Non-Roman fonts used: SPEdessa
This article is also available in transliteration and text-only formats.
George Anton Kiraz, Comparative Edition of the Syriac Gospels: Aligning the Sinaiticus, Curetonianus, Peshitta and Harklean Versions. 4 volumes. New Testament Tools and Studies, vol. 21/1-4. Leiden: E. J. Brill, 1996. ISBN: 90-04-10419-4. Pp. xci+454+257+514+369, 7 plates.
1. "No branch of the Early Church has done more for the translation of the Bible into their vernacular than the Syriac-speaking" (Nestle 1902:645). In fact, within the first six centuries of the Christian era, the Syriac Fathers managed to produce no less than six different versions of the New Testament; viz., the Diatessaron, the Old Syriac Gospels, the Peshitta, the Philoxenian, the Harklean, and the Syro-Palestinian (the latter being in Christian Palestinian Aramaic), not to mention the long series of lost revisions which brought the Old Syriac into closer line with the Greek Gospels in the final form of the Peshitta version.1 The immensity of their scholarship succeeded in keeping modern New Testament scholars very well occupied with discerning the numerous issues related to the interrelationship of these various translations and revisions; in the words of Metzger,
Of all the early versions of the New Testament, those in Syriac have raised more problems and provoked more controversies among modern scholars than any of the others. The reasons lie partly in the multiplicity of translations and revisions of the Syriac Scriptures, and partly in the ambiguity of evidence concerning their mutual relationship [italics mine] (Metzger 1977:3).
2. The purpose of this project is firstly to provide a tool which may be of help in resolving some of the problems and controversies raised by the multiplicity of translations and revisions of the Syriac Gospels. Secondly, it aims to facilitate the study of their mutual relationship. The necessity of such a tool is due to the practical difficulties currently facing the scholar in accessing the desired texts simultaneously and at a glance. This is best exemplified in the case of the Old Syriac texts: although excellent editions of the Sinaiticus palimpsest and the Curetonianus manuscript have been produced by Lewis and Burkitt, respectively (Lewis 1910; Burkitt 1904), both editions pose practical difficulties for the scholar whose main concern is to compare the two texts. Lewis had collated Sinaiticus to the text of the Curetonianus, whereas Burkitt had done the reverse by taking Curetonianus as collational base.2 It has already been suggested by Kahle that the two texts ought to have been edited along the lines of Jülicher's Itala in order that the differences between them be made clear (Kahle 1959: 302; cf. Jülicher et al. 1970-). Even before Lewis and Burkitt had produced their editions, Bonus had taken one step in this direction and compiled the variants of both texts, in addition to those of the Peshitta, in three adjacent columns.
3. An additional difficulty facing the scholar arises because of the poor quality of the standard Harklean text. Whilst a good critical edition of the Peshitta Gospels exists based on forty-two manuscripts edited by Pusey and Gwilliam (Pusey and Gwilliam 1901),3 the poor quality of White's edition of the Harklean leaves much to be desired (White 1778; for the text of John, see Bernstein 1853).
4. The current edition brings together the texts of the Sinaiticus (S), Curetonianus (C), Peshitta(P) and Harklean (H), aligned under each other in such a manner which allows the scholar to access all the versions in one glance. The texts provided in the CESG are based on the aforementioned standard texts of S, C and P. As a reliable standard edition of H is lacking, the CESG provides a new text based primarily on one of the earliest witnesses to the text, Vatican Ms. Syr. 268, thought by Mai to have been written by Thomas of Harkel himself, although Hatch dates it to 858-859 C.E. (Mai 1831:4-5; Hatch 1946). This latter text was kindly prepared for the purposes of this edition by Dr. Andreas Juckel.
5. The lost texts of the Diatessaron and the Philoxenian have no place in the CESG. In the case of the former, this is caused by the absence of an acceptable standard construction of the text. Although there have been attempts to construct the Diatessaron, most notably in the Spanish Polyglot (Ortiz de Urbina 1967),4 New Testament scholarship is far from having produced an acceptable construction, even if that be at all possible. In the case of the Philoxenian, no trace of the Gospels has come down to us, except in the form of a handful of citations in the works of Philoxenos himself.
6. The Syro-Palestinian version (Lewis and Gibson 1899) is also not included, primarily because it represents an independent translation from the Greek and does not seem to have played a role in the history of the development of the Syriac texts,5 and secondly because it is composed in an idiom which differs from Syriac, viz. Christian Palestinian Aramaic. It must be stated that the purpose of this work is but to provide the scholar with a tool; hence, neither investigating the Greek behind the Syriac (apart from few places where it seemed crucial for the alignment), nor providing citations from patristic literature, though crucial for the study of the development of the texts, fall within the scope of this project.
7. The idea of producing this edition sprang out in the academic year 1990-91 while I was reading for the degree of Master of Philosophy in Syriac Studies under Dr. Sebastian P. Brock at the University of Oxford (Wolfson College). During study sessions on the textual relationship of the various Syriac versions of the Bible, Dr. Brock suggested the compilation of a comparative edition of the Gospels based on the standard editions of the Sinaiticus, Curetonianus and Peshitta. The Harklean was initially excluded for the lack of a reliable edition of its text.
8. A preliminary proposal was submitted to E. J. Brill in 1991. Later that year, Prof. Bruce Metzger, editor of New Testament Tools and Studies, responded encouragingly and kindly accepted the CESG into his series. The original proposal aimed at editing the texts of S, C and P in three adjacent columns with special marks to indicate the differences between them. Later, however, inspired by the work of Aland and Juckel, a decision was made to align the texts under each other (Aland and Juckel 1986-). After discussions with Dr. Andreas Juckel, it was also decided to incorporate H into the CESG. Work began in the summer of 1992 by entering the texts on the computer. The texts were proof-read later that year against the original standard editions and were imported into the Syriac Electronic Data Retrieval Archive (see Kiraz 1994:461-475).
9. During 1993, a prototype of CE-EDIT (comparative edition editor) was designed, developed and tested. The program reads a particular verse from multiple textual databases (input), displays the texts on the screen and allows the user to align them interactively. The alignment information (output) is maintained in another database as illustrated below.6
10. The alignment process took place in 1994. The following year, the prototype program, CESG-TEX, was written. The program reads the alignment database and the various textual databases and outputs a file suitable to be processed by the LATEX typesetting system using the sabrâ package, a set of Syriac fonts and macros (Lamport 1994; Haralambous 1995). The process of producing the output is illustrated in the following diagram.
11. The same year, an announcement was placed on the Internet requesting participants to proof-read the texts for the second time. About twenty replies were received.
12. The four texts are aligned under each other as in the following examples. The verse reference (e.g., 7.33) is given in the right margin (abbreviations for the books are included below for clarity, though they are not in the CESG itself). The texts appear in blocks, each block consisting of four lines at most: the top line being the text of S and the bottom line representing H. (Sigla are also given in the Serto script for the benefit of the Syriac reader.) Two signs are used in the alignment: "
" indicates "no correspondence" and"..." indicates lost text (only in S and C).
13. If the whole verse is wanting in S or C, it is omitted in the blocks as illustrated in the following example, where S is partially wanting and C is not extant.
14. However, if a verse is omitted in the text, it is represented by 's throughout (e.g., S in Mk 9.44).
15. The alignment is controlled by two general rules having the following precedence:
16. The next example illustrates the precedence of the two rules. Here rule 2 is abandoned for the sake of rule 1, for if )rbdml in H is aligned with its sisters in the other versions, two alignments (viz. Nm and )xwr) would be sacrificed.
17. There are, however, a number of categories which nullify rule 2. These are (in the following examples, "center" refers to the aligned word[s]) in question):
18. P does not control word order if the words to the right and left of the center are semantically related, but not identical; in the following example, (w$y is the center, surrounded by rm) and llm.
19. P does not control word order if the words to the right and left of the center are particles, prepositions, enclitics, etc. For example, Nyd in Lk 4:40.
20. When word order differs among the versions, an outer expansion from the center is sought, if possible, where the nth (i.e. 1st, 2nd, etc.) word to the right and left of the center correspond. For example, )nks^m forms the center of the following alignment, and the first set of words to its right and left, ryg, correspond; so is the case in the second set, Nwhyty) / ty), and the third set, Nbzlkb.
21. The center is semantically related, but not identical, and no alignment can be found immediately to its right and/or left, for example Nyc^ybx, Nylpn, and Nwlpnd in Mk 3:10.
22. In addition, the following rules have been adopted:
23. If a transposition of two or more words occurs (usually in a conjunction or disjunction form), the transposition in most cases is ignored in the alignment, e.g. )b) and )m) in Mk 10:29.
24. Compounds which occur as one word in some versions while as two words in others are aligned in one column, for example, )crq lk) and )crqlk) in Mt 25:41.
25. The same applies to enclitics such as )n) rm) and )nrm), and to compounds which are semantically related, like )nwr+nq and ))m $yr.
26. Where two phrases do not match, but occur between aligned text, both are aligned under each other, for example, hyl( rm)t) )ld and )tlm Nm rbl in Mk 5:32.
27. In the case of H, consideration was given to the Greek behind the Syriac; thus, $n)l in H is not aligned with its sisters in C and P.
28. The project was completed in November 1995, and the CESG was published in January 1996. The system described above can be used for the alignment of other texts, though that may require minor changes.
© TC: A Journal of Biblical Textual Criticism, 1996.
1For a review of the Syriac translations and revisions of the New Testament, see Bruce M. Metzger, The Early Versions of the New Testament: Their Origin, Transmission, and Limitations (Metzger 1977:3-98). The issue of the Philoxenian/Harklean problem is discussed by S. P. Brock (Brock 1981). 
2In Burkitt's case, this was due to an accident of editorial history: he "took over a commission originally given to R. L. Bensly who had been asked to produce an edition of Curetonianus before the Sinaitic palimpsest was discovered (Black 1972:123-124).
3The base text of this edition was republished by the British and Foreign Bible Society in 1905, 1920 and in subsequent editions. The same text was adopted by the United Bible Society's edition of the Syriac Bible in the 1988 and subsequent editions reformatted in two columns; alas, the Syriac chapter divisions,)xx^c,were omitted.
4For a critical study of the limitations of this work, see Murray 1967:43-49.
5Black, however, concludes that the Palestinian Syriac has some Diatessaronic influence (Black 1941:101-111).
6The original plan was to design a program which automatically aligns the texts to a relatively good degree of accuracy, then have the alignments manually tuned using CE-EDIT. After experimenting with manual alignment, however, it was found that it takes 30-60 minutes to align the texts of one chapter interactively (depending on its complexity and length using CE-EDIT directly.
Black, Matthew 1941. "The Palestinian Syriac Gospels and the Diatessaron." Oriens Christianus 36: 101-111.
Kahle, Paul 1959. The Cairo Geniza. 2nd ed. Oxford: Blackwell.
Lamport, Leslie 1994. LATEX: A Document Preparation System. 2nd ed. Reading, MA: Addison-Wesley.
Mai, Angelo 1831. Scriptorum veterum nova collectio. Vol. 5, Part 2. Rome: Typis Vaticanus.
Murray, Robert 1967. "Reconstructing the Diatessaron." Heythrop Journal 10: 43-49.
Nestle, Eberhard 1902. "Syriac Versions." In Hastings' Dictionary of the Bible.
White, Joseph, ed., 1778. Sacrorum evangeliorum versio syriaca Philoxeniana. Oxford: Clarendon.