CC BY
29
УДК 930.272:2-264
Basil Lourie WHY SO SYRIAN?
A QUANTITATIVE BAYESIAN APPROACH TO THE PERTURBATIONS OF THE TEXTUAL FLOW IN THE SLAVONIC RECENSIONS OF THE PAULINE EPISTLES*
It would be preposterous to claim that thousands of copyists had the same tics nerveux.
William Veder (Veder 2014: 376, n. 12)
Some "Oriental" impact on a very early recension of the Slavonic translation of the five Pauline epistles is demonstrated. The most natural interpretation of this fact is that the Greek originals used for the Slavonic translation were those widespread outside the borders of the ninth-century Byzantine Empire. Moreover, there are some traces of additional editing of the translations from Greek against some Syriac version(s), in the same manner as in roughly contemporaneous Melkite translations of the New Testament from Greek into Arabic. Both I ("Ancient") and II ("Preslav") recensions share the above features that, therefore, are to go back to their common archetype, that is, an even more "ancient" recension. Only the earliest Ethiopic version (EthGr) shows an affinity with the Slavonic material comparable to that of the Syriac material. This fact must be interpreted as an affinity of the early Slavonic version with Greek recensions that were circulating before the seventh-century Arab invasion.
The quantitative method proposed in this paper is dedicated to comparison of two competing hypotheses concerning the textual flow of a highly contaminated tradition. The method has the following preconditions and limitations: (1) the total number of possible hypotheses must be previously reduced to two: that a specific source of contamination existed or not; (2) in the present (simplest) modification of the method, the hypothesis about the presence of a discussed source of contamination must additionally imply a high value of the signal-to-noise ratio (> 0.5), that is, that this hypothetical source, if it actually existed, was the major source of contaminations of a specific kind (defined above as "perturbations").
Key words: Slavonic New Testament, Pauline epistles, Oriental versions of New Testament, Syriac New Testament, Ethiopic New Testament, Bayesian posterior likelihoods, inductive logic.
About the author: Basil Lourie, Doctor of philologycal science, Leading Researcher of the Scientific and Educational Laboratory for Interdisciplinary Empirical Research, Sector for Historical Research of the National Research University "Higher School of Economics" in Perm — branch of the National Research University "Higher School of Economics".
Contact information: 614070, Russia, Perm Region, Perm, Studencheskaya Str. 38, National Research University "Higher School of Economics" in Perm — branch of the National Research University "Higher School of Economics"; tel. +7 (342) 205-52-50, e-mail: [email protected].
В. М. Лурье
ПОЧЕМУ ИМЕННО СИРИЙСКИЙ?
КОЛИЧЕСТВЕННЫЙ БАЙЕСОВСКИЙ ПОДХОД К ВОЗМУЩЕНИЯМ ТЕКСТОВОГО ПОТОКА В СЛАВЯНСКИХ РЕЦЕНЗИЯХ ПАВЛОВЫХ ПОСЛАНИЙ
Доказывается наличие некоего «восточного» влияния на очень раннюю редакцию славянского перевода пяти Павловых посланий. Наиболее естественной интерпретацией этого факта является вывод об использовании для перевода таких греческих редакций, которые были распространены за пределами Византийской империи. Кроме того, отмечаются некоторые следы правки этих славянских переводов с греческого по сирийским переводам — аналогично тому, что сейчас выявлено для мелькитских переводов Нового Завета с греческого на
* Статья поступила в номер 27 октября 2016 г. Принята к печати 02 ноября 2016 г.
© МАИАСК. Археология, история, нумизматика, сфрагистика и эпиграфика. © Basil Lourie, 2016.
Вып. 8. 2016
арабский, которые датируются, приблизительно, той же эпохой. Поскольку отмеченные черты свойственны как I («древней»), так и II («преславской») редакциям Апостола, можно предположить, что они восходят к их общему архетипу, то есть должна была существовать какая-то еще более «древняя» редакция перевода. Из восточных версий, не связанных с сирийской культурой, особую близость к славянскому демонстрирует древнейшая эфиопская редакция, представляющая собой прямой перевод с греческого языка. Это показывает близость славянской версии к редакциям греческого оригинала, имевшим распространение прежде арабских завоеваний VII века.
Также представлен количественный метод сравнения правдоподобия двух конкурирующих гипотез, касающихся текстуального потока сильно контаминированных традиций. Метод имеет следующие предварительные условия и ограничения: (1) общее количество сопоставляемых гипотез должно быть заранее сведено к двум: специфический источник контаминаций текстуального потока либо имел место, либо нет; (2) в предложенной (простейшей) модификации метода налагается дополнительное условие: предполагаемый источник контаминации должен был обладать высоким уровнем (выше 0,5) отношения сигнал/шум.
Ключевые слова: Славянский Новый Завет, Павловы послания, восточные версии Нового Завета, сирийский Новый Завет, эфиопский Новый Завет, байесовы апостериорные правдоподобности, индуктивная логика.
Сведения об авторе: Вадим Миронович Лурье, доктор филологических наук, ведущий научный сотрудник Научно-учебной лаборатории междисциплинарных эмпирических исследований, Сектор исторических исследований Национального исследовательского университета «Высшая школа экономики» в Перми — филиал Национального исследовательского университета «Высшая школа экономики».
Контактная информация: 614070, Россия, Пермский край, г. Пермь, ул. Студенческая, 38, Национальный исследовательский университет «Высшая школа экономики» в Перми — филиал Национального исследовательского университета «Высшая школа экономики»; тел.: +7 (342) 205-52-50, e-mail: [email protected].
1. Introduction1
The knotty problem of the origin of the earliest Slavonic recensions of the Apostolic epistles (Бобрик 2013: 209—273) will not be discussed below in extenso. Instead, I will provide a series of facts so far overlooked and propose a quantitative way of their evaluation.
In 1879, Grigorij Aleksandrovich Voskresenskij (1849—1918) published his monograph dedicated to the history of the Slavonic Apostolos (Воскресенский 1879). His study has been limited to five epistles: Romans, 1 and 2 Corinthians, Galatians, and Ephesians—those preserved in the Interpreted Apostolos of 1220, his basic manuscript. Then, in 1892—1908, he published the whole material he studied according to 51 Russian manuscripts of the twentieth—sixteenth
3
centuries (Воскресенский 1892; Воскресенский 1906; Воскресенский 1908).
Voskresenskij's work has been recently continued by Iskra Hristova-Shomova4. She collated the variant readings of 27 Bulgarian, Serbian, and Russian manuscripts starting from the unique Old
1 The author is very grateful to the colleagues from different fields who have helped him at various stages of research: Alyona Chepel, Irina Gritsevskaya, Iskra Hristova-Shomova, Elena Ludilova, Dmitry A. Morozov, Florent Mouchard, Alexey Ostrovsky, Yana Pen'kova, Alexey Sapkov, Nicholai Seleznyov, Alexander Simonov, Tedros Abraha, Alexander Treiger, Vevian Zaki.
2 With additions by Tatiana Pentkovskaya (Пентковская 2015: 420—421). Cf. (Гауптова 2013) (Russian tr. from Czech of a 1971 paper, with additions by E. Blahova), and (Алексеев 2013). Independently from Zoe Hauptova (her just mentioned 1971 paper) and on another ground, the very idea that the earliest Slavonic recensions of the Apostolos go back not to a single Greek text but different Greek recensions was formulated by Olga Nedeljkovic (following an unpublished thesis by F. Pechuska, 1933): (Nedeljkovic 1972).
3 Thereafter (Воскресенский 1892; Воскресенский 1906; Воскресенский 1908) often quoted without specific references.
4 Cf. (Христова-Шомова 2004). Volume II (2012) deals with the liturgical calendars and the synaxaria. I am extremely grateful to Iskra Hristova-Shomova for having sent me these two volumes as a gift.
Bulgarian 11th-century manuscript of the Apostolos of Enina5 (preserved in a rather poor condition). The South Slavic manuscripts were not taken into account by Voskresenskij. Moreover, Hristova-Shomova did not limit herself to the epistles from Romans to Ephesians but studied the Apostolos as a whole. However, the list of variant readings provided by Hristova-Shomova is far from being exhausting and by no means replaces Voskresenskij's critical edition for the five epistles. It is especially useful as an addition to the apparatus of Voskresenskij.
In his analysis of the variant readings, Voskresenskij specified a number of those that affect the meaning but are apparently unexplainable with referring to the ordinary phenomena for the Slavonic translations from Greek (differences in translation techniques and errors of Slavic translators, editors, and scribes; variant readings in the known Greek manuscripts). However, I noticed that a large part of such difficult variants—that I will define below (section 3.3) as a specific kind of contamination of the manuscript tradition called "perturbation"—is attested to in Oriental, especially Syriac readings unknown in Greek.
The present study has a very precise purpose without aiming at an exhaustive explanation of all the problems related or purportedly related to this fact. This purpose is the following: after having described the phenomenon, to propose a quantitative evaluation of the hypothesis that these various readings reveal the existence of a so far unnoticed (group of) source(s) of contamination of the earliest Slavonic manuscript tradition.
2. The Readings Looking Syrian
2.1. The Slavonic Variant Readings
Let us begin with a review of the readings that I managed to connect, in one or another way, with some Syrian/Syriac material.
The variant readings of the Slavonic text were first discussed in (Воскресенский 1879) and published (according to a larger number of manuscripts) in (Воскресенский 1892; Воскресенский 1906; Воскресенский 1908); then, some additional variant readings were published by Iskra Hristova-Shomova in 2004 (Христова-Шомова 2004). Voskresenskij discerned four recensions of the Slavonic translation. These recensions (in the terms of statistics, clusters of manuscripts) are so sharply distinct that the fact of their distinctiveness is observable even without any specific quantitative methods6. Voskresenskij's classification has been confirmed in later studies with a unique exception: the Apostolos with commentaries (Tolkovyj Apostol "Interpreted Apostolos") is
7
now considered as a separate recension distinct from Voskresenskij's recensions I and II .
Historically, multiple collations with the (different recensions of the) Greek text contributed to divergence of the Slavonic recensions.
5 Cf. (Мирчев, Ходов 1983). All other preserved Bulgarian manuscripts of the Apostolos belong to the Middle Bulgarian period or later.
6 I mean the quantitative methods based on the cluster analysis applied to the Slavonic Apostolos by Ralph M. Cleminson (Клеминсон 2013: 31—61).
7 I retain Voskresenskij's designations for these recensions thus avoiding the modern terms implying their historical interpretation ("Ancient" for I, "Preslav" for II, "Athonite" for IV, and "Chudov"—especially misleading if the corresponding recension is Southern Slavic, as Iskra Hristova-Shomova believe,—for III). However, my using of Voskresenskij's ordinal numbers is unconnected to any presumption concerning the relative chronology of the corresponding recensions. The separation of the Tolkovyj Apostol to a specific recension has no practical value for my study and, therefore, will be ignored. It will turn out that the features we are studying are specific to the recensions I, II, and that of Tolkovyj Apostol, whereas drastically reduced in recension IV, and barely perceptible in recension III.
Вып. 8. 2016
For the Greek variant readings, Voskresenskij consulted systematically the editions by Mill, Scholz, and Tischendorf8 (Millius, Kuserus 1710; Scholz 1836; Tischendorf 1872), who took into account many readings of later Greek manuscripts, Oriental versions, and Greek and Latin Fathers, which were not repeated in the apparatus by later editors who were aiming at reconstructing the earliest Greek text. Generally, Voskresenskij's knowledge of the variability of the Greek text exceeded that of the readers of modern standard critical editions. Of course, the total number of Greek variant readings is even greater, and its real extension could be figured out from preliminary works for the future Editio Critica Maior9.
At the initial stage of the present study, I will simply enumerate the variant readings that look somewhat "Oriental" and, especially, Syriac. Then, we will start to think how to interpret them. As a prerequisite for a review of the relevant readings in the Oriental versions, we need to have at hand a general map of the routes of the Pauline epistles throughout the Christian Orient.
2.2. The Pauline Epistles in the Oriental Versions
Some Oriental versions, especially Syriac, Armenian, Ethiopic, and Coptic, go back to Greek texts available in the Late Antiquity. No wonder, they preserve some variant readings that are not necessarily extremely ancient but, at least, already unavailable in the Byzantine Church after the seventh-century Arab conquest of the most of the former Christian Empire.
For some reasons discussed elsewhere ((Lourie (forthcoming a), Lourie (forthcoming b)), I consider Syrian literary traditions especially important for the earliest Slavonic writing including the New Testament translations. This is why my predominant attention will be focused on the Syrian data (available either directly in Syriac or in Arabic translations from Syriac).
Moreover, I will take into account all other Oriental versions to the extent in which they are published10. This is necessary for any study of the possible non-Byzantine background of the Slavonic text. On the contrary, I will not pay any specific attention to the Latin and Gothic variant readings, because, according to my own impression, Voskresenskij's observations show that they have no specific importance for understanding the Slavonic version.
In the two next sections, I will sketch, as briefly as possible, the data on the Oriental versions of the Pauline epistles, which are to be taken into account in our study as possible witnesses to the Greek variant readings that were lost or marginalised in the ninth-tenth-century Byzantium but are present in the Slavonic.
My introduction will be written in a Syrian-centred coordinate grid. Thus, all the versions will be classified into three categories: Syrian (Syriac and directly translated from Syriac), Syrian-influenced (translated from Greek but in milieux with Syrian spiritual leadership), and non-Syrian but possibly affecting some Syrian Christian tradition(s).
8 Voskresenskij has used as well, whereas less systematically, other scholarly editions accessible to his time. Hristova-Shomova only seldom provides variant readings, normally referring to the text of Nestle—Aland, which is of not too great value for the mediaeval period of the history of the NT text.
9 So far, only the Epistle of James is published in the Editio Critica Maior. For the epistles we are interested in, see: (Aland 1995; Aland et al. 1991).
10 For a (somewhat outdated) introduction to the Oriental versions of NT (van Esbroeck 1998: 399—509).
2.2.1. The Pauline Epistles in Syriac
The greatest part of the surviving Syriac translations of the Pauline epistles is subdivided into two groups: the earliest translation within the Syriac Bible Peshitta (P) and the 616 CE Harqleian recension (H) created as a literal translation from Greek. The Harqleian recension has a long history of its own, and this is why we have sometimes to discern between its different manuscripts (designed from H1 to H4).
Moreover, some material is preserved as quotations in early Syrian authors and translations from Greek into Syriac, and, finally, some important material is preserved as scholia in some manuscripts of H.
Among these authors, Philoxenus of Mabbog has a specific importance, because he provided, in 508, a new translation of NT, which was later taken into account in H (Brock 1981). This translation is available in short quotations only, but it will turn out to be of some importance for our material (s. below, discussion of Rom 11:16).
All these materials are published in parallel by Barbara Aland and Andreas Juckel in the appropriate volumes of their critical edition of NT in Syriac11 (Aland, Juckel 1991; Aland, Juckel 1995).
2.2.2. Direct Translations from Syriac: Sogdian and Arabic
Some parts of the Syriac text tradition are now observable indirectly: first of all, through the medieval direct translations from Syriac. The relevant material is preserved only in two languages, Arabic and Sogdian (almost nothing in two other Christian traditions that have translated from Syriac, Uighur and Chinese).
The Sogdian manuscripts of the Pauline epistles are all found in Turfan, China, in 1902—1914. The bilingual (Syriac-Sogdian) collection of the Pauline epistles (Turfan manuscript C23) is still unpublished. The published lectionary C5 contains some short fragments from the Pauline epistles. They were at first published in 1910 by F. W. K. Müller and, then, republished using some new manuscript fragments by Werner Sundermann (Sundermann 1974; Sundermann 1975; Sundermann 1981).
The whole published Sogdian material of the Pauline epistles is, however, limited to several short fragments. The main source is still unpublished (manuscript C23).
The early Arabic translations of the NT are now at the initial stages of their study. As to the translations from Syriac, two early translations are published, each in a unique manuscript.
A commented (although very briefly) translation made by the Melkite bishop (most probably,
of Damascus) Bishr ibn as-Sirri in 867 (thus according to the colophon) in the manuscript Sinai
12
Arabic 151 (ArSySin), where the Pauline epistles are preserved in full (Staal 1983), and an anonymous translation preserved in a unique 892 CE St. Petersburg manuscript (ArSySpb) only partially and with great lacunae. The date of the latter translation is unknown but, presumably, the
13
early ninth century as the very early (Stenij 1901).
11 Quoted below without page indication.
12 On this manuscript, s. (Griffith 2013: 133—135).
13 No translation provided.
2.2.3. Translations from Greek within Syrian-influenced Milieux: Armenian, Caucasian Albanian, Georgian, and Nubian
These translations are, at least, one or more than one Arabic, the Armenian, one of the two Georgian, and, albeit indirectly, the Caucasian Albanian (through the Armenian). It is still difficult to judge about the Nubian.
An Arabic early (not earlier than the late sixth century, but, most probably, ninth century14) translation contained in the ninth-century manuscript Vatican Arabic 1315, was made from Greek "but not only from Greek", that is, keeping an eye on some Syriac version16.
I do not know whether this conclusion is applicable to the unique published (by Margaret
17
Dunlop Gibson) early Arabic translation from Greek (Gibson 1894) (ArGrSin), but, anyway, I take the latter into account, because its underlying Greek text was certainly acquired by the translator outside the borders of the Byzantine Empire of his time (evidently, in Palestine or Sinai).
The early history of the biblical translations into the languages of the Caucasian/Armenian Churches is now recoverable in a very tentative way.
The Armenian version of the Pauline epistles (Arm) does not have so far a critical edition. The
18
1805 Zohrab Bible remains our main reference . Nevertheless, some liturgical readings from Paul were found, in 1994, in the Caucasian Albanian translation. The Albanian version is still insufficiently studied. According to its first and still the only investigator Jost Gippert, it represents the Armenian text tradition at its earlier stage, where it is sometimes closer to the Syriac and the Georgian19. Unfortunately, the fragments of the partially preserved lectionary containing Pauline epistles are very short.
The Armenian translation of NT goes back to the fifth century, when the Armenian Church was theologically and culturally depending on the Syrian Church of the Iranian Empire—probably, in a greater extent than on the Greek-speaking Church of the Roman Empire.
In Georgian, there are four recensions of the Pauline epistles going back to two different
20
translations from Greek (Childers 2013: 306—307). Thus, the recensions A and D are substantially different, whereas the recensions B and C are somewhere in between (B is close to A, and C is close to D). Normally, the CD text (GeoCD) represents the Byzantine Greek (and, thus, provides nothing new for our purpose; s., however, an exception at 2Cor 1:7), whereas the AB text
14 Cf. criticisms by Sidney Griffiths 2013: 116, of (Kashouh 2011: 169) (who believes that the translation is pre-Islamic and made in Nagran).
15 A digital copy is available on-line on the site of the Vatican Library (www.vatlib.it: 1). I have checked the readings of this manuscript unsystematically.
16 As it was first demonstrated for the Gospels translation by Juan Pedro Monferrer-Sala (Monferrer-Sala 2013). The same author has recently generalised his conclusion on the Pauline epistles: Monferrer-Sala 2015. For the Pauline letters, see also an on-going study by Sara Schulthess, whose first results were presented in the paper "An Arabic Manuscript of Pauline Letters: Vaticanus Arabicus 13" at a conference in Leuven, 24 April 2015 (a PowerPoint presentation is available on-line: s. (wp.unil.ch: 1).
17 No translation provided. Another part of the same manuscript has been published later: (Krenkow 1926), but with no fragments of the five epistles we are interested in. On this dispersed manuscript, s. (Gehin 2006: 38—40).
18 I will quote the Zohrab Bible according to the electronic edition by (Gippert et al. 2008) at the Armazi Project (TITUS Texts: Armenian New Testament) (itus.uni-frankfurt.de: 1).
19 For the photos of the bottom (Albanian) layer of the palimpsest and its editio maior accompanied with Syriac, Armenian, and two Georgian translations of the relevant fragments, s. vol. 2 of the edition: (Gippert et al. 2008). As the most up-to-dated short review of these Albanian materials could be useful (Gippert, Schulze 2007).
20 The problem of the original language of the earliest (4th or 5th cent.) Georgian version of the Pauline epistles is still not resolved definitively; Syriac and Armenian were also proposed. Anyway, recensions AB show many agreements with Syriac P against GeoCD and the Byzantine Greek text.
21
(GeoAB) is often in agreement with the Syriac against the Byzantine Greek . The exact origin of both translations is unknown.
22
In Nubian, the Pauline epistles are preserved in small fragments of a lectionary (Browne 1994). I mention the Nubian in this section, because two of the three Nubian sixth-century kingdoms were converted by the mission of the bishop of Nobadia Longinus, who was not a Syrian himself but was a leader within one of the Syrian Churches. One can say that, in the sixth century, the whole anti-Chalcedonian Church life in Egypt was leaded by competing groups depending on one or another clan within the local Syrian diaspora (Lourie (forthcoming b)).
The translation is made from Greek and certainly not from Coptic. Since 2009, the Palaeoslavists should never forget that the texts from Nubia, in whatever language, could be extremely important for our understanding of the earliest literature in Slavonic. I mean, of course, the short recension of the "Slavonic" 2 Enoch in Coptic, which was identified among the manuscripts found in Nubia (Lourie (forthcoming b)). Thus, I had have to take into account the fragments of the Pauline epistles in Nubian (but, alas, without finding out in them anything important to our purpose).
Finally, the Syriac versions, after having been translated into Arabic, contributed to the creation (in the fourteenth century or somewhat earlier) of the second Ethiopic recension of the NT (s. the next section).
2.2.4. Translations from Greek into Coptic and Ethiopic
From the Coptic translations, only the Sahidic (Sah) and the Bohairic (Boh) are preserved. Both are available in the critical editions by George William Horner (Horner 1905; Horner 1920a;
23
Horner 1920b) . The edition of the earliest Sahidic manuscript of 2 Cor, Papyrus Bodmer XLII is
24
still in preparation . The Sahidic represents a very early recension of the Greek text. The available Bohairic represents a post-Arab recension, even though it goes back to an old translation, too.
The Coptic texts are independent from the Syriac recensions and the Syrian Church traditions. However, on the contrary, the Syrian Christian tradition we are interested in because of its possible influence on the Slavic world was firmly established in Egypt and, in particular, in Alexandria (the place of the revelation to Cyril, the principal character of the Legend of Thessalonica). Thus, the Coptic parallels to some phenomena in the Slavonic texts are to be expected. We have already known a huge piece of evidence, the Coptic version of the "Slavonic" 2 Enoch.
The Coptic tradition is also partially preserved in Arabic translations from Coptic, but these translations of the Pauline epistles are not studied in any details (Kashouh 2011: 258—274).
25
The Ethiopic version of the New Testament is available in three different recensions and many mixed eclectic texts (including the missionary 1830 edition by Thomas Pell Platt (Platt
21 The critical edition of the four recensions: (Dzotsenidze, Danelia 1974).
22 This edition encompasses all the biblical fragments known in Nubian.
23 The Sahidic text contains some lacunae. Horner's apparatus to his Sahidic edition provides (for the Oriental languages, in translation) the parallels from several versions (normally the variants from Greek manuscripts, the Bohairic, the Latin version according to different early manuscripts, the Armenian according to the Zohrab Bible, and the Ethiopic according to the Roman editio princeps and the eclectic edition by Platt; s. (Platt 1830).
24 Sahidic on parchment, according to Wolf-Peter Funk; the edition is in preparation by Rodolphe Kasser. See (Robinson 2013: 183, 190).
25 As a general introduction, s. (Weninger 2003). For more details, s. (Zuurmond 2003; Zuurmond, Niccum 2013).
Вып. 8. 2016
1830)26). Fortunately, the 1548 editio princeps published in Rome27 by the Ethiopian monk and scholar "Petrus Aethiops" (Tasfa Sayon, together with his two fellow-monks from the famous Ethiopian Laura Dabra Libanos) was based on three manuscripts of the earliest recension thus
providing a relatively pure text of the earliest Ethiopian version, even though without meeting the
28
requirements of modern scholarship . Recently, some Pauline epistles appeared in critical editions,
29
including four among the five that are in the focus of our attention.
The earliest Ethiopic version (EthGr) is a direct translation from Greek appeared in the Aksumite kingdom, between the fourth and the sixth centuries.
The second Ethiopic recension (EthAr), as it was mentioned above, is heavily influenced with the Arabic translations from Syriac (whereas perhaps also other Arabic versions). The third Ethiopic recension was a product of intensive scholarship in the sixteenth century but without an independent access to the Greek. There is no normally need, in our study, to distinguish between the second and the third Ethiopic recensions, given that, in their peculiar readings, the two represent some—mostly unknown to us—Arabic recensions. Nevertheless, the readings proper to the third recension will be specified when necessary.
2.3. The Slavonic Readings Looking "Oriental": a List
The peculiar readings that could be explained with some reference to the Syrian/Syriac or other Oriental data are presented in Table 1 (for the abbreviations of the versions, see previous section).
The table does not contain unique Slavonic readings (known from a unique manuscript only). See a discussion of the relevance of the unique readings below (section 3.7).
With the grey fill colour are marked the rows where the Slavonic variant reading is likely to be explained from the Syriac text itself rather than the Greek text underlying the Syriac or other Oriental translation.
For the distribution of the coinciding variant readings among the versions, s. Table 4 (legend: + full coincidence; *+ full coincidence recoverable; ± approximate coincidence; — another reading or lacuna/loss of pages). One can see, from this table, that, after the Syriac versions, the ancient Ethiopic one (EthGr) provides the greatest number of parallels. This version represents a pre-seventh-century Greek text.
For the commentaries, see the next section.
26 No translation provided. Cf. an evaluation by Zuurmond, Niccum, 2013: 231, n. 1: "Platt's edition is even more useless [than the Roman 1548 edition] for text-critical purposes, as it represents a thoroughly eclectic text, with many later elements of Arabic influence".
27 The most easily available edition (where a Latin translation is added, whereas not always very accurate) is Brian Walton's polyglot Bible: (Walton 1657). I will quote it where no critical edition is available (for Gal).
28 According to Rochus Zuurmond's review of Tedros Abraha's critical edition of the Epistle to the Romans, the text "shows remarkably few differences from the Roman [1548] edition": (Zuurmond 2003: 254).
29 Tedros Abraha 2001. No translation for the Ethiopic (Gecez) text, but published is as well an Amharic commented translation (andamta) together with an Italian translation of the latter. Theoretically, the Amharic commented text could reveal some different textual and exegetical tradition than the Gecez one, but, in our case, I have found nothing specific in the Amharic. The Epistle to the Ephesians is published within Uhlig, Maehlum 1993. Finally, the two epistles to the Corinthians were recently published privately by Tedros Abraha 2014; the plural "versions" in the title of this book means that there is a different translation of these two epistles (clearly depending on Syriac Peshitta readings) that is published separately according to an unique manuscript (x = Comboniani S8). I am extremely grateful to Fr Tedros Abraha who generously sent me a copy of this book.
Table 1.
Nr Place Воскресенский 1879 30 page Rec. (mss31) Peculiar reading Normal reading(s) Greek Syriac32 or Other Oriental
1 Rom 207 I and IV оуже не оудолЪгеть к томоу не oÜKexi
6:9 (cf., in II, E10 He oygoB^aeTb) обладает Kupieuei read as *
2 Rom 6:19 225 I [2 mss] в истина "in the truth" в стыню äyiao^ov or H ^^OWD P read as
3 Rom 6:22 225 I [4 mss] в истина "in the truth" в стыню äyiao^ov or H P read as
4 Rom 11:16 225 I [9 mss against 10] и присъпъ стъ [add] и присъпъ [no addition] Kai то фираца Coptic, Ethiopic
5 Rom 11:16 225 I [3 mss] вЪтьвик Сто [add] вЪтьвик [no addition] Kai oi K^aSoi Philoxenus
6 Rom 12:6 210 I and II по числу вере противу / попричту KaTa t^v ävaXoyiav
7 Rom 225 I [5 mss] кльнжщА "those who гонАщая TOU^ Arm
12:14 course" "those who persecute" SiraKovTa^ EthGr
8 Rom 14:7 207 I никыи же нас [add] себЪ оумирагет [no addition] oüSeig ¿auTÖ änoöv^o-Kei Coptic
9 Rom ХШ 204 I [mostly братие моя [add] [no äSe^oi
15:15 South Slavic mss], IV "my brothers" addition] (many mss); cf. parallelism with äSe^oi цои in 15:14.
10 1Cor7: 261 I [only 2 не бо стыдитьсА А2 / не oü SeSou- Arabic
15 mss] не стыдит бо се А27. работить бо са XraTai
11 1Cor7: 37 261 I [save 4 mss] не имыи бЪды нужи ЦП exrav avayK^v Arm, Geo
12 1Cor 11:30 258 I and II и спять мнози доволни iKavoi
13 1Cor 12:29 261 I еда вси силы деють [add] [no addition] Suva^ei^ ^ P
30 And Hristova-Shomova's page when necessarily (marked Xffl).
31 Indicated for minoritary readings of a given recension.
32 The common readings of P and H, unless otherwise specified.
Вып. 8. 2016
Table 1 (Continuation).
Nr Place Воскресенский 1879 Rec. (mss) Peculiar reading Normal reading(s) Greek Syriac or Other Oriental
14 1Cor 15:29 261 I аще оубо мьртвии не въстають аще бо отниудь [om in I] мьртвии не въстають ei öXrag veKpoi OÜK ¿yeipovxai P
15 2Cor 1:7 302 A634 and II [Tolst; cf. ecMt E1, ecMb E2] причастницы есмы страстем есте ¿oxe Oriental parallels
16 2Cor 2:12 293 I, II и двьри ми отвьрзшисА велицЪи [add] о г(оспод)Ъ [no addition] Kai 0upag ЦС1 äveray-^evn? ¿v Kupira Georgian
17 2Cor 2:14 295 I, II, and IV благодать являющемоу победител и нас творящем оУ 0pia^-ßeuovxi P
18 2Cor 5:8 303 II оуповающе дьрзагем öappoß-^ev / öappouv-xeg
19 2Cor 5:14 303 II аще [add] бо любы б(ож)ия сдержить нас любы бо б(ож)ия сдержить нас [no addition] "Л yap äyann xou Хршхои ouvexei Ethiopic
20 2Cor 5:18 296 II [save 5 mss] смирившаго себе смиривша го нас [om] себе xou Kax-aAla^av-тод ¿auxö Ethiopic
21 2Cor 6:7 30535 II [7 mss] в словеси праведнЪ в словеси истиньнЬ / рЪсноти-внЪ ¿v Xoyra äXnöeiag P
22 2Cor 10:10 293 ХШ 262 I (with traces in II and III) яко кпистолик речете соуть тежкы [речете in A17 and A19 and 8 mss in ХШ] яко оубо послаша рече. тяжка öxi ai ¿nioxoXai [9aoiv = P, H], ßapeiai Armen.
23 Gal 1:22 317 I (main reading36) бях же незнаем отиноудь церквам июдеискам лицем хф прооюпю P read as^ta^b
33 This particle (= yap) is added in one ms of P (P17) and, by another hand, in a manuscript of Aphrahat.
34 Karpinsky Apostle, recension I, but "[н^рЪдко представляетъ чгешя 2-й редакцш [often witnesses to readings of the 2nd recension]" (Воскресенский 1908: III).
35 Only the variation of Slavic synonymic renderings of x^^ alnQeio? is discussed.
Table 1 (Continuation).
Nr Place Воскресенский 1879 Rec. (mss) Peculiar reading Normal reading(s) Greek Syriac or Other Oriental
24 Eph 4:13 329 I and II (телесноую) в мероу тела исплъне-нию хвоу възраста ^XiKia^ mg H1.4
25 Eph 4:29 330 I всяко слово злое из оуст ваших да не исходить гнило oanpo^
26 Eph 5:1 330 I, III, IV and Pandecta of Antioch (11th cent.) бываите оубо подобни бгу подражате лие ^i^ntai P H
27 Eph 5:18 333 I [4 mss] д(оу)хом стым [add] [no addition] ¿V nveu^axi Ethiopic
28 Eph 6:7 331 I, III, IV с любовию слоужащ е с приязнью цех'еи-voia^ P
29 Eph 6:22 327 I кже о мънЬ еже о нас / вас xa nepi ■^öv P
2.4. The Slavonic Readings Looking "Oriental": a Discussion
Nr 1 (Rom 6:9). "To prevail" instead of "to possess" (and synonyms); the latter is the only reading of Greek, Syriac, and all other known to me versions (although some of them, as, e.g., certainly the Ethiopic, are not specific enough to provide a clear distinction between "to prevail" and "to possess"). However, the exactly corresponding reading could be provided by the Syriac, if only the ground stem of the same verb is read instead of the reflexive causative one that is present
37
in P, H, and various Syriac authors .
Nrs 2 and 3 (Rom 6:19, 22). Possibly, not a misreading but a variant reading in Syriac, where supposed reading would mean both "righteousness" and "truth". This reading is confirmed
38
with the Georgian AB, where the spectrum of meanings of (Rom 6:22) is the same
as that of the supposed Syriac word. Cf., in GeoCD, "holiness". Lacuna in ArSyrSpb.
All other versions have "holiness / sanctification". However, at 6:19, even GeoAB has "holiness".
Nrs 4 and 5 (Rom 11:16). The addition of the second "saint" in the each phrase of the sentence: si Ss ^ arcapxn ayia, Kai to ^upapa Kai si ^ pi^a ayia, Kai oi K^aSoi. The three Slavonic manuscripts containing the addition at the second part of the verse are among the nine manuscripts containing it at the first part (Xffl 195 adds the tenth such manuscript, the Bulgarian Slepchensky Apostolos, 12th cent.).
36 Including the main ms of rec. I, the Ochrid Apostle (12th cent.). From 39 mss used in Voskr1908 for the rec. I, three are not preserved for this place and 15 contain the "normal" reading. Cf. (Воскресенский 1879: 318): "Охридскш же Апостолъ имЪетъ несомненную связь съ глаголическими памятниками" ("The Okhrid Apostle has a certain connexion to the Glagolitic manuscripts"); this means that it witnesses to a very ancient Slavonic tradition.
37 Cf. (Payne Smith 1903: 579): meaning c) "to take possession, take, occupy", but ^ "to bear rule, bear sway, have the mastery, prevail".
38 I am very grateful to Alexey Ostrovsky for pointing me out this reading.
Among the Oriental versions, we have the following parallels to the Slavonic with additions (Table 2):
Table 2.
Greek Slavonic with additions Syriac (Philoxenus) Coptic Sahidic Coptic Bohairic Ethiopic (EthGr)39
si 5s 'л anapxq ayia, Kai то фираца аще бо начатък стъ. и присъпъ стъ ефле пее^в г^р оу^в. eie пкеоушфм оу^в юле ле ^п^рхн оу^в. ie цоу^в оы ылеп1кеоушфен ЛП : hl^ht : hl-t : : : £ht : Ф-ttfahl : :
Kai si ' piZa ayia, Kai oi K^a5oi и аще и корень стъ. то и вЪтьвиге сто • От г^т.лп о oyog 1сле ^ыоуын оу^в. ie соу^в оы ылешкел^Х : : (DhöfrtVl :
Symptomatic is the absence of the parallels in the Arabic translations from Greek, the Georgian, and the Armenian (not to say of the Syriac and Arabic from Syriac40). The Syriac is attested to only by Philoxenus of Mabbog41. The parallelism between Philoxenus and Coptic and Ethiopic recensions could be not accidental, because Philoxenus' metropolia of Mabbog governed the Syrian missions to the South Arabia (Nagran), the place of the further interference between the
42
Syrian and Ethiopian Christians .
Nr 6 (Rom 12:6). "According to the number of faith" instead of "according to the proportion of faith". Cf., e.g., in the Ethiopic translation from Greek, an idea of counting but still not "number": R^R43 : : "according to counting of the faith" in the early translation, but
RA'P^'z : : "in the measure of faith" in later recensions. Similar readings are preserved
among the variants of GeorgianAB44: [another variant: bib^i^^i^ "sign / miracle"]
8ob "according to the measure/amount of his faith" (with the reading
bi%«8obi in the main text and bi%«8obig&^ in GeoCD, both having the meaning "proportionally"). The same in the Armenian: pus ^mqnjh hmimsng "according to the measure of faith". The same in the Coptic Sahidic: kat^ n^i NTnicTic "according to the measure of the faith", but not in the Bohairic: k^t^ nieoHTeH [ variant oyeoHTeH] hts "according to the/a
likeliness of faith" (clearly, an attempt of rendering Kara x^v avaXoyiav).
The two Syriac translations have the same phrase ^^mms whose main word can be retranslated into Greek in both ways, as avaXoyia as well as apiB^og (Payne Smith 1879—1901: col. 2237). Thus, the Slavonic peculiar reading, if it has something to do with the Syrian material, is to be
39 The words ■ : are accidentally omitted in the main manuscript of the critical edition by Tedros Abraha. The same construction in the later Arabic-influenced Ethiopic recension (EthAr), whereas the wording is slightly different; cf. (Tedros 2001: 108).
40 For this, only the translation of ArSyrSin is available; ArSyrSpb has a lacuna.
41 In his epistle to the Arab (Lakhmide) Christian chieftain Abu Ya'fur, early 6th cent.; s., on this letter, (de Halleux 1963: 203—208). Published by (Harb 1967), for Rom 11:16 see (Harb 1967: 208).
42 On the Philoxenian legacy within the Syrian tradition behind the earliest Slavic writing, s. (Louire (forthcoming
c)).
43 The main manuscripts has n^rtn, but I quote with the etymologically correct spelling.
44 I am grateful to Alexey Ostrovsky for having attired my attention to these readings.
explained as a possible correct alternative rendering of the Syriac word and not as a rendering of a different Greek variant reading.
Nr 7 (Rom 12:14). The Armenian and the ancient Ethiopic (EthGr) versions coincide with the peculiar Slavonic reading: ophhtgtp qmU^&^u Atp, ophhtgtp tL mU^&mhtp: / :
ЛХЛ : : : : "Bless those who curse you (pl.); bless and not curse",
but in EthAr different renderings of xou^ SiroKovrn^ are used: : "those who persecute you
(pl.)" and P^^h^ : : "those who torture you (pl.) and persecute you (pl.)".
Nr 8 (Rom 14:7). Repeated "of us" in "For none of us lives to himself, and no one of us dies to himself'. This reading is the normative one in the Coptic versions, both Sahidic (^ую mmn mmon ы^моу n^q) and Bohairic (оуш£ mmon g\i mmon ы^моу n^q), although is absent in the others known to me (including the Syriac, Arabic, and Ethiopic ones).
I have elaborated elsewhere on the possible relevance of such exclusive agreements with the Coptic in the earliest Slavonic translations46. There is, at least, one such reading in the Slavonic Gospel translation, and an even greater example is the short recension of 2 Enoch as a whole.
Nr 9 (Rom 15:15). The reading was known to Voskresenskij in recension IV only (the Gennadius Bible, 1499, retained in the printed Ostrog Bible, 1581), but is pointed out by Hristova-Shomova in ten manuscripts (mostly South Slavic but including the Russian Christianopolis Apostolos of the 12th cent., which is the earliest among them). This reading is specific to the Syriac versions, whereas the other Oriental versions corresponds to either aSs^oi or omission of this word.
Nr 10 (1Cor 7:15). No additional manuscripts with this reading in Hristova-Shomova. Instead of "the brother or the sister is not under servitude in such cases", two Slavonic manuscripts have "not are ashamed". The most of the Oriental versions follows the Greek text (lacuna in the Sahidic), including Vatican Arabic 13: "there is no power (Ц-^1) in such things on the brother or sister"47. However, other Arabic versions have explicative translations (but lacuna in ArSyrSpb). ArSyrSin: Vj jj^ ( "is not compelled and not enslaved", and—the most interesting to us—ArGrSin: l^i.1 jl ^ Uji.1 (j "and not our brother is to be enslaved or contempted". The last word could be translated also as "to be low, ignoble" etc., which is evidently very close to the Slavonic "ashamed".
It is hardly probable that the explicative translation of ArGrSin follows some Greek text where some word has been added to ou SsSouXrorni. Both ArSyrSin and ArGrSin demonstrate the same manner of an explicative translation. There is no reason to think that they followed some Syriac translation unknown to us, even though such a possibility could not be excluded logically. The most plausible explanation of this similarity is probably the common milieu of the two translations, Arabic-speaking Melkites, where the translators would have tried to anticipate the same understanding problems of their common audience. The Slavonic "peculiar" reading goes back to such an explicative manner of translation attested to in some Arabic Melkite translations only.
Nr 11 (1Cor 7:37). "Peril, disaster" instead of "necessity". The same reading in Armenian "peril, danger") and Georgian (the same word in slightly different spellings: ^^35 AB,
48
■3^350 CD "disaster" ).
45 This spelling in the apparatus of the critical edition (Tedros 2001: 113, n. 109).
46 See: (Lourie (forthcoming c)) and (Lourie (forthcoming b)).
47 I quote this unpublished verse in full (f. 108r, I am very grateful to Dmitry A. Morozov for the following transcription): ^ t^j Jl . JJ-* O-LLi ^l jl ¿VI Jc (jjti. ^¿Jl j* o>j. V ^¿Jl jlS jlj.
48 As Alexey Ostrovsky noticed to me, the meaning "necessity" occurs as well, even though it is secondary, as it is detected by Ilia Abuladze: (Abuladze 1973: 429), s.v. ^35, among other meanings: fyibigofto "(state of) being in need" (cf. gofto "necessity").
Nr 12 (1Cor 11:30). Instead of "and quite a number (ixavoi) [of you] have died" the peculiar Slavonic reading has "and many have died". The Syriac translations do never have kavo^ in
Syriac49) but always "many" (^^i^^): P ^^ii^^o, H ^^^ ^.^ддо [var. The
same in the Vulgate (multi) and the most of the Oriental versions: ArGrSin and ArSyrSin (lacuna in ArSyrSpb), both Georgian (8^535^60), Sah and Boh (respectively, йбчоумннфе and йлеоуннф "a great multitude"), EthGr and EthAr :), with, however, an exception of Arm
that follows the Greek strictly: ^u^. It is not clear whether all these "many" render ixavoi50, or there was, in Greek, another variant reading noXXoi at this place. Anyway, this is an example of a Slavonic variant reading where the coincidence with the Syriac (and other Oriental versions) could be accidental with a high probability (much higher than in the most of cases). Cf. another such example in Eph 4:29 (Nr 24 below).
Nr 13 (1Cor 12:29). The reading of P against the reading of H. The same in GeoAB (56^ У^ЗО^бо 6^ 65^ оузбдб)51. This reading is specific to the Syriac and Syrian-influenced Caucasian versions.
Nr 14 (1Cor 15:29). The Greek оХю^ is never attested to in P (whose text, for this verse, is available also from Aphrahat, early 4th cent., and Babai the Great, early 7th cent., both in the Sasanian Iran). In translations (from Greek) of Severus of Antioch (6th cent.) the Greek adverb is correctly rendered with and, in H, with a calque from Greek mUU. Cf., in ArGrSin, rendering of оХю^ with The same omission of оХю^ in the Ethiopic and in the Georgian AB (whereas GeoCD has у«з^5ф = оХю^). The Coptic Sahidic has рш "indeed" instead of оХю^ (but it reappears in the Bohairic: ^оХшс).
Nr 15 (2Cor 1:7). Paul's second person in "as ye are partakers of the sufferings" changed to the first person in the peculiar Slavonic reading ("as we are partakers...").
The textological history of 2Cor 1:6-7 in the Oriental versions is extremely complicated. In Greek, in the second part of the verse 1:7 (oxi ю^ Koivrovoi sots xrov naGnM-dxrov, ouxro^ Kai x^ ларакХ^аею^), the "partakers (Koivrovoi)" are always "you" (pl.) and never "we": the verse has sots and never so^sv. Moreover, the "passions" meant are obviously those of Christ and not those of Paul himself (cf. 1:5: та лаб'лцаха xou Xpioxou). This structure is rendered correctly by both Coptic, the Georgian AB, and, of course, by the Syriac H. The other versions provide a reach spectrum of readings.
The peculiar Slavonic reading stands firmly on so^sv. It has only one exact Oriental parallel, in the Armenian version:
hpt п(р)щ(Ьи) ^gnpq hUf ^mp^mpmUmgU, UnjU^tu hi tf^pmpni(p)h(mU)U:
inasmuch as we are partaker(s) of the passions, so (we will be) and of the consolation.
The Armenian translation testifies two important things. Namely, that there was, in fact, a Greek reading with so^sv, and that this reading was, theoretically, available in some Syrian milieux having an Iranian background.
49. Cf. (Payne Smith 1879: col. 2704).
50 As it is the common opinion, shared, e.g., by the authors of the Greek retroversion of H: (Aland, Juckel 1991: 591; Aland, Juckel 1995).
51 But not in the Albanian, despite the translation "are all workers of miracles?" in (Gippert et al. 2008: VII—27) (no continuous pagination in the book). The Albanian reads meciqay cexal mil'anunux nahalar. As Alexey Ostrovsky translated for me word by word, meciqay = really, cex + al = all/every + and/also, mil'anun-ux = power-PL, n-ahal-a-r = NEG-be/PART-3PL, that is, "and really are not all powers?", which is the normative Greek reading.
We have not to wonder that the Syriac P apparently follows the Greek:
"... that as you are partakers of the passions, so you will be partakers of the consolation".
It is interesting that this appearance is somewhat misleading. Normally (8 manuscripts out of 11 for this passage), this text is spelled without the vowel signs. In this case, ^Vo^ could be read as plural "partakers", in accordance with the Greek. However, in the three manuscripts
where this word is vocalised, it must be read, both times, as a plural with the pronominal suffix of 1st pers. plur.: ^sVo*. "our partakers / partakers to us". The version H avoids such an ambiguity using another form of plural: ^Vo*. "partakers".
Among the Arabic versions, the unpublished version in the Vatican Arabic 13 follows the Greek 2nd-person text52, but the vocalised recension of P is preserved in ArSyrSpb: ^ jl ^
I^jI ^i ljjl£j^ ^li ^UjVI "as you are our partakers in the pains, you will be our partakers in the consolation also". The same idea is conveyed by the peculiar Ethiopic version x (known from the unique manuscript Comboniani S8 and sharing, according to the editor, many specific Peshitta's readings): ^h^ : Ш9И : : : h^h^ : X^*^ : fr+^X) ■ П^^* :
ШП^гЧОР* : : h6n : : : - "And, behold, we know, as
you are our partakers in the passions and in the tribulations, you will be then our partakers in joy and in endurance".
In another Arabic translation from Syriac, ArSyrSin, both kinds of first-person speech are attributed to Paul: ^Will L^jI ^il ^li Hi ^lilj jij^VI li»- ^^ 53l^2 us Ail
"...as your partaking extends/communicates to us in the passions, and you are our partakers, so also you are in the grace". Here, both the partakers are "ours", and the passions are with "us". The author of this translation, bishop Bishr ibn as-Sirn, prepared a commented philological translation, and, therefore, he obviously tried to encompass the different readings of Syriac manuscripts available to him but already unavailable to us.
In the Georgian recensions CD, "you" become not partakers to Paul personally (as in the Syriac vocalised P version), but to his passions; the "passions" are no longer those of Christ, but of Paul himself: Ъо5^ Ь5^ш збд&5Ш5 85ш В^дбш5, 3й^30Ш5 Б^^зЗобоЬ-^збоЬб "as you are partakers of our passions, so of consoling". In AB recensions, there is no "our" related to "passions" (85ш without В^збш5).
Both Syriac P (vocalised) and Georgian CD readings are interesting to us by preserving, in some form, the 1st person plural.
Finally, an important witness is preserved in Ethiopic. Already Voskresenskij has noticed, from a second-hand reference, that the Ethiopic witnesses for sa^sv (Воскресенский 1879: 302). The situation is not, however, as simple as that. The Ethiopic version, and only in a part of the
52 I would like to quote here the unpublished reading of the ms Vat. ar. 13, f. 120v (read for me by Nicholai Seleznyov): lj*Jl Jj ^l^j'Vl J ljjj ^Sljc J^l ^ L^l lj*jj jlj ^kJ CiAj j* Ijl^jj
53 The edition have which does not result into any meaningful sense. The manuscript, however (available online at (e-corpus.org: 1), has the three diacritical points above the first two letters of the word without a specific distribution of them among the letters, thus allowing both nq- and tf- readings. The editor failed to provide an exact translation corresponding to his choice of nq- ["...as we have shared (?! — B. L.) your fellowship with us in sorrows, while you were our partners, that thus you are also with us in grace" (Staal 1983: vol. 453, t. 41, 95)]. D. A. Morozov considers the alternative reading to be correct, that is, L^' in the meaning "extends, communicates". Even in this case, however, the text is only grammatically correct and understandable but still far from smoothness. I am very grateful to Dmitry Aleksandrovich Morozov for these ideas and for his consultations.
manuscript tradition, contains a lexically different derivate of the hypothetical *Koivrovoi ¿ajusv: "we became equal": : 0¿,el : A^ : hlfr : 1ЧПС : "and as we became equal in
the passions, so we will be united also in our joy" (the main text of the critical edition by Tedros Abraha = ed. princeps). Another part of the Ethiopic tradition, heavily influenced with later Arabic recensions (= ed. Platt), contains a _you-reading but with two "our" applied to both "passions" and "joy": : 0¿,eh^ : A^l : : Ф^-П^ : "and as you became equal in our
passions, so you will be united also in our joy".
Nr 16 (2Cor 2:12). An addition "the great door" is to be read in Georgian AB (not CD): ф5 35^0 8o фофо 800^ "and the great door is opened to me because of/for the
Lord". No trace of this reading in Armenian, Coptic, known to me Arabic, and Syriac versions. A peculiar reading of the second Ethiopic recension—going back to some (unknown to me) Arabic recension—shares, however, the general idea of magnificence expressed with this addition of the word "great".
In the Ethiopic Versio Antiqua (EthGr) the second part of the verse 2:12 (Kai 0upa^ p,oi ^ефуце-уп^ sv кир!ф) is missed, and the second and third recensions fill this lacuna (as a number of other lacunae of the ancient translation), whereas operating with the word "way" instead of "door". One manuscript of the second recension (F = EMML 2198, 15th cent. acc. to the description but later according to Tedros Abraha) reads : : : : AhCh^h : "and is
revealed/opened to me the way of his kingdom to Christ". Moreover, almost the same reading (only with "Christ" instead of "to Christ", with an omission of the preposition) is present in the first recension's manuscript P = Paris, BnF Eth 46 (AD 1419).
Nr 17 (2Cor 2:14). A somewhat difficult place Тф 5s 9еф xdpi? тф ndvxoxs GpiaMPsuovxi ■Лца^ sv тф Хрютф (KJV: "Now thanks be unto God, which always causeth us to triumph in Christ") followed with the phrase, in the same verse, Kai x^v oo^^v xqq yvroosro^ auxou фоу£ройут1 5i' q^wv sv navxi толф, has been first translated into Slavonic correctly, as it seems, in the Elizabeth Bible (1751). Before this, even in the printed Ostrog Bible (1581), the Slavonic reading was normally "Unto God be grace/gratitude, who makes us seen/manifest in Christ", itself not always syntactically clear (and, thus, allowing understanding "who makes seen/manifest the grace"). The recension III (being a fresh translation from Greek) is an exception: блюдущему нас "who preserves us". It is interesting that even this variant reading has an exact Oriental parallel in the mainstream Ethiopic versions (EthGr = EthAr): HHA^ : РоФП^ : "who [sc., God] always preserves us".
The Syriac P reading "who [sc., God] at all time a spectacle makes of us" (^ Л^=>д)
has a parallel only in the Bohairic: "this who manifests us at all time" (ф&д ееоуоы^ ммоы евоХ йсноу швеы). It is interesting whether this reading of Boh is in accordance with that of the unpublished Sahidic Papyrus Bodmer XLII. A similar text in the Ethiopic x (that is often close to P): : HA£ : nhCft^ft : "who [sc., God] always reveals/makes shown in Christ".
Syriac H is a correct rendering of the Greek: , with the marginal glosses in Greek: H1 Gpia^Psuovxi H4 Gpi^aPsuovxi (sic!). The available editions of the Ethiopic present a text that disagrees with any other version; Arm, GeoAB and CD, ArGrSin, and ArSyrSin follow the Greek at the place.
Nr 18 (2Cor 5:8). The Syriac verb used renders both Gappsiv and nsnoiGsiv54. Other languages, especially Semitic ones, would confuse, more or less, these meanings, too. The latter meaning is chosen in the Slavonic recension II. It is, of course, difficult to judge whether the Slavonic language
54 Cf. (Payne Smith 1879—1901: col. 4433).
of the earliest period was equally apt for discerning between the two Greek words as it became in a later period.
Nr 19 (2Cor 5:14). The Slavonic peculiar reading normative for recension II implies something like si at the beginning of the verse (as it was already pointed out by Voskresenskij). Nothing similar in either Greek or Syriac manuscript traditions (both Syriac P and H having m^n«
= 'л yap aydnn, with the possessive suffix "his"). The Ethiopic x also has Ш^Ф^ : hCh^h : "and the love of Christ.".
The exact parallel to the Slavonic is to be found in the earliest Ethiopian version from Greek (EthGr)—not in all manuscripts thereof but in the best ones: X^A : ^Ф^ : hCh^h : "If [or if really] the love of Christ.".
Other Oriental recensions do not contain the same reading but are somewhat relevant. ArGrSin contains м^ u' ¿^ "Truly that the love of Christ.". The phrase jl ¿^ (literally, "in truth that.") looks too heavy for rendering simply ydp, and, therefore, it corroborates Voskresenskij's guess on the presence of both ydp and si.
Two different early Arabic translations from Syriac have: м^ ^'j "And, indeed,
the love of Christ constrains us." (ArSyrSpb) and м- jV "Because/since the love of
Christ." (ArSyrSin). The same reading in the most of the Ethiopic manuscripts of the first recension and all manuscripts of the second and the third: "Indeed (Xh^ :), the love of Christ.".
The unpublished Arabic version of the manuscript Vatican Arabic 13, f. 124r (made from Greek but Syriac-influenced), has »VI м- u' сЫ о* "Because the love of God brings us together."—with "the love of God" instead of "love of Christ" ("God" instead of "Christ" in accordance with H and some Syriac patristic texts; moreover, the manuscript has here, on the top margin, a subheading in Greek HATAnHTOY0(EO)Y = ' aydnn той 0sou "the love of God"55).
Nr 20 (2Cor 5:18). A reading without 'ца^ is preserved by the ancient Ethiopic version (EthGr): НФ®Лт : nhcftf-ft : "who is recompensed by Christ" (with somewhat different wording but also without "us" in some part of the manuscripts of the second recension), whereas the third recension (EthAr) and x follow the Arabic (and other exact versions of the common Greek text): Hh^ni : : nhcft^ft : "who brought us closer to him in Christ". The presence or absence of 'ца^ / "us" redefines the object of the verb, which is either "us" or God (as in the ancient Ethiopic and Slavonic II). The verse seems to be not preserved in Sahidic Coptic. The Boharic, the two Georgian, the Armenian, all Syriac, and ArSyrSin and ArSyrSpb follow the common Greek text.
Nr 21 (2Cor 6:7). "In the word of righteousness" instead of "in the word of truth"; cf. further in the same verse: 5ia xrov onXrov x^ 5iKaioown^. The peculiar Slavonic reading represented with a part of the manuscripts of recension II leads to repetition, in the two cases, of the adjective "righteous". This is a possible disambiguation of the Syriac ^^taa used in P (cf., however, above, Rom 6:19, 22, where the same Syriac word has been presumably understood as "truth"), whereas H uses another synonym that allows avoiding this ambiguity (^лЫд In P, however, the
adjectives accompanying "word" and "armament" are different: г^а^ди ...^^ъопд Normally, the Oriental versions follow the Greek, but there are two important exclusions56.
Bishr ibn as-Sirn's Arabic translation (ArSyrSin) runs as following: jM ... "in
the word of justice/rightfulness. in the armament of righteousness", whereas other Arabic
55 The same reading is the normative one for the four Slavonic recensions (the manuscripts containing "of Christ" occur rarely). In Greek, it is omitted in the Nestle—Aland edition, but s. the apparatus of H. F. von Soden (von Soden 1913: 735).
56 Moreover, there are Ethiopic readings which have very little to do with any known text. No critical edition of the Ethiopic text exists.
translations render the Greek in an exact way, thus differing in the first phrase of our quotation: сИ "in the word of truth" (both ArSyrSpb and ArGrSin as well as Vat. Ar. 13, f. 124v). The learned bishop was certainly perplexed with such a deviation of the Syriac text he has translated. Therefore, he adds the reading as a variant and provides the following scholion: "That is, by the right ( aJILAj) statement which has no corruption in it, in truth (¿^Ц) he draws near to
everyone. Justice is truth -k^lj)" (Staal's tr. with a correction).This gloss testifies that, for as-Sirri, "justice/rightfulness" and "truth" were clearly distinct words.
As-Sirri's understanding of гс'кьсш coincides with our Slavonic reading exactly. Another interesting parallel is not equally exact but still valuable57. In the Georgian AB, we have: Ьофузсоб ^зЗЭббюфд&оЬбособ... Ьб^^ЗОО^особ Ьобб^соо^оЬбособ "with the word of truth... with the armament of righteousness", but the latter word has a variant (in one manuscript) ^зЗбббюфд&оЬбособ "of truth". I do not know any parallel to such variation in the second phrase of our quotation, but it is possible that it is a trace of some uncertainties with the first phrase, that would have eventually affected the whole sentence.
Nr 22 (2Cor 10:10). A number of manuscript, especially South Slavic ones, have exactly the same reading as the Armenian version: яко кпистолик речете соуть тежкы "because you (pi.) say: the epistles are weighty...". The other manuscripts of the recensions I and III contain only "because the epistles are weightly...", without any form of "to say" (contrary to rec. IV, Greek, Syriac etc.). Voskresensky suggested a confusion in reading of Greek, sioiv instead of (pr|oiv. Nevertheless, he has noticed himself that the reading of the An, the Gilferding Apostolos Nr 13 (14th cent., as well as the ms A19, from the same collection), is presented in the Armenian. In fact,
о
the Armenian version contains both речете ("you [pi.] say") and соуть ("are"): Uujigt p, рпщрри huitmitip hii...
Nr 23 (Gal 1:22). "Did not know me altogether" instead of "did not know me by face/personally". Here, both Syriac translations render the same Greek text, but H uses the Greek loanword (rc^c^v^.), whereas P uses a regular Semitic word with the same meaning However, the latter could have easily been misspelled as the composite conjunction formed by .=> "with, in" (in various meanings) and "even if, albeit, although", that is, Such a
construction has probably never used in Syriac but is understandable and corresponds to the peculiar reading of the Slavonic.
Nr 24 (Eph 4:13). Instead of "unto the measure of the stature of the fullness of Christ", the peculiar Slavonic reading has "unto the measure of the body of the fullness of Christ". Both P and H correspond to the known Greek text, and a number of early Syriac authors do the same, but two manuscripts of H (HI and H4) contain the following gloss to the word "of the fullness" (гс^Ьоо*.*): к'г^д "of the body". This gloss proves that the Slavonic translation corresponds to a text tradition known in Syriac but quite marginal, in the two senses of the word. This reading is specific to one of the Syriac versions.
Nr 25 (Eph 4:29). Instead of the normative ооотр6<; (here "unwholesome, corrupt"), the Slavonic recension I has "hateful". This corresponds to the whole Syriac tradition ("any hateful word"): P rdtAJD rdro H к'Ы. к'ЬчЪо Aa^, and the similar or the same in many Syriac authors. The same in the Ethiopic (iT/V : )1C, ■ "no evil word at all"), the Armenian (rnqhq "ill-
favoured, ugly"), ArSyrSin "ugly"), and even the Latin Vulgate (malus); in Coptic, eeooy encompasses the whole spectrum of meanings between ооотр6<; and како<; "what is putrid/evil/wicked" (Crum 1939: 731). Like ICor 11:30 (Nr 12 above), this is another example of a
57 Once more, I am grateful to Alexey Ostrovsky for attiring my attention on it.
Slavonic variant reading where the coincidence with the Syriac (and other Oriental versions) could be accidental with a high probability.
Nr 26 (Eph 5:1). Instead of "be imitators (^i^nxai) of God", the peculiar Slavonic reading is "be similar to God", in perfect correspondence with the Syriac tradition: P ^яд^я or H re^^. example of a Slavonic variant reading where the coincidence with the Syriac (and other Oriental versions) could be accidental with a high probability (much higher than in the most of cases). An exact correspondence in the earliest Ethiopic version (EthGr) only: Ф^Л^ : h^: X^^X^^C : "be similar to God", whereas a later recension has h^ : : AX^X^I^C : "be fitting /
ready / worthy to God".
Nr 27 (Eph 5:18). Only the Ethiopic adds "Saint" to the word "Spirit": ^¿Л : : (at least, in the earliest recension, not in all manuscripts), thus witnessing, with a considerable probability, to the existence of the same reading in Greek.
Nr 28 (Eph 6:7). Instead of serving "with good will/mental attitude", the peculiar Slavonic reading urges to serve "with love". The same in P and ArSyrSin (м^-;), whereas H follows
the Greek literally (^ ^^^^ ^oi.^r.). The Ethiopic versions provide three variants, including : "with love" in the 1548 editioprinceps, but "from the whole your souls" (ПН-fr : 'zihh^ :) in the main text according to the critical edition and a calque of the Greek цех' suvoia^ "with a good mind" : :) in several other manuscripts.
Nr 29 (Eph 6:22). P against H in agreement with ArSyrSin (^¿ь I* "what happened with me"). Also in the Ethiopic (ftif : "my news/story"), with no variants, and in a part of the Bohairic manuscripts: еын етфоп ммо1 "those which are to me" (another part having ммоы "to us").
3. A Quantitative Interpretation
3.1. Introduction
One can easily say that the above results are corroborating our previous conclusion on the earliest layer of Slavic Christian writing: that it was created as a result of a late seventh-century Syriac Melkite but Monothelete mission having in background the Syrian anti-Chalcedonian Church of the adherents of the patriarch of Antioch Paul Beth-Ukkame, who were firmly established in Egypt (Lourie (forthcoming a)). This is not the purpose of the present paper, however. In the present paper, we have to decide what meaning would have the number 29 itself: is this number great or small? Or, in other words, whether its contribution to the hypothesis of the "Paulist" background of the Slavonic writing is significant or not.
In a more formal language of the inductive logic, our question is the following. Regardless of the a priori likelihood of the "Syriac" hypothesis on the origin of the Slavonic writing, what is its a posteriori likelihood resulting from our investigation of the Apostolos? The notion of a posteriori likelihood of a hypothesis allows taking into account the change of its likelihood resulting from observations or experiences.
The inductive logic works with the mathematical formalism of the theory of probabilities and, in particular, with the theorem of Bayes. The theorem of Bayes establishes the link between the probabilities of the two events: that of the event A under condition that the event B occurred and that of the event B under condition that the event A occurred. When we have made our 29 observations, some 29 events occurred. The likelihood of any hypothesis regarding the origin of the Slavonic translation of the Pauline epistles under condition that these events occurred is different from its a priori likelihood, that is, its likelihood before any observations.
From a practical point of view, there are insurmountable impediments preventing us from the calculation of the likelihoods of hypotheses as such. We can compare, however, the likelihoods of any two alternative hypotheses via their ratio.
In our case, the two hypotheses to be compared are, if we call them in the language of statistics, those of accidental (hypothesis A) and of non-accidental (hypothesis NA) provenience of the coincidences between the Slavonic and Oriental readings.
The term "accidental" means, in the present context, that the coincidences between the 29 Slavonic readings in Table 1 and readings of Oriental versions have mostly two sources of origin: (1) a genuine coincidence in the mode of thinking between mutually independent editors, translators, and scribes resulting into identical or similar changes in the text, and (2) occurrence, in the Byzantine Greek manuscripts used by the late ninth- or tenth-century translators into Slavonic, of some readings that were not attested to in Byzantium, that is, very, if not extremely rare, and, moreover—what would have been even more strange—preference to these readings from the side of the Byzantine-oriented translators. These two constituents are to be summarised for giving the total likelihood of the "accidental" hypothesis. This is nothing but the present consensus hypothesis that the Slavonic translation of the Pauline epistles is performed and edited (several times) by the circle of Cyril and Methodius and their direct and/or indirect disciples.
The "non-accidental" hypothesis could be understood in one of the two alternative ways equally legitimate from a logical point of view but sharply different from a point of view of historical interpretation. One of them is our hypothesis that the sources of the "Oriental" readings in the Slavonic are Greek manuscripts accessible outside the Byzantine Empire together with Syriac translations. Another one is the hypothesis that the presently commonly accepted view on the history of the Greek text of the New Testament in Byzantium has very serious flaws, whereas, in fact, the main source of origin of these peculiar readings is one of the major traditions of the late ninth-century Byzantine Greek text.
The latter hypothesis means that our Greek manuscripts dated to the period up to the eleventh century (whose readings are studied and mostly, if not completely, published) do not constitute a representative sampling. One example is Rom 14:11, where the Greek reading тф к^ф (instead of тф беф) is attested to in the 16th-cent. cod. Daventriensis only (Nr 47 in the edition of Scholz58 but completely ruled out from the later critical editions). This is why it became known to Voskresenskij59 and, therefore, it does not figure in our list of "peculiar" readings. However, its very large accessibility in Greek in the middle of the first millennium could be realised from the fact that it first appeared in Syriac not in Peshitta (where accepted is an ancient reading "and to me", До; the same in the ancient Ethiopic) but in H ; symptomatically, with the marginal
readings in mss H1 and H4 "to the God"), that is, in a direct translation from Greek ca 616.
The same situation in the Arabic eighth-century translation from Greek ArGrSin: vj^ "to the Lord".
It would be reasonable to rule out such a historical hypothesis due to its extremely low a priori likelihood60, even though its a posteriori likelihood resulting from our observations will be the same as that of our "Oriental" hypothesis. In any case, for the present study, we will be interested
58 See (Scholz 1836: 196), who indicates the existence of the parallels in the Syriac and the Slavonic.
59 See (Воскресенский1879: 228). Attested to in two mss of recension I according to Voskresenskij.
60 Such an evaluation of this likelihood is not self-evident, however, given the number of preserved but still unpublished variant readings of the Greek manuscripts, which could be approximately realised due to the preliminary works for the Editio Critica Maior. S., however, below (Table 3) two (from the total number of 48) Slavonic "peculiar" readings checked against the totality of the preserved Greek readings, but still not found among them. Another argument against the Byzantine origin of, at least, a part of the "peculiar" reading in Slavonic is the presence, among them, of several reading hardly possible without looking at the Syriac.
only in discerning between the likelihoods of the "accidental" and "non-accidental" hypotheses, without going any deeper into the historical interpretation of the results.
3.2. Basic Ideas for the Computation
Given: m = 29 observations (readings) somewhat favourable to the "non-accidental" (NA) hypothesis.
Show: In what extent these observations are favourable to the NA hypothesis against the competing "accidental" (A) hypothesis.
In order to reach the maximum of objectivity, we can reformulate our task: What level of accidentality is needed to make the two competing hypotheses, NA and A, equally likely? This "level of accidentality" (= probability of accidental appearance, in the Slavonic, of the seemingly Oriental readings) is a calculable magnitude. If we obtain it, every reader will become able to make his or her choice between the two hypotheses, depending on what level of accidentality is, to his or her taste, the maximum allowable.
The results will be presented as a function of the ratio of the two a posteriori likelihoods from the variable representing the probability of accidental coincidence between the Slavonic and Oriental versions against the available Greek text.
Thus, we have to describe the behaviour of the probability of accidental appearance of the "Oriental" readings depending on the number of these readings (m = 29).
Three main parameters are necessarily to perform such calculation. One of them is our m: the number of observations that needed to be interpreted. The second parameter is n: the total number of the comparable observations (readings, in our case) in respect to whom the two hypotheses at stake are somewhat selective (= the respective probabilities of such events must be different for the two hypotheses). Such observations are not only those counted in the number of m, but also some others, where the variant readings are of seemingly the same type (s. next section for the details) but having no Oriental parallels.
Finally, we need to establish the parameter r: the probability of the "Oriental" readings according to the hypothesis NA. The value of r could not be very close to 1.0, because some contribution from the part of accidental readings is inevitable. Nevertheless, it must be substantially higher than 0.5. This means that, according to the hypothesis NA, the "Oriental" readings must be the major, even though not the only source of specific contaminations of the textual flow, which we will call "perturbations" (s. next section). We will perform two series of calculations, for r = 0.8 and r = 0.9.
Otherwise, if we accept r near 0.5 or even lower, we would introduce implicitly an additional supposition of a very high noise level in our text. Indeed, such a situation is possible, but it must be treated in a much more sophisticated way and not in our method of rough evaluation.
In our present method, however, where we will take r substantially higher than 0.5, such a situation (when the real value of r should be about 0.5 or lower) will look as an unambiguous victory of A over NA. This is a limitation of our method, the price of its simplicity.
One can see that r is nothing but the signal-to-noise ratio.
Our limitation for the value of r results from the very nature of the competing hypothesis (A), which implies that even our m observations belong to the noise.
Knowing n, m, and r, we will be able to find the value of q—the probability of the seemingly Oriental readings according to the A hypothesis—corresponding to the condition of the equality of
the likelihoods of the two competing hypotheses. The lower this value of q is, the higher is the likelihood of the A hypothesis in comparison to the likelihood of the NA hypothesis.
In general, there are no specific limitations resulting from the signal-to-noise ratio. There is only a need of sufficient difference between the parameters r and q.
3.3. The Notion of Perturbation of the Textual Flow
The m "Oriental" readings reviewed above are a particular case of the phenomenon that I would like to name "perturbation of the textual flow" (by analogy with the notion of perturbation in physics). Perturbation is, by definition, some change in a process caused by an impact from outside. The notion is, therefore, depending on the way we discern between "outside" and "inside" of the textual flow.
If the process we consider is the textual flow through different cultures and languages, we have to deal with the heavily contaminated textual traditions having a low density of the flow (that is, where the witnesses of the text are relatively far from each other)61. Nevertheless, it is often possible to have an idea of the sources of contamination involved. Some contaminations, however, with a high or a low probability, could not be derived from the sources already known to us. I call these contaminations "perturbations of the textual flow". They witness to—with a high or a low probability—that there are some sources of contamination that we have overlooked.
Formal definition: "perturbation" is an unexpected—from a preconceived point of view—
contamination of the textual flow.
Speaking logically, "contamination" is a notion implying an objective reality within the unique actual world, whereas "perturbation" is its modal generalisation related to the possible world observable from a preconceived viewpoint. "To be a contamination" is a predicate in the first order predicate logic, whereas "to be (or "not to be", or "unclear whether to be or not") a perturbation" are modal states in a modal logic, where "being a perturbation" is a modality. For the sake of simplicity but in the same time, for increasing of the reliability of our evaluation, we will treat the borderline cases as negative (as contaminations that are not perturbations).
For instance, we can be sure that any version of the New Testament in any language is ultimately a translation from Greek. The sources of contaminations that are present in such text by default are mainly the following: those specific to the translation in the language of this version, various possibilities of translating into the corresponding language (various renderings ultimately going back to the same Greek reading), and the variant readings of the Greek manuscripts that would have been affected the version at any stage of either translation or editing. However, if the text of our version A contains, e.g., some forms of the proper names different from that of the Greek New Testament but specific to another ancient version B made directly from Greek, we have to ask whether our version A is a translation of a translation rather than a direct translation from Greek. Such contaminations could be named perturbations from a given viewpoint, namely, under the supposition that version A was produced as a direct translation from Greek. From another point of view—that version A was translated from version B, whereas it was the version B that was translated from Greek directly—these contaminations are normal phenomena, that is, not perturbations.
61 Cf. especially: (Mink 2004; Spencer, Wachtel, Howe 2002).
Our "Oriental" readings are certainly perturbations—from the point of view that the Slavonic Apostolos was translated by Byzantines (and/or their disciples) from a Middle Byzantine Greek text. They are not the only perturbations of our textual flow from the Greek to the Slavonic, however.
Now we need to decide which contaminations we have to consider perturbations, too.
It is already clear that the perturbations are those contaminations of the textual flow that are potentially relevant for comparing the likelihoods of the competing hypotheses. Some of them certainly belongs to the informational noise, whereas some others could be manifestations of an impact of a previously unnoticed source (textual tradition).
3.4. The Total Number of the Perturbations of the Textual Flow (n)
There are two main rules of identifying the perturbations and several additional rules. All these rules are those of screening. The perturbations are those variant readings that remain after the screening through all these sieves.
Main rules:
• The majority of the variant readings in the Slavonic manuscripts appeared certainly—and not simply "possibly"—on the Slavic ground. All of them are irrelevant to us. These are differences in spelling, in morphology (when they do not substantially change the meaning), the errors typical for the Slavonic manuscripts, and, of course, various renderings of the same Greek word—which are not to be confused with the various readings of the Greek manuscripts underlying the Slavonic version.
• The Slavonic variants going back to different readings of the Greek manuscripts are, most often, non-informative either—providing that the corresponding Greek readings be preserved in the accessible Byzantine manuscripts. Such readings are equally compatible with both A and NA hypotheses62. There are no Byzantine readings theoretically inaccessible anywhere in the Christian Orient.
Indeed, there could be difficult cases of the readings attested to in very late Greek manuscripts only and, therefore, lacking from the scholarly editions but widely attested to in the Orient. This is the problem of the representativity of the presently available selecting of the variant readings of the Greek manuscripts, which has been tangentially discussed earlier (section 3.1).
The readings we are interesting in are "peculiar": the differences they represent are meaningful but unexplainable with a reference to any known Greek manuscripts. They are unexplainable in an obvious way on the Slavic ground either, even though their Slavic origin is not to be excluded.
62 Nevertheless, with such tools as the future Editio Critica Maior it will be possible to make the method more selective with taking into account such readings as Rom 14:11 discussed above (section 3.1). The reading xra Kupira was more likely available outside the post-Arabian Byzantine tradition than within. Presently, without considering them, we increase the relative likelihood of the A hypothesis. It is, therefore, safely for our purpose of evaluating the relative likelihood of the hypothesis NA. As an attempt of a more selective approach than the mine, one could mention Olga B. Strakhov's paper (Strakhov 2015). She demonstrates (for Lk 24:43 and 24:36) that some mainstream early Slavonic readings correspond to those Greek readings that, even though being attested to in Byzantium, were rare and, therefore, hardly presenting in the manuscripts used by the translators into Slavonic. However, the following conclusion by the author is not as natural as it appears in her demonstration. The fact of a large distribution of the same readings in the Latin is not sufficient to prove that they penetrated into the Slavonic from the Western rite, because, in both cases dealt with, their distribution in the Syriac is no less large; cf. (Kiraz 1996: 508, 511).
The total number of such readings could not be especially great. For instance, in our selecting which ultimately goes back to about 23,000 Greek words (the total word number of the epistles from Rom to Eph), the part of the 48 perturbations is about 1.7 • 10-4. In a good and well-preserved translation from a known original, the perturbations would be impossible even theoretically. And, indeed, in our material, recension III (a new translation from Greek) is not too far from this ideal. These "irregular"—or, at least, apparently irregular—readings are those to whom the a posteriori likelihoods are certainly selectively sensible.
The above are the main principles to be applied for selecting the perturbations. They are not completely rigorous but sufficient to provide statistical representativity (we do not need to exhaust all the readings classifiable among the lists of n and m, even though the more complete our lists are, the more reliable our comparison of the likelihoods is).
There are some other types of variant readings that are non-informative and should be ruled
out.
Additional rules:
• Short conjunctions (especially и "and") and particles, whose additions and omissions are too frequent in the Slavonic manuscripts. The probability of an accidental change is, in their case, always too high, even if the hypothesis NA and not A is true.
• For the same reasons, we have to exclude those Slavonic readings, whose difference could be explained through a variation, in Greek, of the pronouns "you (pl.)/we" ^цет^/ицет^, ^rov/up-rov, ^Mtv/uMtv, 'лца^/'ица^ regardless of whether the corresponding Greek readings are actually attested to in the preserved manuscripts. Such variations, in Greek manuscript traditions, are typical due to the iotacism and are not specifically related to either Byzantine or non-Byzantine milieu.
• A more delicate problem is the permutations (changes in the word order). Not all Oriental versions preserve the word order of their original. For instance, the Syriac Peshitta translation does not follow the Greek in this matter, whereas another translation into Syriac, H follows63. If we consider a hypothesis where some Oriental influence on the translation from Greek is implied, it is hardly possible to make informative such variant readings as permutations, and it is much safer to factor them out.
• Finally, I have ruled out the unique readings (the variant readings known from unique Slavonic manuscripts), because the rules of statistics are not directly applicable to them. This is not to say that the unique readings have no value for the studies of translations. Indeed, they could be very valuable (s. below, section 3.7)—but normally not for the calculations based on the theory of probabilities.
63 Cf. (Brock 1977). Other analogous chapters of this monograph, which are dedicated to other languages' limitations in the same respect and written by eminent scholars as well, still preserve their importance. Cf. also an interesting study indirectly related to the Slavonic New Testament: (nenrKOBCKaa 2005).
3.5. The Method of Calculation
In the following, I will limit myself to the most necessarily formulae only64. The likelihood of a hypothesis is defined as a specific kind of probability. It features the degree of rigour of the logical inference when the latter is depending on accidental events. If a hypothesis hi is much more likely than a hypothesis h2, this means that
(1) P[e |hi-6-c] » P[e I hxb-c]
where e is the result of an observation under the condition c and with background knowledge b, which could be potentially depending on the competing hypotheses h1 and h2. This direct inference likelihood is a completely objective value.
It is preferable that both b (background knowledge) and c (conditions of observation) were independent of a given hypothesis (otherwise, the computations become much more difficult).
For making the comparison possible, the hypotheses are to be compared under the same b and c. This is not completely trivial.
The background knowledge, in our case, includes the textological data and nothing more. This is the same for both A ("Byzantine") and NA ("Oriental" or "Syriac") hypotheses. In other words, we have, for the time being, to forget the names of Cyril and Methodius and the hagiographical legends about the origin of the Slavic writing.
The conditions of observations, in our case, could be different were we considering the statistical clusters of manuscripts (called in textology "recensions") as separate sets. However, our observations are, in fact, independent from the recensions: we are always considering the whole totality of manuscripts without specifying any subsets (recensions) within.
The formula (2) is a simplified definition of the likelihood of a given hypothesis hr:
(2) P[e I hr-b-c] = —x rm (1—r)n—m m!x(n - m)!
More precisely, this is only posterior probability of this hypothesis, without taking into account its prior probability. The prior probability is not an objective magnitude. It is the likelihood of a hypothesis before any observations are performed. It heavily depends on such things as scholarly opinions. We are not interesting in comparing such likelihoods, and, therefore, we reduce our discussion of the likelihoods of hypotheses to the posterior (aposteriori) likelihoods only.
The comparison of two a posteriori likelihoods of hypotheses is the following ratio (3):
(3)
P en К b - cn = qm (1 - q) n-m
P e" к b - cn _ rm (1 - r ) n-m
64 For the details, s. (Hawthorne 2014). Cf. a good introduction to the inductive logic by Lolita B. Makeeva (Makeeva 2014). Cf. also (Лурье 2014).
where r is the probability of "Oriental" readings according to the hypothesis NA, whereas q is the same probability according to the hypothesis A.
We need to trace the behaviour of q near the points where the value of the ratio (3) becomes close to 1. This means that we have to solve for the variable q the following polynomial equation (4):
q" (1 - qYm = 1
V ) rm (i - r)n-m
The denominator in (4) depends on the value of r, which is to be chosen by ourselves. Thus, the denominator is a constant for given values of n and m. Let us design it a. Then, the equation (4) can be rewritten in a more traditional equation style (designing the variable q as x) as following (5):
xm (i x)n-m (5) x (1 - x) = i
a
The values of a will be calculated for r = 0.8 and r = 0.9.
Then, the real roots of the polynomial equation (5) belonging to the interval ]0, 1[ (the real numbers between 0 and 1 excluding 0 and 1) will be found. These roots will be the probability values corresponding to the equal likelihoods of the hypotheses A and NA. One of these roots, as it is easy to see, corresponds to x = r and is useless to us; it will be rejected for the obvious grounds.
The range of values of q corresponding to the ratio (4) substantially greater than 1, where "substantially" means that it exceeds 1 by several orders, corresponds to the situation when A is more likely than NA, and, vice versa, the ratio (4) substantially lower than 1 means that NA is more likely than A.
After having solved the equation (5) for the variable x, we thus obtain a tool allowing to us to judge, what situation is more realistic. Indeed, if the preponderance of the hypothesis A implies that the values of q must be unrealistically great, this would be a strong argument against this hypothesis and in favour of the competing hypothesis.
3.6. The (Apparently) Non-"Oriental" Perturbations
To proceed further, we need to find n. In fact, this means that we need to find the readings whose amount will be equal to n — m.
These readings are presented in Table 3. They were chosen according to the rules described above (section 3.3).
In two cases, 1Cor 10:9 and Gal 4:7, we were able to check all the readings preserved in the Greek manuscripts, because these places were among the test passages investigated by the team of the Editio Critica Maior (the numbers of the corresponding manuscripts are indicated in the brackets after the Greek variant readings). It is somewhat significant that the peculiar Slavonic readings were not found among more than 600 Greek manuscripts where the corresponding places are preserved and legible.
Table 3.
Nr Place Воскресенский 1879 65 page Rec. (mss66) Peculiar reading Normal reading(s) Greek
1 Rom 1:23 207-208 I в подобьствии тлЪньна тЬла человека образа eiKovog
2 Rom 3:2 ХШ 175 I [10 mss] по всеи/всякои ипостаси всяцЬмь образомь / по всей винЬ ката navTa Tponov
3 Rom 3:2 ХШ 175176 I [5 mss] даровашж са им вЬровашж са им ¿nioTeuQnoav
4 Rom 3:26 228 I [3 mss] от вЪры стыА [from the holy faith] ис Хвы Глоои / I^oou Хрштои / I^oouv
5 Rom 8:28 212 I, IV [= Б7 in II] по прозрЪнию "according to the foresight" по предложению ката np66eoiv [пробега confused with пробегарега]67
6 Rom 11:24 230 I [3 mss] свогеи добрЪи [add A18] / блзЪи [add A17, A20] маслинЪ [cf. above ibid.: eig KaHieXaiov] свогеи маслинЬ [no addition] Tfl i5ta ¿lata
7 Rom 15:14 225 I [12 mss] исплънени плод разоумъных испълнени всего разоума пепХпргам-evoi паалд [т^д] yvraoerag
8 1Cor 3:22 261 I [save 11 mss] аще же всь мир аще же мир [no addition] егте кооцод
9 1Cor 10:9 281 II [save 6 mss] 1са [add] Х(рист)а Х(рист)а / г(оспо)да / Б(ог)а [no addition] Xpiorov (525) / Хрштга (38) / K^rov (43) / 6e6v (16) / om (5)
10 1Cor 14:38 271 I [save 13 mss] аще ли кто не разоумЪгеть. яко б(ож)ия соуть заповЪди [add]. Да не разоумЪвагеть. [no addition] ei 5e тгд ayvoei, а^оеиш [insertion repeats 14:37 on кирши ¿ouv ¿vra^]
11 2Cor 2:4 296 I [5 mss] от печали бо многы сердцоу от печали бо многы и тоугы [om] сердцоу ¿к уар поИ^д бМуегад каi ошохл? rapStag
12 2Cor 6:2 299 I [5 mss] [om] се нынЪ времА бЛгопрiАтно, се нынЪ день спасешА [no omission] i5ou vuv гагрбд еипроо5ектод, i5ou vuv ^epa оraт'Пptag
13 2Cor 7:3 304 II [save 1 ms] and A6 (I rec.) в с(е)рдцих наших есть в с(е)рдцих наших есте ¿v тaíg rapStaig ^rav ¿оте
14 2Cor 7:15 305 II and A6, A20 (I rec.) с(е)рдце (heart) оутроба та onXayxva
15 2Cor 12:1 293 I [2 mss: A6, A7] вь видЪние анГла [add] в видЬния явления гна [om] eig 6magtag кai апога^и^е^ кирши
65 Or Hristova-Shomova's page (ХШ) when necessary (but, even then, the number of variant readings according to (Воскресенский 1892; Воскресенский 1906; Воскресенский 1908).
66 Indicated for minoritary readings of a given recension.
67 The common reading of recension II is по воли "according to the [sc., disposition of the] will".
Вып. 8. 2016
Table 3 (Continuation).
Nr Place Воскресенский 1879 page Rec. (mss) Peculiar reading Normal reading(s) Greek
16 Gal 4:7 (cf. 3:29) 318 I [4 mss incl. A1], II [save 6 mss] причастник б(о)жии д(оу)хом [*0eou 5ia nveu^axo^ ?] Cf. 3:29: post наслЪдници (K^npovo^oi) A1 add б(о)жии д(оу)хом68 наследник б(о)жии исъ хвъ K^npovo^o^ 0eou 5ia Inoou Xpioxou etc. (100 mss), other readings 527 mss.
17 Gal 4:17 316 I [4 mss], II [save 1 ms], III, IV [save 3 mss] цркви вам хотАть "(they wish) to you churches" (sic!) Прельстити хотАть "(they wish) to deceive" ¿KK^eioai ица^ 0eXouoiv
18 Eph 2:18 333 I [12 ms, incl. Aj and A3j] в едино тЬло [A31 дЪло] "into the unique body [A31 erroneously affair] в кдин д(оу)х ¿v ¿vi nveu^axi
19 Eph 5:18 334 I [3 mss + Izbornik 1076], II [save 1 ms] есть несп(а)сение [read as an inexistent word *aoraxnpia] есть блоуд eoxiv aoraxia
68 The two readings are present together in A1 = Apostle of Okhrid, 12th cent.
Notes to Table 3:
Rom 11:24. In the word "good" is repeated with the second mention of the "olive tree": "For if thou... wert graffed contrary to nature into a good olive tree: how much more shall these, which be the natural branches, be graffed into their own [good add] olive tree?"
2Cor 7:3. As Voskresenskij noticed, the reading without the copula at all is known in Greek but
rd '
not with the 3 pers. sg. form sgtí.
2Cor 12:1. "The vision of the angel" instead of "the visions and the revelations of the Lord". This reading is attested to in two manuscripts of the 13th/14th cent. (incl. A6, so-called Karpinsky Apostolos). I have no idea of its origin whatsoever.
Gal 4:17. This erroneous reading recognised already by Voskresenskij (¿KK^oíai pro ¿KK^stoai) is known to me only in Slavonic. However, it is theoretically possible that it first appeared as a scribal error in Greek or in an Oriental translation.
3.7. One Case of Unique Readings
Some unique readings (never included in Tables 1 and 3) could be not only non-accidental but also interesting in some respects. E.g., 2 Cor 3:14 according to the Chudov New Testament (= rec. III): ослепоша (instead of окаменЪша) помышления их. It is obvious, as Voskresenksij has already noticed (Воскресенский 1879: 291), that this variant goes back to a different Greek variant reading: елюрюбп misspelled as елпрюбп. This Greek reading is, however, absent from the editions of the Greek New Testament available even to us, not to say of Voznesenskij, although it is still very probable, of course, that it will reappear in the Editio Critica Maior.
However, the reading елпрюбп is normative for both Syriac versions, as well as the Armenian and ArGrSin. Its presence in H demonstrates that is was quite widespread in the Byzantine Empire ca. 616 (when Egypt, where Thomas of Harqla was working, was still within its borders). Nevertheless, it was certainly absent from the majority Byzantine text of the thirteenth century, the date of the translation preserved in the Chudov New Testament.69
This reading is the unique "perturbation" (in our sense of the word) that I was able to find in the Chudov (III) recension of the Pauline epistles from Rom to Eph. It could be interesting for evaluating the Greek New Testament manuscripts availability in the Second Bulgarian kingdom, the Sitz im Leben of this recension.
3.8. The Quantitative Analysis: Preliminaries
The number of "non-Oriental" perturbations turned out to be somewhat disappointing: it is considerably lower than that of the "Oriental" readings (19 vs 29) and apparently leaves no room for calculations. Indeed, intuitively we can already say that it is not likely that 29 from 48 perturbations (ca 60%) coincide with some "Oriental" readings accidentally. Nevertheless, there is a need of
69 On the Chudov recension, s. the studies of Tatiana Pentkovskaya, culminating in her monograph neHTKOBCKaa 2009. Pentkovskaya considers the translation to be a work of a Russian translator but working outside Russia, in some of "contact zones" influenced by South Slavs and even probably knowing himself some South Slavic language(s). This is logically fragile construction. Hristova-Shomova considers the transanslation to be simply Bulgarian (XpncTOBa-ffl0M0Ba 2004: 785—798). On can add that the third hypothesis—of a South Slavic translator working in a contact zone with the Russians—would be plausible equally to the Pentkovskaya's hypothesis.
performing a quantitative analysis even for such "intuitively clear" situations—for calibrating the method.
Moreover, we will discuss as well, using the same quantitative method, a much less trivial problem of whether the translator into Slavonic had have a look at some texts in Syriac.
We have to solve the equation (5) and to trace the behaviour of the function Ra/na(x) (6)—the ratio of the posterior likelihoods of the competing hypotheses—in the neighbourhoods of the real
70
roots belonging to the interval x e ]0, 1[ .
xm (1 _ x)n_m
(6) Ra/na(x) = —^-}— = a1 xm (1 _ x)n _m
a
Given that the powers of the polynomials we deal with are very high, the ratio of the likelihoods, which is governed with the power law, will be very sensitive to a very small variations of the value of q (or x in our equation (5)), that is, the supposed probability of the accidental coincidences.
3.9. The Quantitative Analysis: Calculations
Let us perform our calculations under conditions (7.1) and (7.2):
(7.1) ri = 0.9; a~l = 2.M020
(7.2) r2 = 0.8; a'1 = 1.2-1016
The real roots of the equation (5) within the interval ]0, 1[ (and after having rejected the root x = r) are the following:
r1 = 0.9; x = 0.238
r2 = 0.8; x = 0.382
These roots are the only real roots within the possible range of probability values, with exception of the roots corresponding to x = r. One can see that, in both cases, the likelihood of the A hypothesis is much lesser than that of the NA hypothesis. That of A would prevail only under condition that, at least, more than every fifth or even every third accidental contamination coincides with some reading of an Oriental version. This is unrealistic.
To figure out the shape of the power law connecting the ratio Ra/na and the presumed probability of the accidental perturbations, it would be useful to consider some plots.
The plots on Figs. 1 and 2 provides general outlines of the corresponding functions. The two figures differ mostly with the scale of R: the scale of the ratio of the two likelihoods is in millions of times in the first case (where the NA hypothesis is put forward in a strong form: r = 0.9) and only in dozens of times in the second case (where this hypothesis is put forward in a weaker form: r = 0.8). One can see that our method would not work at all for week hypotheses (e.g., r = 0.6), because the resulting ratios RA/NA would become too low.
70 These calculations are technically somewhat difficult because of high powers of the equations and functions. The software used is Wolfram Mathematica 10.3.
The two plots are bell-shaped, with two symmetric branches tending asymptotically to zero. The right halves of the plots pass through the value R = 1 at q = r. In fact, only the left halves are informative, and only in the part corresponding to the neighbourhood of the first real root of equation (5), which is symmetric to the point of the plot corresponding to the second real root at q = r. This is why, for the practical purposes, we need large-scale plots such as those at Figs. 3 and 4. These plots show how the power-law function works, and why the method of comparison of the two competing likelihoods is reliable.
The plot of a power-law function contains two almost perpendicular parts connected with each other through the inflection point at a very short period of x-axis. Our task does not consist in an exact calculation of the ratio RA/NA, which would require an exact knowledge of both r and q. Our task is limited to an evaluation of this ratio: whether it has such value that allows insisting that one hypothesis is much more likely that the competing one.
If we can be sure that we are far enough from the inflection point, and our situation corresponds to a plot section belonging to one of the two semi-perpendicular parts, we can be sure that one hypothesis is much more likely than the another one. Otherwise—if we turn out somewhere near the inflection point,—then, the method losses its reliability.
3.10. Looking at the Syriac?
It would be tempting to perform some calculation to evaluate the likelihood of direct checking of the Slavonic against the Syriac, whether the four rows of Table 1 marked with grey fill colour are accidental or not. Such a calculation is, however, hardly possible—at least, without using some sophisticated filters of information noise. The reasons are those explained above (section 3.2). Even under the supposition that some corrections against the Syriac took place, we have no right to claim that the probability of miscorrections resulted from this work is more than 0.5 among the other
71
perturbations of the textual flow. And, indeed, we would have, in our case, 23 perturbations but only 4 among them are supposedly originated from miscorrections against the Syriac. In our case, the signal-to-noise ratio is unacceptably low.
However, we have already obtained serious reasons for a high a priori likelihood of the "Syriac" hypothesis. They are the following.
1. Non-Byzantine ("Oriental") features of the Greek manuscripts used.
2. A similar and roughly contemporaneous method of translation of the New Testament from Greek into Arabic in a similar milieu (the Melkites, whereas rather dyothelete than monothelete): translation from the Greek but adopting some readings of the Syriac.
3. Attestation of the most of the "Oriental" Slavonic readings in either Syriac or direct translations from Syriac (21 from 29).
4. Attestation (direct or indirect) of, at least, five of the "Oriental" Slavonic readings (marked
72
with grey fill colour in Table 4, column "Nr") in Syriac exclusively .
Thus, our four cases of possible mistranslations from Syriac are positioned in such context, where their non-accidental appearance is very likely, even though we are unable to perform the adequate calculations of the a posteriori likelihoods.
71 The four rows of Table 1 marked with grey fill colour together with 19 items of Table 3. The remaining 24 rows of Table 1 could no longer be considered as perturbations, because they could be explained as variant reading of non-Byzantine Greek manuscripts.
72 I exclude from this list Nr 18 (2 Cor 5:8) given that the corresponding double attente is by no means specific to Syriac.
4. Conclusions
4.1. History of the Textual Flow
Our conclusions will be limited to the topics discussed above at length, thus avoiding going deeper into historical interpretation.
1. Some "Oriental" impact on a very early recension of the Slavonic translation of the five Pauline epistles is demonstrated.
2. The quantitative evaluation of this conclusion was, in fact, an evaluation of its logical strength. It was not a necessary mean to derive it from the collected evidences.
3. The most natural interpretation of the previous conclusion (1) is that the Greek originals used for the Slavonic translation were those widespread outside the borders of the ninth-century Byzantine Empire (cf. above, section 3.1).
4. There are some traces of additional editing of the translations from Greek against some Syriac version(s), in the same manner as in roughly contemporaneous Melkite translations of the New Testament from Greek into Arabic.
5. The "Oriental" tradition involved was one of the Syrian ones.
6. Both I ("Ancient") and II ("Preslav") recensions share the above features that, therefore, are to go back to their common archetype, that is, an even more "ancient" recension.
7. Only the earliest Ethiopic version (EthGr) shows an affinity with the Slavonic material comparable to that of the Syriac material (s. Table 4). This fact must be interpreted as an affinity of the early Slavonic version with Greek recensions that were circulating before the seventh-century Arab invasion.
4.2. The Quantitative Method
From time to time, I have permitted to myself occasional references to my earlier papers dedicated to the Syrian Melkite monothelete mission to the Slavs in the late seventh century to which I have attributed the beginning of the Slavic Christian writing, but such historical problems remained, in general, beyond the scope of the present paper.
The quantitative method proposed in this paper is dedicated to comparison of two competing hypotheses concerning the textual flow of a highly contaminated tradition.
The method has the following preconditions and limitations:
1. The total number of possible hypotheses must be previously reduced to two: that a specific source of contamination existed or not.
2. In the present (simplest) modification of the method, the hypothesis about the presence of a discussed source of contamination must additionally imply a high value of the signal-to-noise ratio (> 0.5), that is, that this hypothetical source, if it actually existed, was the major source of contaminations of a specific kind (defined above as "perturbations").
Bbm. 8. 2016
of the Textual Flow in the Slavonic Recensions of the Pauline Epistles
1.5 *106
500000
Fig. 1. Ra/na providing that r = 0.9.
120 100 80 60 40 20
0.2
0.4
0.6
0.8
1.0
Fig. 2. Ra/na providing that r = 0.8.
R
6
6
1.0 x 10
q
1.0
R
q
0.22 0.24 0.26 0.28
R
q
Fig. 3. Ra/na providing that r = 0.9, q e [0.2, 0.3].
Bbm. 8. 2016
0.25 0.30
Fig. 4. Ra/na providing that r = 0.8, q e [0.2, 0.4].
R
q
Tabl. 4
Nr Place Syr P Syr H ArSyrSin ArSyrSpb ArGrSin Arm GeoAB GeoCD Sah Boh EthGr EthAr Others
1 Rom 6:9 *+ *+ — — — — — — — — — — —
2 Rom 6:19 *+ *+ — — — — — — — — — — —
3 Rom 6:22 *+ *+ — — — — ± — — — — — —
4 Rom 11:16 — — — — — — — — + + + + —
5 Rom 11:16 — — — — — — — — — + + + + Syr Philoxenus
6 Rom 12:6 + + — — — ± ± — ± — ± ± GeoAB variants only
7 Rom 12:14 — — — — — + — — — — + — —
8 Rom 14:7 - — — — — — — — + + — — —
9 Rom 15:15 + + — — — — — — — — — — —
10 1Cor 7:15 - — — — + — — — — — — — —
11 1Cor 7:37 — — — — — + + + — — — — —
12 1Cor 11:30 + + + - + — + + + + + + —
13 1Cor 12:29 + — — — — — + — — — — — —
14 1Cor 15:29 + — — — — — + — — — + + —
15 2Cor 1:7 ± — ± ± — + — ± — — *+ — —
16 2Cor 2:12 — — — — — — + — — — — — —
17 2Cor 2:14 + — — — — — — — + — — — —
Tabl. 4 (Continuation).
Nr Place Syr P Syr H ArSyrSin ArSyrSpb ArGrSin Arm GeoAB GeoCD Sah Boh EthGr EthAr Others
18 2Cor 5:8 + — — — — — — — — — — — —
19 2Cor 5:14 — ± ± ± — — — — — — — — ± Vat. Ar. 13
20 2Cor 5:18 — — — — — — — — — — + — —
21 2Cor 6:7 + — + — — — ± — — — — — —
22 2Cor 10:10 — — — - — + — — — — — — —
23 Gal 1:22 *+ — — — — — — — — — — — —
24 Eph 4:13 — — — - — — — — — — — — + H mgg
25 Eph 4:29 + + + — — + — — ± ± + + —
26 Eph 5:1 + + — — — — — — — — + — —
27 Eph 5:18 — — — — — — — — — — + — —
28 Eph 6:7 + — + — — — — — — — + — —
29 Eph 6:22 + — + — — — — — — + + + —
Literature
Aland, B. 1995. Categories and Text Types, and the Textual Analysis of Manuscripts. In Aland K., Aland B. (eds.). The Text of the New Testament. An Introduction to the Critical Editions and to the Theory and Practice of Modern Textual Criticism. 2nd edition. Grand Rapids, MI: "Wm. B. Eerdmans Publ. Co.", 317—337.
Aland, B., Juckel, A. 1991. Das Neue Testament in Syrischer Überlieferung. II. Die Paulinischen Briefe. T.
1. Römer- und 1. Korintherbrief, Arbeiten zur neutestamentlichen Textforschung 14. Berlin; New York: "W. de Gruyter"
Aland, B., Juckel, A. 1995. Das Neue Testament in Syrischer Überlieferung. II. Die Paulinischen Briefe. T.
2. Korintherbrief, Galaterbrief, Epheserbrief, Philipperbrief und Kolosserbrief; Arbeiten zur neutestamentlichen Textforschung 23. Berlin; New York: "W. de Gruyter".
Aland, K., Benduhn-Merz, A., Mink, G. 1991a. Text und Textwert der griechischen Handschriften des Neuen Testaments. II. Die Paulinischen Briefe. Bd. 1. Allgemeines, Römerbrief und Ergänzungsliste, Arbeiten zur neutestamentlichen Textforschung 16. Berlin; New York: "W. de Gruyter". Aland, K., Benduhn-Merz, A., Mink, G. 1991b. Text und Textwert der griechischen Handschriften des Neuen Testaments. II. Die Paulinischen Briefe. Bd. 2. Der 1. und der 2. Korintherbrief. Berlin; New York: "W. de Gruyter" Publ.
Aland, K., Benduhn-Merz, A., Mink, G. 1991c. Text und Textwert der griechischen Handschriften des Neuen Testaments. II. Die Paulinischen Briefe. Bd. 3. Galaterbrief bis Philipperbrief. Berlin; New York: "W. de Gruyter".
Brock, S. P. 1977. Limitations of Syriac in Representing Greek. In Metzger B. M. (ed.). The Early Versions of the New Testament: Their Origin, Transmission, and Limitations. Oxford: "Clarendon Press", 83— 98.
Brock, S. P. 1981. The Resolution of the Philoxenian / Harclean Problem. In Epp E. J, Fee G. D. (eds.). New
Testament Textual Criticism: Its Significance for Exegesis. Essays in Honour of Bruce M. Metzger. Oxford: "Clarendon", 325—343. Browne, G. M. 1994. Bibliorum Sacrorum versio Palaeonubiana. Corpus Scriptorum Christianorum
Orientalium. Vol. 547. Louvain: "Peeters". Browne, G. M. 1994. Bibliorum Sacrorum versio Palaeonubiana. Corpus Scriptorum Christianorum
Orientalium. Vol. 547. Louvain: "Peeters" (Subsidia. T. 87). Childers, J. W. 2013. The Georgian Version of the New Testament. In Ehrman B. D., Holmes M. W. (eds). The Text of the New Testament in Contemporary Research. Essays on the Status Quaestionis. Second Edition. New Testament Tools, Studies and Documents. 42. Leiden; Boston: "Brill" Publ., 293—327. Crum, W. E. 1939. A Coptic Dictionary. Oxford: "Clarendon Press".
de Halleux, A. 1963. Philoxène de Mabbog. Sa vie, ses écrits, sa théologie, Universitas catholica Lovaniensis. Dissertationes ad gradum magistri in Facultate Theologica vel in Facultate Iuris Canonici consequendum conscriptae. Ser. III. T. 8. Louvain: "Imprimérie orientaliste". Ehrman, B. D., Holmes, M. W. (eds). 2013. The Text of the New Testament in Contemporary Research. Essays on the Status Quaestionis. Second Edition. New Testament Tools, Studies and Documents. 42. Leiden; Boston: "Brill".
Géhin, P. 2006. Manuscrits sinaïtiques dispersés. I. Les fragments syriaques et arabes de Paris. Oriens Christianus 90, 23—43.
Gibson, M. D. 1894. An Arabic Version of the Epistles of St Paul to the Romans, Corinthians, Galatians with Part of the Epistle to the Ephesians from a Ninth Century Ms. in the Convent of St Catharine on Mount Sinai. London: "C. J. Clay and Sons" Publ.; Cambridge: "Cambridge University Press". Gippert, J., Schulze, W. 2007. Some Remarks on the Caucasian Albanian Palimpsests. Iran and the Caucasus 11, 201—212.
Gippert, J., Schulze, W., Aleksidze, Z., Mahé, J.-P. 2008. The Caucasian Albanian Palimpsests of Mt. Sinai.
Vol. 2. Monumenta Palaeographica Medii Aevi. Ibero-Caucasica. Turnhout: "Brepols". Griffith, S. H. 2013. The Bible in Arabic: The Scriptures of the "People of the Book" in the Language of
Islam. "Princeton" Publ.; Oxford: "Princeton University Press". Griffith, S. H. 2013. The Bible in Arabic: The Scriptures of the "People of the Book" in the Language of
Islam. Princeton; Oxford: "Princeton University Press". Harb, P. 1967. Lettre de Philoxène de Mabbög au Phylarque Abö Ya'för de Hïrtâ de Bêtn'amân (selon le manuscrit No 115 du fonds patriarcal de Sarfet). Melto 3, 183—222.
Hawthorne, J. 2014. Inductive Logic. In Zalta E. N. (ed.). The Stanford Encyclopedia of Philosophy (Winter 2014 Edition). Accessed at: http://plato.stanford.edu/archives/win2014/entries/logic-inductive/ (accessed 02.12.2016).
Horner, G. W. 1905. The Coptic Version of the New Testament in the Northern Dialect otherwise Called Memphitic and Bohairic. Vol. III. The Epistles of St. Paul. Oxford: "The Clarendon Press".
Horner, G. W. 1920a. The Coptic Version of the New Testament in the Southern Dialect otherwise Called Sahidic and Thebaic. Vol. IV. The Epistles of St. Paul. Oxford: "The Clarendon Press".
Horner, G. W. 1920b. The Coptic Version of the New Testament in the Southern Dialect otherwise Called Sahidic and Thebaic. Vol. V. The Epistles of St. Paul. Oxford: "The Clarendon Press".
Kashouh, H. 2011. The Arabic Versions of the Gospels. The Manuscripts and Their Families. Arbeiten zur neutestamentlichen Textforschung 42. Berlin: 'W. de Gruyter".
Kiraz, G. A. 1996. Comparative Edition of the Syriac Gospels. Aligning the Sinaiticus, Curetonianus, Peshitta andHarklean Versions. Vol. 3. Luke. Leiden; New York; Köln: "Brill".
Krenkow, F. 1926. Two Ancient Fragments of an Arabic Translation of the New Testament. Journal of the Royal Asiatic Society of Great Britain and Ireland 58, 275—285.
Louire B. (forthcoming c). Louire, B. 2016. Syrian Shadows behind the Back of Cyril and Methodius. Vaillant-Jakobson's Hypothesis Revisited. Draft v.3.0. Accessed at: https://www.academia.edu/8721348/Syrian_Shadows_behind_the_Back_of_Cyril_and_Methodius._Vai llant-Jakobsons_Hypothesis_Revisited._Draft_v.3.0 (accessed 02.12.2016).
Lourie B. (forthcoming a): Lourie, B. 2015. The Slavonic Solunskaja Legenda ("The Thessalonican Legend") and Its Syriac Original. In Horn C. B., Villagomez C. (eds). The Syriac Voice in the Dialogue of Cultures: Syriac, Persian, Caucasian, and Slavonic Interlocutors, Eastern Mediterranean Texts and Contexts. Warwick: "Abelian Academic".
Lourie B. (forthcoming b): Lourie, B. 2017. Slavonic Pseudepigrapha, Nubia, and the Syrians. In Nicklas Th. , Moss C., Tuckett Ch., Verheyden J. (eds.). The Other Side: Apocryphal Perspectives on Ancient Christian "Orthodoxies". Novum Testamentum et Orbis Antiquus 117. Göttingen: 'Vandenhoeck & Ruprecht" Publ.
Makeeva, L. B. 2014. An Introduction to Inductive Logic. Moscow: "Idea-Press Publishers".
Millius, J., Kuserus, L. 1710. Novum Testamentum Graecum cum lectionibus variantibus. Amsterdam: "J. F. Gleditsch et fil".
Mink, G. 2004. Problems of a highly contaminated tradition: the New Testament. Stemmata of variants as a source of a genealogy for witnesses. In van Reenen P., den Hollander A., van Mulken M., Roeleveld A. (eds.). Studies in Stemmatology II. Amsterdam; Philadelphia: "J. Benjamins Publ. Co.", 13—85.
Monferrer-Sala, J. P. 2013. An Early Fragmentary Christian Palestinian Rendition of the Gospels into Arabic from Mär Säbä (MS Vat. Ar. 13, 9th c.). Intellectual History of the Islamicate World 1, 69—113.
Monferrer-Sala, J. P. 2015. The Pauline Epistle to Philemon from Codex Vatican Arabic 13 (Ninth Century CE). Transcription and Study. Journal of Semitic Studies 60, 341—371.
Nedeljkovic, O. 1972. Problem stukturnih redakcija staroslavenskog prijevoda Apostola [Nedeljkovic, O. The Problem of Structural Recensions of the Old Slavic Translation of the Apostolos (in Croatian)]. Slovo 22, 27—40.
Payne Smith, J. 1903. A Compendious Syriac Dictionary, Oxford: "Clarendon Press".
Payne Smith, R. 1879. Thesaurus Syriacus. Vol. I—II. Oxford: "Clarendon Press".
Platt, T. P. 1830. Novum Testamentum Domini Nostri et Salvatoris Jesu Christi Aethiopice. London: "R. Watts".
Robinson, J. M. 2013. The Story of the Bodmer Papyri: From the First Monastery's Library in Upper Egypt to Geneva and Dublin. Cambridge: "James Clarke & Co.".
Scholz, I. M. A. 1836. Novum Testamentum Graece. Vol. II. Leipzig: "F. Fleischer".
Sinai MF UCL Arabe 151. Accessed at: http://www.e-corpus.org/notices/99936/gallery/1025829 (accessed 02.12.2016).
Spencer, M., Wachtel, K., Howe, Ch. J. 2002. The Greek Vorlage of the Syra Harclensis: A Comparative Study on Method and Exploring Textual Genealogy. A Journal of Biblical Textual Criticism 7. Accessed at: http://rosetta.reltech.org/TC/v07/SWH2002/ (accessed 02.12.2016).
Staal, H. 1983. Mt. Sinai Arabic Codex 151. I. Pauline Epistles. Corpus Scriptorum Christianorum Orientalium. Vol. 452—453. Scriptores arabici. T. 40—41. Louvain: "Peeters".
Stenij, E. 1901. Die altarabische Übersetzung der Briefe an die Hebräer, an die Römer und an die Korinther aus einem in St. Petersburg befindlichen Codex Tischendorfs vom Jahre 892 n. Chr. zum ersten Male herausgegeben. Helsingfors: "Frenckellska Tryckeri-Aktiebolaget".
Strakhov, O. B. 2015. Cyrillomethodiana and the Vulgate: Luke 24:43 in the Byzantine, Latin, and Early Slavic Traditions. In Diddi C. (ed.). nOAYILTQP. Scripta slavica Mario Capaldo dicata. Moscow; Rome: "Indrik" Publ., 297—315.
Sundermann, W. 1974. Nachlese zu F. W. K. Müller "Sogdischen Texten I". 1. Altorientalische Forschungen 1, 217—255.
Sundermann, W. 1975. Nachlese zu F. W. K. Müller "Sogdischen Texten I". 2. Altorientalische Forschungen 3, 55—90.
Sundermann, W. 1981. Nachlese zu F. W. K. Müller "Sogdischen Texten I". 3. Altorientalische Forschungen 8, 169—225.
Tedros Abraha. 2001. La Lettera ai Romani. Testo e commentari della versione Etiopica Aethiopistische Forschungen 57. Wiesbaden: "O. Harrassowitz" Publ.
Tedros Abraha. 2014. The Ethiopic Versions of 1 and 2 Corintians. Rome: [s.n.].
Tischendorf, C. 1872. Novum Testamentum Graece. Vol. II. Leipzig: "Giesecke & Dervient".
TITUS. Text collection: NT. Novum testamentum. Accessed at: http://titus.uni-frankfurt.de/texte/etcs/ arm/zohrab/armnt/armnt.htm (accessed 02.12.2016).
Uhlig, S., Maehlum, H. 1993. Novum Testamentum Aethiopice: Die Gefangenschaftsbriefe. Aethiopistische Forschungen 33. Stuttgart: "Steiner".
van Esbroeck, M. 1998. Les versions orientales de la Bible: une orientation bibliographique. In Krasovec J. (ed.). The Interpretation of the Bible. The International Symposium in Slovenia. Journal for the Study of the Old Testament. Supplement Series 289. Sheffield: "Sheffield Academic Press" Publ., 399—509.
Veder, W. 2014. Why Wish Away Glagolitic? Ricerche slavistiche 12(58), 373—385.
von Soden, H. F. 1913. Die Schriften des Neuen Testaments in ihrer ältesten erreichbaren Textgestalt hergestellt auf Grund ihrer Textgeschichte. II. Teil. Text mit Apparat. Göttingen: "Vandenhoeck und Ruprecht".
Walton, B. 1657. Biblia SacraPolyglotta. Vol. 5. London: "Thomas Roycraft".
Weninger, S. 2003. Bible. Gecez Bible Editions. In Uhlig S. (ed.). Encyclopaedia Aethiopica. Vol. 1. Wiesbaden: "O. Harrassowitz" Publ., 569—571.
wp.unil.ch: 1: Sara Schulthess. An Arabic Manuscript of Pauline Letters: Vaticanus Arabicus 13. Accessec at: http://wp.unil.ch/nt-arabe/files/2015/04/Louvain2015 -pdf.pdf (accessed 02.12.2016).
www.vatlib.it: 1: BAV — Vatikan Library. Accessed at: https://www.vatlib.it (accessed 02.12.2016).
Zuurmond, R. 2003. Rev. of Tedros Abraha 2001. Aethiopica 6, 253—256.
Zuurmond, R., revised by Niccum C. 2013. The Ethiopic Version of the New Testament. In Ehrman B. D., Holmes M. W. (eds). The Text of the New Testament in Contemporary Research. Essays on the Status Quaestionis. Second Edition. New Testament Tools, Studies and Documents. 42. Leiden; Boston: "Brill", 231—252.
Алексеев, А. А. 2013. Соображения о критическом издании славянского Апостола. [Alekseev, A. A. 2013. Consideration on the Critical Edition of the Slavonic Apostolos (in Russian)]. B: Бобрик М. (ред.). Славянский Апостол. История текста и язык [In Bobrik M. (ed.). The Slavonic Apostolos: the History of the Text and the Language (in Russian)]. Studies on Language and Culture in Central and Eastern Europe 21, 197—206.
Бобрик, М. А. 2013. Аннотированная библиография работ по славянскому Апостолу. [Bobrik, M. A. 2013. An Annotated Bibliography on the Slavonic Apostolos (in Russian)]. B: Бобрик М. (ред.). (ред.). Славянский Апостол. История текста и язык [In Bobrik M. (ed.). The Slavonic Apostolos: the History of the Text and the Language (in Russian)]. Studies on Language and Culture in Central and Eastern Europe 21, 209—273 (in Russian)].
Воскресенский, Г. 1879. Древний славянский перевод Апостола и его судьбы до XV в. Опыт исследования языка и текста славянского перевода Апостола по рукописям XII—XV вв. Москва: Университетская типография [Voskresenskij, G. The Old Slavic Translation of the Apostle and Its Destiny up to the fifteenth century: An Essay in the Study of the Language and the Text of the Slavic Translation of the Apostle according to manuscripts of the twelfth to fifteenth centuries. Moscow: "The University Typography" Publ. (in Russian)].
Воскресенский, Г. 1892. Древнеславянский Апостол. Послания святого апостола Павла по основным спискам четырех редакций рукописного славянского апостольского текста с разночтениями из пятидесяти одной рукописи Апостола XII—XVI вв. Вып. I. Послание к римлянам. Сергиев Посад: 2-я типография А. И. Снегиревой [Voskresenskij, G. The Old Slavic Apostle. The Epistles of St. Apostle Paul according to the Main Manuscripts of the Four Recensions of the Manuscript Slavonic
Text of the Apostle with Variant Readings from Fifty-One Manuscripts of the Apostle of the XII—XVI centuries. Iss. I. The Episte to the Romans]. Sergiev Posad: "2-ja tipografija A. I. Snegirevoj" Publ. (in Russian)].
Воскресенский, Г. 1906. Древнеславянский Апостол. Послания святого апостола Павла по основным спискам четырех редакций рукописного славянского апостольского текста с разночтениями из пятидесяти одной рукописи Апостола XII—XVI вв. Вып. 2. Послание святого апостола Павла к Коринфянам 1-е. Сергиев Посад: Свято-Троицкая Сергиева лавра [Voskresenskij, G. The Old Slavic Apostle. The Epistles of St. Apostle Paul according to the Main Manuscripts of the Four Recensions of the Manuscript Slavonic Text of the Apostle with Variant Readings from Fifty-One Manuscripts of the Apostle of the XII—XVI centuries. Iss. II. The Epistle of the Holy Apostle Paul to the Corinthians First. Sergiev Posad: "Svjato-Troickaja Sergieva lavra" Publ.]
Воскресенский, Г. 1908. Древнеславянский Апостол. Послания святого апостола Павла по основным спискам четырех редакций рукописного славянского апостольского текста с разночтениями из пятидесяти одной рукописи Апостола XII—XVI вв. Вып. III—V. Послания святого апостола Павла къ Коринфянам 2-е, к Галатам и к Ефесянамъ. Сергиев Посад: Свято-Троицкая Сергиева лавра [Voskresenskij, G. The Old Slavic Apostle. The Epistles of St. Apostle Paul according to the Main Manuscripts of the Four Recensions of the Manuscript Slavonic Text of the Apostle with Variant Readings from Fifty-One Manuscripts of the Apostle of the 12th—16th centuries. Iss. III—V. The Epistles of the Holy Apostle Paul to the Corinthians Second, to the Galatians, and to the Ephesians. Sergiev Posad: "Svjato-Troickaja Sergieva lavra" Publ.]
Гауптова, З. 2013. К вопросу о соотношении служебного и дополнительного текстов старославянского Апостола [Hauptova, Z. 2013. To the Question on the Correlation between the Liturgical and Additional Texts of the Old Slavic Apostolos (in Russian)]. B: Бобрик М. (ред.). Славянский Апостол. История текста и язык [In Bobrik M. (ed.). The Slavonic Apostolos: the History of the Text and the Language (in Russian)]. Studies on Language and Culture in Central and Eastern Europe 21, 11—30.
Клеминсон, Р. 2013. Количественный метод в текстологии (на материале Соборных посланий) [Cleminson, R. 2013. The Quantitative Method in the Textology (on the Material of the Synodal Epistles) (in Russian)]. B: Бобрик М. (ред.). Славянский Апостол. История текста и язык [In Bobrik M. (ed.). The Slavonic Apostolos: the History of the Text and the Language (in Russian)]. Studies on Language and Culture in Central and Eastern Europe, 21, 31—61.
Лурье, В. М. 2014. Солунская легенда: оригинальное славянское произведение или перевод с сирийского? Количественный подход к оценке правдоподобности альтернативных гипотез. МАИАСК 6, 190—215 [Lourie, B. 2014. In Materialy po arheologii i istorii antichnogo i srednevekovogo Kryma. Arheologija, istorija, numizmatika, sfragistika i jepigrafika = Materials in Archaeology and History of Ancient and Medieval Crimea. Archaeology, History, Numismatics, Sigillography and Epigraphy 6, 190—215].
Мирчев, К., Ходов, Хр. 1983. Енински Апостол. Старобългарски паметник от XI в. София: БАН [Mirchev, K., Khodov, Khr. Praxapostolus versionis palaeobulgaricae ad fidem codicis Eninensis saeculoXIscripti. Sofia: "Bulgarian Academy of Sciences" Publ. (in Bulgarian)].
Пентковская, Т. В. 2005. Пандекты Антиоха в славянском и арабском переводе: опыт сопоставительной характеристики [Pentkovsakaya, T. V. 2005. The Pandekta of Antiochus in Slavonic and Arabic Translations: An Essay of Comparative Characterisation (in Russian)]. Преславска книжовна школа [The Preslav Literary School] 8, 270—281 (in Bulgarian)].
Пентковская, Т. В. 2009. К истории исправления богослужебных книг в Древней Руси в XIV веке: Чудовская редакция Нового Завета. Москва: Московский государственный университет [Pentkovskaya, T. V. 2009. Towards the History of Correcting the Liturgical Books in the Old Rus' in the XIV century: the Chudov Recension of the New Testament. Moscow: "Moscow State University" Publ. (in Russian)].
Пентковская, Т. В. 2015. Славянский Апостол: история текста и язык. Составитель Марина Бобрик. Труды Института русского языка им. В. В. Виноградова 5, 415—422 [Pentkovskaya, T. V. 2015. In Proceedings of the V. V. Vinogradov Russian Language Institute 5, 415—422].
Славянский Апостол. История текста и язык. 2013. B: Бобрик М. (ред.). [The Slavonic Apostolos: the History of the Text and the Language. 2013. In Bobrik M. (ed.). (in Russian)]. Studies on Language and Culture in Central and Eastern Europe 21 (in Russian)].
Христова-Шомова, И. 2004. Служебният Апостол в славянската ръкописна традиция. Т. I. Изследване на библейския текст. София: Университетско издателство «Св. Климент Охридски» [Hristova-Shomova, I. 2004. The Liturgical Apostolos in the Slavonic Manuscript Tradition. Vol. I. A Study of the Biblical Text. Sofia: "Sofia University «St. Klement Ohridski»" Publ.]. ¿5^(,b9g, o. 1973. dgj(£>o jd&m^^o j6ob (jfoojcaSo a>5o(obo: 9д^бод6д55 [Abuladze, I., A
Dictionary of the Old Georgian Language. Tbilisi: "Metsniereba" Publ. (in Georgian)]. 9ro^g6o9g, j., 3. 1974. 3^3(Qb j3ob(ca(ja>d jd&m<£)((o gj&bojbo, ¿30(o
g6ob 35OTg^6ob 36ro9g5o, 16; rn5o(obo: ot5o(. ЬбЬд(9^. ^6-(ob £¿9-55 [Dzotsenidze, K. Danelia, K. The Georgian Versions of the Epistles of Paul. Proceedings of the Chair of the Old Georgian Language 16. Tbilisi: "Tbilisi State University" Publ. (in Georgian)].
