press, 19892). Kneis, Philipp, The Emancipation of the Soul: Memes of Destiny in American Mythological Television (Frankfurt am Main: Peter Lang, 2010). MacDonald, Nathan, Deuteronomy and the Meaning of ‘Monotheism’ (Tübingen: Mohr Siebeck, 2003). Michel, Jean-Baptiste et al., ‘Quantitative Analysis of Culture Using Millions of Digitized Books’, Science 331, 6014 (2010), 176–182. Middleton, Guy D., ‘Mycenaeans, Greeks, Archaeology and Myth: Identity and the Uses of Evidence in the Archaeology of Late Bronze Age Greece’, Eras Journal 3. 2002. http://www.arts.monash.edu.au/publications/eras/edition-3/middleton.php, last accessed 14 January 2013. Varela, Julio A, ‘Vortex to Virus, Myth to Meme: The Literary Evolution of Nihilism and Chaos in Modernism and Postmodernism’, Electronic Theses, Treatises and Dissertations (2004), paper 4584. http://diginole.lib.fsu.edu/etd/4584.

internet networks and academic research: the example of new testament textual criticism Claire Clivaz 1. Introduction According to Michael Nielsen in Reinventing Discovery,1 we are now facing the most important turning point in sciences for three centuries, but ‘it is a slow revolution that has quietly been gathering steam for years’.2 Today all scholars can have the impression that their work is evolving in so fast a way that it becomes really difficult to adapt themselves efficiently. But, as I will argue in this chapter, scholars in other fields of research, who already have extensive experience of the advantages and risks of Digital Humanities can offer valuable insights to researchers in the related fields of Biblical, early Jewish, and Christian studies. In order to show this, I will begin by reviewing some points about the Digital Humanities landscape and the emergence of open scientific networks. Then, I will present some examples from the online networks used in my own field, New Testament textual criticism field, and I will evaluate them in the light of my previous observations. Finally, in my conclusion, I will offer suggestions about what we could do to improve the ways in which we use technology to collaborate in our work in New Testament textual criticism (NTTC). The way in which scholars conduct their research is often studied by specialists with an interest in the sociology of knowledge. I think it is imperative that subject specialists should also be involved in the analysis of what happens now in the production of their own knowledge. Both perspectives are complementary. If a sociological inquiry can offer a critical point of view on the process of transformation of knowledge, only the scholars belonging to the particular field of study can claim exactly their needs in the development of tools and new ways of research communication. An interdisciplinary research is therefore required, as I will argue in my conclusion. 1  Nielsen, Reinventing Discovery. Thank you to my sociologist colleague Olivier Glassey for this reference. 2 Nielsen, Reinventing Discovery, 10.

156

claire clivaz 2. The Open Scientific Network in Contemporary Digital Culture

The history of the interactions between Humanities and computers is quite long, and was inaugurated in the 1940s for the purposes of US military research, as Vannevar Bush observes in a 1945 article, ‘As We May Think’. In an avant-garde way, he describes a hypothetical proto-­hypertext system called the ‘memex’ (memory extender): Consider a future device for individual use, which is a sort of mechanized private file and library. It needs a name, and, to coin one at random, ‘memex’ will do. A memex is a device in which an individual stores all his books, records, and communications, and which is mechanized so that it may be consulted with exceeding speed and flexibility. It is an enlarged intimate supplement to his memory.3

It is impressive to read today Bush’s description of ‘wholly new forms of encyclopedias [that] will appear, ready made with a mesh of associative trails running through them’, and the announcement of ‘a new profession of trail blazers, those who find delight in the task of establishing useful trails through the enormous mass of the common record’.4 A particularly important point is that Bush announces a collaborative conception of diverse jobs in society, from the historian to the lawyer: everyone will be invited to work with the competences of his/her colleagues, he claims. Yet although Bush announced this idea very clearly as early as 1945, it is only in the last few years that scholars in the Humanities have begun to develop large scale opportunities for collaborative workspaces in the context of digital culture. The first humanist scholar to build a computing tool was a contemporary of Bush, the Jesuit Father Roberto Busa, who developed the Index Thomisticus.5 Indexing, listing, classifying: in the two generations that followed Busa’s work, Humanities considered computational resources as essentially a way to ‘list’ knowledge, to produce every kind of catalogue and classification, according to a logic of association, ‘Humanities’ AND ‘Computing’.6 The emergence of the label ‘Digital Humanities’, born in 3 Bush, ‘As We May Think’. 4 Bush, ‘As We May Think’. 5 See http://www.corpusthomisticum.org/it/index.age, last accessed 10 March 2013. On Busa as pioneer, see McCarty, ‘What is Humanities Computing? Toward a Definition of the Field’, 8. 6 For this thematic, see Clivaz, ‘Common Era 2.0. Mapping the Digital Era from Antiquity and Modernity’, 43–48.



internet networks and academic research

157

Figure 1. Index Thomasticus.

an oral conversation in April 2001,7 signals a turning point in the field, with the appearance of DH centers, academic jobs and degrees in Digital Humanities (Bachelors, Masters, PhD).8 One can easily observe that definitions of the DH are quite general and flexible, even in this rather clear definition by Claire Warwick: Digital Humanities is an important multidisciplinary field, undertaking research at the intersection of digital technologies and humanities. It aims to produce applications and models that make possible new kinds of research, both in the humanities disciplines and in computer science and its allied technologies. It also studies the impact of these techniques on cultural heritage, memory institutions, libraries, archives and digital culture.9

In my opinion, the reason for this generality and flexibility is that Digital Humanities goes largely beyond a specific field of research: the field exists, of course, as show the numerous institutions and degrees called ‘DH’. But they may be transitional. Therefore I suggest that the label Digital Humanities is better understood to designate a development that is more likely a one-generation period of mutation for the entire field of the Humanities,

7 Kirschenbaum, ‘What Is Digital Humanities and What’s It Doing in English Departments?’. 8 A non-exhaustive list of undergraduate programs in DH is available here: http:// tanyaclement.org/2009/11/04/digital-humanities-inflected-undergraduate-programs-2/, last accessed 19 March 2013. 9 See http://www.ucl.ac.uk/dh/courses/mamsc, last accessed 19 March 2013.

158

claire clivaz

than an area that will become entrenched as a particular field of study. It illustrates the turning point of the dissemination of the ‘Humanities and Computing’ knowledge throughout all the Humanities. During this transition, this label will remind us that DH are first of all Humanities ‘made’ with the fingers, if we think of the Latin word digitus at the roots of digital.10 Robert Darnton, president of the Harvard Libraries, refers to the German word Fingerspitzengefühl in his essay The Case for Books to underline this ‘fingers’ dimension: We find our way through the world by means of a sensory disposition that the Germans call Fingerspitzengefühl. If you were trained to guide a pen with your finger index, look at the way young people use their thumbs on mobile phones, and you will see how technology penetrates a new generation, body and soul.11

So, at a deeper level, this new label signals the emergence of the awareness among scholars in the Humanities that something is changing deeply in the very nature of their disciplines: Humanities can now be ‘touched’, and ‘made’, or re-made, re-invented. In this ‘Humanities digital factory’, the collaborative work is one of the points that is evolving very fast. Scholars have begun to express themselves on blogs as a new forum of discussion. They present a public discourse, written in a fast way, and generally not peer-reviewed or otherwise endorsed except by their name and reputation. Academic blogs are sometimes grouped to get more visibility and a kind of certification: one can refer to hypotheses.org,12 an interesting French initiative, or to culturevisuelle.org, a ‘collaborative scientific media’.13 But generally speaking, new ways of peer review or quality control for such academic discourses have not yet been developed in the Humanities. As Guédon and Siemens emphasized in 2000, ‘it used to be that being printed was ‘the’ distinction; electronic publishing changes this and leads us to think of the distinction

10 Clivaz, ‘ “Humanités Digitales”: mais oui, un néologisme consciemment choisi!’. 11  Darnton, R., The Case for Books: Past, Present, Future, XIII. 12 I joined hypotheses.org to post my own academic blog, Digital Humanities Blog: http://claireclivaz.hypotheses.org/, last accessed 19 March 2013. 13 See their presentation: ‘Bienvenue sur Culture Visuelle, média scientifique collaboratif, proposé par le Laboratoire d’histoire visuelle contemporaine (Lhivic/EHESS). Appuyée sur une ferme de blogs de chercheurs, d’enseignants et d’étudiants qui travaillent et dialoguent à ciel ouvert, la plate-forme propose une publication collaborative, placée sous le pilotage d’un comité d’édition. Elle dispose d’un réseau social ouvert aux lecteurs et expérimente divers outils de recherche, de veille et d’édition multimédia’; http://­ culturevisuelle.org/a-propos, last accessed 19 March 2013.



internet networks and academic research

159

phase completely separately from the publishing phase’.14 We can read in this sentence all the opportunities and the difficulties of open scientific expression online: from the perspective of its users, is such communication ‘serious’ and ‘approved’? From the perspective of the scholars, is such a communication useful, if it does not contribute to their output of peerreviewed work and the recognition to which it leads? Experiments in other fields than Humanities can offer us some early answers that may help us to push further in our reflection. The natural sciences were the first to experiment in forms of collaborative work and even in the collaborative production of scientific knowledge. Certainly, the examples given by Nielsen in Reinventing Discovery are impressive; they show that collaborative endeavors clearly work. First, in 1996, biology researchers decided at a historical meeting in Bermuda to open data on the ADN research: Although many attendees weren’t willing to unilaterally make the first move to share all their genetic data in advance of publication, everyone could see that science as a whole would benefit enormously if open sharing of data became common practice. So they sat and talked the issue over for days, eventually coming to a joint agreement—now known as the Bermuda Agreement—that all human genetic data should be immediately shared online.15

One can wonder about the effects of a parallel decision with regard to, for example, all the digital images that scholars have made of papyri and other manuscripts. Even if the effects would be less obvious in term of impacts for the human society, such a shift would be a tsunami in classical studies. Second example: in January 2009, the mathematician and Fields Medalwinner Tim Gowers opened a collaborative blog, the Polymath project, to try to solve a difficult mathematical problem. This experiment offered the opportunity to ‘see mathematicians making mistakes, going down wrong paths, getting their hands dirty following up the most mundane of details, relentlessly pursuing a solution. And through all the false starts and wrong turns, you see a gradual dawning of insight’.16 At the moment when Nielsen wrote his book, he could announce that the ‘core mathematical problem’ was solved, even if some ‘cleanup work’ remained to 14 Guédon and Siemens, ‘The Credibility of Electronic Publishing: Peer Review and Print’. 15 Nielsen, Reinventing Discovery, 7. 16 Nielsen, Reinventing Discovery, 1–2.

160

claire clivaz

be done. The digital media of communication allows one to ‘show mistakes’ made by the scholars during their collaborative work: the mistakes become a way to improve the collective result. This is a very important turn in our common perception of scholarly work and productions, the consequences of which remain to be considered fully. Thirdly, on the collaborative tool Galaxy Zoo,17 a range of people— volunteer astronomers—classified 150 million galaxies; the tool produced twenty-two scientific papers and even ‘discovered an entirely new class of galaxy, the ‘green pea galaxies’—so named because the galaxies do, indeed, look like small green peas’.18 Galaxy Zoo has allowed a very important turn in the edition of the Oxford library papyri, with the project Ancient Lives19 that I present below. In this example, the production of twenty-two scientific papers begins to offer a way to a solution to the difficult question of the recognition and the certification of such a ­collaborative work. Fourthly, ethical and technological problems appear also to be solved by the collaborative work: the InnoCentive platform helped to find a solution to bring Internet access into regions of India that were without electricity, by launching a contest.20 As Nielsen explains, such examples of micro-collaborations provoked by the Internet relationship illustrate a ‘designed serendipity’: ‘Instead of being an occasional fortuitous coincidence, serendipity becomes commonplace. The collaboration achieves a kind of designed serendipity’.21 He is here looking for an intermediate notion between chance and will, coincidence and design. The notion of ‘serendipity’ in all the applications that are made with it about the digital culture remains to be fully ­evaluated.22 What is interesting here for our analysis is that Nielsen’s use of the concept of serendipity23 implies that he has a positive perception of the value of online collaborative work. But his analysis demonstrates also that at least three conditions have to be brought together in order to allow such a collaborative process to produce knowledge. 17  See http://www.galaxyzoo.org, last accessed 22 April 2013. 18  Nielsen, Reinventing Discovery, 5. 19  See http://ancientlives.org/, last accessed 19 March 2013. 20 Nielsen, Reinventing Discovery, 251, footnote 222. 21  Nielsen, Reinventing Discovery, 27 and 251, footnote 228: Nielsen borrows the concept of ‘designed serendipity’ to Jon Udell, ‘Sam’s encounter with manifactured serendipity’. 22 See in this sens Paveau, ‘Ce lumineux objet du désir épistémique’. 23 See the Oxford Dictionaries online: ‘serendipity means an occurrence and development of events by chance in a happy or beneficial way’. http://oxforddictionaries.com/ definition/english/serendipity, last accessed 23 March 2013.



internet networks and academic research

161

1. Famous scholars enroll themselves in such collaborative workspaces, if they can get some advantages: in Galaxy Zoo papers were produced, for example, which offers a way to get some academic distinction through such work. This point cannot be erased, since it determines all the academic process. It has to be present at some point in any new online collaborative ways of academic production. 2. Open scientific communities have to share a specific focus to be efficient and productive. As Nielsen puts it: ‘For collective intelligence to be successful, participants must be committed to shared body of methods for reasoning so disagreements between participants can be resolved, and do not cause permanent rift’.24 That is eminently the case for the ADN open resources, or for the Polymath project, as well as for Ancient Lives (see below). 3. Openness is a requirement for research: ‘For networked sciences to reach its full potential, it must be open science, based on a culture in which scientists openly and enthusiastically share all their data and their scientific knowledge’.25 As Eysenbach pointed already in 2006, in certain cases, articles are more quoted if they are available in open access, than if they are kept in a journal, even a famous one.26 Since 2006, such an evolution is more and more obvious, including scholarly blogs activities, and even the use of Twitter. One of the most famous illustrations of the importance of open science is probably the birth and fast dead of Nupedia27 (2000–2003), the peer-review ancestor of Wikipedia.28 Open access has always side effects, sometimes predictable, sometimes unexpected. In the New Testament field, the possibility to have full free 24 Nielsen, Reinventing Discovery, 33. 25 Nielsen, Reinventing Discovery, 88. 26 Eysenbach, G., ‘Citation Advantage of Open Access Articles’. Quoted by Glassey, ‘L’internationalisation de la recherche face aux outils “ouverts” de la collaboration scientifique de masse’, 69. Thank you to Olivier Glassey for this reference. 27 See http://en.wikipedia.org/wiki/Nupedia, last accessed 19 March 13. 28 See http://en.wikipedia.org/wiki/Wikipedia:About, last accessed 03/19/13: ‘Wikipedia was founded as an offshoot of Nupedia, a now-abandoned project to produce a free encyclopedia. Nupedia had an elaborate system of peer review and required highly qualified contributors, but the writing of articles was slow. During 2000, Jimmy Wales, founder of Nupedia, and Larry Sanger, whom Wales had employed to work on the project, discussed ways of supplementing Nupedia with a more open, complementary project. Multiple sources suggested that a wiki might allow members of the public to contribute material, and Nupedia’s first wiki went online on January 10, 2001’.

162

claire clivaz

access to the Codex Sinaiticus29 has given it de facto a special status among the New Testament codices. I have met New Testament colleagues—not particularly involved in textual criticism—using now the Nestle-Aland 28th edition30 and the online Codex Sinaiticus as their basic Greek texts for their teaching and research. Such an evolution was probably not intended as the result of this online project, but it has nevertheless had such a practical effect on New Testament research. Online use comes to matter more and more, and perhaps even more than the academic habits of scholars who were trained in print. When openness is used without a focus, it does not seem to be efficient, as may be seen from the example of Philica, presented by Kathleen Fitzpatrick in Planned Obsolescence. Philica is an ‘open publishing network, co-founded by British psychologists Ian Walker and Nigel Holt, which invites scholars from any field to post papers, which are then made freely available for reading and review by any interested user’; this totally open ‘journal of everything’—as it names itself—has got ‘only 164 articles or notes published [. . .] between March 2006 and August 2009, a mere 4 of which were in the humanities’.31 Only 319 articles or notes are announced in March 2013 on Philica,32 although all fields are invited to publish freely in the journal. Everything means nothing, at a certain point. Even Google begins to suffer from its choice to offer openness without either distinction or focus. As Frédéric Kaplan underlines, ‘Google’s weakness is its desire of universality. [. . .] The scholarly knowledge in Humanities is composed with singularities and specificities’.33 Research communities in Humanities who will maintain a focussed approach at the methodological level but openness at the point of publication have great potential to produce new knowledge and discoveries. Consider for example the two-year old Ancient Lives tool produced by the University of Oxford. Over the course of a century, since their discovery in Egypt, only about one quarter of the Oxyrhynchus papyri have

29 See http://codexsinaiticus.org/, last accessed 19 March 2013. 30 Institut für Neutestamentliche Textforschung, Novum Testamentum Graece, Stuttgart, Deutsche Bibelgesellschaft, 201228. 31  Fitzpatrick, Planned Obsolescence: Publishing, Technology, and the Future of the Academy, 17. Copyright (c) 2009 New York University, This text may be distributed in part or in whole on condition that (1) distributed text is not sold, whether or not such sale is ‘for profit’ and (2) distributed text bears this notice in full. Except as permitted by law, all other uses are prohibited without written permission of the publisher. 32 http://www.philica.com/, last accessed 19 March 2013. 33 Kaplan, F., ‘Géostratégie des humanités digitales’. My English translation.



internet networks and academic research

163

been ­published. But now the team responsible for editing these texts has opened an online tool, Ancient Lives,34 that allows volunteers to help them to decipher the papyri. It is included in the Zooniverse real science online,35 the birth place of Galaxy Zoo. At the moment, it is the only project announced in the rubric ‘Humanities’ on Zooniverse.36 The project belongs to the Citizen Science Alliance and presents itself in this way: The Citizen Science Alliance is a transatlantic collaboration of universities and museums, who are dedicated to involving everyone in the process of science. Growing out of the wildly successful Galaxy Zoo project, it builds and maintains the Zooniverse network of projects, of which Ancient Lives is part. Ancient Lives is a collaborative tool between a diverse collection of Oxford Papyrologists and Researchers, The Imaging Papyri Project, The Oxyrhynchus Papyri Project, the Egypt Exploration Society and the following institutions. The papyri belong to the Egypt Exploration Society and their texts will eventually be published and numbered in Society’s Greco-Roman Memoirs series in the volumes entitled The Oxyrhynchus Papyri.37

As may be seen, the idea of endorsement or recognition is present in the tool, since, at the end of the process, a collaborative work could be published in the Oxynrhynchus Papyri volumes. So, the three conditions to get an efficient online collaborative work are gathered together here: academic distinction, a focused approach and openness. A tutorial is provided for each potential volunteer to test his/her own capacities. As users themselves complain,38 there are no statistics on the tool to evaluate its success, but, as far as one can see, it works: texts are deciphered, transcribed and translated. Users engage in discussion online, and the work expands. Some topics / objects are related to New Testament topics, such as this discussion on the ‘number of the beast’: ‘χις or χξς (616 or 666)’.39 As the evidence suggests, such a tool should be evaluated in terms of the research network to which it leads: it represents an important

34 See http://ancientlives.org/, last accessed 19 March 2013. 35 See https://www.zooniverse.org/, last accessed 19 March 2013. 36 See https://www.zooniverse.org/projects, last accessed 19 March 2013. 37 See http://ancientlives.org/about, last accessed 19 March 2013. 38 See litonatlas: ‘It would be nice to have some global statistics on the project accessible on the website. For example how many characters in total have been identified, what proportion of the 500,000 papyri have been reviewed’; http://talk.ancientlives.org/ users/4e55e8ab686b984b500005dc, last accessed 19 March 2013. 39 See http://talk.ancientlives.org/chat/discussions/DAL10030mr, last accessed 19 March 2013.

164

claire clivaz

Figure 2. Ancient Lives.40

development in an academic field used to working mainly in the closed world of libraries. 3. New Testament Textual Criticism and Online Networks If we turn now to specific examples of online forum discussions in the field of NTTC, do we find among them focus, open science and distinction/recognition? With regard to focus, there is no doubt that this condition is filled. These discussion groups are really focused and specific, because they are related to a community, where scholars share the same field (textual criticism), methodologies and objects of study (New Testament Greek manuscripts). But, as I will show, this scholarly milieu needs to go in a more decisive way in the direction of openness, and needs also to develop ways of gaining scholarly distinction and recognition from their collaborative work. There are now several online forum and lists of discussion in the NTTC field, and one can also observe the emergence of collaborative working environments. I do not claim here to give a full picture of these initiatives, but rather to focus on some illustrative examples. I classify these lists, forum and tools in three categories, according to their openness and their purpose. 40 http://ancientlives.org/tutorial/transcribe, last accessed 23 March 2013.



internet networks and academic research

165

3.1. Private Lists or Lists with Limited Subscriptions A first category is illustrated by private lists or lists with limited subscriptions, which cannot be read in open access. They do not have as their purpose the dissemination of knowledge outside their own circles. The most recognized scholars in the field belong to these lists and use them to exchange information and to engage in discussion. For example, the Cbgm-list41 is moderated by Klaus Wachtel (Institut für neutestamentliche Textforschung (INTF), Münster). Its name corresponds to a methodology, ‘Coherence-Based Genealogical Method’,42 and underlines the coherence of the focus of the list. It is not accessible in open access, but only to the scholars enrolled.43 Opened in 2008, it offers only some dozen of messages to date, with fewer than twenty authors writing, and almost nothing new during the last year. It is not possible to find on its website a description of the purpose of the list (an ‘about’ section). Some messages underline the feeling of some authors of speaking to a ‘distinguished forum’. The exclusive nature of the list ensures quality control, but it is useful only to a very few number of people and does not have a real impact on research. Another example is the ‘NTEditions’ list, hosted by JISC44 and moderated by Hugh Houghton (Institute for Textual Scholarship and Editing [ITSEE], Birmingham). This list was set up at an editorial meeting of those working on the Editio Critica Maior and related projects in March 2006 for the exchange of information and ideas. As for the Cbgm-list, quality and certification are represented by virtue of the membership, but it remains only an internal means of communication (primarily for committee members of the International Greek New Testament Project)45 and has relatively low traffic. 3.2. Forums or Tools with Restrictions A second category is represented by semi-open forums for discussion or collaborative tools, with two different kinds of limitations. The first kind of limitations are requests and conditions imposed by external institutions, such as libraries. That is the case for a very interesting ­collaborative

41  See http://listserv.uni-muenster.de/mailman/listinfo/cbgm-list, last accessed 19 March 2013. 42 See for an introduction by Gerd Mink, http://egora.uni-muenster.de/intf/service/ downloads.shtml, last accessed 19 March 2013. 43 The names and number of these scholars are accessible only to the members of the list. 44 See http://www.jiscmail.ac.uk, last accessed 22 April 2013. 45 See http://www.igntp.org/, last accessed 19 March 2013.

166

claire clivaz

Figure 3. Virtual Manuscript Room.46

­ roject, still in its early stages and representing an important turning point p for all in the NTTC field. This is the project of a ‘workspace for collaborative editing’, described in this way on the ITSEE website:47 ‘the creation of an online transcription facility, [is] leading to the development of a Workspace for Collaborative Editing’.48 This collaborative editing workspace is developed in the Virtual Manuscript Room, in common with the INTF in Münster.49 The INTF provides a ‘New Testament Virtual Manuscript Room’, with the possibility for users to participate in indexing and transcribing manuscripts.50 A call for contributions to the collaborative work has been made. If the purpose is similar to that of Ancient Lives, the conditions related to the exploitation of the New Testament manuscripts are quite different. Indeed, the University of Oxford has the possibility to open the work on the Oxyrhynchus papyri, as it wants, on Ancient Lives. ITSEE and the INTF are working with manuscripts images provided by a range of diverse libraries: this fact imposes a constraint and limits the collaborative editing workspace to approved scholars, or scholarly collaborators, who have to .

46 http://vmr.bham.ac.uk/, last accessed 19 March 2013. 47 See http://www.birmingham.ac.uk/research/activity/itsee/index.aspx, last accessed 19 March 2013. 48 See http://www.birmingham.ac.uk/research/activity/itsee/index.aspx, last accessed 19 March 2013. 49 See http://egora.uni-muenster.de/intf/index_en.shtml, last accessed 19 March 2013. 50 See http://ntvmr.uni-muenster.de, last accessed 19 March 2013.



internet networks and academic research

167

announce themselves and must be accepted as participants in the ­project. Even if this limitation is hardly a serious impediment to the success of the entire project, it leads nevertheless to questions about the status of ancient manuscripts, generally speaking: is it not time to claim, as did the biologists in 1996 for ADN, free access to these documents from the past? The revolution would not be more difficult to achieve than it was for ADN, but probably more difficult to explain in terms of societal challenges and importance. Nevertheless, in my opinion, free access to all the material kept in academic and private libraries—open access—has to be claimed: I agree with the recommendation of the European Commission ‘on access to and preservation of scientific information’, edited on the 17th of July 2012.51 The examples evoked in part 1 speak for themselves about the methodological importance of achieving real open science. The second kind of limitations that produce semi-open collaborative tools or online discussion forums are limitations chosen by the people in charge of the tool. I give here the example of the Evangelical Textual Criticism blog. It explains under the title of the blog: ‘This is a forum for people with knowledge of the Bible in its original languages to discuss its manuscripts and textual history from the perspective of historic evangelical theology.’52 Unlike the three other NTTC examples given above (the two internal lists and the collaborative editing workspace), this blog is not related to a specific institution, but to a group of people recognizing themselves in the ‘perspective of historic evangelical theology’. At the academic level, it is a place of high-quality discussion, often offering the most recent information: it was the first place where I have read, for example, the announcement of the publication of the NA28 on the 25th of June 2012,53 even if the blog assumes of course ‘mistakes’, as one can see on this screenshot below, with some information on this topic that has been struck through because it was found to be incorrect. We have seen with the Polymath project example above that the ‘mistake culture’ is typical of collaborative work in online networks.

51  See http://ec.europa.eu/research/science-society/document_library/pdf_06/recom mendation-access-and-preservation-scientific-information_en.pdf, last accessed 19 March 2013. 52 See http://evangelicaltextualcriticism.blogspot.nl/, last accessed 19 March 2013. 53 See http://evangelicaltextualcriticism.blogspot.nl/2012/06/nestle-aland-28-in-press .html, last accessed 19 March 2013.

168

claire clivaz

Figure 4. Evangelical Textual Criticism blog.54

This blog is moderated by Peter Head and Tommy Wasserman, and I have qualified it as ‘semi-open’, because to participate in its discussion, potential contributors are ‘requested to respect the blog’s ethos’. Further, to initiate a discussion and to be a full participating member of the blog, one has to follow these requests: Those applying for membership must indicate that they have read either the OT or the NT in its original language(s), should be actively involved in textcritical research, and should be already contributing to the blog through comments. They should give e-mail details of an academic and a pastoral referee, a summary of their academic and/or ministry involvement, a statement of their doctrinal commitment (which may be by reference to various 54 http://evangelicaltextualcriticism.blogspot.nl/2012/07/nestle-aland-28-at-sbl-in­chicago.html, last accessed 19 March 2013.



internet networks and academic research

169

classic evangelical statements of faith, e.g. 39 Articles, Westminster Confession), and an indication of their area of interest within textual criticism.55

One can see the list of the nineteen contributors members (currently all men)56 accepted on the website. Despite this limitation, the blog is particularly active, for there have been some 1942 posts since 2005, which means a high level of productivity among the few members who can initiate discussions. The blog is consulted on a frequent basis: one can have access to the statistics for the twelve last months.57 As the evidence shows, this blog is a very interesting example of new developments in academic networks, and I will conclude this article with my hope to develop a research project on that topic. For that to happen, the field of NTTC would need a sociological inquiry to evaluate as fairly as possible what might be at stake. The quality of the ETC blog is clear: many scholars in the field consult it. But its conditions of membership are really questionable and concern different levels of preoccupation. Regarding its potential scientific impact, in the light of my arguments above, it is clearly restrictive to limit the contributions of the scholars, even if the contributors are so productive, and produce work of such a satisfying quality. Beyond this general remark, questions may be asked about the reason, in a Western scholarly context, for requiring a ‘doctrinal commitment’, whatever that commitment might be. A full sociological inquiry would be here needed to evaluate the institutional impact of Christian theology in Europe, and the topic would promise to be not easy to analyze. As I have remarked in a previous article,58 the example of the protestant scholar Gerd Lüdeman and his academic point of view that denies the resurrection offers another example that might be considered in such a debate: in 2008, after fifteen years of due process, the most authoritative German tribunal has sided with the Faculty of Theology in Göttingen against Gerd Lüdeman.59 So, it appears, a full analysis of the decision to impose a 55 See http://evangelicaltextualcriticism.blogspot.nl/, last accessed 19 March 2013. 56 I refer to the list on the welcome page of the blog: http://evangelicaltextualcriticism .blogspot.nl/, last accessed 19 March 2013. 57 See http://www.sitemeter.com/?a=stats&s=s18div327, last accessed 19 March 2013. For example, there were 2,931 visits during the week, in which I end this article. 58 Clivaz, C., ‘Why Were the Resurrection Stories Read and Believed? And What Are We Making of Them Today?’, 561. 59 See Leitsätze zum Beschluss des Ersten Senats vom 28. Oktober 2008—1 BvR 462/06, point 3: ‘Die Wissenschaftsfreiheit von Hochschullehrern der Theologie findet ihre Grenzen am Selbstbestimmungsrecht der Religionsgemeinschaft und an dem durch Art. 5 Abs. 3 GG geschützten Recht der Fakultät, ihre Identität als theologische Fakultät zu wahren und ihre Aufgaben in der Theologenausbildung zu erfüllen’. http://www.bverfg.de/ entscheidungen/rs20081028_1bvr046206.html, last accessed 19 March 2013.

170

claire clivaz

confessional restriction on a scholarly blog in a European context would require a careful sociological and theological inquiry. From the perspective of digital cultural, the evaluation is more simple and clear: any kind of limitation to open science tends to decrease the development of knowledge, as already illustrated in part 1. Moreover, ETC blog does not offer any means of academic distinction. As we have seen in the first part of this article, all these three conditions have to be gathered together to facilitate a sustainable collaborative network that produces knowledge: academic distinction, open science and a clear focus. So, it would be very interesting for the editors of the blog to consider producing some articles or other academically recognized form of publication that draws on the content of their blog, such as Galaxyzoo did, for example. Moreover, as we will see, even the third category does not offer all of them today. 3.3. Online Open Forums To the third category belong online open forums of discussion, such as the Textual Criticism of the Bible on Yahoo, opened in 2004;60 7749 messages have been posted on the 24th of March 2013, and the forum counts 593 members. It is not totally open, however, even if it is probably the most used online forum of discussion in NTTC today: one has to apply to the moderator in order to become a member. Nevertheless, as may be seen from the large membership, many applicants have been accepted. The moderator is named on the welcome page—Wieland Willker—, but to discover his professional identity, one has to go to another NTTC forum of discussion, the TC forum (132 users),61 and to log in as a member. At that step, one learns that Wieland Willker is not a professional scholar of textual criticism, but a chemist at the University of Bremen (Germany).62 One of the most impressive features on this online forum of discussion is the wide range of participants: one can find non-professional but very well informed people (like Wieland Willker), very famous scholars in the field (like Bart Ehrman), a range of diverse scholars in the field of NTTC, pastors working in parishes, and also non-professional people. But, as far as I can see, even in the short time between when I presented my analysis as a conference paper at the international SBL meeting 2012 in Amsterdam, 60 See http://groups.yahoo.com/group/textualcriticism/, last accessed 19 March 2013. 61  See http://tcg.iphpbb3.com/forum/portal.php?nxu=64774768nx21631, last accessed 19 March 2013. 62 See http://www-user.uni-bremen.de/~wie/, last accessed 19 March 2013.



internet networks and academic research

171

and the moment in March 2013 when I have revised it for this chapter, there has been a great deal of evolution in the way that this Yahoo forum is used. A full inquiry could be made here to understand the (short) history and the present evolution of this online forum, but I will offer just some preliminary observations about two points: the ­commitment to the online forum of a famous scholar in the field, and the content of the messages. If we consider the example of Bart Ehrman, one can see that his involvement in the forum of discussion has more or less stopped during the last few months. He has posted 125 messages since he began to participate in 2006, but his last message is dated from the 27th of October 2012. Meanwhile, in April 2012, he opened a ‘public forum’ on his own blog,63 and seems now to prefer to use that platform when he wishes to express himself online. His own ‘blog’ apparently offers a more ‘controlled’ space for public debate by a well-known scholar. Therefore it is quite different from the online Yahoo forum, where we can see that Bart Ehrman had to explain a mistake he had made in a video.64 Moreover, he had a quite hard exchange with Dan Wallace about the notion of ‘inerrancy’, such as used at the Dallas Theological Seminary.65 Given these developments, it would be useful to be able to ask Ehrman himself to comment on the way in which he wishes to pursue his involvement in conducting academic debates on online forums. The decision to be made is not insignificant. For, as we have seen already in regard to the Polymath Project and the ETC blog, willingness to make academic statements online implies that scholars are willing to risk making mistakes in public. Furthermore, it suggests not only a willingness to accept that mistakes may be made in public, but also the need to accept this possibility within academic discourse, which will require good grace and flexibility on the part of researchers who will have to accept that they can no longer be understood as producing faultless academic work. Regarding the content of the Textual Criticism of the Bible on Yahoo, it is a mixture of very fine, new and important information, and useless messages, including clear mistakes. Certain topics are particularly prominent:

63 See http://ehrmanblog.org/category/public-forum/page/28/, last accessed 19 March 2013. 64 Message 7261, http://groups.yahoo.com/group/textualcriticism/message/7261, last accessed 19 March 2013. 65 Messages 7264 and 7266, http://groups.yahoo.com/group/textualcriticism/­ message/7264; http://groups.yahoo.com/group/textualcriticism/message/7266, last accessed 19 March 2013.

172

claire clivaz

at least 848 messages mention Mark 16 and the problem of the short/ long ending of this canonical Gospel, but they come often from the same members, who monopolize the debate, not always usefully. The absence of academic distinction in the tool could, in the medium term, prove a demotivating factor for scholars who use the list. We can also observe the emergence of new topics, such as the Quran (about 14 messages with the mention of the Quran). This allows international and intercultural contacts, as Sara Schulthess’ article illustrates in the present volume.66 A lot of useful information is present in the messages: bibliographical information, hyperlinks to some images of manuscripts, etc. Even today this knowledge, although available, is not yet fully catalogued or fully exploited. At the end of this part 2, it is possible to see that even a relatively small field of research such as NTTC produces a number of online collaborative tools, or forums for discussion. They are all focused on the same field, with shared methodologies, and they show a necessary diversity of points of view. But, as yet, there is no widely recognized way in which they all contribute to the free dissemination of knowledge, or through which their participants receive appropriate academic recognition for their work. In what follows, I will sketch some ways in which those two goals might each be achieved. 4. What Could be the Next Steps? This overview of the current position of academic communication and collaboration in NTTC illustrates the general transformation of the scholarly production of knowledge that arises as the consequence of digital culture. In my opinion, we are facing here a turning point similar to the moment preceding the creation of scholarly journals or reviews. The sociologist of sciences, Dominique Vinck, reminds us that shortly before these journals were established, the seventeenth century scholar Father Mersenne (France) was in epistolary contact with more than 200 other scholars.67 In 1665, the French Journal des Savants began to facilitate systematic communication of news and researches among scholars.68 I add that two

66 See the article in this volume: Schulthess, ‘The Role of the Internet in New Testament Textual Criticism: the Example of the Arabic Manuscripts of the New Testament’. 67 Vinck, ‘Les transformations des sciences en régime numérique’, 50, footnote 1. Thank you to my colleague Dominique Vinck for this reference. 68 Vinck, ‘Les transformations des sciences en régime numérique’, 46.



internet networks and academic research

173

months later in 1665, the English Philosophical Transactions69 started a similar task, but with a more focused perspective, offering only scientific contents, whereas the French Journal des Savants contained also obituaries and diverse documents. If we draw a parallel across the centuries, we can consider, for example, that the 7749 messages on the Yahoo forum are our present epistolary network. What should be or could be the new form of ‘reviews’ in the next years? How could we improve the ability of NTTC field to produce collective science in this intermediate period of transition between different forms of scholarly communication?70 First of all, research needs be done to understand the present situation with regard to at least the following two points: 1) a sociological inquiry that gives more information about the different actors and networks of the NTTC online communication and collaboration, and establishes a full map of these new scholarly collaborative tools and online forum of discussion; 2) an analysis of the content of online discussion forums such as ETC blog and the Textual Criticism of the Bible on Yahoo. At the methodological level, one could note here the useful research conducted by Geoffrey Rockwell and Stefan Sinclair, using Voyant Tools71 on the Humanist Discussion Group, a journal that had been established online for fourteen years; a summary of the inquiry is available online.72 Secondly, one should encourage the participation of some NTTC scholars in the general discussions on open access that are already taking place in Europe and in the USA, in the hope that we may one day achieve full open access for all digitized ancient manuscripts. Everything has to be done to help the ITSEE and the INTF to achieve even greater access in their development of a collaborative editing workspace. As Olivier Glassey points out, the keeping back of information in a context of competition can stop the maximal development of collective knowledge.73 New Testament textual critics should try to reach our own ‘Bermuda Agreement’, as 69 See http://rstl.royalsocietypublishing.org/, last accessed 19 March 2013. 70 About an analysis of the digital communication about the Jesus’ Wife Gospel, see my two blog’s articles: ‘The real “scoop” of the so-called Jesus’ Wife Gospel: the transformation of the peer-review process’, http://claireclivaz.hypotheses.org/147; ‘Jesus’ Wife Gospel and academic research: how could we join the blog culture, the scholars and the publishers world?’, http://claireclivaz.hypotheses.org/189, last accessed 19 March 2013. 71  See http://voyant-tools.org/, last accessed on 19 March 2013. 72 Rockwell and Sinclair, ‘The Swallow Flies Swiftly Through: An Analysis of Humanist’. 73 Glassey, ‘L’internationalisation de la recherche’, 67. See also Schroeder, ‘e-Research Infrastructures and Open Science: Towards a New System of Knowledge Production?’, 2 and 13. Thank you to Olivier Glassey for these references.

174

claire clivaz

the biologists did. If such open access to knowledge can be agreed, our focus will need to turn next to how best to manage the involvement of professional scholars and researchers. As noted above, Textual Criticism of the Bible on Yahoo does not offer any kind of academic recognition or endorsement, and already this forum may be becoming less attractive for scholars, since too many people are using it as a platform for writing about too many topics with greatly varying levels of knowledge and expertise. Yet such an online forum of discussion has the potential to succeed in this lively milieu, if only we can improve it. Further research could lead to new and improved ways of organizing the exchange of information and the collaborative work. We need to agree on ways in which to provide academic recognition for the collective and individual work produced on these collaborative tools. Finally, computational methods should be used to exploit the content that has already been put online by New Testament textual critics. As I have noted, a lot of bibliographical information is available on discussion forums. It might be possible to collaborate with the French team of OpenEdition, who employ an automated system to look for bibliographical references.74 Other information could also be searched automatically, in order to be exploited for NTTC. Last but not least, the question of archiving and storage the research material on the collaborative tools and online forum of discussion should be considered in a global context. At the end of this chapter, I feel that I have just started an Odyssey in a new scholarly world of collaboration and communication. Other fields of research can teach us a lot and stimulate NTTC to improve what it does already: producing knowledge in a collaborative way. I am convinced that NTTC is one of the most efficient test-cases for helping Humanities to enter in the new mode of scholarly production of science. References Bush, V., ‘As We May Think’, The Atlantic Magazine (July 1945). http://www.theatlantic .com/magazine/print/1945/07/as-we-may-think/303881/, last accessed 19 March 2013. Clivaz, C., ‘ “Humanités Digitales”: mais oui, un néologisme consciemment choisi!’, http:// claireclivaz.hypotheses.org/114; last accessed 19 March 2013. ——, ‘Why Were the Resurrection Stories Read and Believed? And What Are We Making of Them Today?’, in Resurrection of the Dead. Biblical Traditions in Dialogue (BETL 249), ed. by G. van Oyen and T. Shepherd (Leuven: Peeters, 2012), 555–577.

74 See http://bilbo.hypotheses.org/, last accessed on 19 March 2013.



internet networks and academic research

175

——, ‘Common Era 2.0. Mapping the Digital Era from Antiquity and Modernity’, in Lire Demain. Des manuscrits antiques à l’ère digitale / Reading Tomorrow. From Ancient Manuscripts to the Digital Era, ed. by C. Clivaz, J. Meizoz, F. Vallotton and J. Verheyden, with B. Bertho (Lausanne: PPUR, 2012), 23–60 (ebook). Darnton, R., The Case for Books: Past, Present, Future (PublicAffairs, 2009). Eysenbach, G., ‘Citation Advantage of Open Access Articles’, Biology 5 (2006/4). http:// www.plosbiology.org/article/info:doi/10.1371/journal.pbio.0040157#s3, last accessed 19 March 2013. Fitzpatrick, K., Planned Obsolescence: Publishing, Technology, and the Future of the Academy (New York: NYU Press, 2011). http://mediacommons.futureofthebook.org/mcpress/ plannedobsolescence, last accessed 15 January 2013. Glassey, O., ‘L’internationalisation de la recherche face aux outils “ouverts” de la collaboration scientifique de masse’, in Recherche et enseignement supérieur face à l’internationalisation: France, Suisse et Union européenne, ed. by J.-P. Leresche, P. Larédo and K. Weber (Lausanne: PPUR, 2009), 65–84. Guédon, J.C. and R. Siemens, ‘The Credibility of Electronic Publishing: Peer Review and Print’. http://web.viu.ca/hssfc/Final/PeerReview.htm; last accessed 03/19/2013. Kaplan, F., ‘Géostratégie des humanités digitales’. http://fkaplan.wordpress.com/2011/04/01/ geostrategie-des-humanites-digitales/, last accessed 19 March 2013. Kirschenbaum, M.G., ‘What Is Digital Humanities and What’s It Doing in English Departments?’, ADE Bulletin 150 (2010), 56–57. http://mkirschenbaum.files.wordpress .com/2011/03/ade-final.pdf, last accessed 19 March 2013. McCarty, W., ‘What is Humanities Computing? Toward a Definition of the Field’ (1998), 1–9. http://www.cch.kcl.ac.uk/legacy/teaching/dtrt/class1/mccarty_humanities_computing .pdf, last accessed 19 March 2013. Nielsen, M., Reinventing Discovery: the New Era of Networked Science (Princeton: Princeton University Press, 2012). Paveau, M.-A., ‘Ce lumineux objet du désir épistémique’, in Louis de Mailly. Les aventures des trois princes de Sérendip. Suivi de Voyage en sérendipité, ed. by A. Volphillac, D. GoyBlanquet and M.-A. Paveau (Vincenne: Thierry Marchaisse éd., 2011), 225–243. Rockwell, G. and S. Sinclair, ‘The Swallow Flies Swiftly Through: An Analysis of Humanist’. http://www.dh2012.uni-hamburg.de/conference/programme/abstracts/the-swallowflies-swiftly-through-an-analysis-of-humanist/, last accessed 19 March 2013. Schroeder, R., ‘e-Research Infrastructures and Open Science: Towards a New System of Knowledge Production?’, Prometheus: Critical Studies in Innovation 25 (2007/1), pp. 1–17. Udell, J., ‘Sam’s encounter with manufactured serendipity’, Jon Udell’s Radio blog, March 4, 2002. http://radio-weblogs.com/0100887/2002/03/04.html, last accessed 19 March 2013. Vinck, D., ‘Les transformations des sciences en régime numérique’, Hermès 57 (2010), 45–51.

New Ways of Searching with Biblindex, the Online Index of Biblical Quotations in Early Christian Literature Laurence Mellerin 1. Introduction: The Stages of the Project1 Biblindex is a project led by the French Institut des Sources Chrétiennes with the support of two computer-science labs in Rhône-Alpes and two biblical institutes, the Peshitta Institute of Leiden University and the Institut für neutestamentliche Textforschung in Münster. It is in receipt of funding from the French National Research Agency for the period from 2011 to 2014 and we, together with other partners, are hoping to obtain European funding in the coming years. Its goal is the constitution of an online index of biblical references which are found in Christian and, eventually, Jewish literature, drawing on both Western and Eastern sources from Late Antiquity and the early Middle Ages. Two particularly innovative features of Biblindex are its spatiotemporal breadth and its multilingual corpus: patristic texts in Greek and Latin will be analyzed, along with oriental languages, in particular Syriac, Coptic, Ethiopian, Armenian, Georgian. These new developments have been made possible by the significant progress in critical editions of oriental texts. Access to the data is completely free. 1.1. Corpus, Metadata, Interfaces At present, Biblindex consists only of a comprehensive inventory of biblical quotations and allusions in Early Greek and Latin Christian literature, providing bibliographical information about the edition used for each work. Each entry on the website offers a series of numbers indicating the chapter and verse of the biblical text, its location in the patristic writing and the corresponding page and line numbers in the reference edition. This data is merely the first step of Biblindex. It derives from the Biblia Patristica index of biblical quotations in the early Christian 1 More information can be found on our online research notebook, http://www .biblindex.hypotheses.org, last accessed 8 May 2013, as well as in the first part of Mellerin, ‘Biblindex, online Index of Biblical Quotations in Early Christian Literature: methodological issues’.

178

laurence mellerin

Fathers, a project which the Centre for Patristics Analysis and Documentation (CADP) in Strasbourg started in 1965 and developed until 2000. The published volumes of Biblia Patristica cover the first three centuries, along with part of the fourth. They form an inventory of 270,000 entries, which have all been scientifically verified.2 Each entry in these volumes indicates the book, chapter, section, and inclusive line numbers for the quotation or allusion. It follows the order of the biblical books. No search is possible using alternative criteria. There were around 100,000 more references in the database inherited by Sources Chrétiennes, unchecked but in electronic form: these are already online. Works of several authors of the fourth and fifth centuries had been entered. In addition to this material, there are about 400,000 unverified hard-copy references in boxes still unscanned, comprising very different texts that were written in different places and at different times. Chronological and geographical incompleteness makes the corpus less relevant for statistical searches. Augustine’s works for instance have not been analysed yet. The more the number of quotations increases, the more useful the index will become for researchers. This is why Biblindex intends to put online the whole content of the CADP archives and to increase the database with data from biblical indexes on the largest possible scale. In addition, a large amount of metadata relative to each work has been added: dates, locations and biographical information about the authors, along with the definition of a precise taxonomy of works, sorted by genres, topics and keywords. Since December 2008, a search interface has been available online in beta version, which allows simple searches in this huge corpus. The search form allows investigation of the corpus by entering one or more biblical references, and sorting of the results by ancient authors, works, dates, geographical areas or bibliographical references. A more user-friendly interface will be implemented in 2014, providing a greater ability to structure the data and visualise results. 1.2. Giving References to Biblical Texts This first version of the website gives references from only one composite Bible which includes the books of the Hebrew Bible, plus seven further 2 Beginnings to Clement of Alexandria and Tertullian, vol. 1; The third century (apart from Origen), vol. 2; Origen, vol. 3; Eusebius of Caesarea, Cyril of Jerusalem, Epiphanius of Salamis, vol. 4; Basil of Caesarea, Gregory Nazianzen, Gregory of Nyssa, Amphiloque of Iconium, vol. 5; Hilary of Poitiers, Ambrose of Milan, Ambrosiaster, vol. 6; Didymus of Alexandria, vol. 7, 2000; Philo of Alexandria, Supplement.



new ways of searching with biblindex

179

texts which are not found in the Hebrew Bible but are included in Rahlfs’ edition of the Septuagint, and the 27 books of the Greek New Testament; the verse numbering of one modern translation of the Bible, the French Jerusalem Bible, is always used. The terms of reference were thus defined not from the objects to be dealt with, but from the user’s constraints: while the canon accepted by the Fathers was varied and fluid, the division of the Hebrew Bible results from a consensus of modern scholars. It was just more practical to use a single, albeit inadequate, scheme.3 However, this practice raises problems: the form of the biblical text to which the website refers is usually a form of text unknown to Fathers who quoted the Bible; quotations in different languages have to be referred to different ‘Bibles’. Furthermore, for several reasons, no one version of the Bible can be used as the sole source for referring to biblical citations: some books, whether canonical or not, are not included in some versions; some books, chapters or verses are only included in some versions; the order, names and numbers may differ for books, chapters and even verses. Finally, no modern canonical definition can be adopted as the rule for studying ancient texts. The new Biblindex will use several ways of referring to biblical texts by chapter and by verse, at least one for each language. Initially, in addition to translations into modern languages, five ancient versions of the Bible, each with its own system of referencing, will be used: the Hebrew Bible, the whole Greek Septuagint and the Greek New Testament, the Latin Vulgate, the Peshitta for the Syriac texts and the Zohrab Armenian Bible. These referencing systems are in no way intended to be prescriptive. Their only function is provisionally to relate any patristic text to a shared point of comparison. We provide a data model allowing auto­matic switching between chapter and verse numbering in the different Bibles.

3 Another difficulty in referring to our sources must be pointed out: although it is tempting to think of modern critical editions of patristic texts as fixed entities, no two ancient handwritten copies of any text were ever the same as one another, even when they were written in the same language, and versions in different languages also exhibit differences from one another. Unfortunately Biblindex does not solve this problem and uses only modern critical editions, considering that the user who wants to go a step further may read the critical apparatus of the text he or she is interested in. It would have been too complicated to take all manuscripts into account. But each time an analyst notes something relevant about biblical variants in the transmission of the patristic text, a comment is added in the database.

180

laurence mellerin 1.3. The Textual Phase

We want to go a step further by starting with what we call ‘the textual phase’ of the project, in which use of XML-TEI and data mining will be required. Our aim is not to build a new textual databank, but to build partnerships with those already existing: Thesaurus Linguae Graecae (TLG) at the University of California, Irvine, the Library of Latin Texts by Brepols Publishers (LLT), and the Syriac database currently developed at Brigham Young University. An analyst will receive an electronic text from a textual databank and tag it through a web interface on the Biblindex site. Her or his work will be converted into XML-TEI files, which can be used either on the website of the textual databank or on the Biblindex website. Metadata (information about authors and works) are stored in the SQL database4 using UTF-8 characters, while quotations are encoded in the XML-textfiles.

Textual databases (TLG; LLT, etc.)

1

L

[Ἐν ἀρχῇ ἐποίησεν ὁ Θεὸς τὸν οὐρανὸν καὶ τὴν γῆν.] Πρέπουσα ἀρχὴ τῷ περὶ ???..... Web interface used by the analyst

XM

XML-TEI encoded texts

3 2 Ἐν ἀρχῇ… XML-TEI files XM

L

3

Biblical referencing systems METADATA BIBLINDEX database

Figure 1. IT-architecture.5

4 SQL (Structured Query Language) is a special-purpose programming language designed for managing data in relational database management systems. 5 All images are under the copyright of Biblindex.



new ways of searching with biblindex

181

2. Examples of Research Results I now come to what I see as the main point of this paper: giving some examples of research results which may be obtained with such a powerful tool. The first is for biblical scholars and the second for specialists in patristics. 2.1. The Utility of Biblindex for Biblical Scholars 2.1.1. Studying the Constitution and History of the Canon Because of the accuracy of its data, Biblindex will make it easier to see and compare the pre-canonical forms of the Scriptures. Ancient Christianity did not know a global definitive list of books, and even less a list of editions or authorized versions. Biblindex can throw light on the spread of biblical writings in such or such a part of the world to such or such a period, and thereby highlight the synchronic and diachronic dimensions of the texts’ transmission. For instance, the Book of Ruth has always been known as canonical, but was not very often quoted, and not everywhere. Biblindex shows that it was used since the third century in Alexandria and Caesarea. The Antiochene Fathers seem not to refer to it as a canonical work before the fifth century, when Theodoret of Cyrrhus writes his Commentary on Ruth.

Figure 2. Use of the Book of Ruth in the fourth century / at the beginning of the fifth century.

182

laurence mellerin

2500

Figure 3. Use of Wisdom Writings.6

2000

Ecclesiastes Proverbs Song of Solomon Ecclesiasticus Wisdom

1500 1000 500 0

Alexandria

Antioch, Syria

Palestine, Cyprus

Researchers might have believed that it was excluded from the canon until four quotations of it were discovered in a work of John Chrysostom, his Commentary on Matthew, whose authenticity was absolutely proved. This simple fact is evidence of its inclusion. Another example shows that using Biblindex may open new research fields. Why do these Antiochene Fathers so rarely quote the Book of Ecclesiastes, whereas in Alexandria there are hundreds of references? The three books attributed to Solomon, Proverbs-Song of Solomon-Ecclesiastes, were not used very much in this area, but the references to Ecclesiastes are at a particularly low level. 2.1.2. Studying Biblical Apocrypha Biblindex has established the following principle: from the moment we know that a text was considered biblical by a Church Father at a given time, the statement of its occurrences in Biblindex is relevant. Our compilation of biblical referencing systems will therefore include all possible books quoted in a specific language. It will thus be easier to study the survival and the diffusion of apocryphal texts, today lost, such as the books of Pseudo-Ezekiel. If we search for quotations of Ezekiel 37:7 in Origen, Gregory of Nyssa or Epiphanius

6 Palestine and Cyprus: including Epiphanius.



new ways of searching with biblindex

183

of Salamis, we read a phrase, ‘bone with bone and joint to joint’,7 which does not entirely match the well-known Hebraic text: ‘and the bones came together, bone to its bone’. Besides, these quotations more closely match a text from the first century before the Christian era, found in Hebrew in the Dead Sea Scrolls and called Pseudo-Ezekiel. In this way we can demonstrate that Epiphanius, at the end of the fourth century, still knew this text and that he probably read it in Greek, just as Gregory of Nyssa did. So there was a Greek translation of it.8 More generally, the influences and the geography of the history of ideas are better known, in qualitative terms—thanks to the text of the quotations—as well as in quantitative terms—according to the number of occurrences. 2.1.3. The Textual Criticism of the Bible And finally, when our textual phase is completed, Biblindex will become a great tool for the textual criticism of the Bible. The comparative search for the quotations of the Fathers, often rich in variants and unexpected readings compared with earlier or modern standard texts, can contribute to the reconstruction of the original text. Indeed, even though passed on by late witnesses, quotations found in authors older than most of the preserved biblical manuscripts allow us to go further back into the history of the biblical texts. 2.2. The Utility of Biblindex for Patristics Scholars The data have to be improved and checked and the corpus has to be enlarged, but some results can already be obtained. The main fact of bringing together so much data highlights global patterns of which we were not necessarily aware before this data was collected. In this paper a total of some 500,000 references, in texts from the first six centuries, are taken into account. 2.2.1. The Most Quoted Texts 48% of the quotations refer to the Old Testament and 52% to the New Testament, although there are of course many more verses in the former. Thus we should not overestimate the percentage of New Testament references by a specific Father. 7 ὀστέον πρὸς ὀστέον καὶ ἁπμονία πρὸς ἁπμονίαν. 8 See Vianès, ‘Vers un texte grec du Pseudo-Ézéchiel? La vision des ossements desséchés’.

184

laurence mellerin Torah

Historical Books

Prophets

Psalms

Text Length Number of Quotations

Wisdom Writings

Figure 4. Patristic Quotations of the Old Testament. Number of References

70000 60000 50000 40000

Number of References

30000 20000 10000 0

s

alm

Ps

sis

ne

Ge

iah

Isa

us

od

Ex

ut

De

y

m

no

o er

b

Jo

l s rs ue iah erb be m am ov S r u 1 P N

m

e Jer

Figure 5. The 10 most quoted books of the Old Testament.

In the chart above, the number of references to each section of the Old Testament has been weighted according to the text length of each section. We can see very clearly that Torah and Psalms are quoted disproportionately; prophetic texts are very often referenced, but in relation to their text length this number of references is not as significant as for the Psalms. Historical texts are very clearly underquoted. If we go a step further, we see that the Church Fathers’ three favourite books are the Psalms, Genesis and the book of Isaiah. As far as the New Testament is concerned, the Gospels and Pauline Epistles are the most quoted texts, both absolutely and in relation to their text length. Matthew’s Gospel is overrepresented, whereas Mark is not very often quoted. This is due to the actual practices of the Fathers, but also from a selection bias due to the systematic choice of Matthew in the database in case of identical synoptic parallels.



new ways of searching with biblindex

185

Gospels Johannine Writings

Acts Text Length Number of Quotations

Catholic Letters

Pauline Epistles Hebrew

Figure 6. Patristic Quotations of the New Testament. Number of References

60000 50000 40000

Number of References

30000 20000 10000 0 M

ew

th

at

hn

Jo

ke

Lu

o

1C

s

ian

th

rin

s

an

m

Ro

ts

Ac

ws

re

b He

o

2C

s

ian

th

rin

E

ns

sia

e ph

k

ar

M

Figure 7. The 10 most quoted books of the New Testament.

A top ten list of the most quoted verses can also be established. After the integration of Cyril of Alexandria’s works into the corpus, the first verse of John’s Gospel is the most quoted by far. We can see the predilection of the Church Fathers for christological verses: equality between God and his Son and the reality of the incarnation are the main issues debated with pagan, Jewish or heretical adversaries. These are not surprising results, but the establishment of such global comparison points is completely new.

186

laurence mellerin

John 1523 In the beginning was the Word, and the Word was with God, and 1:1 the Word was God. Gen 1304 Then God said: Let us make humankind in our image, according 1:26 to our likeness; and let them have dominion over the fish of the sea, . . . John 1215 And the Word became flesh and lived among us, and we have seen 1:14 his glory, the glory as of a father’s only son, full of grace and truth. John 1091 Jesus said to him, ‘I am the way, and the truth, and the life. No one 14:6 comes to the Father except through me.’ Gen 1087 then the Lord God formed man from the dust of the ground, and 2:7 breathed into his nostrils the breath of life; and the man became a living being. Phil 597 (Christ Jesus) who, though he was in the form of God, did not 2:6–7 regard equality with God as something to be exploited, 876 but emptied himself, taking the form of a slave, being born in human likeness. . . . 1 Co 1:24

853 but to those who are called, both Jews and Greeks, Christ the power of God and the wisdom of God.

Col 1:15

712 He is the image of the invisible God, the firstborn of all creation.

Gen 3:1

609 Now the serpent was more crafty than any other wild animal that the Lord God had made.

Hebr 585 He is the reflection of God’s glory and the exact imprint of God’s 1:3 very being, and he sustains all things by his powerful word. Figure 8. Top ten of the most quoted verses.

2.2.2. The Distortion Between the Bible and its Patristic Reconstruction Insofar as the text of the patristic quotations will be gradually integrated into the database, we shall ultimately be able to reconstruct the Bible of a particular author. One would then be able to compare the Bible of two authors, or one author’s Bible at different periods of his life. Studying the use of a specific pericope can also lead to very interesting results. If we take as an example the opening of John’s Gospel, which contains the most frequently quoted verse, John 1:1, we have a sample of more than 6,000 quotations or allusions. Although it is not exhaustive, we can assume that such a huge corpus may be a representative sample from the period that we are studying. We can draw a curve showing the number



new ways of searching with biblindex 1600

187

In the beginning was the word . . .

1400

And the Word became ��esh and lived among us ......

1200 1000 He himself was not the light, but he came to testify to the light.

800 600 400 200 0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

Figure 9. Opening of John as quoted by the Fathers.

of references to each verse of this biblical section, very far from an equal use of each verse. Verses about John (6–8) are far less quoted than the verses about the Word (whom John identifies with Jesus): less than 20 references, whereas verse 1 is quoted more than 1,500 times! 2.2.3. Association of Verses As an instrument of textual comparison and hermeneutics, Biblindex makes visible thematic relationships, similarities in the grouping of references and echoes between the authors. Without this tool, such comparisons could only be perceived with difficulty. For instance: Augustine identifies the blasphemy against the Holy Spirit in Matthew 12:31—‘Therefore I tell you, people will be forgiven for every sin and blasphemy, but blasphemy against the Spirit will not be forgiven.’—with the unforgivable sin of the First Epistle of John 5:16—‘If you see your brother or sister committing what is not a mortal sin, you will ask, and God will give life to such a one [. . .] There is a sin that is mortal; I do not say that you should pray about that’.9 Is this an original idea, or did some Fathers associate these two verses before him? We can find 160 references to these two verses, but only four works which quote them together and near from each other: the treatise De Pudicitia by Tertullian, The Commentaries on John and Matthew written by Origen, and Ambrose’s 9 Augustine never makes explicit the link between these two verses, but the evolution of his exegesis of 1 Jn 5:16, from the De sermone Domini in monte (I, 22, 73) to his Retractationes (I, 19, 7), cannot be explained without referring to his reflection about final impenitence in his Sermo 71 devoted to the exegesis of Mt 12:31.

188

laurence mellerin

treatise De Paenitentia. As a matter of fact, Tertullian and Ambrose are Augustine’s sources on this precise subject. 2.2.4. Textual Criticism in the Patristic Field Last but not least, the biblical references provide information about the history of the texts which quote them. They give us valuable information about the textual family or the biblical manuscripts to which the author had access; about the debates on which he was dependent. They supply him with quotations he does not always verify or reinterpret. These elements can be deduced from the shape taken by the quotation (i.e. the biblical variants of which it gives evidence), or from the interpretation given. The precise form of the text helps to classify texts. Biblical references may enable to date a work: thus we find in John’s first Apocryphal Revelation10 biblical groups containing typically heterodox eschatological views, based on which we can guess it was written in the Byzantine Middle Ages. Isolated or organized together, these are a classical criterion to authenticate a work. Similarly, a previously unidentified allusion to the Letter of Jude helps, for example, to remove the paternity of a text ascribed by some authors to John Chrysostom who, as an Antiochene, would not quote this biblical book. The links between two or several texts, not just as far as their authenticity is concerned but also their literary or doctrinal influence, come out more clearly; it is particularly true as regards the ‘jungle’ of the anthologies or chains, exegetical compilations containing many fragmentary, badly attributed or deformed texts. Duplicates will be spotted by establishing that a particular fragment comes from a work transmitted and known in another way. 3. Methodological Issues Regarding Textual Analysis in Biblindex11 3.1. Definition of ‘Text Units’ In the input interfaces, the critical text of the work to be analysed will be available in a text editor. The analyst will select parts of this text. Then he or she will validate and define the type of the selected biblical 10 Tischendorf, ‘Apocalypsis Iohannis Apocrypha’. 11  This part is an abstract of a more developed paper given in Oxford at the Patristics Conference in July 2012, to be published in Studia Patristica.



new ways of searching with biblindex

189

reference. Search criteria are determined by the distinctions made during the analysis. So we have to think ahead about every conceivable way to sort references. The ‘object’ considered in Biblindex is an intersection of variable range between the biblical corpus on one hand and the patristic corpus on the other, called a ‘text unit’. A text unit is a part of a patristic text in which an author refers, explicitly or implicitly, through quotation, mention or reminiscence, to one or more parts of the biblical text. Multiple ‘biblical references’ may be found in a single unit. In each text unit, the author’s thought and the biblical contents are meshed together in such a way that they are often impossible to separate. 3.2. Typology of Biblical Citations Two types of biblical citations have been established, depending on whether or not they are to be linked to precise verses. Each of these types falls into two cases. First, biblical citations that do not refer to specific verses may be allusions to a biblical stretch of variable range, or they may be a mention of customary biblical word or phrase or of a biblical proper noun. Second, biblical citations may be related to one or more verses. 3.3. Critical Analysis of Biblical Citations Once the first step has been completed and a biblical reference has been allocated to a precise verse, features of the quotation may then be specified. It may be an exact quote of a complete verse which appears only once. But the situation is rarely so simple. A quotation may correspond exactly to the form of text in the biblical referencing system or may be slightly different. If the form of a quotation differs from the form in which it is found in the Bible, a reason for these differences may be specified. We also have to take into account texts with parallels, explicit quotation, quotations being used several times, authors quoting authors quoting the Bible. 4. Conclusion Even if Biblindex already provides possibilities of finding new results and, above all, new ways of searching, from now on we have four main goals. First, a massive enlargement of the corpus; second, the improvement of

190

laurence mellerin

the metadata; third, the development of innovative interfaces and IT tools. Our work is at present strongly ‘manual’. If we could make a computerassisted comparative analysis between a biblical source and a patristic text, the analyst would only have to check and refine what this tool would have prepared, instead of tracking each reference to the Bible himself or herself. There is here specific work to be done on each ancient language. The Turgama project at the Peshitta Institute in Leiden already works on such topics regarding Hebrew and Syriac; the LIRIS lab will soon start working on our Greek corpora. But there is also room for generic issues, independently of language, about text re-use. Our fourth goal is the constitution of a global network of collaborators. About ten researchers are already entrusted with the project at Sources Chrétiennes, and around fifty more French researchers from other labs. But each researcher using the website could take part in the improvement of Biblindex, once the collaborative interfaces have been implemented. References Biblia Patristica, Index des citations et allusions bibliques dans la littérature patristique. Beginnings to Clement of Alexandria and Tertullian, vol. 1 (Paris: Editions du Centre National de la Recherche Scientifique, 1975). Biblia Patristica, Index des citations et allusions bibliques dans la littérature patristique. The third century (apart from Origen), vol. 2 (Paris: Editions du CNRS, 1977). Biblia Patristica, Index des citations et allusions bibliques dans la littérature patristique. Origen, vol. 3 (Paris: Editions du CNRS, 1980). Biblia Patristica, Index des citations et allusions bibliques dans la littérature patristique. Eusebius of Caesarea, Cyril of Jerusalem, Epiphanius of Salamis, vol. 4 (Paris: Editions du CNRS, 1987). Biblia Patristica, Index des citations et allusions bibliques dans la littérature patristique. Basil of Caesarea, Gregory Nazianzen, Gregory of Nyssa, Amphiloque of Iconium, vol. 5 (Paris: Editions du CNRS, 1991). Biblia Patristica, Index des citations et allusions bibliques dans la littérature patristique. Hilary of Poitiers, Ambrose of Milan, Ambrosiaster, vol. 6 (Paris: Editions du CNRS, 1995). Biblia Patristica, Index des citations et allusions bibliques dans la littérature patristique. Didymus of Alexandria, vol. 7 (Paris: Editions du CNRS, 2000). Biblia Patristica, Index des citations et allusions bibliques dans la littérature patristique. Philo of Alexandria, Supplement (Paris: Editions du CNRS, 1982). Mellerin, L., ‘Biblindex, online Index of Biblical Quotations in Early Christian Literature: methodological issues’, in Biblical Quotations in Patristic Texts, H.A.G. Houghton, L. Mellerin (eds), Studia Patristica LIV (Leuven: Peeters, 2013). Tischendorf, C., ‘Apocalypsis Iohannis Apocrypha’, in Apocalypses apocryphae, C. Tischendorf (Leipzig, 1866, repr. Hildesheim, 1966), XVIII–XIX (introduction), 70–94 (Greek text edition). Vianès, L., ‘Vers un texte grec du Pseudo-Ézéchiel? La vision des ossements desséchés’, in L’Antiquité en ses confins. Mélanges offerts à Benoît Gain. Hors série 16 de Recherches et Travaux (Grenoble: Université Stendhal-Grenoble 3, 2008), 163–175.

Aspects of Polysemy in Biblical Greek. A Preliminary Study For a New Lexicographical Resource Romina Vergari 1. Introduction In the field of Biblical Studies, as in all other fields of research, the development of a new tool cannot be justified by the mere fact that nowadays technology allows the development of increasingly sophisticated equipment. On the contrary, such technology is genuinely innovative and truly humanistic, only if it is able to respond more effectively to the actual needs of the scholar, leading to the broadening and enhancement of his knowledge. As recently pointed out by van der Meer, ‘some forty years after the first volume on Septuagint lexicography appeared in the Septuagint and Cognate Studies Series, the state of that area is better than ever before’.1 Suffice to mention the lexicons published in the last ten years: the Louvain lexicon by Lust, Eynikel and Hauspie2 or the Leiden lexicon by Muraoka,3 the richness and value of which cannot be underestimated. Due to their editorial structure, however, the idea conveyed is that the vocabulary of a language or of a corpus of texts consists in a list of entries, and that the meaning or contextual sense for each entry can be represented as fixed entity, which is permanently associated with the lemma. Research in the field of Lexical Semantics, in its recent development, emphasized how this theoretical framework reveals inadequacies in representing the meaning of a word, which turns out to be far more protean by nature. The semantic structure of a given word is assumed to be largely flexible and context-dependent, graspable only in the light of the combinations in which it occurs. This change in theoretical perspective also involves rethinking and updating means to describe the meaning.

1  van der Meer, ‘Problems and Perspectives in Septuagint Lexicography: The Case of Non-Compliance (ΑΠΕΙΘΕΩ)’, 65. 2 Lust, Eynikel and Hauspie, eds., Greek-English Lexicon of the Septuagint. 3 Muraoka, A Greek-English Lexicon of the Septuagint.

192

romina vergari

A lexicon can therefore be said to correspond more to a dynamic system of relations: a) syntagmatic sense relations between lexical units in the same string, i.e. philonymy, xenonymy, tautonymy; b) paradigmatic sense relations between lexical units occurring in a given combination, and a set of possibilities provided by the language, relations of identity and inclusion, as synonymy, hyperonymy and hyponymy, holonymy and meronymy; and relations of opposition and exclusion, as taxonymy, co-meronymy and antonymy. Moreover, the values attributed to a word (i.e. the senses stored in the lexica) exhibit different degrees of entrenchment;4 some of them are more stable and inherent, others are generated ad hoc, triggered and dissolved by context, due to processes of sense modulation and sense generation which govern the combination of words in their actual use. Both senses and relations are affected by a high degree of flexibility. How is it possible to represent such an organization? How can we grasp and describe the limits of this variation? Can new technologies offer help in this task? Is it possible and desirable for Septuagint Lexicography to take advantage of tools that are already available elsewhere, thanks to the progress of Digital Humanities in other fields? The main aim of the present paper is to draw to the attention of readers the preliminary results of the project ‘Septuagint Word-Clustering Database’ (hereafter SWCD), a new tool available for the study of the biblical lexicon. The presentation begins with a corpus-based analysis of a portion of the biblical Greek lexicon. Attention is focused on the nouns associated with the semantic sphere of ‘Law’ and their Hebrew equivalents, within a well-defined and homogeneous textual corpus. The analysis is underpinned by a theoretical study of word-meaning representation, carried out within a lexicological framework which takes into account the flexibility produced by the context, in a range from inherent to selectional polysemy.5 It assumes the syntagmatic processes 4 The notion of entrenchment, developed by Ronald W. Langacker, is one of the foundational insights of cognitive linguistics. According to his theoretical framework, linguistic structures are more realistically conceived as falling along a continuous scale of entrenchment in cognitive organization: ‘each linguistic structure, as the meaning associated with a lexeme, has some degree of entrenchment, which reflects the frequency of its previous activation and determines the likelihood of its subsequent activation’ (Langacker, Foundations of Cognitive Grammar 49). 5 Croft and Cruse, Cognitive Linguistics; Pustejovsky and Rumshisky, ‘Between Chaos and Structure: Interpreting Lexical Data through a Theoretical Lens’.



aspects of polysemy in biblical greek

193

of sense modulation, which contribute in determining the meaning of complex structures.6 Following the principles of the Dynamic Construal Approach, it can be argued that word-senses as such are not directly encoded in the lexicon. Rather, they can be described as textual readings arising from the stereotypical syntagmatic patterns associated with a given word. We made use of formalism, namely a system of semantic-tagging for encoding contexts that helps to determine the contribution of a single word occurring in a larger unit. Further, we applied to the study of the biblical Greek lexicon a methodology by means of which such stereotypical contexts can be identified and subsequently structured in a selection context inventory, encoding both stereotypical syntactic and semantic information. Such analysis led us to the development of SWCD, able of storing and representing all collected data. Some preliminary results produced at the first stage of the research project will be presented here. 2. Preliminaries for a Genuine Corpus-Based Approach Implementing the project, the first step consisted in defining the limits of the textual corpus under investigation. The structural homogeneity of the Greek linguistic variety and the similarity of the techniques adopted by translators (for those texts for which a Hebrew Vorlage is preserved), led us to include in the corpus the books of Pentateuch, the book of Isaiah and the book of 1 Maccabees.7 Secondly, the limits of the lexicon to be analysed had been established. Among the following (in order of frequency), some words have been intuitively included from the start, others have been added over the analysis:

6 Pustejovsky, ‘Type Theory and Lexical Decomposition’. 7 The choice of limiting the corpus to this portion, meets the needs of carrying out the investigation within a language, which presents, as much as possible, homogeneous features. Therefore, we had to take into account two aspects: the structural, functional and stylistic features of the Greek as well as the translation technique chosen by translators. Such methodological requirements have led us to base the analysis on the classification of the books of the Septuagint drawn up by S.J. Thackery, as part of his still fundamental study A Grammar of the Old Testament in Greek according to the Septuagint. According to Thackery, among the translations, only the books of Pentateuch, Joshua (only partially) and Isaiah express a good κοινή Greek (13–14).

194

romina vergari

μαρτύριον, διαθήκη, νόμος, κρίσις, ἐντολή, δικαίωμα, νόμιμον, πρόσταγμα, κρίμα and φύλαγμα.8 After the distributional analysis of such nouns in the corpus was carried out, we decided to limit the semantic tagging procedure to the relation between the noun and the adjective within the Noun-Phrase domain (hereafter NPh); and the relation between the verb and the direct object within the Verbal-Phrase domain (hereafter VPh). 3. Designing SWCD: Theoretical Background The SWCD consists of a spread sheet in which lexical data can be processed by means of filters. Each individual syntagmatic feature constitutes a filter able of sorting homogeneous word-classes. 3.1. Selectee Each nominal item (nouns, adjective, and participles), stored as lemma, is accounted for as a selectee. Taking into consideration the directional properties which the mechanism of the lexical entries selection shows within a syntagmatic domain, it has to be considered as a semantic dependent.9 3.2. Nominal Class Parameter

Values

Nominal Class

Countable (designating a class of instances) Countable (designating a single instance) Uncountable

8 The Greek text investigated is based upon the edition of the LXX edited by Alfred Rahlfs (Rahlfs, ed., Septuaginta. Id est Vetus Testamentum graece iuxta LXX interpretes). The large-scale analysis here presented has been made much easier by the use of the electronic version of this text, available in Accordance, which includes the Kraft/Taylor/ Wheeler Septuagint (LXX) Morphology Database (v. 4.10a Copyright © 2008, Bernard A. Taylor and Dale M. Wheeler). 9 ‘The direction, in which selection operates, is correlated with grammar. The relevant generalization is that adjectives select their head nouns and verbs select their complements; nouns, in general, are always selectees’ (Cruse, Meaning in Language, 223).



aspects of polysemy in biblical greek

195

Every selectee10 is described as regards to its nominal class. The values associated have to be considered as an inherent property of the lexical item, i.e. not context-dependent. – Countable (designating a class of instances). Each noun that occurs in the corpus both in the singular, and in the plural is included in this class, even though a distinction should be drawn between nouns which exhibit a quantitative plural (‘one’ as opposed to ‘more than one’ in terms of multiplicability,11 e.g. γυνή ‘woman’, ‘wife’, τρυγών ‘turtle dove’, στολή ‘dress’, ῥάβδος ‘branch’) and nouns which exhibit a qualitative plural (e.g. ἀγαθός ‘right/good’, ἀγαθά ‘goods’, γένημα ‘harvest’, ‘produce’ γενήματα ‘fruits’, νόμος ‘law’, νόμοι ‘rules’). – Countable as single instances. Proper nouns (that is nouns whose reference is to a class constituted by a single element) are considered in this class (e.g. Γαλααδ, Θαμαρ, Μωυσῆς). – Uncountable. No noun belonging to this class occurs in the corpus in plural (e.g. βοτάνη ‘grass’, αἷμα ‘blood’, κρέας ‘flesh’, χοῦς ‘mud’, δικαιοσύνη ‘justice’, λατρεία ‘service’, ‘worship’). 3.3. Type Parameter

Values

Type

Natural Artifactual Complex

A type12 value is assigned to each selectee in the specific syntagmatic context considered. The semantic type is the level representation of a nominal affected by the mechanism of selection.13 The noun is specified with reference to an ontology of reference. Following Pustejovsky,14 the

10 Cf. Corbett, Number; Acquaviva, Lexical Plurals: A Morphosemantic Approach. 11  Belardi, ‘La questione del numero nominale’, and Gobber, ‘Numerabilità, culminazione semantica e categorizzazione’. 12 Pustejovsky, ‘Type Theory and Lexical Decomposition’. 13 Pustejovsky, The Generative Lexicon. 14 Pustejovsky, ‘Type Construction and the Logic of Concepts’; Pustejovsky, ‘Type Theory and Lexical Decomposition’.

196

romina vergari

ontology divides the domain of individuals into three levels of structure type: natural types, artifactual types and complex types. 3.3.1. Natural Type The selectee is specified only with regard to the identification of an individual within a broader domain, e.g. ὄνος ‘donkey’ [animate], κέδρος ‘cedar tree’ [plant], ἄνθρωπος ‘man, human being’ [person], φῶς ‘light’ [energy], στέαρ ‘fat’ [substance]. 3.3.2. Artifactual Type The selectee is specified not only with regard to its position in the ontology, but also with regard to its function, origin or location. In other words, a predicative force is associated with the nominal item members of this class. This predicative force becomes evident only thanks to the syntagmatic processes of semantic composition in context, e.g. the noun στήλη ‘stele’ designates a bounded physical entity equivalent to ‘stone’, and in this respect can be described as quasi-synonym of λίθος. Furthermore it has a predicative force which specified the function of its referent (μιμνῄσκεσθαι τινα/τι). The word κτῆνος refers to ‘a head of cattle’, like πρόβατον ‘sheep’, αἴξ ‘goat’ or μόσχος ‘calf ’, but κτῆνος is able also of transmitting information about the nature of ‘property, possession’ associated to these animals, the word ὁλοκαύτωμα on the other hand gives information about their function as sacrificial victims. Other examples of this type are: ἱερεύς ‘priest, minister’ compared to ἄνθρωπος; ἀργυρώνητος ‘slave bought’ and οἰκογενής ‘slave born in the master’s house’ compared to δοῦλος and παῖς; ὑποζύγιον ‘donkey, beast of burden’ compared to ὄνος. 3.3.3. Complex Type A complex nominal can be selected with more than only one type, different one from the other. The phenomenon of vagueness (which arises if the change in representation is correlated to the context) is connected to complex type nouns. Within the corpus the item γῆ ‘portion of land’ is a countable noun. Its referent can be represented as a bounded physical entity. The verbs δεικνύναι ‘to show’, λαμβάνειν ‘to conquer’, παρέρχεσθαι ‘to cross’; or the NPhs Ἀιγύπτου, τῶν Χαλδαίων select this sort of representation. The word γῆ ‘earth’ can be selected also as a mass physical entity in opposition to οὐρανός ‘sky’; this type arises in combination with the selectors κρίνειν ‘to judge’ and ποιεῖν ‘to create’ (and the subject κύριος ‘the Lord’).



aspects of polysemy in biblical greek

197

The type-complexity can be even greater, e.g. the word βδέλυγμα occurs in association with a cluster of different configurations. In combination with the selectors θέλειν ‘to pursue’, εἶδον ‘to know, to take part in’, ποιεῖν ‘to perform’, it designates a definite process and the sense ‘abomination’ is triggered in context; however, in combination with the verb τιθέναι ‘to put’, the noun βδέλυγμα is able of admitting a bounded physical entity type, i.e. ‘idol’, object of abomination par excellence. The noun θυσία exhibits similar syntagmatic behaviour. The expression διδόναι θυσίαν triggers the reading ‘to give an offering’ [bounded physical entity], whereas the phrase ποιεῖν θυσίαν ‘to make an offering’ [definite process]. The parameter ‘type’ is evidently correlated to the more classical distinction between referential nouns and eventive nouns, i.e. between those nouns which have the primary function of designating an entity (both physical or mental), and those (mostly de-verbal), which project in the NPh their complements and share some semantic features as duration, dynamicity, telicity with verbs (examples are διατήρησις ‘preservation’; εὐχή ‘prayer, vow’; διαθήκη ‘covenant’; ἀρά ‘prayer, curse, imprecation’; γογγυσμός ‘groaning, grumbling’). Nevertheless, the eventivity in nouns should be regarded as a contextdependent property, since it can be triggered by context in referential nouns and vice versa. 3.4. Qualia Structure Parameter

Values

Qualia structure

Formal Constitutive Agentive Telic

The notion of qualia roles has been elaborated by James Pustejovsky,15 although it is also possible to trace correlation with notions developed within other theoretical frameworks.16 In the database four values are

15 Pustejovsky, The Generative Lexicon, 76–83. 16 See, for instance, the notion of ways-of-seeing developed by Cruse. Although he accepts the parallel with Pusteyovsky’s qualia roles, he is quite skeptical about the possibility of

198

romina vergari

associated with this parameter: formal, constitutive, agentive, and telic. According to Pustejovsky ‘these four factors drive our basic understanding of an object or a relation in the world. They furthermore contribute to our ability to name an object with a certain predication’.17 The constitutive quale structures a lexical item with regard to the relation between the referent and its constituents or proper parts (as material, weight, component elements); the formal quale allows the dinstinction of the object within a larger domain (its orientation, magnitude, shape, position); the agentive quale is related to factors involved in the origin of an object (its creator, the causal chain); and the telic quale conveys information about its purpose and function. It should be pointed out that the qualia structure plays a significant role not only in the syntagmatic processes of sense exploitation, sense modulation and sense coercion in context, but also in morphological processes as compounding and derivation. On the one hand, the predicative force is inherent to the meaning of a complex word; on the other hand, it is rather present in potentia in the noun, and activated only in context. Generally speaking, it can be said that every lexical item presupposes a formal quale, which governs the lexical relation of hyper/hyponymy (ξύλον ‘tree’ vs. κέδρος ‘cedar’, κυπάρισσος ‘cypress’, λευκή ‘poplar’); and also a constitutive quale, which governs the lexical relation of holo/meronymy (e.g. μόσχος ‘calf ’ vs. βραχίων ‘shoulder’, ὀσφῦς ‘hip’, στηθύνιον ‘breast’, ἔνυστρον ‘stomach’, λοβός ‘lobe (of the liver)’, σιαγόνιον ‘jawbone’). 3.4.1. Morphological Level In the following examples, it is possible to pinpoint the action of the qualia roles: ἅρπαγμα ‘spoils’ designates assets acquired by means of violent actions (derivated from ἁρπάζειν ‘to seize, to plunder’ agentive quale); λυχνία ‘lamp-stand’ indicates an artifact by means of one of its component parts (derivated from λύχνος ‘torch light’ constitutive quale); δακτύλιος ‘ring’ suggests the location where the artifact must be placed (deriving from δάκτυλον ‘finger’ formal quale [position]), περίζωμα ‘apron, belt’ (περί ‘around, round about’ and ζωννύναι ‘to gird’ formal quale [position]); ὑποδύτης ‘tunic, robe’ (ὑπό ‘under’ e δύειν ‘to wear’ formal quale [position]);

limiting the number of this semantic discontinuity to four. For a more detailed discussion, see Cruse, ‘Aspects of the Micro-Structure of Word Meanings’, 49–50. 17 Pustejovsky, The Generative Lexicon, 85.



aspects of polysemy in biblical greek

199

φωστήρ ‘star, great light’ (φῶς ‘light’ telic quale); πρωτότοκος ‘firstborn’ (πρῶτος ‘first’ and τίκτειν ‘to give birth’ agentive quale); ἀργυρώνητος ‘slave bought’ (ἄργυρος ‘money’ and ὠνεῖσθαι ‘to buy’ agentive quale); οἰκογενής ‘slave born in the master’s house’ (οἶκος ‘house’ and γένος ‘offspring’ agentive quale); κτῆνος ‘cattle’ (κτᾶσθαι ‘to purchase, to obtain’ agentive quale); κατακάρπωσις ‘ashes of the sacrifice’ (κατά ‘down’ and κάρπωσις ‘sacrifice’ formal quale [position] and agentive quale); ποτήριον ‘cup’ (πίνειν ‘to drink’ and then ποτήρ ‘container for drinking’ telic quale); ὑποζύγιον ‘donkey’ (ὑπό ‘under’ e ζυγός ‘yoke’ formal quale [position] and telic quale); περισκελής (περί ‘around, round about’ and σκέλος ‘leg’ formal quale [position]). 3.4.2. Syntagmatical Level As for the noun βιβλίον ‘book’, the agentive quale is exploited by the selectors like γράφειν ‘to write, to compile’ and ἅγιον ‘holy’; the telic quale is exploited by the selector ἀναγιγνώσκειν ‘to read aloud’, its formal quale plays a significant role in the combinations as διδόναι βιβλίον, τιθέναι βιβλίον and its constitutive quale in contexts as σφραγίζεσθαι βιβλίον; moreover the constitutive quale is also presupposed in the paradigmatic relations with words as ἀντίγραφος ‘copy of document’ or expressions as ἀναγιγνώσκειν βιβλίον, where the noun can be accounted as the hyperonym of a set of nouns designating the content of the text, e.g. ἀρά, and ᾠ δή. It is possible to establish a hierarchy among type, qualia structure and Aktionsart/Seinsart. On the basis of the data gathered, we can say that: – Every selectee of the natural type embedded at least the formal and the constitutive quale; such roles govern the paradigmatics relations of hyper/ hiponymy and holo/meronymy. – Every selectee of the artifactual type embedded at least an agentive and/or telic quale; such structure governs the syntagmatic relations of phylonymy. – Every selectee of the complex type exhibited in context more than one Aktionsart or Seinsart. The following are the most frequent patterns of variation: – Bounded abstract entity ~ bounded physical entity ἅγιος βιβλίον ‘holy book’ ~ λαμβάνειν βιβλίον ‘to take the book’; δεικνύναι βραχίων ‘to show the power’ ~ διδόναι βραχίων ‘to offer the shoulder (in cultic offerings)’.

200

romina vergari

– Mass abstract entity ~ state ἀναγγέλλειν δικαιοσύνην ‘to announce the justice’ ~ διατηρεῖν δικαιοσύνην ‘to keep the justice’. – Mass physical entity ~ definite process δεικνύναι αἷμα ‘to show some blood’ ~ ἐκζητεῖν αἷμα ‘to demand account of the killing (lifeblood)’. – Bounded physical entity ~ state διδόναι βασιλείαν ‘to give the kingdom’ ~ διδόναι βασιλείαν ‘to assign the sovereignty’; λαμβάνειν ῥάβδον ‘to hold the rod’ ~ λαμβάνειν ῥάβδον ‘to take authority’. – Bounded physical entity ~ Mass abstract entity ἀνταποδιδόναι ἀγαθόν ‘to give in exchange for good’~ ποιεῖν ἀγαθόν ‘to practice good’; αἴρειν χείρ ‘to raise the hand’ ~ δεικνύναι χείρ ‘to show the power’. – Definite process ~ state ποιεῖν ἀγαλλίαμα ‘to celebrate’ ~ ἀγαλλίαμα αἰώνιον ‘everlasting jubilation’; ἀνταποδιδόναι ἀδικίαν ‘to return the injustice’ ~ ἀφαιρεῖν ἀδικίαν ‘to take away the injustice’. – Mass abstract entity ~ Mass physical entity ~ Bounded physical entity ~ state δεικνύναι δύναμιν ‘to show the power’ ~ λαμβάνειν δύναμιν ‘to acquire goods’ ~ τιθέναι δύναμιν ‘to place the army’ ~ μεγάλη δύναμις ‘great power’. – Mass abstract entity ~ Mass physical entity ἀναγγέλλειν δόξας ‘to proclaim the glory’ ~ ἀφαιρεῖν δόξας ‘to steal the wealth’. – State ~ definite process φυλάσσειν τὴν εἰρήνην ‘to keep the peace’ ~ συντιθέναι εἰρήνην ‘to decide a truce’. 3.5. Valency Parameter

Values

Valency

Monovalent Bivalent Trivalent Avalent



aspects of polysemy in biblical greek

201

As only an eventive nominal type complex is able to show a specification for the valency parameter,18 it is possible to establish a hierarchy between type and valency. Avalent: the event expresses by the noun head does not require the projection of arguments e.g. [ὑετός [0]] ‘rain’, [πῦρ [0]] ‘fire’. Monovalent: the noun head projects one argument on the NPh; e.g. [ἡ ἡμέρα [τῶν σάββατων]] ‘the day of Sabbath’, [ἡ ἱερωσύνη [Φινεες]] ‘the priesthood of Pinhas’. Bivalent: the noun head projects two arguments on the NPh; e.g. [ἡ ἔχθρα [ἀνὰ μέσον σου] καὶ [ἀνὰ μέσον τῆς γυναικός]] ‘enmity between you and the woman’; [ἡ εὔνοια [τοῦ Δημετρίου] [τῷ Ιωναθαν]] ‘the benevolence of Demetrius to Jonathan’; [τὸ δόμα [τῆς ἱερατεία [ὑμῶν]] ‘the gift of your priesthood’. Trivalent: the noun head projects three arguments on the NPh. In the data collected so far, there is not yet any example of such a noun. 3.6. Aktionsart/Seinsart Parameter

Values State

Aktionsart

Process Instantaneous event

definite undefinite

Every referential noun is specified for the Seinsart; every eventive noun is specified for the Aktionsart.19 It is useful pointing out that ‘referents of NPhs are not objects in the real world, but rather mental constructs that are created, stored, and retrieved in the minds of the speech participants’.20 The Aktionsart and Seinsart should be regarded as context-dependent features, not as properties inherent in the meaning of the noun. They are available in potentia in the semantic structure of a given noun, triggered in actu by the syntagmatic context.

18  Giorgi and Longobardi, The Syntax of Noun Phrases. 19  Comrie, Aspect (Cambridge Textbooks in Linguistics). 20 Rijkhoff, J., ‘Order in the Noun Phrase’, 326.

202

romina vergari

3.6.1. Aktionsart The Aktionsart is a specification concerning the internal structure of an event. These are the parameters used to determining the Aktionsart: Semantic Category

Duration Dynamicity Definiteness Telicity

Nouns Referential

State

Indefinite Process

Definite Process

Instantaneous

– – – –

+ – + –

+ + – –

+ + + –/+

– + – +

The following are the tests used in assigning a value for the Aktionsart: – Duration test: ‘for xduration’: ἔλεος αἰώνιον ‘everlasting favour’; πολὺς κλαυθμός ‘long cry’ (vs. μέγας κλαυθμός ‘a deep cry’); – Dynamicity test: ἐγένετο ‘it happened that/there was/were’: ὕδωρ ἐγένετο ‘there was rain’; κρίσις ἐγένετο ‘there was a dispute’; – Event definiteness: ‘x ἡμερῶν’, ‘x μηνῶν’: ὁδός τριῶν ἡμερῶν ‘a three days’ journey’. – Telicity test: ἡμέρα ἐκδικήσεως ‘day of vengeance’; ἐν ὥρᾳ ἐπισκοπῆς ‘at the hour of visit’. In general, nouns denoting a state have the features of duration and definiteness (θλῖψις ‘anxiety’, ἀγαλλίαμα ‘jubilation’, πικρία ‘bitterness’, εὐφροσύνη ‘cheerfulness’); indefinite process nouns have the features of duration and dynamicity (πονηρός [NEU] ‘the evil’, ὅρμημα ‘violence’, γογγυσμός ‘groaning, grumbling’); definite process nouns have the features of duration, dynamicity, definiteness and telicity (ἐκδίκησις ‘vengeance’; ἐπισκοπή ‘visit’). 3.6.2. Seinsart The referential selectees are specified for a Seinsart. Every possible Seinsart configuration has to be regarded in reference to ontology. In the database, a portion exemplified can be found. Nevertheless a correlation (in terms of frequency and naturalness) between formal properties (morphological and syntactical) and Aktionsart/Seinsart can be established:



aspects of polysemy in biblical greek

203

– Between uncountable noun and mass entity. However, unexpected combinations can be found: e.g. the word χοῦς ‘mud’ is associated to a mass physical object Seinsart, nevertheless we found the word in this combination: χοῦς δίκαιος, in such context the word is evidently used to designate a bounded physical entity whose meaning is ‘unit of measure for liquids’ (in opposition to σταθμίον ‘unit of measure for solids’). The word βούτυρον ‘butter’ conveys a mass entity Seinsart, but in the VPh λαμβάνειν τὸν βούτυρον this semantic type does not match the type required by the verbal selector ‘to take (to seize)’, so that the value ‘bounded entity’ is introduced syntagmatically.21 – Between countable and bounded entity. However unexpected combinations can be found, as the collective neuter plurals κτῆνη ‘cattle’, πρόβατα ‘flock’, σκευή ‘furnishing’, ξύλα ‘wood’, ὀστέα ‘bones’. – Between uncountable nouns and state or indefinite process. Examples of this correlation are the nouns denoting a state γογγυσμός ‘groaning, grumbling’; ἀγαλλίαμα ‘jubilation’, θυμός ‘anger’, μακροθυμία ‘forbearance’, εἰρήνη ‘peace’, ταπείνωσις ‘humiliation’, ἀπώλεια ‘destruction, desolation’. – Between countable nouns and definite process nouns. Predictable examples of this correlation are: φωνή ‘cry’, πόλεμος ‘war’, ἑορτή ‘feast’, ἐτασμός ‘disaster’, γάμος ‘wedding’. – Between countable nouns and instantaneous event nouns. Predictable examples of this sort are: κρίσις ‘sentence’, ἀδίκημα ‘damage’, ἀνταπόδοσις ‘requital’, μῶμος ‘disgrace’. However, unexpected combinations can be found in the database. Let us dwell for a moment on the noun εἰρήνη (uncountable in our corpus) in the combinations φυλάσσειν τὴν εἰρήνην ‘to keep the peace’ and ἱστάναι εἰρήνην ‘to establish the peace’. In the first phrase, the object designates the state ‘lack of physical strife and armonious state’, or ‘lack of mental, inner turmoil, peace of mind’.22 In the latter, the ‘definite process’ reading ‘truce’ or ‘treaty’ emerges due to pressure of context. This process affects and changes also its paradigmatic relations, to the point that εἰρήνη, designating relationships among groups or individuals ruled by written documents, becomes 21  In this regard, see Prandi, The Building Blocks of Meaning, 127: ‘the specific way of instantiating a mass, connected with the use of determiners, consists in selecting a given portion out of the mass, and sometimes in imposing a shape or container on it’. 22 Muraoka, A Greek-English Lexicon of the Septuagint, ‘εἰρήνη’.

204

romina vergari

part of the taxonomy: διαθήκη ‘covenant’, ὁρισμός ‘obligation’, συμμαχία ‘military alliance’. The word ἀγαλλίαμα is another interesting example of this kind of syntagmatic behaviour: in the NPh αἰώνιον ἀγαλλίαμα the word indicates a state of ‘jubilation’ or ‘joy’; on the other hand, the combination ποιεῖν ἀγαλλίαμα triggers the definite process reading ‘to celebrate (a feast)’. 3.7. Compositional Operations of Sense Modulation Every syntagmatic combination is tagged in the database with one of the following values: – Pure selection: the semantic structure of the selectee suits the requirements of the selector. – Accommodation: the semantic structure of the selector changes as a function of selectee’s requirements. – Type coercion: the semantic structure of the selectee adjustes to that of the selector; this alteration is accomplished by either: • Exploitation: only a subcomponent of the selectee’s type is accessed and exploited by the selector. • Introduction: the selector introduces features absent in the structure of the selectee. 4. Testing SWCD: Some Outcomes In the following section, the main features of SWCD will be shown. These tables are generated through the queries that from time to time will be presented. 4.1. Cluster Words according to their Syntagmatic Properties By searching the database for any verb, the direct objects with which it combines in the corpus are sorted:



aspects of polysemy in biblical greek

205

Sub-Type

Type

Selectee

Selector

Token

Class

– ἀνταποδιδόναι

Reading

1 Macc 10:27

V ἀνταποδιδόναι ἀγαθός

complex action

to return favour

Lev

18:25

V ἀνταποδιδόναι ἀδικία

complex action

to return evil

Isa

66:4

V ἀνταποδιδόναι ἁμαρτία

complex action

to repay sin

Gen

50:15

V ἀνταποδιδόναι ἀνταπόδομα complex cost

to requite