Bryn Mawr Classical Review

Stephanie Roussou, Pseudo Arcadius' Epitome of Herodian's 'De Prosodia Catholica': Edited with an Introduction and Commentary. Oxford Classical Monographs. Oxford; New York: Oxford University Press, 2018. Pp. xxxviii, 596. ISBN 9780198805588. £120.00.

Reviewed by Filippomaria Pontani, Università Ca' Foscari (f.pontani@unive.it)

Version at BMCR home site

Publisher's Preview

Two years ago, through an unprecedented public crowdfunding, the citizens of Belluno (Italy) acquired on the market a copy of the rare second edition (1512) of the Institutiones Graecae grammaticae by their fellow countryman, the important humanist Urbano Bolzanio (1442-1524). This tribute to Bolzanio, a pupil of the Greek émigré Konstantinos Laskaris in Messina (himself the author of the famous Erotemata, the first Greek book published in the West in 1476, and the first book published by Aldo Manuzio in 1495), was more than an act of civic pride: Urbano's Institutiones was the first Greek grammar written in Latin, and thus represents a highlight in the process of systematisation and dissemination of grammatical thought that had started with Manuel Chrysoloras in late-14th-century Florence.

Bolzanio and Laskaris are the scribes of two out of five extant manuscripts of Ps.-Arcadius' Epitome of Herodian's General Prosody (also known as On Accents), a complex work that Stephanie Roussou has now made accessible in a learned and careful critical edition (with introduction and commentary) that represents a major improvement over its outdated predecessors, the editions by E. H. Barker (1820) and M. Schmidt (1860). Roussou is the first to make use of the two oldest witnesses preserved to our day: ms. Matritensis 4575 (M, written by Laskaris some time before 1482) and ms. Bodl. Baroccianus 179 (O, written by Laskaris' collaborator Leon Chalkiopoulos in 1495): both derive from the lost παλαιὰ βίβλος seen by Laskaris in the monastery of San Salvatore in Lingua Phari in Messina. The importance of these manuscripts has long been recognised, and Roussou's stemma codicum differs from that reconstructed by R. Schneider only in details of contamination: she plausibly contends that the Baroccianus was copied from both Laskaris' Matritensis and its antigraph (the παλαιὰ βίβλος), and that thus, despite its much more careless text, it must be reckoned an independent witness. On the other hand, Roussou's claim that Bolzanio's manuscript (now in Copenhagen, Hauniensis GkS 1965: siglum A) also shows signs of contamination with the παλαιὰ βίβλος, is less compelling (the coincidence in variants at 343.8 can be explained by way of polygenesis).

Roussou's description of manuscripts does not give a well-rounded historical picture of the transmission of this work. For one thing, it lacks the indication of watermarks, which might better clarify and articulate the two chronological stages of the tradition,¹ one in late 15th-century Messina (mss. M, O, A) and one in 16th-century Italy (mss. Par. gr. 2102 and 2603, sigla C and B). Where and when were these latter manuscripts produced? If C was in the hands of the well-travelled copyist and forger Iakobos Diassorinos, what is its place in his activity as reconstructed by Carmen García Bueno?²

Precise descriptions of manuscripts are not just erudite quisquiliae, or elements in the history of the work's reception. Take mss. B and C: the very attribution of this treatise to Arcadius rests on the titles of these two codices, and it is therefore essential to ascertain that—as Geppert had already pointed out—Ἀρκαδίου γραμματική in ms. B (f. 17r) is indeed the fruit of a correction of an earlier title that probably, if the extant traces can be trusted, contained the genitive Θεοδοσίου, which also appears in the headings of mss. MOA. Furthermore, the problematic status of book 20 of Ps.-Arcadius' Epitome (probably the fruit of Diassorinos' concoction from pre-existing material, and thus judiciously left out from the present edition) has been intensely explored by scholars from Pfeiffer to Nagy with respect to the much-debated—and culturally inspiring—passage about the origin of Greek προσῳδία (pp. 211.8-216.12 Schmidt).³

More importantly, the reader ought to be reminded that Konstantinos Laskaris—on whom most of what we know of this Epitome ultimately rests, since mss. M, O and A were produced in his entourage—was not a simple scribe, but an experienced grammarian who understood well the text he was copying, to the point that he profited greatly from it in his own work (just like his pupil Bolzanio after him), most notably in book III of his Erotemata.⁴ Roussou herself (p. 52) credits Laskaris with a major textual intervention in book 19 of the Epitome, where he added an entire section on adverbs in order to fill a gap in his source. A more thorough study of Laskaris' autograph ms. Matritensis 4689, which was completed in 1488,⁵ would have provided more evidence on the dating and the purpose of Laskaris' study of Ps.-Arcadius' text, which, as we have seen, he ascribed to Theodosius and considered on a par with Herodian's lost masterpiece (see f. 113v περὶ τόνων τὰ τῶν ἄλλων καὶ Ἡρωδιανοῦ τὴν μεγάλην προσῳδίαν καὶ Θεοδοσίου ἐν βιβλίοις εἴκοσι).⁶ On ff. 114-117, that manuscript contains an epitome of book 16 of Herodian's General Prosody that is largely dependent on Ps.-Arcadius' text, and that should therefore be collated side by side with ms. M for that one book.

This state of affairs helps understand the strange expression "Lascaris apud Lentz" repeatedly occurring in Roussou's critical apparatus to pp. 309-34: this definition concerns the readings of Laskaris' epitome of Book 16 as reported by Lentz in the apparatus to his edition of Herodian's General Prosody; but a modern editor of Ps.-Arcadius should verify those readings by collating directly Laskaris' manuscript Matr. 4689, especially in light of their importance (e.g. the right reading H for I in p. 311.9, the longer supplement in p. 318.2, the correct verb μέμβλω for μέλπω in p. 319.7, the intelligent supplement of ἀρχόμενα in p. 320.10, etc.).

Roussou's Schwerpunkt concerns the role of Ps.-Arcadius' Epitome as a source for the reconstruction of Herodian's lost masterpiece Περὶ καθολικῆς προσῳδίας, a foundational work on Greek orthography, prosody and grammar, whose idiosyncratic and arbitrary edition by A. Lentz has not yet been superseded and has misled generations of scholars. Chapter 2 of Roussou's introduction is a succinct and masterly overview of what we know about the nature, length, arrangement, and extant sources of Herodian's treatise, above all its epitomes—not only the two extant ones, namely, John Philoponus' Praecepta tonica, edited by G. Xenis in 2015, and the very work by Ps.-Arcadius edited here, but also the two fragmentary ones: the 4th-century parchment codex PAnt 2.67, and the 10th-century palimpsest Vind. Hist. gr. 10, which H. Hunger wrongly thought to be a witness of Herodian's very treatise. Passages such as the one on words ending in -ιος (pp. 159-60 R.) show the degree of overlap between Ps.-Arcadius and PAnt 2.67, and Roussou's apparatuses prove invaluable in assessing the relationship between the different sources. The peculiarity of Ps.-Arcadius' Epitome is that, while largely omitting the examples and quotations taken from ancient literary sources, it preserves closely Herodian's book-numbering and matter subdivision, even if—as stated in the work's preface—it digests the material in a more "reader-friendly" manner, namely by developing shorter sub-sections rather than by piling up doctrine in lengthy and heavy chapters.

Roussou discards the attribution of this Epitome to the imperial grammarian Arcadius, for which the only textual basis, as we have just seen, is the titles of two 16th-century codices descripti; on the other hand, she discusses at length the attribution to Theodosius, which emerges from Laskaris' manuscripts and references.⁷ The absence of any hint of the work in the lists of Theodosius' scholarly output, as well as some overt contradictions with doctrines we know as positively Theodosian, undermine this attribution. The work thus hovers in a vacuum: it was written by an anonymous author some time between the 2nd century CE and the early Byzantine age. The language of grammarians is too stereotyped and conservative to allow an exact dating, and even the more apparently "recent" words, such as p. 239.7, βοῦρκος, may belong to younger additions rather than to the original epitome.⁸ At times, however, one wonders if Roussou's text is compatible with a dating even in late antiquity: see e.g. the odd juxtaposition of hypothetical clauses in p. 237.2 (εἰ μὲν δισύλλαβα εἴη καὶ μακρᾷ παραλήγει), and the contradictory verbal concordance of neuter plurals in p. 329.9 (ἐὰν δὲ μὴ οὕτως ἔχῃ βαρύνονται).

A major feature of Roussou's edition is the attention to parallel sources: it should be stressed, however, that the texts listed by Roussou on pp. 75-80 (from Stephanus of Byzantium to the Etymologicum Magnum) are in fact indirect sources of Herodian's General Prosody rather than of the epitome here at stake. Hence, the many (too many?) references we find in the apparatus testimoniorum of Roussou's edition, do not represent testimonia to our text, but rather quotations from grammatical sources more or less directly indebted to Herodian (above all Theodosius, Theognostus, John Philoponus and Choeroboscus); that is, they are parallel comparanda that may occasionally help restore the corrupt text of Ps.-Arcadius' Epitome.

In some cases, these comparanda are listed without any indication of their mutual dependence.⁹ In other cases, some important evidence for Herodianic materials is neglected:
- p. 280.4: in the treatment of genitive Θῶνος (Ps.-Arcadius' text has been heavily restored by modern philologists) why not add a reference to the scholia to Od. 4.228 (the only occurrence of this term Θῶνος), where Herodian's doctrine on the issue is attested and discussed?
- p. 260.18: the toponym Κάρφη is corrected into Σκάρφη, but it also appears in schol. Od. 3.270a;
- p. 330.30: Roussou transposes καὶ τοῦ παρακειμένου, but neglects the fact (already discussed in Lentz's apparatus) that ἀκάχημαι and ἀλάλημαι were considered by Herodian as "Aeolic" forms of the present rather than the perfect, see e.g. schol. MaTY Od. 4.807c (ἐνεστώς ἐστιν Αἰολικός);
- p. 130.10: Ps.-Arcadius argues for the oxytone accentuation of words composed from bisyllabic verbal stems such as νεκροβαστάξ: Roussou refrains from tackling the flagrant contradiction with Choer. in Theod. 287.6-9 (quoted in the apparatus, but not further discussed), a passage that invokes the paroxytone accent for these same compound words: what was Herodian's view on the issue?

Roussou's introduction is rich in interesting and original materials: the learned essay on the two versions of Book 15 on enclitics (pp. 26-49) might have made the object of a separate article, and the insightful glossary of problematic or ambiguous terms occurring in Ps.-Arcadius, from βαρύτονος to κύριον to προσηγορικόν (pp. 58-75), will benefit every reader interested in grammatical terminology.¹⁰ The introduction understandably refrains from discussing the development of accentuation and orthography from the Alexandrian through the imperial age: these issues would deserve a book on their own, and they will rather concern the future editor of Herodian's fragmentary General Prosody. Still, in Roussou's commentary—mostly devoted to the critical discussion of single words—one finds some notes about the epitomator's modus operandi that would more naturally have found their place in the general introduction: I point exempli gratia to the perturbations of the alphabetical order, or to the sometimes baffling association of grammatical forms and phenomena that are not entirely consistent or homogeneous with each other (a recurring feature discussed on pp. 489-90).

Roussou's policy in establishing the text is prudent and judicious, above all in grounding it on a firmer basis. There is a tendency to posit many lacunae¹¹; occasionally the supplements could be better motivated¹²; sometimes the indication of literary sources behind the glosses is inconsistent.¹³ But on the whole Roussou's choices in such a difficult (at times very difficult) text take due account of the interventions of her predecessor M. Schmidt, and add new, well-considered and useful conjectures, which often heal smaller textual troubles. For example, in p. 196.7 the supplement ἀγανός provides just the right word that could be glossed as πρᾷος; in p. 174.11 νόθος for transmitted ὄθος and p. 214.7 τιλτός for MO's non-existent τικτός are brilliant corrections, which solve the problem of the very inclusion of these adjectives within the respective paragraphs.¹⁴ In p. 130.10 previous editions had διασφὰξ ὀξύνεται, whereas the complete text now shows a reference to the rare noun σφάξ, otherwise attested only in paraphr. Lycophr. 317 Leone and in schol. Opp. hal. 1.744: if we see here a saut du même au même, one might restore the text as δια<σφὰξ μᾶλλον ἀπὸ ῥήματος ἢ ἀπὸ τοῦ> σφὰξ ὀξυτόνου.¹⁵

The critical apparatus is encumbered with a large number of trivial orthographical peculiarities. While it can be argued that these are far from irrelevant details in a work dealing precisely with orthography and accentuation, I believe that the innumerable cases where ms. O omits a rubricated initial (typically ὰ instead of Τὰ) could safely be obliterated, and so could the hundreds of omissions of iota mutum (save of course in passages where the iota mutum is at stake), and the many cases where normal Byzantine accentuation differs from ours (e.g. p. 313.1 ἄλλό τι rather than ἄλλο τι, in a passage not dealing with enclitics).

The book is well produced: the Greek is remarkably correct; mistakes and blunders are comparatively rare¹⁶; there remain occasional slips in the Latin¹⁷; a random check on manuscripts shows that Roussou's readings are generally reliable.¹⁸ In such a poorly studied text, one might have expected a fuller coverage of earlier conjectures.¹⁹ And the hasty reader will regret the lack of a conspectus siglorum (the sigla of manuscripts can be found on pp. 83-86). But minor quibbles aside, this edition is an important achievement, and will certainly become the standard text of Ps.-Arcadius for decades to come.

---

Notes:

1.   Manuscripts of excerpts are rapidly dismissed on the ground that they "do not offer any good readings for the reconstruction of the text" (p. 87): but e.g. the excerpt on f. 64v of ms. Royal 16.D.XIV (a codex most probably older than either B or C) anticipates Schmidt's corrections Ἀρκεσίλαον and κέ ποθι for Ἀκεσίλαον and κέν ποθι in p. 299.18-20. No explanation is given of the sudden appearance of ms. Casanatensis 1710 in the apparatus to p. 307.7.
2.   El copista griego Jacobo Diasorino (s. XVI): estudio paleográfico y codicológico de sus manuscritos (diss., Madrid 2017).
3.   See Ph. Probert, "Ancient Theory of Prosody," in S. Matthaios, F. Montanari, A. Rengakos (eds.), Brill's Companion to Ancient Greek Scholarship (Leiden 2015), 923-48: 924-25, with earlier bibliography.
4.   The studies by Antonio Rollo are pathbreaking: see esp. A. Rollo, "La grammatica greca di Urbano Bolzanio," in P. Pellegrini (ed.), Umanisti bellunesi fra quattro e cinquecento, (Firenze 2001), 177-209.
5.   See T. Martínez Manzano, Constantino Láscaris. Semblanza de un humanista bizantino (Madrid 1998), 156-62, a useful discussion largely neglected in Roussou's introduction to the benefit of 19th-century scholarship.
6.   Incidentally, it is hard to believe that Laskaris' description of Herodian's book as πολύστιχος should be credited with any first-hand authority, as Roussou does on p. 6.
7.   Roussou's decision to mention the titles carried by the manuscripts in the apparatus' note to the first word of the text is odd: if the editor believed the title of mss. MO to be the fruit of Laskaris' conjecture, then he could have put it in brackets in the text.
8.   See p. 75, although it should be remarked that ms. C (i.e. its scribe Diassorinos) proved acutely aware of this lexical problem when replacing βοῦρκος with the more "ancient-looking" word βόρβορος.
9.   On p. 279 virtually the same sentence is quoted from the Etymologicum Gudianum, from the Etymologicum Magnum and from the Epimerismi Homerici (in this order), but the Epimerismi are here clearly the source of the etymologica; on p. 278 the text is restored through a reference to passages of Choeroboscus that are not included in the apparatus comparandorum; on p. 158 (see the discussion on p. 421) a single passage is quoted from Theognostus and from three Byzantine etymologica, whereas Theognostus and at most the Genuinum would have been enough (and even so, why does Roussou print ἀρχῆος on l. 7 rather than ἀρχηός as in most of the parallel sources?).
10.   The intepretation of κοινῶς as referring to a "colloquial" Ephesian gloss in p. 180.11 is prima facie problematic, and deserved some more comments on the linguistic side.
11.   E.g. p. 126.12 the name Πολυσπερχῶν is evoked simply because Σπέρχων has just been quoted, there is no need to think of a lacuna.
12.   p. 332.2 the saut du même au même does not really tally with the two lacunae posited by Roussou.
13.   If p. 115.16 ἀτμήν is identified as Callimachus' fr. 178, then p. 116.13 τιβήν should also be identified as Lycophr. 1104.
14.   It should be added that τικτός was omitted by BC, and that the impossible θικτός is printed by Schmidt not as his own conjecture but as the reading of ms. A, i.e. of Urbano Bolzanio: is this true?
15.   In fact, Roussou's statement (p. 385) that mss. BC omit the entire sentence ought to be corrected, because B has ὀξυτόνου, and C has ὀξύνεται.
16.   E.g. in p. 84 note 213, RGK I 237 is in fact II 237; p. 487 Ep. Hom. 357 is in fact Epim. Hom. A 357, and the attribution to Herodian is not in the text, but conjectured by Egenolff; on p. 1 Roussou is unaware that a new edition of the scholia to Euripides' Hippolytos has been published by J. Cavarzeran, De Gruyter 2017.
17.   p. 120 "statuerem" should be "statui"; p. 122 "librorum" should be "operum"; p. 125 "nominorum" for "nominum"; p. 128 "eundum" for "eundem"; p. 148 "scriptor" for "librarius"; p. 166 "hic" for "huc"; p. 177 "referre" for "referri"; p. 189 "inversam" for "inversum"; p. 311 "vel" for "an"; p. 330 "praebet" for "praebent".
18.   Ms. O has πής (not πῆς) in p. 275.4, εἶμαι (not εἷμαι) in p. 330.29, and it omits βιβλίον ιε´ in p. 276.13 (the app. crit. is confusing here).
19.   In p. 129.1 the lacuna had already been indicated by Schmidt; in p. 318.6 Lentz had conjectured κράζω for the unattested τράζω; in p. 329.5 the transmitted ἐγκαταλιπῶ presupposes a never attested aorist ἐγακτελίπην, and perhaps Lentz's correction ἐγκαταλειφθῶ deserved a mention.