Textual Introduction by Claudio Moreschini
Textual Introduction by Claudio Moreschini THE manuscript tradition of the Poemata Arcana falls into two main groups, one comprising the direct tradition ofGregory's poetry, which in turn forms two families Q and W,' the other the 'FourteenParaphrase Corpus' of Nicetas David (IT). This corpus, the object of previous scholarly investigation,' consists of fourteen poems (seventeen in the arrangement of M. 37), accompanied by a paraphrase, or rather commentary, by the Byzantine commentator Nicetas David
(ss. ix-x). The poems on which he wrote are (in Migne's numeration) I. I. I+5, 7-Q; I. 2.9, 14, 15+16, 17(+2. 1. 2), 31, 33. Nicetas gave them the title of ta daéppnra; we do not know whether he invented it or found it already in use. In the sixteenth century it was translated, in circumstances we shall consider in due course, as Arcana. The title
Poemata Arcana used in the present edition for Migne's 1. 1. 1-5, 7-9, in our numeration Arc. 1-8, is therefore derived from Nicetas, even if strictly speaking it applies not only to this collection but also to the other poems noted above. All the same, our Poemata Arcana, the first part ofNicetas' corpus, must have constituted a homogeneous grouping even before his day, since they stand together in the same order in the manuscripts of the two families Q and WV, albeit with the addition of 1. 1. 11. Nicetas may have omitted this latter and added the other poems, which are on moral subjects.'
This introduction summarizes the findings set out in C. Moreschini, 'La tradizione manoscritta dei Carmina Arcana di Gregorio Nazianzeno', Atti della Accademia Pontaniana, XS 44 (1995), 99-120.
' These symbols have been in general use ever since they were first proposed by L.
Sternbach, 'Cercidea', Eas, 30(1927), 347-60, in connection with his projected edition of Gregory of Nazianzus' works for the Polish Academy ofSciences in Krakow, on which see now M. Plezia, £ 'incomptuta edizione delle opere di Gregorio Nazianzeno progetiata a Cracovia (Naples, 1991): cf. e.g. Gregor von Nazianz:
De vita sua, ed, Ch.Jungck (Heidelberg, 1974), 39; Gregor von Nazianz:Carmina de virtute Ja/Tb,ed. R. Palla, tr. and comm. M.
Kertsch (Graz, 1985), 39.
? See Fr. Letherz, Studien zu Gregor von Nazianz: Mythologie, Uberlieferung, Scholiasten (Bonn, 1958), 161-8 et alibi; Palla, edn. 9.
* His commentary on Arc. 1-5 is available in the recent edition by C. Moreschini and I. Costa, Niceta David: Commento ai Carmina Arcana di Gregorio Nazianzeno (Naples,
1992), that on the rest in the older edition by E. Dronke (Gottingen, 1840), reprinted in M. 38. 681-842. ix TEXTUAL INTRODUCTION 1. The oldest manuscript in the Q family is Bodleianus Clarkianus 12
(C), s. x, fos. 1517-166", which owing to the loss ofone leaflacks Arc. 1.
1-24. It divides the longer poems into sections: thus it breaks Arc. 4 into (a) vv. 1-88, entitled wepi xécuov, and (b) wv. 8g—100, entitled wepi ayyéAwy, and similarly Arc. 8 into (a) vv. 1-30 wepi d:anKav, (b) wy.
31-81 mepi éemidaveias Xpiorod, (c) vv. 82-99 epi Bawriopaztos.
The same division of Arc. 8, into the same sections with identical titles, is found in Neapolitanus Graecus 24 (N), ss. xiv—xv, fos. 148*- 165", but this manuscript does not divide Arc. 4.7 Like C and N, Vaticanus Graecus 482 (Va), s. xiv, fos. 79'~96", divides Arc. 8, into wv.
I-30, 31-81, and 82-99, but not Arc. 4; it omits Arc. 7.
Marcianus Graecus 82 (Ma), s. xii, also belongs to the Q2 family. It includes our poems (fos. 86-127"), accompanied by a paraphrase (not Nicetas David's), but omits Arc.
8. 19-33, so that we cannot tell whether it exhibited C's first break, between wv. 30 and 31. The second is indeed indicated, but in a later hand (possibly s. xiv), which adds a marginal note zepi tov Geiov Bazriopatos; however, the beginning of section (c) at v. 82 constitutes the first line of a new pericope.
Within the 92 family we may observea series ofagreements in error between Ma and N* Arc. 2. 1 82] 84; 2. 51 pw) jew addy (an interlinear gloss entered into their hyparchetype); 3. 7 wévos] aBévos; 5. 67 prius re om.; 6. 44 pév] Av; 7. 58 odpavds} odpavoi MaN™; 7. gt évi] év; 7. 126 ds Hyeis] as Kai Hueis. The tradition from which N derives appears from its readings to be radically contaminated from the V tradition of Laurentianus vi. 10 (L), a manuscript to be examined later: Arc. 5. 39 76 wév... 76 8° év] ra wev... ra 8' €v LN, corr. N?; 5.
Jo dviyszev SVa?N?? |dvower alii, dvvozev CMaVa, iouev LN; 6. 69 €8¢dafev IILN] éfediSaferv CMaVa; 7. 6 ye TILN] re CMaN? (Va deest).®
Furthermore, N seems to be contaminated froma tradition detectable in Va™: thus only N and Va know the title wepi zpiddos for Arc. 1 wepi apydv, which is found both in the rest of the Q tradition (Ma and Va itself; C deest) and in one branch of the manuscripts containing
Nicetas' commentary (the other, like WY, omits the titles): Arc. 4. 36 éym I1VareN] e'gvge CMaVaL,; ibid. éuvedos IILVaveN] éuzredov
* N is a composite MS, of prime importance for the Arcana, but for other poems of Gregory's embodying no independent tradition; cf. in greater detail Palla, edn. 54 ff. Cf. also C. Moreschini, 'Studi sulla tradizione manoscritta di Gregorio Nazianzeno: la
Gedichtgruppe IV, in Miscellanea in onore di Giovanni Tarditi (Milan, 1995), 1235-55- > Cf. Palla, edn. 61.
* CE. Palla, edn. 61-2. Itis probable that contamination by a MS ofthe W family took place in the exemplar from which N was copied.
X TEXTUAL INTRODUCTION CMaVa; ibid. od«ér' TILVaeN] ovdev CMaVa; 5. 33 7} yap 59 Beds T1Va?N] 78n yap beds CMaVaL.
Agreement between C and Vais found in Arc. 2. 41, where both read véov instead of déos as in the other manuscripts, and in Arc. 8, where both transpose v. 26 to follow v. 28. Moreover, in both manuscripts v. 66 is omitted, and then added in the margin by the scribe; which indicates that the hyparchetype of C and Va, designated » by Palla, exhibited this line in the margin. However, for the Poemata Arcana, as for Gregory's other poems, Va appears to be akin to L.' In Arc. 3. 77 both L and Va (like I) read evre instead of the oizve of CMaN, and pvyjorw instead of their pviorov; in 4. 6 éoav (with IIL) against the other 92 manuscripts' éav; in 1. 1. 11. 2 (not part of the Poemata Arcana) both L and Va read ov céBer' icoféws opaviov Adyov contra metrum.
Va also presents a certain number of readings derived from the I family; but Palla has observed, precisely on the basis of certain of the Poemata Arcana (3. 31, 8. 98), that the marginal readings and corrections inserted into the text ofVa over erasures in a different ink derive specifically from 6, the hyparchetype ofS and Mb, to be examined ina moment.* At all events, to complete our demonstration that the marginal readings and corrections in Va derive from the II family, we may adduce the following instances: Arc. 1. 19 re om. Q, add. Va'; 1. 26 perérecta I1LVa?] perémofev Q; 2. 10 om. MaNVa, add. Va"™; 2. 19 dvapyos GQ] avdpyou IVa'; 3. 36 adrés éeA@dv TTVa™] €Abay éxixoupos Ma, éA@ éxrixoupos CNVaL; 3. 63 dé IT] om. Q, add. Va*™; 5.
68 dtpodos TIN'*Va*] dzpowos MaNVa; 8.
14 tpédaro IVa?) otpégar' MaNVa; 8. 35 Aevoowy TINVa*] Aeiowr CMaVa. We may also cite the two passages indicated by Palla to show the derivation of Va* from 6, Arc. 3. 31 cawrjp Va (with the other
2 MSS)] dwrjp SMbVa"; 8. 98 a0' dipia KtixdAos EAiooe: Va cett.] dua 7' HerAiw cedrvy re SMbVa™.
2. In the second family ofmanuscripts, that designated 'VP, the Poemata Arcana are present only in Laurentianus vu. 10 (L), s. xi; this is one of the most authoritative witnesses to the tradition, and the one containing the largest number of Gregory's poems. The Arcana are found on fos. 45'-53°. However, L shows closer kinship to Q than to IT in that * Va, as Palla, edn. 46 observes, is a particularly contaminated MS, with a corresponding disorder in the sequence of poems; proof lies, so far as the Arcana are concerned, in the omission of Arc. 7, tepi buys. Cf. Moreschini, 'Studi', 105-6.
* Palla, edn. 48-50. xi TEXTUAL INTRODUCTION (i) it adds 1. 1. 11 after Are. 8, (ii) after Arc. 8. 30, 81 it signals the beginning ofa new poem (albeit without inserting headings), dividing Arc. 8 into three parts like C, N, and Va. For the rest, L (i.e. VY) agrees in turn with Q and II, as is evident from the apparatus criticus. A further characteristic of L, already noted by previous scholars,' is that it demonstrates a marked reworking ofform and grammar. So far as the Poemata Arcana are concerned such scholarly revision is evident in L from the following examples:
Arc. 4. 97 Ovnrés] bOird6s NC', Bpozdés L, Ovards Caillau; 5. 5 éreryépevor] éAtaoduevov L; 5. 22 duds udpos] dudomopos L; 5. 39 Hyiv] duper L; 5. 50 éroiae] éAicaer L; 6. 5 Avogévorai] Svouévoin: L (or is this an error in transcribing from capitals?); 6. 44 wév TT] wav Q, rw' L; 6, 50 peydAct] weydAov L;
7-38 téiovos KixAocow IT] iéiovos KixAov tw' Q, i£iovds twa Kd«Aov L; 7. 42 ovdé Odyvor] ov Bayvor L; 7. 49 765€] rade L; 7. 53 Gptorov] dprore L; 7. 78 déars] dvacs L; 7. 112 waAaiopaciv] caxddpovos L; 8. 4 erecta] perérerta L; ains] yains L; 8.85 épeider] yu' épeiber L.
A further feature of L should perhaps also be ascribed to a similar learned initiative: the insertion on fo, 52°' of sixty lines, not found in the rest of the manuscript tradition, within Arc. 8, after v. 18. They have been transcribed by Bandini,'° again by Vari,'' and most recently by Wyss;'? the last-named has sought to demonstrate their authenticity by comparison with other poems of Gregory's, and by the observation that he habitually repeats himself or reuses in different contexts passages, phrases, or syntagms already employed elsewhere. But in that case, why should Nicetas and the 92 hyparchetype independently omit these lines? Indeed, I am quite ready to believe that this substantial insertion, stitched together in cento fashion from other poems by Gregory, was also due to scholarly activity of this sort.
3. By virtue of its antiquity a considerable significance also attaches to an ancient Syriac translation ofsome ofGregory's poems,' preserved
* Cf. e.g. Jungek, edn. 39.
Bandini, Catalogus codicum manuscriptorum Bibliothecae Laurentianae, i (Florence, 1764), 219-20.
'"! R. Vari, 'Sancti Gregorii Nazianzeni codicis Mediceo-Laurentiani celeberrimi collatio', Egyetemesphilologiai kizlany, 20 (1896), 759-72, esp. 770-2.
2 B, Wyss, 'Zu Gregor von Nazianz', in Phyllobolia fiir Peter Von der Mull (Basel, 1946), 153-53, esp. 159-72.
3 This translation appears to be the work of Theodosius, later metropolitan of xii TEXTUAL INTRODUCTION in Vaticanus Syriacus 105.'* As regards the disposition of the Poemata
Arcana, however, the translator seems to have allowed himselfremarkable freedom. Thus he singled out only a short section ofArc. 7 (vv. 125) to form poem 129 (p. 164). Moreover, in order to give the poem a coherent sense he altered the text ofvv. 24-5 as we read it in the rest of the manuscript tradition, namely duyy zAalopévy re St'Hépos. wd' dv éuoin | mdaot wéAot, rendering instead 'anima pervadens per aerem eum qui ei similis est. In other words, his Greek text is something like w@ 8' dv duoin wéAor, and waa appears to be left untranslated; of course we cannot say whether the change was his own, or was already present in the Greek source he was using. Arc.
1 ('Aliud Sancti Gregorii Theologi e libro poematum de principatibus') and 2 ('Aliud, de Filio'), are moved elsewhere (pp. 8 ff.). The translation seems to be by turns very free and extremely literal, not that I am in a position to assess the potentialities of Syriac for translating in accordance with criteria we generally define as 'literal', 'free', or 'paraphrastic'. It is therefore difficult to judge the readings that we should call isolated in the Greek tradition, among them Arc. 1. 6 woAAd«is om. Syr.; 1. 13 pnyvijevor ~ cadentes; 1.32 wovwtatos ~ solitarie; 2.7 Beou| deus; 2. 32 ayévaros sive dyévyntos] ingenerationis; 2. 45 Qeoio OM.;2. 53 Eu') te; 2.
59 wAeiov OM.; 2. 63 owdmaros] carni.
One of the hardest passages to interpret is Arc. 2. 6, where the manuscript tradition gives us dmjopos ot6ev TIMaNVa] amiopos ovdev CVa?L (dvéprepov ovdév Caillau). The Syriac translation reads
(vv. 5-7) 'nihil [sc. erat] quod praecederet Patrem magnum, hic enim omnia tenet, et alienum a Patre etiam nihil', giving the overall sense
'nothing is detached, separated from the Father. He who is begotten of the Father (is) the Logos of the great God.' The translator presumably read dmjopor neuter, agreeing with ovdév, and in his understanding of this rare word gave the prefix dzmore weight than the verb aipw.
Edessa, ¢.804: cf. A. Baumstark, Geschichte der syrischen Literatur (Bonn, 1922), 276. On Syriac translations see Sebastian Brock, "The Limitations of Syriac in Representing
Greek', in Bruce M. Metzger, The Early Versions af the New Testament: Their Origin, Transmission, and Limitations (Oxford, 1977), 83-98, and the articles reprinted in the Variorum collections Syriac Perspectives on Late Antiquity (London, 1984), nos. I-IV and Studies in Syriac Christianity (London, 1992), no. X.
Sancti Gregorii Theologi liber carminum iambicorum: Versio syriaca antiquissima e cod Vat.
CV, i, ed. P. J. Bollig, SJ (Beirut, 1895).
'§ We quote the readings of this Syriac version according to the translation kindly supplied to us by Fr J. Guirau of the Istituto Patristico Augustinianum, Rome, whose help we are very glad to acknowledge. xili TEXTUAL INTRODUCTION This points to agreement between Syr, L, and C(Va?): nihil = oddév. In other instances too the Syriac version agrees with L against 92: Arc. 1.6
8' TILVQ*, quidem Syr.} om. MaNVa (deest C); 2.
10 habent TILVa?Syr.] om. Q; 2. 80 dadpxov IILN?', incorporei Syr.] dvdpyov Q; 7.16 dAéyoroww] dAdyoto L, harmonia. . .formae non rationalis Syr. This last, an agreement in error, leads us to suppose that the family is older than Q.
4. The commentary on Gregory's Poemata by Cosmas ofJerusalem takes us back to a very early stage of the tradition. It was published by Mai from the codex unicus Vaticanus Graecus 1260, s. xii as Collectio et interpretatio Historiarum quarum meminit Divus Gregorius in Carminibus suts tum ex S. Scriptura tum ex profanis poetis atque scriptoribus:
Opus Cosmae Hierosolymitani cognomento Philogregorii = M. 38. 341-680.
The part that interests us here comes in cols. 452—66 (fos. 78°-87° in the Vatican manuscript).!© The Arcana group, however, is incomplete, for Cosmas does not comment on Arc. 4 and 6 (which is not to say he did not have them at his disposal). The title is wepi apyav; the regrettably few textual pericopae, comprising very few lines or passages, are:
Arc, 1. 11-13 (12 Aatropévov Cosmas cett.] Aayrduevor L); 2. 62-3; 2. 65-6 (65 yuw codd. ~ pév Cosmas); 2.
67-8 (vmépoyefe 5° ws Cosmas CMaNCuCoMb] izdvoyebe 8° ws S, trépoyed' ws LVa); 2. 68-9 (68 deme] dvdyeuse Cosmas); 2. 7o-1 (xvetuart ~ mvedatos Cosmas); 2. 72; 2. 73-4
(youva xaéyev] yotvara wév Cosmas); 3. 30-1 (31 67' éx] 6ve Cosmas); 3. 37-9; 5. 27-33 (29 ef Se TT] ef wey QL Cosmas; 30 éAifers Cosmas; 31 71s] 7
Cosmas); 5. 32-40 (33 9 yap 89 TINVa'] #5y yap CMaVaL Cosmas; 40 75 om.
Cosmas); 5. 53-7 (56 60cwv] Gowv Cosmas); 7. 38-9.
More remarkably, he too makes three poems out ofArc. 8, just like the @ manuscripts CNVa (see §1 above). Thus Cosmas' commentary affords usa series of isolated readings (most likely attributable to the copyist ofVaticanus Graecus 1260) and general agreement with the Q family. Significant agreement with L, by contrast, cannot be established.
© Mai's edition is, like all his editions, very much wanting from a philological standpoint. Despite the labour involved, a new edition of the text would be extremely useful. xiv TEXTUAL INTRODUCTION 5. Even earlier is the anthology generally known as Doctrina Patrum de incarnatione Verbi, which goes back to ¢.7o00. It includes two excerpts from the Poemata Arcana:
1. 25-35 (= Doctrina Patrum 1. 3, p.2
Diekamp), 2. 47-50 (= 40. 16, p. 302). In 1. 26 the Doctrina Patrum exhibits the reading peréaiofev as in Q, instead of weréwerra as in Nicetas and L.
6. The oldest manuscripts in the II family number two: Vaticanus Graecus 488 (Vl), s. x, which includes our portion of the Arcana together with Nicetas' paraphrase at fos. 1'-63", and Cusanus 48 (Cu), also s. x, 1-124.'' They are two completely independent copies, each showing isolated readings not shared with the other. Those in Cu include Arc.
1. 14 wae] woae; I. 32 OM.; 2. 22 péyar] péya; 3. 39 rexTés] TiKTds; 3. 58 pilav] pilyv; 4. 52 Kkparepoio] Kaprepoio; 5. 29
Beds} Gedv; 7. 16 Kai] re xai. In V1, note e.g. 1. 17 waporbe] mdpos; 2.71 vids] watpds3 3. 55 Geioror] Geiors; 5.61 Hépos] Hpeos; 6. 31 kai om.;8.
A relationship can also be established between V1 and the manuscripts Marcianus Graecus 494 (Md), s. xiii (as Palla has demonstrated), fos. 203'-206","* and Parisinus Coislinianus Graecus 262 (Co), s. xii, fos. 164°-175*, each of which contains a short paraphrase. Co's text is incomplete, breaking off abruptly after Arc. 7. 12. It is above all the paraphrase that links Co to Md: however, it is not Nicetas', even though Md attributes it to him. This leads us to believe that Co and Md descend from the same hyparchetype, which, moreover, may be derived from the source of V] (Palla's y), on the basis of agreements between V1, Md, and Co, as in the following instances: Arc. 2. 71 vids] matpés; 4. 70 dr') az'; 6. 11 rewwapecba] rewadpeba VIC, rewoueba Md. Agreement between Co and Md can be found in many other places as well: 3. 43 tpeccois] tpiotv; 3. 55 Geroddporoww]
Geoddporow; 4. 67 Acioowr] Acvawv; 6. 2 Tepiwyais] mEpiwzais; 6.
41 rw] re; 6. 55 réa0v] to0c0r; 6. 98 ad Beds] adfis; 6. 99 rocodrtov] roodtiov; 7. 1 §'] 8€.
For this group we must also consider another manuscript of the same period, Monacensis Graecus 488 (Mn), s. xiii, fos. 160°—162", which, however, exhibits only Arc. 1. 1-3. 68, all set out continuously as though in prose. This manuscript has no paraphrases; it must belong to the class whose earliest representative is V1, for like Co and '' On these I MSS cf. Palla, edn. 1-17.
XV TEXTUAL INTRODUCTION Md it contains the error Arc. 2. 71 vids] watpés. Since, moreover, Mn shares certain isolated readings with Md and Co (Arc. 1. 6 8' om.; 2. 32 dyévntos| ayévynros), it may be taken to derive from the same hyparchetype. Indeed, it seems to have more in common with Md than with Co, as appears from the following instances: Arc. 1. 2 rreptyeout] wreptyeot; 2. 25 weoonyt] weonys. For all that they are independent copies, we have not relied on Md and Mn because their texts are so poor.
To II also belong Marcianus Graecus (Mb), written in 1327, fos.
88*—108", which omits Nicetas' paraphrase, and Oxoniensis Baroccianus 96 (5S), s. xiv, fos. r'~42", which retains it. These two manuscripts have a common bond ina series oferrors and isolated readings that set them apart from the other manuscripts in the I family. Shared errors include Arc. 1. 17 dvdpes] iSpes; 2. 59 dynrés] dyamnrés; 2. 66 dwpoddpor] Swpoddpos; 3. 31 sawryp] dwrip; 4. 49 dvdpodédvor0] iSpoddvoco; 6. 2 dvrouévy| avropévny; 8. 98 a8' dpia KdKAos Edicoet]
Gua 7 neAiw ceArjvy ve. Among the isolated readings in S are 2. 12 pEepotrecoww] pepdmreawy; 2. 17 ef] ef ye; 2. 19 6 ypdvos] ov ypdvos; 2. 67 vaépoxefe] vadvoyebe; 3. 42 GxtioTov] Axrirov; 4. 10 Stéxpever] 8° éxpwwev; 5. 46 Cwoddpous] Cwnddpovs; 7. 22 dé om.; 8. 45 avdpouéotot] ddpopéoroc; among those in Mb are 2. 37 émdégerai] émdéEqrat; 2. 53 €u'] éudv; 2. 80 ef xeiva] éxeiva (ut VaCo); 3. 20 zupéds] mp6s3 3. 30 0M.; 4. 9 dxpoor] Gxpovor; 5. 13 xopds] yods; 6. 58 wAccev] wAeoas; 7. 103 OAKxe] OyjKe Se.
We follow Palla in employing the symbol to indicate the hyparchetype of S and Mb." Palla was obliged to admit that there were only a few conjunctive errors in y and 6 opposed to correct readings in Cu." On the basis of an error common to y and6 in 1. 2. 3. 47
(dv7idbw] dvidwv), he postulated a genealogy along the lines shown in the stemma in Fig. 1.
Parisinus Graecus 1220 (P)), s. xiv, fos. 191°—202", lacking the paraphrase, also belongs among the 6 manuscripts, concurring in its errors
(e.g. Arc. 1. 34 vwpets] vopets; 2. 59 dyntés] dyamnrés; 4. 49 dvdpodoévoio] époddvoio), and within it generally agreeing with Mb rather than S (2. 53 beds] Beot; Eu') éudv; 2. 80 f xeiva] éxeiva; 3. 20 mupds] mp6s; 330 OM.; 4. 9 dxpoor] dypovov; 6. 58 wAecev] wAcoas. On the other hand, the absence from Pj of many of the errors in Mb (e.g.
1. 2 areptyeoo| wreptyeo; 1. 28 povvoyevots] povoyevots; ' Palla, edn. 26 ff.
0 Thid. 30. xvi TEXTUAL INTRODUCTION MN / NS yt Cu V1 )J\ SMb Fic. 1. Stemma ofI family according to Palla 4. 22 éyeipew] dyeipew; 4. 53 attoddixros] avtoddixrov; 5. 12 ott dv] obr' dp'; 5.
13 xyopdés] xods) proves that it is not descended from Mb, but probably from the same manuscript from which the latter was copied. In his editio princeps of the Arcana, published in Paris in 1575, Billius is known to have made use of a manuscript belonging to Cardinal Sirleto (the 'Codex Sirletanus'), now lost; he compared it where necessary with the 'Codex Reginae Matris', no further identified.' Palla has demonstrated that we have very good grounds for taking Pj to be the Codex Reginae Matris used by Billius to correct, so far as possible, the errors in his copy-text, the Codex Sirletanus."
However, the reading a¢@:ros at Arc. 5. 55, common to Pj and the Codex Reginae Matris, merits attention: it is typical not of1], which is agreed on dypovos, but of Q (WV has dypovos like I). Influence from Q on Pj may also be detected in Arc. 4. 36 éya TT] éywye QL, éyed Pj, ye yp. S.¥.; 4.67 Aedacev LVa] Acicey CMaN, Actcowy IT, Aediooery Pj; 6.
2 avropévy QLCuVICoPj] avrowévyy SMb; 7. 97 redfe véov QP§] rev&éer ov IILN? and, significantly, the omission of Arc. 2. 10 from Pj and the entire Q tradition. At all events Pj is of mixed origin; we have not made use of its readings in establishing the text, since it adds nothing new with respect to Mb.
There is another sixteenth-century edition, that of D. Hoeschel, 5S.
Gregorii Nazianzeni, Theologi, Arcana; seu de Principiis, versus CCCC " Palla, edn. 32 ff. gives more details on the Sirletanus and its place in the tradition.
= This results from Palla's own recent researches, 'Alle fonti della prima edizione billiana dei carmi di Gregorio Nazianzeno', in Polyanthema: Studi di letteratura cristiana antica offerti a Salvatore Costanza (Messina, 1999). xvii TEXTUAL INTRODUCTION XXCH cum Paraphrast Graeca. Eiusdem Carmen contra Apolinarium. Item
.. Graecé nuncprimum publicata studio etoperaDavidisHoeschelij, A.M... .
LugduniBatavorum ... MDCXT; this rests on the I] tradition, but on a different branch from Billius' text. Palla has recognized that the pericopae of Gregory's text and the accompanying paraphrase in Hoeschel's edition are very similar to those in Md.* It is unlikely, however, that Hoeschel used Md or a manuscript closely resembling it, since the paraphrase is there explicitly ascribed to Nicetas, even though falsely. Hoeschel, by contrast, attributes it to him with hesitation, and even then not on the title-page ('cum Paraphrasi Graeca') but in the preface ('Hunc tractatum, cum perspicua paraphrasi, quam
Nicetae arbitror esse', 7; my emphasis). Probably, knowing the existence of a paraphrase by Nicetas, he attributed to him the paraphrase he had discovered, which he could not have done unless it had been anonymous. Now the same paraphrase is given anonymously in Co, a manuscript of the same subfamily as Md. There are several other factors besides that persuade us to attach Hoeschel's edition to the Co tradition:
(i) Like Co, Hoeschel's edition contains only Arc 1-6. To be sure, Co includes the first twelve lines of Arc. 7, which Hoeschel did not, but there is a logical explanation: he did not see fit to publish a manifestly mutilated poem." He occasionally splits a pericope in two, but that was one of the licences permitted to an editor; the fact remains that none of his pericopae is longer than in Co. That is to say,
Hoeschel felt entitled to a degree of licence only within a pericope already extant in his model.
(ii) Co and Hoeschel agree in a great number of errors, against discrepancies attributable easily enough to a sixteenth-century editor.
Chief among the errors is the omission within Arc.
1. 38-9, which makes the text run 7ucdaeis dbovepoi re, AUyvos oKoriots evi KéATOLS.
Another, to be attributed purely to a false reading found in Co, is the saut du méme au méme in 5. 62-3, reducing the two lines to 'EBpatwv
7éyvys. Other shared errors include Arc. 1. 3 dvadaivew] dudaivery; 3 Edn. 24 n. 22.
* This also explains why Hoeschel published a total of only 482 (continuously numbered) lines ofthe Arcana, as stated on his title-page. By adding up the lines in Arc.
1-6 we obtain 485 lines; from this total we deduct those omitted outright (3. 71, 5. 31) or reduced from two to one (1. 38~, 5. 62-3), which brings us down to 481 lines. But in the printed text, Arc. 1. 9 is misnumbered 10, so that an extra line is assigned to that poem; we thus arrive at a total of 482 lines.
XVili TEXTUAL INTRODUCTION 2. 73 mapérois] wapéras; 4. 13 order] éordvat; dvdcya wavra oM.; Hoeschel frequently used another manuscript ('alter') to correct the text he was printing, or to indicate variant readings. Since as a rule he was extremely scrupulous and precise in his work, it should be possible to identify this supplementary 'alter'. At Arc. 3. 67 Hoeschel notes in the margin 'Hactenus alter recentior'; this remark is very valuable, directing us at once to Mn, which as we have seen breaks off precisely at Arc. 3. 68. This hypothesis is confirmed by the agreements in isolated readings or in error between Mn and the alter codex in the margin of Hoeschel's edition, e.g. Arc.
1. 31 xivoupeévy in textu, ot kwounévy alter (= Mn); 2. 7 éxzeduais in textu, éxreduxes alter (= Mn); 2. 8 yevvyjrop: ion in textu, yevynrjpos fon ve alter (= Mn); 2. 17 yevvyoves dAAat in textu, yervjoves GAAa: te alter (= Mn); 2. 21 axpovov in textu, dvapyor alter, mendose (= Mn); 2. 55 woAA6v in textu, woAA@ alter (= Mn).
Nevertheless, we cannot set down Hoeschel's edition as a straightforward transcript of Co, for a certain number of the latter's errors have not passed into the printed text. We may of course ignore those obvious mistakes in grammar or metre that should (and above all could) have been corrected by a modern editor. Hoeschel would, however, have found it difficult to correct certain errors in Co such as the following, which are not found in his edition: Arc. 4.61 éo7:v Hoeschel cett.] om. Co.; 5. 23 cvvédnoav éudv réAos od Hoeschel cett.] om. Co;
6. 24 émoracinv Hoeschel cett.] émoraciay Co; 6. 54 76 tpiror Hoeschel cett.] tpérov Co. From this it follows that Hoeschel used a manuscript closely akin to Co, but not identical with it; furthermore, he states (p. 66) that he used a 'codex Augustanus'. His edition is in any case ofno value in establishing the text of the Arcana.
7. The Poemata Arcana were published for the first time together with others by Billius (Jacques de Billy) in 1575. They occupy a separate section within his edition; it may be noted that he published not only
Arc. 1-8, but also the remaining poems in the Nicetas corpus, in the same order as indicated in §1. It has been rightly inferred, therefore, that the manuscript used by Billius for this part of the edition—that given him by Cardinal Sirleto, now lost—belonged to the IT tradition.
This has been demonstrated by Palla, who has allocated the Codex Sirletanus to a class within this tradition created during the humanistic era. Billius himself informs us that where necessary he has
XIX TEXTUAL INTRODUCTION corrected unsatisfactory readings in this manuscript from another copy, the Codex Reginae Matris, which Palla has shown can most probably be identified with Pj. The origin of the editio princeps in a manuscript of the Nicetas tradition explains why the text of the Arcana, as we have read it so far, is essentially a IT text; it also explains why this group ofpoems bears the title ofArcana, the Latin translation of ra dxdéppyra. No subsequent editor, indeed, has made systematic use ofmanuscript sources, which is why the text most often cited up to now, that in M. 37. 397-429, 438-64, preserves the same characteristics, departing only slightly from Billius. The next edition, by Morellus, is merely an aggregate of the sixteenth-century editions and reprints; the textual modifications are few and insignificant, for the most part suggested by Billius' own notes. Caillau, the last editor of Gregory's poems, relied, so far as the Arcana are concerned, on certain manuscripts from the Bibliotheque nationale, and then only sporadically (Parisinus Graecus gg1, a late and worthless apograph, and also Co), and probably also Va.
The apparatus criticus presented here in support of our edition is selective: the isolated readings are omitted, except for those in L, which as indicated above may often be interpreted as the result of scholarly revision during the Byzantine era. We have ignored other manuscripts that were not in the strict sense descripti, if they offered what was clearly a mediocre or interpolated text (e.g. Pj and MnMqd).
Besides the isolated readings, we have omitted variants of spelling, accentuation, or punctuation.