Bryn Mawr Classical Review

J. Uría has just issued the first part of a complete translation of Charisius' Ars grammatica.¹ Although this volume contains only Charisius' Book I, no more than half the whole Ars is left to come. Book I, except for the concise first chapters presenting some basic notions of grammar (voice, letters, syllables, words), is a detailed study of Latin nominal morphology. The book, published with the high regard for readers' comfort of the Biblioteca clásica Gredos, is classically set in two parts: an introduction (pp. 7-73) and the annotated translation (pp. 77-395). One can only hope that volume two will include the indices so useful for studying that kind of text.

The introductory part concerns itself with the whole treatise. All expected headings are to be found here : the author's identity, content and context of the grammar, its fortune and transmission. On all these questions, the translator makes a clear and concise survey of current scholarship before taking a mostly prudent and sometime understated stance.

Virtually nothing is known of the life and whereabouts of Charisius, a fourth-century gentleman of letters whose only extant work is the compilatory treatise on Latin grammar he dedicated to his son. Uría handles with sckepticism, and rightly so, the fictitious character scholars have struggled to lend to Charisius out of scarce evidence. All that can be said with confidence is that, in accordance with the titles of uir perfectissimus magister (here translated caballero, supervisor) which he awards to himself in the epistolary preface, he occupied some kind of official position, presumably in the eastern part of the empire.

Charisius' Ars grammatica is, avowedly, a compilation, relying on two kinds of sources. On the one hand, it follows an elementary, unoriginal grammar, possibly that of Cominianus. On the other hand, it makes abundant use of pre-third century works of more speculative inclination. These works themselves are apparently known to Charisius through a number of citations and reworkings. Karl Barwick, the last editor of Charisius, gained a reputation in Quellenforschung for his attempt at disentangling the sources handled by this author. Following this traditional path, Uría's analysis of content is largely focused on matters of sources and authorship. There is a discussion of the fundamental notions of grammar according to Charisius, such as analogia, ratio, consuetudo, auctoritas. Still, a comprehensive survey of problems of translation and adopted solutions is missing. As translating Latin grammars is still a new field in classics, such a methodological statement would have served both the specialized scholars and the reader unfamiliar with that kind of technical text.

We are then reminded of Charisius' Fortleben. His popularity with other late antique grammarians is striking, given that the grammar was all but forgotten during most of the Middle Ages and enjoyed only a brief early modern revival before being dismissed along with all such compilatory works. Hence the lacunose state of the extant text. Uría follows Barwick's Teubner 1925 edition corrected by Kühnert in 1964. The page and line numbering from the Teubner text accompanies Uría's translation, which stands by itself without the Latin text. Such is the practice of the chosen collection, originally designed for translations of widely available, well edited classics.

Nevertheless, it is to be expected that a translator will depart on some points from the editor, and even more so in this particular case, since Uría is giving the very first full translation of this grammar. Differences from Barwick's text are summarized pp. 52-57, between the introduction and the bibliography. Only the corrections on Book I are listed here. More dramatic changes are to come in volume two, especially in the fifth and last book, if Uría stands by his condemnation of Barwick's choices stated on pp. 16-17.

Changes to the text come in no small part from emendation made by the editors of those works which are known to us through compilations such as Charisius'. Besides, Uría has adopted some suggestions made by previous scholars and, occasionally, has decided to come back to the manuscript text. That leaves only a few personal conjectures the translator has argued for elsewhere.² Only one of them, in my view, cannot stand the test of translation: the hypothesis of a confusion between ratio and oratio at the beginning of Chapter 15 (p. 62 l. 6 Barwick) : paucis admodum partibus [o]rationis normae suae dissentientibus, translated by dado que muy pocas partes discrepaban de la norma de la razón. As suae was left hanging, the translator could not help but arbitrarily delete it as well. On the contrary, at the beginning of Chapter 17 (p. 149 l. 24 Barwick), the reading nunc in rebus nunc in [o]rationibus occupata (scil. analogia), is an elegant emendation, consistent with the following Greek definition : συμπλοκὴ ... ἐν λέξει.

Translating the whole grammar of Charisius is certainly an exhausting and unrewarding task. It is mostly made up of paradigm expositions repeated under various presentations and lexical listings in alphabetical order. On each topic, Charisius has usually gathered excerpts from two or three previous authors, differing in purpose and ambition. Hence the numerous redundancies, contradictions and brutal changes in tone and style. The technical terminology itself varies in scope, depending on whether it is taken from a simplistic textbook or discussed in an essay by a scholar well aware of its theoretical implications. Words such as ratio, analogia, norma, regula, obseruatio, for instance, are keenly distinguished in the highly theoretical introductions to Chapters 15 and 17 but elsewhere, they offer roughly the same meaning. Thus, in most of the book, they are treated by Uría as near synonyms and rendered alternatively by razón, norma and regla.

In Book I, three items stand apart for their high level of abstraction and their engagement with the general rules of language: the compiler's preface and the introductory parts of Chapters 15 and 17, respectively anonymous and attributed to the third-century Julius Romanus. These are the most discussed parts of the book, and the rare passages to have been granted a translation before Uría's. Still, the translation is not always as accurate as it should be, beginning with the prefatory letter, where Charisius promises his son that, thanks to this grammar, he will learn quatenus Latinae facundiae licentia regatur aut natura aut analogia aut ratione curiosae obseruationis aut consuetudine, quae multorum consensione conualuit, aut certe auctoritate, quae prudentissimorum opinione recepta est.³ The Spanish translation reads en qué medida gobiernan la elocuencia latina la licencia, la naturaleza, la analogía, la razón de escrupulosa observancia, el uso, afianzado por acuerdo de muchos, o, por supuesto, la autoridad, tomada del juicio de los más sabios. It is unfair to Charisius, who distinguishes neatly between the 'licence of Latin eloquence' (licentia, a nominative) and the five principles by which it is ruled (natura, analogia, ratio, consuetudo and auctoritas in their ablative forms). Certainly, Charisius is not proving an original thinker here : he is merely following his unnamed source for Chapter 15 who thought Latin first developed in an unruly way before being disciplined by grammar. Except for this straightforward idea, the beginning of Chapter 15 is problematic and no attempt at translating it can be conclusive. But Uría is too cautious in his treatment of the introduction to Chapter 15. On the one hand, he makes (and argues in his notes for) conservative choices of text (the aforementioned correction of orationis into rationis notwithstanding). On the other hand, he follows Schenkeveld's understanding of the passage.⁴ But Schenkeveld, despite all his interpretative skills, had to alter the text repeatedly in order to make sense of it. These are in my view the major shortcomings in Uría's translation of the theoretical parts of the grammar.

The abundance of examples is a striking feature of this grammar: Charisius might have intended to provide his son with a lexical as well as grammatical tool. Leaving the illustrative Latin words without translation would come down to not translating half the book. But, as the point of all these examples is either phonetical or morphological, the original forms must be cited for the sake of clarity. Uría has treated the word-lists and literary (mostly poetical) quotes in different ways. When dealing with single words, he gives the italicised Latin form, followed by the Spanish equivalent in regular script. Many of Charisius' word-lists aim at contrasting Latin and Greek. Consequently, they already offer a translation, albeit a Greek one, which Uría transliterates and puts into round brackets. Such a presentation accounts for the difference in length between Barwick's edition (192 pp.) and the translation (318 pp.). Still, it proves as readable as might reasonably be expected.

The treatment of literary quotations is less successful. Those found in chapters on morphology are directly translated into Spanish, with only the form under discussion added in Latin. As an unintended consequence, it highlights the argumentative function of these auctorial quotes: in most cases they exhibit an agreement as proof of a grammatical analysis. For instance, on p. 259, Charisius quotes the Virgilian ferrata calce fatigat to establish the feminine gender of calx. The translation fatiga con el herrado talón (calce [abl. fem.]) obscures the point of the example. On the contrary, verses illustrating matters of phonetics and, in particular, metrics are left in Latin, accompanied only with one mark of quantity (in Aeneid 3.211, p. 95, the breve sign is misplaced on Ionio whereas it should appear on the last syllable of insulae). As there are not so many of these cases, they could be quoted in Latin and Spanish without major change to the total length of the book.

The minimal scansion remains the least understandable decision, though. It makes the chapter on syllables (for all purposes, an introduction to metrics) unmanageable to the non-expert reader. It even forestalled the correct understanding of one paragraph. The difficulty lies in what appears to be a digression towards the end of the chapter (p. 13 l. 20-26 Barwick; the scanning is mine): plerumque etiam structuram prosae habet, ut 'non te nullius exercent numinis irae' (-- -- -- -- -uu --). item 'et uersa puluis inscribitur hasta' (-- -- -- -uu --). quis enim non uideat structuram 'nullius exercent' (--u ---) et 'puluis inscribitur' (-u -- uu)? In note 58, Uría draws far-fetched parallels in an unsuccessful attempt to clarify the sense of structura prosae. What he does not do is scan the quoted phrases together with the grammarian's instructions. That would have shown, first, two hexameters (with short syllable lengthening before the caesura), and then (keeping the linguistic quantities), two metrical clausulae of the most ciceronian kind, which makes it plain that, here, structura prosae means 'prose metrical clausula'. That said, this note falls in the middle of an argument about the quantity of final syllables and looks as if it was interpolated into the text. It is worth noticing then that the table of contents of the Ars grammatica shows in book V the plan of a complete metrics (verse and prose), now lost altogether. That could be the origin of the digression: there are in the grammar of Charisius instances of such internal displacements (see e.g. p. 145 l. 13-16 Barwick).

This last case leads to a concluding remark. Until recently, ancient grammatical treatises were never entirely translated, in particular those written in Latin. Thus, most of the problems left unsolved by the editor, including textual problems, could be neglected by generations of scholars. Uría's translation grants the whole book of Charisius a much needed close reading (although this is, among Latin grammarians, one of the most commented on), raising unsuspected questions. One can only regret that the translator sometimes shies away from solving the problems his work has just raised.

1.   I thank Sandra Mouton for correcting my English. All remaining errors are of course my responsibility.
2.   In four textual notes : 'Textual Criticism and Source Study in Ancient Latin Grammar: Charisius Ars Grammatica 46.5 Barwick; Excerpta Bobiensia, Grammatici Latini 1.547.35 Keil', Classical Philology 95, 2000, 61-71; 'Charisiana I...', Exemplaria Classica 10, 2006, 245-252; 'Charisiana II...', Exemplaria Classica 11, 2007, 133-143; 'Charisius (Ars Grammatica p. 158.6-12 Barwick): On Cinna, fr. 1 Blänsdorf', Mnemosyne 60, 2007, 116-120.
3.   I am well aware that this text rests only on the editio princeps. Still, since Uría did not make any emendation to this sentence, I assume he accepts it as printed here and in all editions.
4.   The translator makes use of both articles by Schenkeveld on Charisius I, 15, one published in P. Swiggers, A. Wouters edd, Ancient Grammar: Content and Context, Leuven, Peeters, 1996, 17-35 and the second one in the American Journal of Philology 119, 1998, 443-459.