Redaction in Old Slavic Texts. The Paradosis of O Pismenex Text Family γ before ca. 1300

Krystyna Szczesniak, Halina Wa˛tróbska (eds.) Tematy. Ksie˛ga jubileuszowa w 70. rocznice˛ urodzin professora Leszka Moszyn´skiego. Gdansk: Wydawnictwo Uniwersytetu Gdan´skiego 1998: 432–450

REDACTION IN OLD SLAVIC TEXTS The Paradosis of O Pismenex Text Family g before ca. 1250

The paper presents a reconstruction of the text of the Old Slavic treatise on the origin and the status of the Glagolitic alphabet as transmitted by its text family g before ca. 1250 (14 MSS, including the only MS known to be preserved in Poland). It examines the role of accidence vs. editorial intervention in the constitution and subsequent development of this text and evaluates the linguistic competence of its author and its transmitters. It argues that all manuscripts are equivalent, but that none can be equated with the text as such. The treatise O Pismenex=1 explains, within the space of one quire in Glagolitic script, the features and the history of the Glagolitic alphabet and the Bible translations written in it, and argues their greater holiness by comparison to the Greek. It is thought to have been composed between 893 and 921,2 partly on the basis of Greek grammatical treatises (Ps.-Theodosius, scholia to Dionysius Thrax3). Its author has been identified as a monk named Hrabr, but the attribution cannot be substantiated.4 In earlier papers I have examined its text families b,5, a6 and e7 and found their Cyrillic archetypes to represent independent witnesses of the Glagolitic original; in spite of the latter being called a ‘Russian redaction’,8 I found it to contain no traces of either a Russian accent or an editorial intervention in the constitution of its text.

Redaction does play a part in the text family g,9 but to an extent yet to be established with precision.10. The family text is preserved by 14 MSS: Ch 1400-1500, Russian: Sbornik Moskva, GIM, Cµudov. 269 (Kuev 1967:197) Kt 1650-1700, Russian: Sbornik Kostroma, GAKO, F.558, op.2, 231 (Kuev 1967:355) Mol 1500-1600, Bulg./Mold.: Sbornik Athos, Hilandar 481 (Kuev 1967:210) P1 1600-1700, Russian: Sbornik S.-Pb., RNB: Pogod. 1297 (Kuev 1967:283) P2 1600-1700, Russian: Sbornik S.-Pb., RNB: Pogod. 1300 (Kuev 1967:286) Pl 1625-1675, Russian: Sbornik S.-Pb., BAN, 21.5.15 (Pligin 57) (Kuev 1967:316) Ps 1600-1700, Russian: Sbornik Moskva, RGB, F.228 (Pisk.) 150 (Kuev 1967:275) Rzh 1400-1500, Russian: Sbornik Moskva, RGB, F.274 (Rogoz¬sk.) 253 (Trembovol'skij 1989) Sh 1625-1675, Russian: Sbornik S.-Pb., RNB, F.344 (Sµiban.) 193 (Kuev 1967:264) T 1600-1700, Russian: Apocalypse Moskva, RGB, F.299 (Tixonr.) 5 (Kuev 1967:246) U1 1 600-1625, Russian: Chronograph Moskva, GIM, Uvar. 16 (Kuev 1967:349)

1 Fundamental study of the text and publication of transcripts of the MSS: Kuev 1967; additions by Damjanova 1978, Jazykova 1979, Kuev 1988, Trembovol’skij 1989 and Kossova 1980. The latter attempts to establish the relationship of the MSS among each other and to reconstitute the author's text; alternative reconstruction by Marti 1981. 2 Canev 1994, ch. 2, Trembovol’skij 1985: 6 proposes a dating before the advent of the Cyrillic alphabet. 3 On the Greek sources cf. most recently Ziffer1995a: 60-62. 4 Canev 1994, ch. 4, following Hanus¬ 1859 and Il’inskij 1915, points out that in all MSS save one (text family s) the G Cµernorizca Hrabra modifies pismena and, therefore, indicates the authorship of the Slavic alphabet, not of the text. 5 Veder 1996a. 6 Veder 1996b. 7 Veder forthcoming 1. 8 Jazykova 1979. 9 In my earlier papers I considered the senior hyparchetypes d and k to represent different text families, but now have to concede that Kossova 1980 is right in deriving them from a common Cyrillic archetype, as the concordance of their readings in contrast to b, e or a proves. In my earlier studies of the text (notes 5-7), I proposed to distinguish 11 different text families, but these must now be reuced to no more than ten. 10 The extent of editorial intervention has so far been established satisfactorily only for text families b and e.

2 – TEXT FAMILY g [122]

Vl1 1500-1600, Russian: Paleja Moskva, RGB, F.113 (Volok.) 551 (Kuev 1967:221) Vl2 1500-1600, Russian: Sbornik Moskva, RGB, F.113 (Volok.) 573 (Kuev 1967:225) Wr 1500-1600, Ukr./Mold.: Trebnik WrócΩΩaw, BU, 318 (Kuev 1967:214) Tanja Ivanova (Sofia) made machine-readable recordings of the MSS and subjected them to a global collation under radical reduction of the number of graphemes (a q (→ a, e & :→ e, j jd → j, z ™ → z, i " )" |" → i, o w → o, u ou ' \ £ → u, ) | § → §, superscripts → ø). This revealed highly distinctive individual features in three groups of MSS,11 Vl1 Vl2 = h, T Pl Sh = l and Ps P2 Kt = m. For each group we established the paradosis (West 1965:53f.) by reconstructing the text of its hyparchetype according to the following general rules: 1. readings are not restricted to any specific linguistic level, but if necessary selected even down to the level of single graphemes; 2. concordant readings of all MSS are assigned to the hyparchetype; 3. if readings diverge, to the hyparchetype are assigned those that are less likely to be the result of contraction, corruption or obliteration, i.e. as a rule the more complete or correct reading; 4. if readings conflict, a reading is sought to explain and resolve the conflict. Repeated recollation of these reconstructions with the MSS at a lesser degree of abstraction revealed that l and m were, in fact, only witnessed by T and Ps, respectively, while the other two MSS of each group were separated from them by the junior hyparchetypes lí and mí, respectively. When these had been reconstructed and differentiated from l and m, recollation of the hyparchetypal texts with the MSS left unclassified revealed the following: i. the text selections transmitted by Rzh could be united with m, ii. l and m could be united in a senior hyparchetype k, iii. Ch U1 P1 could be united in an hyparchetype q,12 iv. Mol Wr could be united in an hyparchetype z, and v. z and q could be united with h in a senior hyparchetype d. Finally, a collation of both d and k with the reconstructed text of the Cyrillic archetypes b, a and e led to the conclusion that they could not be separated from one another but were both derived from a common Cyrillic archetype g, distinct from the other three Cyrillic archetypes established so far. After this inductive step by step reconstruction, we put the resulting model of text development to the reverse test in order to evaluate the plausibility of derivations13 and to correct the reconstructions wherever necessary. After that, I established the following stemma (see next page):14

d z q and h present essentially faithful renditions of the g text, q being distinguished from z mainly by its elimination of the etymological nasals and :, and h by the lack in its antigraph of a bifolium containing the text 8:22-14:1. The MS Mol shows traces of contamination with a MS belonging to text family a.15 k as well as its junior hyparchetypes l and m, by contrast, overlay over the text of their respective antigraphs editorial interventions in t:1, 10:4-11:4 and 14:17, 26. The MS Rzh contains a selection only from the m text.

950 1100 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700

11 Already established by Kossova 1980: 41-45, whose sigla I use (z q are introduced by me). 12 There are no arguments to derive the MSS U1 P1 directly from Ch, as Kossova 1980: 74 does; the preservation in the MSS P1 U1 of of 7:3 me¢e and 9:1 bo, as well as of many -| desinences rather demonstrates their independence from Ch. 13 At the same time I verified the probability of coincident variation and contamination in the transmission. 14 Shaded = not attested; the terminus quo ante of the reconstructed hyparchetypes (which may represent more than a single MS) is arbitrarily set at 50 years before the earliest date of the oldest MS or the junior hyparchetype, respectively. 15 The contamination was recorded by Kossova 1980: 38-39, but imprudently related directly to the b MS M. In fact, the non-g readings of Mol can all be traced to either an a or a b MS; of the double readings, which are most indicative of contamination, 14:26 kwcel( + kwstel( can be traced either to a or b , while 1:2 k¢iŸ + pismeì can only be traced to a, as noted by Ziffer (who uses the siglum b for a), but imprudently interpreted to mean that Mol is primarily a witness of a, not g.

[122] TEXT FAMILY g – 3

Mol [B] * z Wr [U] Ch [R] d q U1 [R] P1 [R] Vl1 [R] h Vl2 [R]

w g T [R] l Pl [R] lí Sh [R] k Rzh [R] Ps [R] m P2 [R] mí Kt [R]

* Contaminated with an a MS

Stemma of Text Family g

Below, I present the reconstructed text of g, highlighting its main differences from e, b and a in a commentary to the unit header. Wherever k, l or m deviate from it by editorial intervention, I print their text in parallel. Comments as to non-editorial developments of the text are given in footnotes (in principle not to single MSS, but to hyparchetypes only). t:1* b slovo w k¢igah|, a o pisme¢eh) ~r|¢oriz‘ca hrabra

g SKAZAÚIE16 l O AZ‘B3K/ SLOVEÚ‘SKOI. I O ElIÚ‘SKOI. SIR/^6 GRE^ESKOI. SKAZAÚ2E

2-3 1-3b kako (zyk| v|z(l| est| pisa¢ie k¢igami, a om, 1-2e filosof| az|b'k'. po (z|"k' slove¢|sk', e add t:4-6 i k‘¢ig|" pr:vede. ot| gr|~|sk|"h) ¢a slove¢|skii (zyk| KAKO S7STAVI ST–62 KÒRIL7 1SLOV/ÚOM7 PISMEÚA PROTIV32 4Z62K3.17 1:1-4 1b pisa¢iq, a k¢ig), 2b r:j|mi, a r:za¢‘mi, e anticipates 1:5 before 1:2 Pr:jde 'bo slov:¢e ¢e im:(h\ 1pisme¢). ¢) ~r)tami i 2¢ar:za¢|mi ~|t(h\. i gadah\18

5-7 1a delays ¢\jdaah\ s( after pisme¢|", 2b k¢igami ee s\e poga¢i. kr|stiv‘e je s(. 1¢\jdaah\ s( rim|sk|"mi i gr|~|sk|"mi 2pisme¢|"19

8-10 1e b add slov:¢|sk|" pisati slov:¢|sk\ r:~|. bez 'stroe¢i(. ¢) kako mojet| 1s( pisati dobr:.20

16 h omits the title. 17 k punctuates after kiril) (eliminating, as elsewhere, £) and pisme¢a, l corrupts (z|"k' → qz|"ka. 18 d contracts, as elsewhere, the ipf suffix :( → (, k and h independently eliminate the : in slov:¢e (q m read slove¢:) and harmonize ~r)tami → ~r)ta¢|mi to fit ¢ar:za¢|mi, k and h independently eliminate, as elsewhere, \. h punctuates after slove¢e, ~erta¢mi and ¢ar:za¢mi. 19 k abbreviates the ipf suffix, h punctuates after pisme¢|". 20 h m trivialise slov:¢|sk\ → slov:¢|sk'<, k omits the punctuation of 1:9, l omits s( pisati.

4 – TEXT FAMILY g [122]

11-17 e b add 1:16 ~a(¢ie gr|~|sk|"mi pisme¢|". b)–. ili jivot). ili ™:lo. ili cr–kv|. ili ~l–k).21

18-23 a b om edrot|" and put the (-word in the last place, e b a put the <-word after the \-word ili irota. ili edrot|". ili <¢ost|. ili \d\. ili (z|"k). ili qd|.22

24-25 1a b:( i i¢a podob¢a sim). i tako 1b|"( m¢oga l:ta.23

2:1-3 1e add ¢a pol|™\ Po tom) je ~l–kol<b|c| b)–. stro(i 1v)s(. i ¢e ostavl(( ~l–~a roda bez raz'ma.24

4-5 1-2a privod(i spase¢i<, 3e a pomilovav| ¢) v)s( ¢a raz'm| 1privod( i 2spase¢ie. 3pomilova ro∂ slove¢esk|.25

6-7 1e poslav|, 3e add v) m‘¢ieskom| ~i¢' i 1posla im) st–go ko¢|sta¢ti¢a filosofa. 2¢aricaemago k£ril‘la.26

8-9 1e b st–a, 2b slov| m\ja praved¢a i 1isti¢‘¢a. i s)tvori im) 2pisme¢). l–. i. i–.27

10-11 ova 'bo po ~i¢' gr|~|sk|"h) pisme¢). ova je po slove¢|st: r:~i.28

12 1e gr|~|sky w= pr)va je ¢a~e¢) po 1gr|~|sk'.29

3:1-3 1-2e ¢a~(t)k| oboego edi¢|, b om 3:1-7:14 o¢i 'bo alfa. a s|i az). 1w= aza ¢a~(t) 2oboe.30

4-5 1-2e takoje i o¢) i qkoje o¢i. podobl|e s( jidovsk|"m) pisme¢em). s)tvori(. 1tako i 2s) gr|~|sk|"m).31

4:1-3 1e skaz'et|, 2a ou~i¢e¢ie jidove bo pr|voe pism( im\t). al‘f|. eje s( 1skazaet| 2'~e¢ie. s)vr|a\e32

3-5 1e qkoje ~, 2e add '~iti s(, 3e gl–<t|, 4e a eje 1v)vodimom' 2d:ti'. i 3gl–\e. '~i s(. 4se est| alf|.33

21 k omits 1:11, l eliminates, as elsewhere, |" after velars and misreads ™:lo → ™lo. 22 g d k had badly legible nasals in \d\: z and m read (d', l oud(. h omits the punctuation of 1:21. 23 h trivialises i¢a podob¢a → i¢a( podob¢a(, l → i¢a podob¢a( (harmonized by lí → i¢a( podob¢a(), d and independently m trivialise b|"( → b:a, lí misreads m¢oga → m"r¢a(. 24 k and independently q h trivialise stro(i → stro(, d innovates, as elsewhere, initial o → w. Mol reads ~l–~a roda slov:¢ska, of which slov:¢ska is attested as an independent anticipation of 2:5 in many MSS of different text families without any traces of contamination. 25 m anticipates the punctuation of 2:4 before i. 26 h capitalizes the initial i, h and independently lí m change ko¢|sta¢ti¢a → ko¢|st(¢ti¢a, lí omits st–go. m punctuates after im). 27 h and independently m Ch write the number in full. h punctuates after im). 28 All MSS save Ch Sh Ps trivialise slove¢|st: → slove¢|st:i. 29 h and independently U1 Pl trivialise pr)va → per|vago, l trivialises gr|~|sk' → gre~eski, m → gre~eskom'. 30 k vocalises s|i → sei. m punctuates after 'bo and sei. 31 k trivialises s) → s)i, m s( → je s(, l and independently Wr corrupt pisme¢em) → pisme¢om). h and independently l P1 omit the punctuation after o¢i, d and independently m that after sotvoria. 32 m omits the punctuation before al‘f|, lí before s)vr|a\e, h and independently Mol punctuate after s)vr|a\e. 33 l omits se est|. m omits the punctuation after d:ti', z and independently l that after gl–\e, h punctuates before alf|.

[122] TEXT FAMILY g – 5

6 1e a podob(e i gr|ci 1podobl|e s( tom'. alfa r:(.34

7 1-2e spodobi(, 3-4e gre~eskom' i 1spodobi 2s( re~e¢ie skaza¢i( jidov|ska. 3gr|~|sk' 4(z|"k'.35

8 da re~et| d:ti' v) '~e¢i( m:sto. ii.36

9 1e om, 2-3a s( re~et|, 4e add eje est| ii al‘fa bo. 1ii. 2gl–et 3s( gr|~|sk|"m) 4(z)"kom|.37

5:1-2 t:m| bo podob( s( st–|"i k£ril). s)tvori. pr|voe pism(. az).38

3-4 1e a pr|vom', 2e az' , 3a da¢ou ¢) (ko i 1pravom' s\' pisme¢i. 2az). i w= b–a 3darova¢' rod' slove¢|sk'.39

5-6 1e add d:tem|, e add 5:7 to &st) statikvam) ¢a w=vr|ze¢ie 'st). v) 1raz'm) '~aim‘ s( b'kvam).40

8 velik|"m) razdvije¢iem) 'st| v)zglasit| s(.41

9-10 1e ispov:d(t) a o¢a pisme¢a malom| razdvije¢iem) oust| v)zglas(t‘ s( i 1pov:d(t) s(.42

6:1-2 1a (, 2e a gl–ati Se je s\t| pisme¢a slove¢|ska. sice 1ih) podobaet) pisati i 2glaati.43

3 1e add g d, a add g, 2-3e i pro~a( a. b. 1v. 2daje i do 3(.44

4 1-2e si( je, 3-4e om, a om 6:4-8 1i w= 2sih) s\t| 3~et|"re mejd' 4des(t|ma podob¢a gr|~|sk|"m) pisme¢om). 3s\t| je 4si.45

5 a. v. g. d. e. z. i. #. ". k. l. m. ¢. ¶. o. p. r. s. t. u. f. h. $. w.46

34 m trivialises tom' → tom' je. 35 k omits the punctuation after jidov|ska. 36 k omits the punctuation after m:sto, h after ii. 37 l anticipates the punctuation before bo, z and independently q k omit the punctuation before gl–et s(. 38 l capitalises T:m|, l and independently Wr Ch omit bo. z and independently k P1 V2 omit the punctuation before pr|voe. 39 k omits i, z and independently q k trivialise az) → az', l and independently Wr trivialise darova¢' → darova¢¢', d and independently m trivialise slove¢|sk' → slove¢|skom'. m omits the punctuation after az', l and independently Wr U1 P1 after slove¢|sk'. 40 lí and independently U1 omit the punctuation after oust). 41 z and independently k trivialise v)zglasit| → v)zglas(t| (by anticipation from 5:9), Wr U1 V1 independently jump to 5:10. 42 d and independently m trivialise pov:d(t) → pov:d'\t (z → spov:d'\†, T Pl → pov:da<t). q h punctuate after pisme¢a. 43 q and independently h m Mol trivialise slove¢|ska → slove¢ska(. 44 d and independently m omit i. 45 l abbreviates ~et|"re → d–, z and independently k trivialise des(t|ma → des(toma (m further → des(tom'), h and independently lí m trivialise si → s"i. 46 d and independently l write digraphic ou, m omits $.

6 – TEXT FAMILY g [122]

6-7 1-2e qje s't| a ~et|"re ¢a des(te po slov:¢|sk' (z|"k'. 1ije s\t| 2si(.47

8 b. j. ™. c. ~. . . ). |". |. :. <. \. (. 7:1-2 1a drou™ii, 2e a om, 3e a est|... s)tvoril| 1Ú:cii je 2'bo gl–\t|. po ~|to. li–. pisme¢) 3s)tvoril).48

3 1a om, 2e me¢|i, a m|¢|im| mojet 1bo s( i 2m|¢e togo pisati.49

4-5 1-2e kd–.mi, a kd–, 3e ¢o, 4a om qkoje i gr|ci. 1dv:ma bo des(t|ma i 2~et|"r‘mi pi\t|. 3a ¢e v:d(t| 4izv:st¢o.50

6-7 1e om, 2e ' ¢ih| kolic:mi pi\t) 1gr|ci. est| bo 2'bo. kd–. pisme¢).51

8-9 1a ¢| ¢) ¢e ¢apl)¢(\t s( t:mi k¢ig|". 1i prilojili s\t| dvoglas¢|"h). a"–.52

10-11 1e ~isl:h| v| 1~isme¢eh) je tri. estoe. i dev(† des(t¢oe. i dev(† s)t¢oe.53

12-13 1e s)beret‘, 2e toi je, a t)jde i 1s)bira\t| s( ih). l–. i. i–. t:m|je tom' podob¢o. i v| 2t) wbraz)54

14 s)tvori st–|"i k£ril). l–. pisme¢) i. i–.55

8:1-2 1a Drou™ii, 2e om, 3a om, 4a gl–\t|, 5a ~esomou, 6e om, 7a b k¢igy 1S\t| 2je 3¢:cii 4gl–\e. 5~em' 6s\t| slov:¢|ska 7pisme¢a.56

3-5 1a add togo, 1-2b ¢:ß ih|, 3e s|tvori, a est|... s)tvoril|, 4a add to, 5e agg–l|, 6e zako¢|¢|", b isko¢i slov:¢|sky( k¢igy, a ijdeko¢‘¢i 1¢i bo 2t:h) b)– 3s)tvoril). 4¢i 5agg–li. ¢i s\t| 6ijdoko¢|.57

6-7 1e ije ot| zako¢a, b (je isko¢i, a ijde w= ko¢a, 2e b pr"(t‘¢y qko jidov|sk|". i el‘li¢|sk|" 1iz¢a~(la s\t|. i 2pri(t|" s\t| bg–m|.58

8-9 1e takov"i, 2e edi¢:m| im|, a ¢am|, 3e s)tvori, 4b pisa¢iq a 1dr'™i m¢(t). qko bg–) 2sam) est| 3s)tvoril) 4pisme¢a.59

47 l abbreviates the number, q and independently h m Mol trivialise slov:¢|sk' → slove¢‘skom'. z q omit the punctuation after slove¢‘skom'. 48 z and independently P2 omit je, l omits 'bo, d and independently m write the number in full. 49 h and independently Mol Ch U1 trivialise m|¢e → me¢ee (h further → me¢:e, m independently → me¢e(, Wr → meìe). 50 l abbreviates the number. d and independently l omit the punctuation after izv:st¢o. 51 h and independently U1 Pl omit bo, h and independently m Wr Ch write the number in full. 52 m trivialises dvoglas¢|"h) → dvo<glas¢|"h). 53 l abbreviates the numbers. 54 q and independently h m write the number in full (q h retaining the punctuation), m capitalises T:mje, l adds i to tom', d and independently l mí trivialise t) → toi. 55 h and independently m write the number in full, l reduces it to li–. 56 d and independently mí trivialise gl–\e → gl–<ei, l ~em' → k ~em'. 57 l replaces bo → bg–) and trivialises sotvori¬ → sotvori¬ est| and ijdoko¢| → izjeko¢‘¢a, d trivialises ijdoko¢| → izjoko¢| (z further → isko¢i). lí omits the punctuation after sotvori¬ est|, lí q after agg–li, q h after izjoko¢|. 58 m trivialises jidov|sk|", el‘li¢|sk|" → jidov|sk|"i, elli¢‘sk|"i, k trivialises pri(t|" → pri(t¢|". 59 z and independently h k trivialise dr'™i → dr'™ii, k trivialises est| s)tvoril) → sotvoril) est|, l omits sam) and the punctuation after pisme¢a.

[122] TEXT FAMILY g – 7

10-12 1e b om, 2e b gl–<t|, e b a add oka(¢|¢"i, 3e b a om, 4e add t)k)mo, 5e povel:, 6e add r:( i ¢e v:d(t| 1s( ~to 2gl–e. qko tr|mi 3li 4(z|"k|" b)– est| 5povel:l) k¢igam) 6b|"ti.60

13-15 b om 8:13, 1e gl–et| s(, 2e ije, 3e a b:, 4e tit)la, 5e add ¢a kr–st: g¢–i, 6e evr:isky., 7e b lati¢|sky ,8e b gre~esky qkoje v) e£aŸlii 1piet|. 2i 3b:(e 4d|ska 5¢apisa¢a. 6jidov|sk|". 7rim|sk|". i 8el‘li¢|sk|". 16-17 1e a ¢:st|, 2e i togo radi gl–<t|, b togo radi, 3e b a k‘¢igy a slove¢|sk| 1¢e b: t'. 2t:m)je ¢e s\t| slove¢|ska 3pisme¢a w= b–a.61

18 1e v)zgl–em|, 2e a tac:m|, 3e b bez'm‘¢ym| k) t:m| ~to 1gl–em|. ili ~to re~em|. k) 2takov|"m) 3bez'miem).62

19-20 1b a om, 2e add k‘ ¢im|, 3-4e st–ago pisa¢"(, 4b a k‘¢ig| oba~e 1da 2re~em| w= 3st–|"h) 4pisa¢ii. qkoje ¢a'~ihom‘ s(. 21-22 1-2e v|zmoj‘¢a s't|, 3-4a i¢ogdo\, 4b ~l–k|, e add 8:22b-27 i i¢o. qkoje i $alom¢ik| gl–et|. hvalite g–a v‘si (zyci. i pohvalite ego v‘si l<d"e. a ¢e edi¢:mi tremi pisme¢y. i (zyky. qkoje o¢i bas¢oslov(t|. i qko v)s( 1po r(d' 2b|"va\t| w= b–a. a ¢e 3w= 4i¢ogo.63

9:1-3 1-2e ¢e s|tvori 1¢:st| bo 2s)tvoril) bg–) pr:jde jidov|ska (z|"ka. ¢i rim|ska. ¢i elli¢|ska.64

4-6 1b rouskyi, 2e add toi s( (zyk| gl–ae ¢) 1sir|sk|. im)je adam) gl–a. i w= adama do 2potopa.65

7 1-2e po potop:, 3-4b do outvr|jde¢iq stl|pa i 1w= 2potopa do¢deje bg–) razd:li (z|"k|". 3pri 4stolpotvore¢ii. 8-10 b om 9:9, 1e om, 2-3b razd:li (zyky qkoje piet|. razm:e¢om) je 1b|"v)em) (z|"kom). i qkoje 2(z|"ci razm:si( 3s(.66

11-12 1b a ob|"~a(, 2e add ko(jdo, b add ky( g tako i ¢ravi. i 1ob|"~ai. i oustavi. i zako¢i. i h|"trosti 2¢a (z|"k|". k tako razd:lia s( i ¢‘ravi. i ob|"~ai. i 'stavi. i zako¢i. i hitrosti ¢a (z|"k|". 13 1e b a eg£p‘t(¢om|, 2b dast|, 3b a zemlem:re¢ie g 1eg£ptom) 2oubo 3zemlem:rie. k Qkoje se gl–< est| qz|"k) egipet|sk"i. em'je s( 'bo dostalo zemlem:rie.67

14 1b rousom| a per|som). i hald:om). i 1as£riom). 15-16 1b dast| ~, 2b zv:zdo~|te¢ie, 3b vl|hvova¢ie, 4e b om, 5e ~aro†vore¢"e, a ~arova¢ia, 6e b a v|s(, 7b hytrosti g 1,2™v:zdo~|tie. 3vl)ve¢ie. 4vra~eva¢ie. 5~arove. i 6v|s(ka 7h|"trost| ~l–~a. k marg. astro¢omie ™v:zdo~etie. vl)ve¢ie. vra~eva¢ie. ~arova¢ie. i vs(ka hitrost| ~l–~a.68

60 l omits s(. l and independently U1 P1 Ps Kt omit the punctuation after b|"ti. 61 l trivialises slove¢|sk| → slove¢|ski, lí trivialises slove¢|ska → slove¢|ska(. 62 z and independently k trivialise bez'miem) → bez'm¢em), l further → bez'm¢|"m). m punctuates after re~em|. 63 q and independently k omit po r(d', h puts b)– after vsq in order to finish the syntagm, mutilated by the lack of a bifolium in its antigraph (the text resumes at the end of 14:2). 64 k capitalises Ú:st|, z and independently P1 P2 corrupt ¢:st| → ¢:. 65 q and independently lí m trivialise sir|sk| → sir"isk). l and independently Wr U1 Ps omit the punctuation after sir"isk), k and independently U1 after potopa. 66 l and independently Wr trivialise b|"v)em) → b|"vim), k capitalises I. 67 m trivialises s(... dostalo → dostalo s(. 68 m adds the marginal note to the text, l omits it as well as vl)ve¢ie and ~arova¢ie, l corrupts v|s(ka → i¢a (lí adds s'et¢a().

8 – TEXT FAMILY g [122]

17-18 1e evr:om|, 2b daß marg. e£reom| 1jidovom| je 2st–|"( k¢ig|". v ¢ihje est| p|sa¢o.69

19-21 1b add isko¢i, 2b add s\t|, 3e b a ¢ei, b om 9:21 1qko bg–) s)tvori ¢–bo i zeml\. i v)s( 2qje ¢a 3¢e<. i ~l–ka.70

22 1-2b om 1i vs( po 2r(d' qkoje piet|.71

23 1b daß, 2b A -i\, 3b ritori\, 4b om g a elli¢om) 1je 2gramatiki(. i 3ritoriki(. i 4filosofi(.72 k a el‘li¢om) je dast) s( gramµatiki(. i ritoriki(. i filosofi(.73

10:1 Ú) pr:jde sego el‘li¢i ¢e im:(h\ svoim| (z|"kom| pisme¢). 2-3 1e om, 2a b:( ¢o fi¢i~|sk|"mi pisme¢|" pisah\ svo\ 1si r:~|. i tako 2b|"( m¢oga l:ta.74

4 1a posl:jde g pa¢amid) je 1posl:di pried).75 k B|"st| ¢:kto v‘ ¢ih) filosof) ime¢eµ pa¢amid|. ije posl:di pried). 5 1b i s|tvori im| ~, 2e om g 1¢a~e¢). w= alf|". 2vit|". k s)tvori el|li¢om) az‘b'k' ¢a~e¢) w= alf|". vit|". 6 1b slov| g ™"–. 1pisme¢) t)kmo. el|li¢om) obr:te. k i poloji im). ™"–. pisme¢). t)kmo el|li¢om) obr:te.76

7 1e kad‘mos|, b om 10:7 g priloji im). 1kad)m| milisii. pisme¢a tri. k Po tom) je po ¢:kolic:Δ∆ l:teh). i¢) k¢ij¢ik) ime¢em) kadom) milisii. i sii priloji im) tri slova pisme¢‘¢a.77

8-9 1-3e i s|stavi( s( . #"– . pisme¢|. i pisaah\ t:mi pisme¢y l:ta m¢oga, 2b sloves|" g 1t:m)je m¢oga l:ta. dev(ti< ¢a des(te 2pisme¢|" 3pisah\. k t:m)je m¢oga l:ta. dev(ti< ¢a des(te slov)mi svo< azb'k' svoimi pisme¢|" ¢'jdah' s( pisati.78

69 d and independently mí put the marginal note in the place of jidovom|, l adds it, preceded by i, to the text, m copies it in the margin (Rzh leaves it out), l omits est|. 70 z misspells zeml(, l omits the punctuation after zeml< and trivialises ¢e< → ¢ei (as independently Wr). l omits the punctuation after ~l–ka. 71 z and independently k trivialise piet| → pisa¢o, l further → ¢apisa¢o and explicits by adding v‘ b|"t"i. 72 d omits je and the conjunctions i and changes the nasals to produce A-forms. 73 l changes dast) → izobr:toa and corrupts ritoriki( → ritor"(, m omits a and changes, as elsewhere, f → #. 74 k omits si (as independently Wr) and trivialises b|"a → b:a (as independently z Ch), m corrupts m¢oga → po m¢oga. 75 l corrupts ije → (je, lí further (koje. 76 z m Ch U1 write the number in full. 77 l and independently P1 abbreviate the number, m trivialises pisme¢‘¢a → pisme¢¢a(. 78 l abbreviates the number, m trivialises slov)mi → slov).

[122] TEXT FAMILY g – 9

10 1e pisme¢a, b slov:, a pisme¢i g i po tom) simo¢id) obr:t) priloji. v–. 1pisme¢e. k Po tom) je i¢) gramµatik) ime¢em) simo¢id). obr:t) i ee priloji im) dva slova pisme¢¢a(. i 'je s)lagaet s( ~islo sloveß ih). ka–.79

11 1b slova g epiharii je skazatel|. tri 1pisme¢a wbr:te. k Epiharii je ¢:kto skazatel| k¢igam). sei iz)obr:t) i priloji im) tri slova pisme¢¢a(.80

12 g i s)bra s( ih). kd–. k I tako s)bra s( azb'ka gre~eska(. slov) ~islom). kd–.81

13-14a g po m¢o™:Δ∆ je l:t:h|. diw¢is) gramµatik).82 k Po tom) je po m¢o™:Δ∆ l:teh| dio¢is) ¢:kto ™:lo gramµatik). 14b-15 1b add slova, 2b add slova g est| 1dvoglas¢|"h) wbr:te. po tom je dr'g|"i. 2e–. k est| dvoglas¢|"h) slov) iz)obr:te. Po tom je dr'gii filosof). e–. slov) priloji.83

16 1a drougyi, 2b slova g 1i¢) je. g–. 2~isme¢ita(. k i i¢) k¢ij¢ik). g–. slova. imije ~isla pi't s( estoe. i dev(† des(t¢oe. i dev(† sot¢oe.84

17 1e m¢o™:mi, b m¢o™:m| i tako m¢o™i 1m¢og|"mi l:t|". 18 1-2b s|stavie gramat' g edva 1s)bra(. l–. i i–. 2pisme¢).85 k edva s)braa azb'k' gr|~|sk'<. l–. i i–. pisme¢). 11:1-2 1e m¢o™:m|, a m¢ogom|, b om 11:1-5 g po tom) je 1m¢og|"m) l:tom) mi¢'vem). bj–iim) povel:¢iem). k Po tom je m¢og|"m) l:tom) mi¢'vim). bj–iim) povel:¢iem)86 m Po tom je m¢og|"m) l:tom) mi¢'vim). bj–iim) prom|"slom)

3 1e m'jei g wbr:te s(. o–. 1m\j|.87 k obr:te s(. o–. m'j| m'drec|. 4 1e add k‘¢igy, e add 11:5 povel:¢"em| eg£p|t|skago cr–( p‘tolom:( ije pr:lojia k¢ig|" w= 1jidov|ska ¢a gr|~|sk|"i (z|"k)..88

79 l trivialises obr:t) → obr:te, lí omits sloveß, mí omits the punctuation after simo¢id) and corrupts i ee → ee, ~islo → ~islom) and sloveß → slov). 80 lí and independently P1 abbreviate the number, m trivialises izobr:t) → izobr:te. 81 l fails to write the capital, m and independently Wr Ch write the number in full. 82 d and independently l omit the second m of gramµatik). m omits the punctuation after gramatik). 83 l adds initial i toi, mí and independently Ps trivialise dvoglas¢|"h) → dvo<glas¢|"h), l fails to write the capital, m writes the number in full. 84 l omits ~isla and abbreviates the numbers. 85 q m write the number in full. 86 All MSS save Mol Ch read def mi¢'vim). l trivialises bj–iim) → i bj–iim). 87 d and independently l mí trivialise m\j| → m'jei, l trivialises m'drec| → m'dr|" Δ∆, mí writes the number in full. 88 d omits k¢ig|".

10 – TEXT FAMILY g [122]

12:1-21e add pisme¢a b a add k¢ig|" , 2b om, 3e add v| m‘¢ieskom) ~i¢', 4e add filosof) a 1slove¢|sk|" edi¢) 2st–|"i ko¢|sta¢ti¢|. 3¢aricaem|"i 4k£ril|.89

3-4 1a pr:loji, e add 12:5 ot| gr|~|skyh) ¢a slov:¢|sk"i (zyk| i pisme¢a s)tvori. i k¢ig|" 1pr:vede. v) mal:h) l:t:h).90

6-8 1e om, 2a m¢og|", 3-4e sed§mi< m'ji, 5e 'stroi(, 6e m'jei k‘¢igy pr:vedo(, a pr:loje¢ie a o¢i 1m¢o™i i 2m¢og|"mi l:t|". 3sedm| 4ih) pisme¢a 5'stroi. a. o–. 6pr:vojde¢ie.91

9-10 1b k¢iga, 1b s|tvori \ t:m)je slove¢|ska 1pisme¢a. st–:ia s\† i ~ßt¢:ia. st–| bo m\j) 2s)tvoril) ( 3est|.92

11 1b add s|tvori( a gr|~|ska( el|li¢i 1poga¢ii. 13:1-3 1e po ¢em|, a po¢eje, 2b potvara\, a postraq\t|, 2-3e b om Ae li kto re~et|. (ko ¢:st| oustroil| dobr:. da 1t:m s( postra(et|. 2i 3ee.93

4-5 1-2e i otv:aem| im|, b to r|ci k| ¢im|, 3e grecy , 4e b om, 5-6b potvori( s( 1w=v:t) re~em) 2sim). i 3gr)~esk| 4takojde m¢ogajd|" 5s\t| 6postra(li.94

6-7 ak£la. i sim|mah). po tom) i i¢i m¢o™i.95

8-9 1a pr|voe, e om 13:8-9, b om 13:9 oudob:e bo est| posl:di potvoriti. ¢eje 1pravoe s)tvoriti.96

14:1 1b om, 2e vam|, 3-5b s|tvori, 4b gramatou, 5e 'stroil| Ae bo voprosii k¢ig‘~i( gr|~|sk|"( 1gl–(. kto 2v|" 3est| 4pisme¢a 5s)tvoril|.97

3-5 1e a pr:lojil|, 2-3e malo ili k¢ig|" 1pr:vel). i v) koe vr:m(. to 2r:d‘ci 3w= ¢ih) v:d(t|.98

6-7 1e a b'kar(, 1b om, 2e vam|, 3-4b s|tvori ae li v)prosii slove¢|sk|"( 1b'kare gl–(. kto 2v|" pisme¢a 3s)tvoril| 4est|.99

8-10 1e a pr:lojil|, b s|stavi, 2b w=v:a\t| i ili k¢ig|" 1pr:vel). to vsi v:d(t|. i 2w=v:ave rek\t).100

89 All MSS save Mol P1 trivialise slove¢|sk|" → slove¢|sk|"i, k changes ko¢|sta¢ti¢| → ko¢|st(¢ti¢), d and independently m omit the following punctuation. 90 m and independently Pl trivialise mal:h) → mal|"h). q and independently lí omit the punctuation after s)tvori, k after prevede. 91 l abbreviates the first number, d and independently m write the second in full, l trivialises 'stroi → 'stroia. k omits the punctuation after l:t|". 92 k capitalizes T:m)je, undone by l, and omits the punctuation after pisme¢a, l and independently P1 trivialise slove¢|ska → slove¢|ska(. 93 m trivialises postraqet) → postroqet) and omits the punctuation after it. l omits the punctuation after re~et) and i ee (as independently d). 94 z trivialises gr)~esk| → gr)œsk|"i. k omits s\t|. 95 m trivialises i¢i → i¢ii. From here on, k retains £. m omits the punctuation after akula, q after si|ma Δ∆. 96 z and independently l trivialise potvoriti → tvoriti, l corrects pravoe → pervoe, m miscorrects it → prava(. 97 k anticipates the punctuation before gl–( and omits v|" (as independently q h Wr). The defect in h ends with s)tvoril), joined directly to 8:21. 98 h and independently k Ch U1 trivialise r:d‘ci → r:d‘cii. 99 l omits the auxiliary. lí m omit the final punctuation. 100 l trivialises w=v:ave → w=v:a<t) ", m capitalises I.

[122] TEXT FAMILY g – 11

11-14 1b om, 2b add solou¢|skyi, 3e a add t| ¢am|, b t|i ¢y, 4b gramatou, 5b a s|tvori, 6e pr:lojil|, a pr:loji, b om 1st–|"i 2ko¢sta¢ti¢| filosof|. ¢aricaem|"i 3k£ril|. 4pisme¢a 5s)tvoril|. i k¢ig|" 6pr:vel|.101

15-17 1e add episkop) morav|sk)"i g i me#odii brat) 1ego. marg? s\t| bo ee jivi. ije s\t| vid:li.102 k i me#odii brat) ego. sostavista az|b'k' gramot|" slove¢|sk|"(. m i me#odii brat) ego. sostavista az|b'k' gramot|" r'sk|"a. 18-19 1e v|pros(t|, 2b add v|si, 2-3e i otv:aem|, add 14:20 po sed‘mom) s)bor:. v). md–. l:to i ae 1voprosii v) koe vr:m(. 2to v:d(t| i rek\t). 3(ko

21-22 1b vr:m(, e add 14:22-23 i mt–re ego #eodwry. ije pravov:r)¢'< v:r\ 'tver‘dista g v) 1vr:me¢a mihaila cr–( gr|~|ska.103 k vo vreme¢a mihaila cr–( gre~eska. pri patriars: fotii.104

24-25 g i borisa k¢(™( bl)gar|ska. i rastica k¢(™( morav|ska.105 k v‘ l:ta borisa k¢(™( bl)gar|ska. i rastica k¢(™( morav|ska.106

26 1b a kocel( g i 1kostel( k¢(™( blat|¢|ska k i kostel( k¢(™( blat|¢|ska. vo k¢(je¢ie k¢(™( velikago vse( r'si r<rika. poga¢a s'a ¢e kree¢a. za. rk–. l:t) do kree¢iq r'sk|"( zeml(.107

27 1a l:ta, 2a mira, b om 14:27-15:7 v 1l:to je o† s)zda¢"!( 2mir'. ‰™t¶g–.e.108

15:1-3 S\t| je i i¢i o†v:ti. (je i¢‘de re~em|. a ¢¢–: ¢:st| vr:m(.109

4 1a tak|, e add est|, 2a dal| 1takov| raz'm| bratie bg–) 2daroval| slov(¢om|.110

5 em'je slava. ~|st|. dr)java i pokla¢(¢ie.111

6-7 1e om, 2a om ¢¢–: i pris¢o i v) 1besko¢e~¢|"( v:k|" 2v:koµ. ami¢|.112

101 lí omits 14:12, l omits the punctuation after sotvoril). 102 All MSS save Mol Rzh abbreviate me#odii → me#o ∂. l omits the punctuation after ego. 14:16-17, preserved only by the MSS M (text family b, with additional ih) in fine) and Mol, have been the subject of some controversy, cf. Kuev 1967(note 1):28 and Ziffer 1993(note 15): 87-88. Mol could have got these syntagms either from the a MS used to verify its text or, more likely, from d via its antigraph z, considering their replacement in k by a chronologically neutral statement; either supposition lends weight to Kuev's (and Ziffer's) premise that the syntagms were eliminated at an early date as anachronisms. 103 h and independently Wr trivialise gr|~|ska → gre~eskago. 104 l trivialises pri → i pri. 105 d omits k¢(™( after rastica, h and independently Rzh trivialise bl)gar|ska → bol‘gar|skago. 106 lí mí trivialise borisa k¢(™( → k¢(™( borisa. 107 l trivialises k¢(™( velikago → velikago k¢(™(, all MSS save Rzh modernise r'si → r's"i . 108 d writes out the ordinal endings ‰™–.¢o i t–. i ¶–. i tret"!e, z and independently h l mí Ch Pl trivialise mir' → mira. 109 All MSS save P1 trivialise i¢i → i¢ii, h and independently Ch T P2 trivialise i¢|de → i¢§d:. l omits the punctuation after w=v:ti. 110 q and independently h m trivialise daroval) → est| daroval|, h and independently k U1 harmonise slov(¢om| → slove¢om). 111 l omits the rest of the doxology after slava. 112 z and independently Ps omit besko¢e~¢|"(, U1 V1 add the Cyrillic alphabet from a to :.

12 – TEXT FAMILY g [122]

The tradition of the g text differs markedly from that of the text families b, a and e by the fact that it exhibits non-trivial developments of the text beyond its Cyrillic archetype. In it, we find three types of editorial intervention, viz. a. verification of one antigraph against another (MS Mol), b. expansion by integration of marginal notes, explicitation and updating (hyparchetypes klm) and c. reduction by selection and adaptation (MS Rzh), as well as four successive stages of intervention, viz.1. g, 2. k, 3. l and m, respectively, and4. Rzh. These interventions have two features in common: none of them affects the text in its entirety, and all of them degrade the language of the text. We shall examine them separately.

1. The text of the Cyrillic archetype g is an idiolectal rendering of the Glagolitic original w without discernible traces of editorial intervention. It is marked by problems in at least 36 syntagms, which persist in the text of the senior and junior hyparchetypes and the MSS: (a) problems created by the author: the syntagm 9:9, apparently a Biblical quotation, but unparalleled in the known Slavic versions, could well be a trace of a draft version, rejected by the author in favor of 9:10 (cf. its lack in b); (b) cyrillisation: the word order < \ ( q (1:20-23) and the addition of #, x and $ to the list of letters (6:5) are sure signs of cyrillisation; (c) reading and interpreting scriptio continua: i ista (corrupted to i st–a in be) → i isti¢‘¢a (2:8 = a); w= a → w= aza (3:3); ¢a~(t)k) oboego (= e) → ¢a~(t) oboe (3:3); ijdeko¢|¢)" (a) → ijdoko¢| (8:5); a–.m) (cf. e edi¢:m|) → sam) (8:9, cf. a ¢am|); (d) oversight: li (8:11) add; i rim|sk|" (8:15) om; (e) unfamiliarity of names and concepts: palamid) (= be) → pa¢amid) (10:4 = a) and koc|la (= b) or kocel( (= a) → kostel( (14:16 = e); bouk)stavam) (corrupted to a quasi-gloss bouk)vam). to &st) stati k) vam) in e) → bouk)vam) (5:7 = a); k)¢ig|" (= ea) → pisme¢a (8:2 = e, 17); (d) discursiveness: the s)kaza¢i& kako-title with information from 2:6-11 and 12:1-4. (= be) conflicts with the more economical o-title (= a) and may well be a secondary explicitation; the doublet conjunctions bo 'bo (7:7) and a... je (9:23) are surely corruptions of single conjunctions in the original; v:koµ (15:6) surely is an addition from liturgical practice (cf. its lack in a); (e) language problems: (1) definiteness: of the three participles (2:2-4) only the first is preserved in the def form; the single t) (7:13) surely is a corruption of an original reading t) je (e a); the def forms dr'g|"i (10:15) and m¢og|"m- (10:17, 11:1, 12:6) surely are corrupt; (2) verbal tense and syntax: the pf-participle s)tvoril) (7:2) is surely a corruption of the original pf with auxiliary; the pl s)bira\t) (7:12) with the num subject surely is a corruption in view of the sg in 10:12 and 12:7; the lack of a verb leaves the syntagms 9:6-7 dangling; (3) unfamiliar forms and miscellaneous errors: pisme¢em) → pisme¢om) (6:4); m|¢|i → m|¢e (7:3); ijde w= ko¢a (a) → iz¢a~(la (8:6); tak) (a) → takov) (8:18, 15:4); po ¢em| (e) → t:m (13:3) is most certainly erroneous if 14:16-17 is original; (4) Serbisms: the corruption of original pr|v- → prav- (5:3, 13:9) surely betrays a Serbian accent, as may the elimination of the original adnominal D s)pase¢i< (2:2-4 = a, Lk 1:77) and the denasalisation b'kar( (= ea) → b'kare (14:6).

1a. The MS Mol deviates from the g text, transmitted faithfully, but not always competently via d and z, in no more than six instances by adding text from a: pismeì + k¢iŸ (1:2), add i rimsk|"( (8:6), add wkaa¢"!i (8:10), zemlem:rie → zemlem:re¢ie (9:13), add je (10:7), kwstel( + kwcel( (14:26); the first and the last addition degrade the text (in 11 more instances the original text is degraded owing to the individual competence of the transmitter: 1:7 om i gr|~|sk|"mi, 2:3 ~l–~a + slwv:¢ska from 2:5, 4:1 bo → je, 6:5 om x, 7:8 ¢e ¢apl)¢:\† → ¢e pl)¢:\†, 7:9 om s\†, 7:14 om pismeì, 9:21 om i ~l–ka, 12:7 'stroi → s)tvori, 14:5 to r:d§ci → to mal"!i + posr:di, 15:5 om i pokla¢(¢ie).

2b. The hyparchetype k deviates from the g text by outright editorial intervention in the sections 9:11 - 11:4 and 14:11-27, a section that has attracted the attention of compilers from before

[122] TEXT FAMILY g – 13

ca. 1100 to after ca. 1400.113 Its main aim seems to be explicitation (a) by adding to the identification of the people mentioned:114 ¢:kto v§ ¢ih) filosof) ime¢em) pa¢amid| (10:4), i¢) k¢ij¢ik) ime¢em) kad)m| milisii. i sii (10:7), i¢) gramµatik) ime¢em) simo¢id) (10:10), epiharii ¢:kto skazatel| k¢igam). sei (10:11), diw¢is) ¢:kto ™:lo gramµatik) (10:14), dr'g|"i filosof) (10:15), i¢) k¢ij¢ik) (10:16), m'j| m'drec| (11:3); (b) by specifying the contribution and/or the destination: astro¢omie (9:15, from marginal note), el|li¢om) azb'k' (10:5), im) (10:6), slova pisme¢§¢a (10:7 = b), slov)mi (not sloves|", b) svo< azb'k' svoimi (10:9, thereby creating an anacolouth with pisme¢|"), slova (not du slov:, cf. b) pisme¢¢a( (10:10), slova pisme¢¢a( (10:11), azb'ka gre~eska( slov) ~islom) (10:12), slov) (10:15), slova. imije ~isla pi't s(. ™–.e. ~–.e. c–.e (10:16, from 7:11), azb'k' gr|~|sk'< (10:18), k¢ig|" (11:4, = e), (z|"ka (11:4, = be); (c) by specifying the circumstances: po tom) je po ¢:kolic:Δ∆ l:teh) (10:7), i ee (10:10), tako (10:12), po tom) je (10:13), ¢'jdah' s( (10:9, from 1:7); (d) by adding verbs to opaque syntactic constructions: razd:lia s( (9:11), qkoje se gl–< est| qz|"k) egipet|sk"i. em'je s(... dostalo (9:13), dast) s( (9:23 = b), b|"st| (10:4, with subsequent reinterpretation je → ije), s)tvori (10:5), i poloji (10:6, thereby creating an anacolouth with t)kmo... obr:te), priloji (10:15); (e) by added subtotalling: i 'je s)lagaet s( ~islo sloveß ih). ka– (10:10, following 10:9); (f) by adding information: pri patriars: fotii (14:22). A secondary aim seems to be updating (g) by replacing the anachronism s\t| bo ee jivi. ije s\t| vid:li → sostavista az|b'k' gramot|" slove¢|sk|"( (14:16-17); (h) by adding a Russian chronicle note vo k¢(je¢ie k¢(™( velikago vse( r'si r<rika. poga¢a s'a ¢e kree¢a. za. rk–. l:t) do kree¢iq r'sk|"( zeml( (14:26). The two expansions wbr:t- → iz)obr:t- (10:11, 14) seem to be no more than harmonisations with the g usage from 10:11 onward, the first, in part form, creating an anacolouth with i priloji im), added from 10:7.

3b. The hyparchetype l deviates from the k text, apart from accidental variants, by a single editorial addition before the title: o az§b'k: slove¢§skoi. i o el¬i¢§skoi. sir:~| gre~eskoi. Its source must be sought in the o-titles of the treatise, known not only from text family a (full text) or b (compilation), but also the title o gramot: the excerpt versions referred to in note 112 above, with which it shares the attention for the history of the Greek, rather than the Slavic alphabet.

The hyparchetype m deviates from the k text, apart from accidental variants, by two editorial interventions: 11:2 povel:¢iem) → prom|"slom), and 14:17 slove¢|sk|"( → r'sk|"a, the latter obviously degrading the sense. If it was not the scribe's habit to identify slov:¢|sk) with r'sk) and, consequently, replace it, its source may be an erroneous interpretation of the chronicle note appended to 14:26 in k.

4c. The MS Rzh omits from the m text t:1 - 9:10 and starts with an adaptation of 9:7, omits 9:18-

22, restarts without adaptation 11:4, omits 12:1 - 14:14, restarts with an adaptation of 14:11, omits 14:18-19, restarts with an adaptation of 14:21, and omits 15:1-7. The editorial interventions of Rzh can be characterised as follows: (a) adaptations to compensate for prior omissions: do¢deje bg–) razd:li (z|"k|". pri stolpotvore¢ii → po stl)potvore¢ii je b|"ß eg∂a razd:lia s( qz|"ci (9:7), st–|"i ko¢sta¢ti¢| filosof|. → po toµ takoª l:toµ m¢og|"µ mi¢oe∂im) ko¢sta¢ti¢| (14:11), and v) vr:me¢a mihaila cr–( gr|~|ska → pri cr–i mihail: gr|~|st:m| (14:21); (b) harmonisations with m usage: dast| s( → dosta s( (9:23 following 9:13), pisme¢- → gramot (10:1-2 following 14:17), pisme¢) → slov) ~isloµ (10:6, 18 following 10:12), obr:te → iz)wbr:te 113 The first (Russian) excerption of the Azbuc¬na Molitva (cf. Veder 1998) contains a summary of an excerpt, known in at least three other different versions from the Paleja and Sborniki (Veder forthcoming 2). The MS Rzh demonstrates that the attention of compilers for the history of the Greek alphabet had not abated even 300 years later. 114 The g text items are printed in small type.

14 – TEXT FAMILY g [122]

(10:6 following 10:11), and m¢oga → po m¢oga (10:8, 17 following 10:3); (c) the 14 remaining deviations from the m text seem to be rather more of an accidental nature: 9:15: om vra~eva¢ie, 10:1: ¢o preªde → ¢| ispr|va and svoim) (z|"kom) → ou sebe, 10:2: fi¢i~esk|"mi → a#i¢eisko< and pisah' → ¢'j∂ahou s( pisati, 10:7: om milisei, 10:15: po tom → po siΔ∆, 10:16:dittography qv: dev(t†, 10:18: s)braa → s)brali s'†, 11:1: mi¢'vim) → mi¢oe∂iµ (with ¢ as in 14:11), 11:3: obr:te → obr:toa (with num subject), 11:4: prelojia → prelojili, 14:11: om st–|"i and #ilsof), 14:27: v) l:to je → a.

At this point arises the question of dating and localising the various stages of text development. Proceeding in reverse chronological order provides the best chances to reach reliable conclusions.

1a. The edition in the MS Mol must, for want of parallels, be considered an occasional contamination and ascribed to the scribe of the MS.115 The shorter the text, the less the chance to locate a parallel in the thoroughly untidy Sborniki of the Slavia orthodoxa, and the greater the probability that the contamination will remain unique, unless, of cours, somebody copies it, again by chance. Thus the text must be dated to the 16th century, but the featureless contribution the author made to its development leaves us without any possibility to even make a guess at the area of provenance; we should probably not venture too far from the Moldavian monastery of Bistri®a, whose hegumen had it in its possession until his death at Hilandar (Athos) in 1606.116

4c. Like the preceding, the version in the MS Rzh must be considered an occasional excerption and ascribed to the scribe of the MS.117 His contribution to the development of the text is almost just as featureless, save for his use of gramot and slov- for pisme¢-, but we can be relatively certain that these items did not belong to his active lexicon, since they can most satisfactorily be explained as part of his effort to reduce the lexical variety of the m text by harmonization to a limited number of arbitrarily selected lexemes. There is, in sum no obstacle to dating this text, like the MS, to the 15th century, nor to a localisation in Russia.

3b. The l and m texts, by contrast to Mol and Rzh, show a marked Russian accent in their editorial interventions, viz. the absence of second palatalisation, and the semantic Russism, respectively. Their origin, too, must be sought in Russia, yet there is no way of dating l other than prior to the MS T and the junior hyparchetype lí, nor m other than prior to the MS Rzh and the junior hyparchetype mí. The arbitrary dating proposed in the presentation of the stemma, i.e. 50 years prior to the earliest possible date of the oldest MS, i.c. no later than 1350 for m, and possibly, but not necessarily somewhat later for l, is as good as any.

2b. The k text, like that of the MS Rzh, presents us with the puzzle of the repeated insertion of slovo in the (Serbian) sense of «letter» hand in hand with the addition of a Russian chronicle note. The latter definitely points to Russia as place of origin; but does the latter point to Serbia, i.e. to a different author? I think not. The editorial interventions of k are essentially coextensive with the short versions mentioned in note 112, which know no other word for «letters» than slovo, and k's addition sostavista az|b'k' gramot|" slove¢|sk|"( could very well be explained as a reflex of their final statement a slove¢|sk|"( k¢ig|" pr:loji kiril| filosof| s me#odiem| v) l:to ‰™txg–; a Russian chronicle note i v) d¢i k¢(™( r<rika ¢ov)gorodskago. egoje s¢–om) r'ska( zeml) preide forms a regular complement of one of the short versions and could have served as a model for the corresponding addition in k. In sum, I think that the entire editorial

115 This assumption must, of course, be verified by an in-depth study of the textual and linguistic habits of its scribe. 116 Ivanov 1931:442. 117 The assumption must, of course, be verified as thoroughly as the preceding.

[122] TEXT FAMILY g – 15

activity of k was triggered by the desire to integrate the text of one of the short versions into the g text and that his use of the word slovo is, like in the MS Rzh, merely testifies of an effort at harmonisation. Thus the origin of the hyparchetype k must be sought in Russia, and the dating proposed in the stemma, i.e. before ca. 1300, is as good as any.

1. The g text, not unlike k, confronts us with the puzzle of two indubitable Serbisms within a mass of featureless variants, which do not seem to derive from a Glagolitic archetype different from that of b, e or a, but merely from a different reading of that archetype. Considering the probability that, like the metacarakthrismov" of Greek MSS from uncial to minuscule script,118 the transcription from Glagolitic to Cyrillic in Slavic was a process limited in time to a central period of ca. 150 years (after which a parchment quire like ours, unbound for ease of copying, would have definitely perished), i.e. lasted in Slavic from ca. 900 to ca. 1050, after which only incidental and peripheral transcriptions may have occurred, two explanations are possible: i. the variants including the Serbisms belong to the same author, which raises the question of access to the Glagolitic archetype, independently read by b e and a at its repository (did a Serbian transport himself to join the community at Preslav, or that of Ohrid, and make his transcription there, or was the Glagolitic archetype transported?); ii. the Serbisms and the other variants (or most of them) belong to two different hands collaborating to produce a MS, which raises the question of heterogeneity in the variant readings of g relative to bea (are there any criteria to dissociate two - or more - possible layers within their mass?). I have no criteria by which to decide which explanation is the more probable. For the time being, these questions remain an object of speculation within the realm of textual palæontology, i.e. the area devoid of MSS witnesses. We can, however, be reasonably sure that the Glagolitic original would not have survived three centuries until ca. 1200 and that g may have originated somewhere on the Balkans at least a century and a half before that date.

As to the quantity and the quality of the variation, it should be noted that the non-trivial text changes recorded in the four stages of development of the text are neither numerous, nor consistent, nor even significant, except for the filiation in transmission. They are indicative of a type of textual transmission that was current in the Slavia orthodoxa during at least eight centuries: faithful to the extreme to the text, if not to the letter (save, of course, the accidents of human frailty), and almost free of any earnest endeavor to check the text's correctness, not to speak of its contents.

After the author of the Cyrillic archetype g had produced as best he could one of the many conflicting readings of the Glagolitic original w, four hyparchetypes and at least 6 MSS simply copied his work (one hyparchetype and 2 MSS even without noticing, or at least attempting to fill in a large gap which affected almost one third of the text). One other author (k) edited a section of the text, prompted by an excerpt version: he integrated its text but, being unable to make a consistent choice of readings, often produced double readings (resulting in rudimentary plete¢i& sloves)) and anacolouths; the overall effect of his labour is no more than emphasis. Three other authors (of hyparchetypes, not of MSS; they just copied what they found in their respective antigraphs) added notes before the doxology (kí) and to the title (l), respectively, or consciously changed two words (m). One scribe of a MS (Rzh) decided to excerpt from his antigraph essentially the emphasised sections and to adapt the close context to their isolation, again usually taking his words from elsewhere in the text and not from beyond it. Only one of the scribes of the 23 hyparchetypes and MSS in well over four centuries (Mol) found and used the opportunity to

118 Dain 1975: 126-7.

16 – TEXT FAMILY g [122]

verify the text of his antigraph against another antigraph belonging to a different text family, but he sorrowly failed in his task: his verification lacks both depth and consistency, and save for the six changes he made, he seems to have been satisfied with his antigraph, unaware of any other problems in its text.

Such conservatism in transmission is untypical for an environment, where written and spoken language are supposed to have been to a large extent coextensive and interactive, as the traditional language model of the Slavia orthodoxa would teach us. In view of the lack of considerate linguistic updating, both text-internal harmonisation and introduction of doublets must be considered indications of lack of linguistic security, just as introduction or neglect of anacolouths and obvious misinterpretations and misreadings are indications of lack of linguistic proficiency of the transmitters. If, indeed, none of the transmitters of text family g of O Pismenex¿ was really conversant with the language of his antigraph, not to mention its sources, then none of the MSS can be trusted on its own, not even that source MS which had direct contact with the original author's text. The MSS then are in principle equal in their testimonies, yet none of them can in principle be equal to he author's text as such (unless, of course, it is an autograph).

Given such type of transmission, we must put out trust into the author (or translator) of any given text to have known how to put his intentions into words, to have handled the language of his text with confidence and consideration; if we do not trust him, we lose whatsoever base to pursue whatsoever philological investigation. We are, in fact, obliged to endeavor to reconstruct his text from the rubble of transmission, i.e. from the paradosis of the various text families, because it is the sole positive set of data on which to base our research into the Old Slavic language.

Bibliography Canev 1994: Canko Canev. Aspekti v izsledvaneto na ‘O Pismenex¿ Cµernorizca Hrabra’. Sµumen (Dissertation) 1994. Dain 1975: Alphonse Dain. Les manuscrits. Paris 1975. Damjanova (= Spasova) 1978: Marija Damjanova. Ob odnom interesnom momente v neissledovannom spiske

soc¬inenija Cµernorizca Xrabra ‘O pismenax’. Palæobulgarica 2(1978)3: 80–86. Ivanov 1931: Jordan Ivanov. Ba˘lgarski starini iz Makedonija. Sofia 1939, repr. 1970. Jazykova (= Stepanova) 1979: Anna Ju. Jazykova. Russkie redakcii Skazanija O pismenex Cµernorizca Xrabra.

Novosibirsk (MA Thesis) 1979. Kossova 1980: Alda Giambellucca–Kossova. Cµernorizec Xraba˘r. O Pismenex¿. Sofia 1980. Kuev 1967: Kujo M. Kuev. Cµernorizec Hraba˘r. Sofia 1967. —— 1988: Cµernogorskijat prepis na Xrabrovoto skazanie. Palæobulgarica 12(1988)3: 103–112. Marti 1981: Roland W. Marti. Stilistic¬eskie osobennosti apologii Hrabra. Staroba˘lgarska literatura 10(1981): 59–70. Trembovol’skij 1989: Jaroslav L. Trembovol’skij. Drevnebolgarskij pamjatnik ‘O Pismenex¿’ Cµr¿norizca Xrabra na

Rusi. Palæobulgarica 13(1989)4: 68–90. Veder 1996a: William R. Veder. The Earliest Attested Text of the Treatise O Pismenex. Studies in Slavic and

General Linguistics 23(1996): 221–254. —— 1996b: Linguistic Problems in the Transmission of O Pismenex. The Paradosis of Text Family a. Slavica

Gandensia 23(1996): 7–29. —— 1998: Zas¬to i kak da se rekonstuira ca˘rkovnoslavjanski tekst. Paradozata na Prologa kam Uc¬itelnoto evangelie

na Konstantin Preslavski. Preslavska kniz¬ovna s¬kola 3(1998):3–25 —— forthcoming 1: La «redaction russe» d'O Pismenex. La paradosis de la famille de texte e. Ricerche Slavistiche,

forthcoming.

[122] TEXT FAMILY g – 17

—— forthcoming 2: Kratkite versii na O Pismenex. Paradozata na razklonenieto m (do 1050 god.). Sbornik Iv.

Ga˘la˘bov, t. 2. Veliko Ta˘rnovo 1997, forthcoming. Ziffer 1993: Giorgio Ziffer. Sul testo e la tradizione dell’Apologia di Chrabr. Annali dell'Istituto Orientale di Napoli:

Slavistica 1(1993): 65–95. —— 1995a: Zur Komposition des Traktats ‘Über die Buchstaben’ des Mönchs Chrabr. Die Welt der Slaven

58(1995)1: 58–75.