You are on page 1of 18

Published following peer review in Revue d'Histoire des Textes, n.s. vol. V, 2010, p. 25-42.

The transmission of the text of the P scholia to Hermogenes


Rui Miguel Duarte Centro de Estudos Clssicos (University of Lisbon) Researcher FCT Fundao para a Cincia e Tecnologia (Portugal) Post-doc scholar

I. PROLEGOMENA

Of the popularity of Hermogenes rhetoric throughout the eras we have as evidence a production of commentaries by several school rhetors, since Harpocration, a contemporaneous of his and a critic of his issues theory1. Among those commentaries there is a plentiful corpus, whose textual tradition is inseparable from the text they comment, figuring in a restrict family named as P by H. Rabe2, one among the significant number of codicological families of codices that transmit all the treatises attributed to the rhetor of Tarsus. Those commentaries (or scholia) had till now only one edition, by Christian Walz, in the volume 7 of his series Rhetores Graeci3. Some other scholars have done in the last century researches on these scholia and on the manuscripts that bear witness of them. Those are names as S. Glckner or H. Rabe4,
1 On the other hand Harpocration would have influenced the hermogenic theory of the types of style.

See Michel PATILLON, Anonyme de Sguier. Art du discours politique, Paris, 2005, p. LVIII-LXV. On the problem of the identification of rhetors with this name, see Malcolm Heath, Porphyrys rhetoric, in Classical Quaterly 53 (2003), p. 144-166, sp. 147, and idem, Theon and the history of the progymnasmata, in Greek, Roman and Byzantine Studies 43, 2003, p. 129-160, sp. 132 and following. 2 Cf. H. RABE, cf. Rhetoren Corpora, in Rheinisches Museum 67, 1912, p. 321-357, sp. 323. See below the explanation of the manuscript sigla. The sigla used are RABEs, except for the one referring to Monacensis graecus 8, taken from Stephan GLCKNER Die Handschriften der P-Scholien zu Hermogenes (Breslau nowadays Wroclaw 1928), p. 5, and for the ones identifying lost and conjecturally reconstructed codices, which are mine. 3 C. WALZ, Rhetores Graeci vol. VII, Stuttgart & Tbingen, 1836, p. 104-696. 4 From S. GLCKNER, cf. the work quoted above n. 2; Quaestiones rhetoricae. Historiae qualis fuerit aeuo imperatorio capita selecta, Breslauer philologische Abhandlungen 8.2 (1901), p. 1-115; ber den Kommentar des Johannes Doxopatres zu den Staseis des Hermogenes, in Wissenschaftliche Beilage zum Jahresbericht des Kniglichen Gymnasium zu Bunslau I, Kirchhain, 1908, and II, Kirchhain, 1909; Zur Komposition der P.-Scholien zu Hermogenes , in Satura Viadrina altera, Festschrift. Zum 50 jhr. Bestehen d. Philol. Vereins zu Breslau, Breslau nowadays Wroclaw 1921), p. 1-11. As for H. RABE, see op. cit.; Hermogenis opera, in Rhetores Graeci II, Leipzig, Teubner, 1913 (reimpr. 1985) [introduction]; Prolegomenon sylloge, in Rhetores Graeci XIV, Leipzig, Teubner, 1931 [praefatio]. 1

to whom an edition of Hermogenes and comprehensive studies on the manuscript tradition are due. Later on, by the mid-century, G. Kowalsky5 and his disciples L. Rychlewska6 and V. Borzemska-Lesnikowska7, and later the Italian R. Romano8, continued this study. In 2006, at the University of Aveiro (Portugal), I presented my Pd.D. thesis. In the thesis, I gave an account of the history of the texts transmission throughout the manuscript tradition up to Walz edition, and justified the need of a new critical edition and the criteria for it. A section out of the P scholia was chosen: the chapters 1-16 (consisting of a general introduction to the system), corresponding to Walz 7 104-245. The new critical text of these scholia was provided with a Portuguese translation9. The research I was able to do led me to some conclusions regarding the manuscript filiation. For the constitutio textus, only Pa and Pc were taken into account. The apographs, as well as Walz, were considered mainly for the purpose of establishing a stemma, even though they have been used occasionally, whenever they borne good conjectures. These conclusions can be summed up as follows: (1) Pc would be the only of the extant witnesses of its branch; (2) all the other codices were included in the branch represented by Pa, which I named ; in this branch, two filiation lines, Pa and , were determined, the last one a conjecturally lost codex; (3) the net of dependencies of Pa was also complex. My former research depended upon a limited sample of texts, the scholia I studied, and neither upon the Hermogenic text itself nor upon other texts borne by the manuscripts. Obviously, the conclusions that I came up to could be either confirmed or infirmed. In fact, further and closer researches provided enough evidences to make me reconsider some of them. And that is exactly a research update that I shall propose in the following pages. Its goal is to set the point of the present knowledge of the tradition of the text of the scholia.

5 See his edition Hermogenis de Statibus, Travaux de la Socit des Sciences et des Lettres de Wroclaw,

series A, nr. 1 (Wroclaw 1947). 6 In Anonymum Hermogenis Statuum interpretem (Rh. Gr. VII, 397-442 WALZ) cum Nilo (Par. gr. suppl. 670 sqq. 36v-65r) collatum observationes criticae, in Eos 41, 1 (1940-1946), p. 173-184, and 42, 1 (1947), p. 195-211. 7 De Anonymo Hermogenis Statuum interprete (Rh. Gr. VII 320-397 W.) cum Nilo (Par. suppl. gr. 670 ff 1r-36r) comparato, in Analecta Hermogenica, Travaux de la Socit de Sciences et Lettres de Wroclaw, srie A, 42 (1951), p. 17-43. 8 Il commentario a Ermogene attribuito a S. Nilus di Rossano (Par. suppl. gr. 670, ff. 1-179), in Epeteris Etairias Byzantinon Spoudon 47, Athens, 1989, p. 253-274; Niliaca ab Anonymo Rh. Gr. VII W. non expressa, in Vichiana 3rd series, 2, 1991, p. 263-264; Un nuovo capitolo del commentario a Ermogene attribuito a S. Nilus di Rossano, in Vichiana 3, 1992, p. 189-198; Nuove ricognizioni sul commentario a Ermogene attribuito a S. Nilo di Rossano, in Orpheus 21 1-2, 2000, p. 84-91. 9 R. M. O. Duarte, Comentrios ao tratado sobre os Estados de causa de Hermgenes de Tarso por autor annimo, Aveiro, Universidade de Aveiro, 2006. 2

II. THE EDITION OF WALZ VOL. VII (W7)

Now, a critical analysis of the Walz edition (defective regarding several aspects) shall be provided; this analysis is justified by the fact that this work has been until today the reference edition of the text of the scholia and the sole intermediary between this and us. The codices used by Walz were Pa, Pc (named by the abbreviation Par.1), Pb (Par.2), Aa, Ne and Mb. In the introduction (introduction p. III), Walz states that he benefited from a descriptive study of Mb, having hence done a scarce confrontation with Ne and having corrected the commentaries based on Pc; so that, whenever the source of a correction is not named, one should infer that it is Pc. It could also be perceived that W7 deeply relies on Mb. As for Pb, he states that this codex was not profoundly checked, but only in some pages. Pa, however, was not used but occasionally. Two examples of omissions in Pc were, as it is mentioned, restored through Pa. The W7 edition is, for several reasons, an imperfect work, mainly because it almost despises Pa, depending especially on apographs. There is no real apparatus, but footnotes to the text, in which he sporadically quotes variants and transcribes scholia minora without noticeable criteria. On the other hand, he presents plentiful omissions. The most of these are due to homeoteleuton: 1.34.1-2 - omitted by W7 120.5 after || 1.35.11-12 - omitted by W7 121.3 after ) || 2.48 all the scholium omitted by after W7 127.11 || 3.94.2 : omitted by W5 159.31 || 9.2.6-7 - omitted by W7 202.25 after || 13.6.5-6 - omitted by W7 215 4 after . As for some other variants and omissions W7 relies directly on apographs as the source of the text: 1.19.2-4 - omitted in Mb after due to homeoteleuton, whence also W7 113.26 || 2.73.1 omitted in Mb after , whence W7 145.7 || 2.b.1 P: Mb whence W7 130.17) || 2.75.5-6 - omitted in Mb after whence W7 145.30) || 3.94.2 : omitted in Mb (cf. W7 159.31) || 3.96.5-7 - omitted in Mb due to homeoteleuton (cf. W7 161.17 after ) || 9.1.5-6 omitted in Mb due to homeoteleuton (cf. W7 202.5 after ).

There are still cases of misreading by Walz of the lesson he used: 1.2.1 Pc whence W7 and I: because of iotacism Pa Pb and Ne, in which W7 erroneously read (cf. 105.19 n. 18) || 1.17.13 transferred after in Pc, having W7 read instead of (cf. 113.8 n. 22) || 1.21.8 P whence W7 and I: reads W7 in Pc (cf. 114.13 n. 5) || 2.79.27 P whence I: W7 perhaps based on Pb (cf. W7 148.17) || 11.1.2 codices whence I: W7 reads erroneously in Mb (cf. 205.21 n. 1) || 13.10.48 Pc whence W7 and I: Pa, having W7 in Pb misread (cf. 218.12 n. 27) || 16.a scholium to be transferred after 16.2: after the scholium 5 in Pa after the scholium 1 in Pc and not omitted in this codex as W7 declares (cf. 234.20 n. 1) || 16.a.23 P whence W7 and I: Pb where erroneously W7 reads (cf. 235.24 n. 16). W7 however, gives some good lessons, from his own conjecture or taken from the manuscripts. It is worthwhile to quote all of them, since they were all taken as valid lessons for the establishment of the text: 1.49.4 correctly restored W7 127.16 and I from Demosthenes First Olinthiac 1.23: in all the codices || 2.59.6 before correctly added by VhMb, whence W7 136.4 whence I, cf. below 2.59.7-8 : omitted in P || 2.68.13 PaNe thus correctly W7 142.1 cf. n. 47 (though he did not check Pa) whence I: in the other codices || 2.85.9 W7 152.3 whence I: in the codices || 4.7.11 W7 181.11 whence I cf. the source of the scholium (W4 210.6): codices || 5.9.2 corr. Pb1 whence W7 186.12 and I: in genitive in the other witnesses possibly by homeoptoton with || 6.12.3 whence W7 191.23 and I: Pa || 7.9.13 correctly W7 199.2-3 whence I: Pa Pc || 9.2.20, 32 correctly W7 203.18, 204.2 and I: Pa Pc || 14.10.9 Pa equally W7 225.13 and I: in the other testimonies || 15.2.12 corrected by W7 227.26 based on Demosthenes Third Philippic 9.5 and equally I: Pa . Pc || 15.2.12 suppressed by W7 228.1 based to Demosthenes loc. cit and equally || 15.2.12 Pa whence and equally W7 loc. cit and I.: in the remaining codices || 16.2.1 corrected by whence W7 231.8 and I: in the other codices. However, other conjectures, corrections and additions of his were not considered acceptable. Because thy were not relevant for the establishment of the text, it is not useful to give here but a sample (over seventy-five variants)10:
10 See a more exhaustive list in my thesis p. 52-53.

1.6.1 after added by W7 107.11 || 2.54.19 P whence I: W7 132.25 || 2.84.4 P whence I: W7 151.5 || 2.87.54 P whence I: W7 154.29 || 3.98.17 after added by W7 165.3 || 3.16.9 P whence I: W7 174.14 || 4.8.6 before added by VhMb whence W7 182.14-15 || 6.10.2 P whence I: W7 191.7 || 7.8.1 P whence I: W7 198.11 || 9.2.20 before added W7 203.17 || 12.1.3 P whence I: W7 206.18 || 12.1.22 before added by W7 207.19 || 13.1.16 P whence I: Mb whence perhaps W7 210.8 || 13.7.5 before possibly by homeoteleuton added by W7 215.12 || 13.10.39 P whence I: W7 217.30-31 || 14.8.18 after added by W7 223.13-14 || 15.1.14 P whence I: W7 227.5 || 15.4.12 before added by W7 228.28 || 16.5.10 P whence I: W7 233.17 || 16.d scholium signed with number 15 by W7 242.8 and f. || 16.d.9 P whence I: W7 ib. || 16.f.14 P whence I: Pc || 16.f.15 P whence I: W7 244.28. Other W7 errors relate to the numbering of the codices: 3.90-99 as 100-109 || 3.20 marked with number 8 || 4.3 scholium not numbered. In conclusion, the W7 edition is on the whole an imperfect work, in spite of the good conjectures. It relies essentially on the apographs of Pa and on Mb as well, almost neglecting Pa itself, and omitts a great number of textual materials. Therefore, as far as the text's reconstitution is concerned, this work is not but seldom useful.

III. THE TRADITION OF THE TEXT

THE CODICES

The family of codices designated P by Hugo Rabe has, as its most ancient known witnesses, two codices, both from the eleventh century: Parisinus graecus 1983 (Pa); Parisinus graecus 2977 (Pc). Still more recent codices belong to this family, dated from the thirteenth to the sixteenth centuries, out of which six were examined: Parisinus graecus 2916 (Pb = Par.2 W7); Neapolitanus (= Farnesinus) II.E.5 (Ne); Ambrosianus P 34 sup. [graecus 617] (Aa); Vaticanus graecus 2228 (V); Palatinus Vaticanus graecus 23 (Vh); Monacensis graecus 8 (Mb).
5

The P scholia had up till now one only edition, published by Christian Walz, in the volume 7 of the Rhetores Graeci collection, pp. 104-696 (W7). The manuscripts that bear the scholia will be examined right away. As Pa and Pc were the codices the constitutio textus in my thesis was mainly based on (regardless of a few good corrections provided by the apographs), they are worth a more detailed account. Parisinus graecus 1983 (Pa), Paris, Bibliothque Nationale de France, parchment, 260 x 215 mm, 295 folios, tenth-eleventh century11. The hermogenic text, which occupies the folios 44-284 (the scholia edited in my thesis, corresponding to W7 104-245.4, being found in folios 44r-59v), is written in a circumscribed space, vertically centred on the page, and close to the line of margin of the spine, the rest of the space being reserved for the writing of the scholia, which occupy 55 to 62 lines, from the top to the bottom, around the text of Hermogenes. This one occupies circa one sixth of the total written space, which show that the folios were especially prepared to receive not only the text of the rhetor but also the scholia (known as scholia maiora or greater scholia). Both the hermogenic text and the text of the scholia were apparently copied by the same hand; the size of the scholias handwriting is smaller than that of the hermogenic text and the ink colour is light brown, whereas the hermogenic text is dark brown. The handwriting is well taken care of, well drawn and regular from the top to the bottom of the folios. The scholia relate to the text through numerical note references marked with red ink in the interlinear space of the hermogenic text, and retaken at the margin of it. The space between the text of Hermogenes and that of the scholia was left blank in order to receive other scholia, with characters of inferior size. The note references of the later ones consist of graphic signs. These facts, in addition to the disposition of these scholia (between the main hermogenic text and the greater scholia), make us think that the scholar community of the Hermogeness readers of that time would consider these scholia of secondary importance. Due to this, these scholia are known as scholia minora (smaller scholia)12. The text of the scholia is divided in several parts, which in my edition are considered as chapters, similarly to what has been done in W7. The titles of the chapters, when they appear, stand out detached, in darker ink and in bigger size, either laterally to the hermogenic text, or over him, or in the scholia text itself, separated by the rest of the text in the line by larger spacing. Over the line corrections were made here and there by the same hand that copied the scholia (Pa1), some others were made by a second hand (that I designate as Pa2) using darker ink. Still there are marginal notes and corrections
11 For a summary of the contents of the codex see our thesis, p. 39-41, M. PATILLON, Corpus rhetoricum, Paris, Les Belles Lettres, 2008, p. L-LIII and G. AUJAC, Recherches sur la tradition du de Denys dHalicarnasse, in Revue dhistoire des textes 4, 1975, p. 32-35. 12 See the edition of WALZ.

from a third hand (Pa3). To this hand are due text restorations in several places, because of severe deterioration of the material. This hand uses an ink colour similar to the one used in Pa2, but with different character drawing, less careful and bigger in size. The second hand dates probably from the twelfth-thirteenth centuries, so it is necessarily earlier to the copy of the codices Pb and Ne, which depend on it. One can date the third hand, based on the handwriting, from the thirteenth-fourteenth centuries. The work of Pa3 is necessarily later than the copy of the model of , for in none of them are there evidences of having used versions of Pc, which in the texts tradition are considered to be unique. The interventions of the third hand are the following: 1.5.3 in the other witnesses whence I: Pa3 || 1.21.4 Pc whence I: Pa3 unreadable Pa || 1.21.11 in all the witnesses: Pa3 || 2.53.1 : Pa3 || 3.97.10 : Pa3 || 16.3.2 Pa3. Parisinus graecus 2977 (Pc = Par.1 W7), Paris, Bibliothque Nationale de France, parchment, 232 x 169 mm, 344 folios (with a second folio 305), eleventh century. The contents of this codex are similar to those of Pa. Compared to this one, the preparation of the codex, as well as the handwriting, is less cared for. One can notice the intervention of one hand only, which used brown ink. To this hand are due as well some supralinear corrections. The mode of edition is similar to the one of Pa. The Hermogenes text (copied seemingly by the same hand), which occupies the folios 60-326r (the scholia edited in my thesis being found from folios 60r-79v), is written in a circumscribed space vertically centred on the page, and close to the line of the spine margin. The rest of the space of the page is reserved to the writings of the scholia, which occupy about 48-55 lines, from top to the bottom, around Hermogenes text. Hermogenes text occupies only about one sixth of the total written area. The handwriting of the scholia has smaller dimensions compared to the hermogenic text. The scholia have also numerical note references. Besides these scholia, there are the scholia minora, that occupy an empty space between the text of Hermogenes and the greater scholia. The text of the scholia is equally divided into chapters as in Pa. The titles of the chapters stand out detached in darker ink and larger size, either in the text of the scholia itself separated from the rest of the text with larger spacing, or at the either margin. Parisinus graecus 2916 (Pb = Par.2 W7), Paris, Bibliothque Nationale de France, paper, thirteenth century13. The edition mode is similar to that of Pa and Pc. The scholia surround the frame of the text reserved to the text of Hermogenes. The system, however, is not strict: though
13 For a more summary of the contents of the codex see our thesis p. 41-43, and M. PATILLON, op. cit.,

LIII. 7

the Hermogenes text frame has been always prepared close to the spine margin, its position relating to the pages height, as well as its dimension, vary. There are cases of pages exclusively with text of scholia (71r, 86v, 90r, 92r). The text of the scholia edited in my thesis is found in the folios 68r-95r. Neapolitanus (= Farnesinus) II.E.5 (Ne), Naples, Biblioteca Nazionale Vittorio Emanuele III, paper, thirteenth century. The edition mode is similar to that of Pa and Pc. The text of the scholia is written around the frame of the text reserved to the text of Hermogenes, which occupies a space of about one sixth of the total written area, vertically centred on the page, from top to bottom, and limited by the line of the spine margin. The space left over is reserved to the writing of the scholia. The preparation of the folio, the rulings and the handwriting are more cared for than in Pb. The text of the scholia edited in my thesis can be found in folios 58r-77v. Palatinus Vaticanus graecus 23 (Vh), Vatican, Biblioteca Apostolica Vaticana, paper, end of the thirteenth century. The edition mode is similar to the Pa and Pc. The text of the scholia is written around the frame of the text reserved to the text of Hermogenes by the edge of the spine margin. However, the delimitation of areas for the text and the scholia is not strict, as it can be seen in other manuscripts. The text positioning, relating to the pages height, as well as its dimension, vary. On folio 86r, the frame of the hermogenic text is situated at the bottom of the page. The text edited my thesis can be found in folios 73r-88r. In the other codices only the scholia were copied, not the text of Hermogenes. Vaticanus graecus 2228 (V), Vatican, Biblioteca Apostolica Vaticana, paper, fourteenth century. A compilation of continuous commentaries, in which selected P scholia (entitled ) were mixed up with others scholia from another commentator, which can be found on the Vindobonensis phil. graecus 130 (Wc) as well14. As far as an exam could be made, the scholia are introduced by quotations from the of Hermogenes, which could be identified as the lemmata of the scholia. Occasionally the titles of the chapters appear, in larger handwriting and darker ink. The selection criteria
14 Cf. H. HUNGER, Katalog der griechischen Handschriften der ster. Nationalbibliothek, I, Wien, 1961, 148-149);

S. LILLA, Codices Vaticani Graeci, codices 2162-2254, Vatican, 1985, p. 307-313S. GLCKNER, ber den Kommentar II 8-12; Zur Komposition sp. 6; for a description, H. RABE, Aus Rhetoren-Handschriften. 5. Des Diakonen und Logotheten Johannes Kommentar zu Hermogenes , Rheinisches Museum 63, 1908, p. 127.151, sp. 128-130. 8

of the scholia, as in the case of Ambrosianus sup. P 34 [graecus 617] (see the following commentary on this manuscript), are not understandable. The textual materials corresponding to the scholia 1.3 (= W7 106.9), etc., until 16.13 (= W7 241.28) can be found between the folios 122r to 147r. Ambrosianus P 34 sup. [graecus 617] (Aa), Milan, Bilioteca Ambrosiana, paper, 212 x 155 mm, fifteenth-sixteenth centuries. Between the folios 285-298, there have been compiled scholia from 1.1.9-10 (= W7 104.15-16), 1.2.5-1.3.5 (= W7 106.5-13), until 9.2.7-8 (= W7 202.26-203.1), in a rapid and incoherent way, proceeding from this point on in Latin. Monacensis graecus 8 (Mb), Munich, Bayerische Staatsbibliothek Mnchen, paper, sixteenth century. It is an apograph of Vh, as it will be shown further on. Out of all apographs, this is the most recent one and the only one, among the ones that only transmit the text of the scholia, where there was a concern in copying the text as it was, without contaminating it with others of different origins. The text corresponding to that edited in my thesis occupies the folios 140r-192r.

THE TRADITION BRANCHES

I have resumed above the threefold conclusion I could formerly get to, concerning the tradition of the text. I shall now take up the discussions on these subject matters, with the reformulations that need to be done. (1) After a comparative examination of the codices Pa and Pc, I came to the conclusion that these manuscripts represent each one of them a distinctive and parallel branch of the tradition depending on a common archetype P. The collation of all the above listed manuscripts, and the divergences found between them, allowed me to group them under two main divergent lines descending from P, each group depending necessarily on a lost hyparchetype, and : to the -branch belong Pc and V, the rest of the witnesses are grouped in . These divergences are related to a great number of variants and other aspects. In my thesis15, it was assumed that V belonged to the -branch, given the variants taken into
15 See in our thesis 45, 48 and the stemma 50.

account, but a closer comparative analysis of the witnesses has imposed another conclusion: it belonged to the same branch as Pc. Evidences can be found not only in the scholia but also in other texts borne by the manuscripts. Separative variants between and are, for instance, the following: 1.48.4-5 ... ... whence I: -... -... - 16 || 2.76.1 whence I: || 2.76.9 whence I: thus || 3.92.3 I from the source (see below) and thus corrected by Pb: PaVh thus probably unreadable in Ne omitted in Pc V || 3.17.9 whence I: || 6.1.18 Pc whence I: omitted in V . The case of 3.92.3 (alternation / ) is due to the confusion between the respective abbreviations. Indeed these sometimes may be confounded in a faster trace. A similar case is the lesson 2.79.30 : W7 presents , maybe based on Pb. A digitalized image of the former in Pa is here presented, where can be seen:

. Pb has clearly (with a rude spirit), perhaps based on (unsure, due to the unreadableness in Ne):

. The examination of the context and the consultation of the source (Syrianus R2 40.17-2017) led to the conclusion that the correct lesson is and not : Lemma St. 34.2-8]
3.92.1-3 <sc. > . , , . This type of matter <the ill-forged> is classified in the impossibility and is the contradiction of the history. It is constituted by all the parts, but it only includes a defective exam relatively to the persons, considering that in the given example the dead appear as being alive.
16 The sigla P, , and are used whenever the lesson of a lost manuscript can be reconstructed

whether upon the accord of all the witnesses of the respective branch that bear the given lesson, or upon the most part of the witnesses, regardless of corrections made in any of their descendants. 17 Syrianus, Syriani in Hermogem commentaria, ed. H. RABE, Rhetores Graeci vol. 2, Leipzig, Teubner, 1893. See apparatus ad locum. On this variant and other questions related to the practice of palaeography and textual critic of the scholia P, see R. M. O. DUARTE, Aventuras de um editor de textos crticos gregos, in gora 3, 2001, p. 25-49 sp. p. 33. 10

The error in reading and the confusion therefore was probably of the model , hence the agreement of Pa and Vh. Some of the cases above listed of oppositions versus are not uncommon: /, or even indicative/conjunctive (thematic vowels / /). By themselves, variants like these are insufficient as separative. Nevertheless, what is striking is the accumulation of them. Such an accumulation is probably not casual, but seems rather evidence of distinct tradition branches. As for evidences from other texts borne by the manuscripts than P the scholia, I rely on Rabe and in his edition of Prolegomenon sylloge (R14). In the preface18, he expressly states that Pc and V have the same model (even though the later is a worse witness). The manuscripts Rabe used were mainly Pa, Pc and V, among others. On the following pages, I present the variants registered in the critical apparatus on the texts 15 (238.1255.3) and 16 (255.4-258.12) of Rabes edition of the prolegomena to the Hermogenic , that corroborate the separation between and : Text 15: 238.2-3 Pa whence R14: || 238.6 whence R14: Pa || 238.7 Pa whence R14: || 238.11 whence R14: Pa || 238.13 Pa whence R14: || 238.13-14 whence R14: omitted in Pa || 239.1 Pa: omitted whence R14 || 239.4 Pa whence R14: || 239.12-13 whence R14: Pa || 239.13 the first Pa whence R14: omitted in || 239.16 the first Pa whence R14: omitted in || 239.20 Pa whence R14: || 241.6 whence R14: Pa || 241.17 Pa whence R14: omitted in || 241.23 Pa whence R14: 241.23 omitted in || 241.25 Pa whence R14: omitted in || 242.4 Pa whence R14: || 242.8 before omitted in Pa whence R14: added in || 242.10 whence R14: Pa || 242.14 Pa whence R14: V Pc || 242.15 Pa whence R14: || 243.5 : Pa || 243.23 whence R14: Pa || 245.5 Pa whence R14: || 246.2 Pa whence R14: || 246.5-6 whence R14: Pa || 246.13 whence R14: omitted in Pa || 246.20 whence R14: Pa || 247.6 whence R14: Pa || 250.1 whence R14: Pa || 250.6 whence R14: Pa || 250.11 Pa whence R14: || 250.13 whence R14: Pa || 250.18 Pa whence R14: || 250.22 Pa: whence R14 || 251.3 Pa whence R14: omitted in || 251.12 R14: Pa || 251.14 Pa whence R14: || 251.18 whence R14: omitted in Pa || 252.13
18 R14 p. LXIX.

11

Pa whence R14: || 252.26 Pa whence R14: || 252.27 inc. - 253.11 des. omitted in || 253.26 Pa whence R14: || 254.25 whence R14: Pa. Text 16: 255.12 Pa whence R14: || 256.13 whence R14: Pa || ib. whence R14: Pa || 256.14-15 Pa whence R14: || 256.18 Pa whence R14: || 257.15 Pa whence R14: || 257.22 Pa (cf. Herm. Id. 225.10) whence R14: || 258.2 Pa whence R14: . Though some of the above cases are examples of trivial error types such as iotacism (one case), anastrophe or inversions of the order of two words or syntagmata (one case), for and vice versa (five cases) , they were quoted for the purpose of clearly demonstrating the divergence between the traditions branches, specially when Pc and V altogether (= ) agree against Pa. There are nevertheless others cases in which whether V or Pc disagrees with one another and agrees with the variant of the other branch: Scholia P: 1.3.7 P corrected over the line by Pa1 whence - also Pc - V. Prolegomena to the Hermogenic (R14): Text 15: 239.1 PaV whence R14: Pc || 244.24 PaPc whence R14: V || 247.22-23 PaPc whence R14: V || 248.14 PaV whence R14: Pc || 248.15 PaV whence R14: Pc || 249.14 PaPc whence R14: V || 251.11 PaV whence R14: Pc || 254.12 PaV whence R14: Pc. Text 16: 255.15 PaV whence R14: Pc || ib. PaPc whence R14: V || 256.1920 PaV whence R14: Pc19. Since V does not transmit of the P scholia but some extracts, the analysis that follow of the opposition between and is all based on evidences from Pc. These other differences are the frequent omissions of Pc comparing to Pa (one hundred and seventy one cases), but few and not very significant are the opposite cases (thirty-eight). Indeed, some of the omissions of Pc are very extensive: eight cases correspond to two or more lines in my text: 1.22.11-13, 1.27.2-6, 2.83.7-10, 2.87.53-55, 3.92.2-4, 3.98.20-21, 4.6.15-16, 13.4.4-7. Some of the omissions (four) are due to homeoteleuton: 2.83.7-10, 4.6.15-16,

19 The tradition of the texts was seemingly complex and, so to speak, multigenic. In the cases where Pa

and V agree against Pc one could think of contamination of V with a tradition of the text represented by the branch . See e.g. variants for R14 254.12, 255.15, or 256.19-20. On the other hand, there are cases where Pa and Pc oppose to V; in such cases, the contamination would have been of Pc with a tradition of the text represented by the branch . See e.g. variants for 244.24, 247.22-23. Cases of iotacism 248.14 /) singulier/plural 251.11 /, 255.15 /, or even feminine/masculine 249.14 / should no be considered as strong evidence, because they are common variant alternatives. 12

13.1.7; 13.4.4-720. Other differences come from common error types (iotacism, inversions). Others are related to the edition of certain textual materials among the scholia maiora, in one of the witnesses, and among the minora on the other: such as the cases 1.19-20, 2.53, 3.14, 11.2, 16.d. There are also cases of transposition of textual material, sometimes of whole scholia: 2.c, 16.a, 16.b, 16.c and 16.d. (2) Rabe suggested that Vh would not depend on Pa, but on a manuscript close to this one21. The textual materials among the folios 36-41 (previously listed with the numbers 8-12) are in the wrong order in Pa and Pb, but in the correct order in Vh. For Glckner, on his turn, Vh would depend on Pa, but with at least one intermediary22. His conclusion is based on the textual extract 2.75.5-6 . , . , that is, the three inferior lines of f. 48r in Pa. Originally omitted in Vh, it is further on added to the line and to the margin of the scholium. Another evidence is the omission of the material W7 307.14-16 in Vh. Rabes insight, as I believe, is correct. Vh effectively depends on a manuscript close to Pa and with the same model as this one. I named that manuscript . Its date is uncertain. The divergence of the two tradition lines Pa and (meaning the agreement of all the three manuscripts or, at least, of Vh and Aa) can be conjectured on the basis of the sum of separative variants like the following ones: 1.11.2 P whence I: omitted in || 2.52 title of the chapter correctly before the 2.52 in Mb: at the margin of the in Pc omitted in Pa over the text of Hermogenes in Vh || 3.95.2 Pc whence I: thus Pa equally || 3.16.10 Pc whence I: Pa || 3.19.8 P whence I: || 3.20.1 Pa whence I: equally . The strongest evidence of the separation between the traditions branch represented by Pa and the one represented by Vh is the omission of the title of the chapter 2 in Pa and their apographs. In fact, if in Pa the title were missing, the copyist of would necessarily have before him another codex, in spite of being close to Pa. However, another lesson could be accounted as a counter-example, leaving by now opened the hypothesis that Vh as well as would depend, in fact, on Pa: 1.6.3 P whence I: Vh omitted in the rest of the manuscripts. This form is an aorist participle of shiver with cold. A digitalized image of that place of Pa is presented here

20 See apparatus ad loca. 21 H.RABE, Rhetoren-Corpora 324 and n. 1. 22 Op. cit. 6-7.

13

. Over the letter one can see a straight horizontal stroke (with accent). This is one of the possible ways to abbreviate . I would have , a hapax. Pc has an arched stroke with the form of a tilde over the , an abbreviation that can be used for , whence 23. Would Pa have mistaken? In general, the copyist of this manuscript worked carefully, therefore the reading appears as strange and unusual and resulting from confusion between the abbreviations and . This situation is neither frequent nor unique. In the following folio and very close to each other, one can find the lessons and . In these situations the accentuated was similarly written with straight horizontal stroke over the previous letter. Both cases are obvious, for they are very common words and of a basic Greek lexicon, so that under no circumstance would they be read and . This variant is not enough so as to set forth the hypothesis of Vh depending on Pa because, as it could be seen, the evidences that Vh depends instead on , are stronger. This is a mere confusion between abbreviations probably committed by the copyist of . From Vh obviously derives Mb, because it reproduces variants and errors of Vh, to which still it adds its own errors. Rabe had already noticed this filiation24. (3) The readings shared by Pb and Ne against Pa allow sustaining the conjecture of an intermediary between Pa and these codices25. I identify that codex, written between the twelfth and thirteenth centuries, with the abbreviation . This codex can be reconstituted upon the evidence provided by the following variants: 1.35.11 Pc and whence I: Pa || 1.35.17 Pa whence I: all the other witnesses || 1.48.4-5 Pc whence I and thus corrected by : -... -... - || 2.c after scholium 2.59 Pc whence I and thus corrected by : after 2.61 in || 2.76.1 Pc whence I and thus corrected by : Pa || 2.76.9 Pc whence I thus corrected by : Pa || 2.87.86 P whence I: || 3.94.4 Pc whence I: omitted in Pa over the line Pb as well seemingly Ne || 3.97.10 I based on Herm. St. 34.15: Pa3 || 4.1.1 Pa whence I: equally Vh omitted in || 5.14.1 Pc whence I: Pa || 15.2.12 I based on Demosthenes
23 As well as for , which is improbable in this context. 24 H. RABE, Rhetoren-Corpora, loc. cit. 25 On the dependencies of the apographs to Pa cf. the studies of RABE, Rhetoren-Corpora loc. cit. n. 1

and GLCKNER, loc. cit. 14

On the Halonnesus 7.5: Pa || 15.4.10 4.10 P whence I: || 16.a after 16.2: after 16.5 Pa among the minora in || 16.2.1 whence W7 231.8 and I: in the other codices || 16.b after 16.4 my conjecture: after 16.a Pa among the minora || 16.c after 16.b in my conjecture: among the minora || 16.d after 16.7 my conjecture: after 16.14 after 16.12 . These codices follow the second hands interventions (Pa2) in Pa and some corrections done by the first hand (Pa1). The dating of Pa2 (twelfth century) forms the terminus post quem of the dating of . The variants dependent on Pa2 are: 2.54.21 Pa whence I: corrected over the line by Pa2, whence || 16.5.33 Pc whence I: thus Pa2 whence seemingly Pa. The variants depending on Pa1 are: 1.3.7 the vulgate whence I: thus corrected over the line by Pa1 whence . Pa || 1.5.7-8 the vulgate whence I: thus corrected over the line by Pa1 whence . Pa. As for the rest, depends on Pa. The hypothesis the comparative examination of the shared readings of Pb and Ne allows is that both depend directly and separately on the same model. No variants were found as evidences to sustain the possibilities either that Pb would be an apograph of Ne, or the reverse.

THE P ARCHETYPE

P represents the lost model of the tradition of the scholia. This codex would be a minuscule and not an uncial. That is evidences found through comparative analysis of the separative variants of Pa and Pc allow to conclude. Such are the cases of confusion between similar abbreviations, for instance, the already analysed 1.6.3 and 3.92.3 the alternation / in Pa and its apographs. Other situations are the result of the confusion being this one more common between the abbreviations of and : 2.76.1 Pc whence I: || 3.17.9 Pc whence I: || 6.1.18 Pc whence I: . It is also the confusion between similar abbreviations that explains the following case: 14.6.1 Pc: whence I Pa.

15

Another case is a misreading, because of the similarity in the minuscule writing between the palaeographic traces of and : 2.65.2 Pa whence I: Pc. In the archetype the previous stick of would probably be linked to and the copyist of Pc would have read it as a ligature of and . The posterior stick would naturally have been read as a .

THE NEW STEMMA


The conclusions drawn together throughout the exam of the tradition of the scholia can be reduced to the following new stemma codicum:

16

P Pa 11th cent. Pa1 Pc 12th cent. Pa2 13th cent. Pa3 Pb 14th cent. V 15th cent. 16th cent. Aa Mb Ne Vh

17

SIGLA

W4 = Marcellinus, Sopater and Syrianus, , ed. Christian WALZ, Rhetores Graeci vol. 4, 39-846. W5 = Sopater , ed. Christian WALZ, Rhetores Graeci vol. 5, Stuttgart e Tbingen, 1-211. W7 = Anonymous, , ed. Christian WALZ, Rhetores Graeci vol. 7, Stuttgart e Tbingen, 1832-1836, 104-696. An. = Anonymous. R2 = Syrianus, Syriani in Hermogem commentaria, ed. H. RABE, Rhetores Graeci vol. 2, 1893. R14 = Prolegomenon sylloge, ed. Hugo RABE, Rhetores Graeci vol. 14, Leipzig, Teubner, 1931. Pa = cod. Parisinus graecus 1983, parchment, tenth-eleventh century. Pa1 = corrections by the first hand. Pa2 = corrections by the second hand. Pa3 = corrections by the third hand. Pc = cod. Parisinus graecus 2977, parchment, eleventh century. Pc1 = corrections by the first hand. P = accord of and , i.e. the supposed lessons of the archetype. = supposed model of Pa and . = supposed model of Aa, Vh and Mb, and an apograph of . = supposed model of Pb and Ne and apograph of Pa. = supposed model of Pc and V. Pb = cod. Parisinus Graecus 2916, paper, thirteenth century = Par.2 Walz. Aa = cod. Ambrosianus P 34 sup. (graecus 617), paper, fifteenth-sixteenth centuries. Mb = cod. Monacensis graecus 8, paper, sixteenth century, ff. 140r-192r. Ne = cod. Neapolitanus (= Farnesinus) II.E.5, paper, thirteenth century, ff. 58r77v. Vh = cod. Palatinus Vaticanus graecus 23, paper, end of the thirteenth century, ff. 73r-88r. V = cod. Vaticanus graecus 2228, paper, fourteenth century, ff. 122r-147r.

18

You might also like