To read through the seven large and closely-printed volumes of Keil’s Grammatici Latini is (crede experto) one of the more mind-numbing experiences that the study of classics affords. But within that large and repetitive corpus two works in particular stand out as affording genuine interest and valuable material. One is the huge and complex grammar of Priscian which stands almost at the end of the tradition, important as the only Latin grammatical text that deals systematically with syntax; the other is the fourth-century ars grammatica of Charisius, an awkward compilation from three earlier grammars of different types which tells us, precisely because of its awkwardness, more than any of the others about the origin and development of Roman grammatical thought. One of these earlier texts, the work of C. Iulius Romanus in the third century CE, is a major source for fragments of Republican Latin literature, particularly those containing odd adverbs and interjections; and one six-page passage of Romanus, the introduction to his book on adverbs, is the object of Dirk M. Schenkeveld’s attention in this monograph.
What initially aroused S.’s interest in Romanus is the mention in this passage of an otherwise unfamiliar Stoic term for the adverb,
The argument of Romanus’ preface is contorted and difficult, with two major topics that are not sufficiently explained or distinguished, namely the question of whether interjections are a subset of adverbs (Romanus rejects that idea firmly and at length) and the question of whether adverbial endings in -o or in -e are correct, a problem in the balance between analogy and consuetudo. S.’s attempts to unravel the tangled skein of Romanus’ argument are the best part of this book; but it must be confessed that I still remained unclear about some important issues after reading S.’s analysis, his translation, and his commentary. S. at times assumes knowledge of the technical terminology of Roman grammarians; his arguments are often inconclusive; and — perhaps the signal failing of his discussion — he tends to judge the linguistic/grammatical argument by a standard which he himself knows Romanus was not trying to meet. Romanus was, almost certainly, not a ‘professional’ grammarian, and the work of which we have large sections quoted, the
Despite the peculiar expectations that S. brings to Romanus, however, he also brings a genuine interest and admiration for Romanus’ quirky vigor, and a genuine desire to sort out an extraordinarily difficult piece of Latin. After his analysis, both the logic of Romanus’ argument and its perplexities are far more apparent. S. is honest and clear about the difficulties of this text, which is welcome; but it does not seem to me that he solves many of them. S. is at his best in dealing with questions of formal grammatical theory, which he knows very well; but his chapters on Charisius and Romanus are often no more than intelligent summaries of standard material and often seem not to reach any decisive result. His account of scholarship on Charisius’ sources is accurate (although he seems to look for ways to criticize Barwick’s groundbreaking work on the subject), but he seems vague and uncertain on the question of intermediate sources and the possible two versions of Charisius. He wants to believe the identification of Charisius as a grammarian of Constantinople, but he simultaneously recognizes that that identification depends on a very shaky emendation. He wants to date Romanus fairly late in the third century, but at one point he seems to date Marcius Salutaris (on whose identity the dating of Romanus depends) before 200 CE instead of after 250. He sets out the evidence about the title and scope of Romanus’ work clearly, but his argument again seems inconclusive. One has the sense, in these two chapters, that S. is being dutiful in setting out background rather than interested in the problems themselves.
As for the text and apparatus themselves. S. sets out the Introduction far more clearly than Barwick did in his Teubner text: he introduces mostly sensible paragraphing, and he uses italics and quotation marks in a way that makes the basic sense and structure of the sentences much clearer. He disagrees with Barwick’s text in several dozen places, some of them minor, through making careful use of Cauchius’ annotations and the emendations of others. Readings that seem to me distinct improvements over Barwick include dixerint (246.27), occulto (247.6), maxime (247.15), Titinius (251.7), Aufinum (251.8), and idoneis (252.6). But not all his changes are so likely, and in several places he alters Barwick’s text without any note in the apparatus: thus etenim to enim (248.18), expedit to expeditur (249.4), putauit to putaui (249.19). The first and last of this are probably typographical errors; the second he discusses in the commentary: it is his own emendation, and certainly possible, but needs to be included in the apparatus as well as the commentary. The apparatus, indeed, is much less useful than the text itself: it is very confusing in typography and often organization, and different notes on the same line are run together without division. He offers a translation to show what he thinks the text means, and that is in itself a very brave thing to do with such a recalcitrant text. But it is clumsily literal without necessarily making any sense at all: nostri (meaning ‘Romans’ or ‘Roman grammarians’) at 246.28 is translated as ‘our men’; at 252.16-17 we find the peculiar phrase ‘you contend me this usage’ translating contendere a nobis … ut … examen (sc. sit). This is neither English nor Latin, and leads to an extraordinarily contorted interpretation, where Kroll’s simple emendation of examen to examinemus makes the passage much clearer.
For the most part, S.’s text is good and his commentary relatively useful when he is dealing with the topics that most interest him, relating to formal grammatical theory. But in other passages, where Romanus goes off on tangents or uses strained and strange language, he seems to go widely off the mark. Thus, in a passage on whether interjections are a subset of adverbs, Romanus proposes, and then rejects, a defense of this structure based on Stoic terminology. The text as printed by S. begins ‘Quorum siquis defensionis ineat tutamen, quod idcirco …’ (247.13-14). The grammatical function of quorum is ambiguous (does it depend on siquis or on defensionis, either of which is possible in terms of sense?), but S.’s explanation is not easily comprehensible, and he links it — for reasons unclear to me — to a textual problem. tutamen is Cauchius’ reading (C); the main manuscript tradition, represented by N, reads tamen, and Barwick here accepts Keil’s emendation tramitem. S.’s argument is that tutamen in C “is a variant of the MS tradition, and for that reason to be preferred.” That is a specious argument, on two counts: one is that one can not tell if many of the readings in C, including this one, are Cauchius’ conjectures or readings that he found in his manuscript; the other is that manuscript evidence is no defense of nonsensical readings. S. himself translates his text as ‘devises in their defense the plea that …’ and, more literally in the commentary, as ‘enter into the protection of the defense’ and for inire tutamen suggests as a parallel the phrase inire gratiam — which is supremely irrelevant. He has not bothered to check that tutamen appears only three times in classical Latin (at least from my search of the PHI disk), always in poetry and the attached genitive is concrete, not abstract, and never in any legal or grammatical context, while trames appears in a very similar context, meaning ‘line of argument’ in Varro, de Lingua Latina 5.5. Keil (as followed by Barwick) knew what he was doing here, and S., to judge from his note, does not.
Another passage in which S. is badly lost occurs at 248.26ff., where Romanus digresses from his long list of types of adverb to make an analogy from law (S.’s text):
optandi (sc. adverbia), utinam veniat, velim currat. Quam voluntatem gaudio proximam esse cum dicimus, ut ( Aen. 1.626) ‘seque ortum antiqua Teucrorum a stirpe volebat,’ votis esse comparem credimus. Itaque neque fidei commissorum voluntas sequius accessit, ut non in ea praeter iudicium, quamvis tacitum nec adscriptum, votorum solis nisibus et beneficiis crederemus.
S. interprets Romanus to be talking about the relationship between voluntas and votum as comparable to that between scriptum and sententia in the interpretation of wills. But it is surely impossible to make votorum … beneficiis a periphrasis for ‘written testament’: the contrast is between the volition of the testator, however expressed, and vague hopes or prayers that have no legal standing. If that is true, then S.’s printing of the emendation ut non for ut haec in the manuscript is probably right (although I confess that I find the sentence very hard to understand in any case), but his rejection of Lachmann’s emendations of gaudio to dubio and beneficiis to indiciis is almost certainly wrong. His commentary on this passage — admittedly very obscure — is extraordinarily confused and unhelpful, with an almost total absence of relevant parallel passages and an equal lack of common sense. Wishing is not equivalent to joy, and Romanus is not saying that it is: wishing is comparable to prayer — a sense of uncertainty as to whether something will happen. Hence gaudio makes no sense, and dubio is not so far from the paradosis as would be, for instance, spei. And it might also be noted that S. does not even bother to comment on the one thing that would strike a modern reader as most odd in Romanus’ comments, that velim is interpreted as an adverb — try that on a modern undergraduate.
Another area, and an important one, in which S. does not work hard enough is in explaining the genre and form of Romanus’ work. Although he is well aware of Romanus’ separation from the main line of the ars grammatica and enjoys precisely those elements — odd language, rhetorical style, jokes — that make Romanus different, he is very weak in explaining the context of the work and its function. He gives a long and not very useful list of stylistically developed works on technical subjects, but then decides that they are not relevant, as they are all in poetry. He dismisses as parallels the prefaces of Vitruvius, Pliny, and Cicero’s de Oratore, and concludes that Romanus’ Introduction to the Adverb is “totally unique in Latin (and Greek) literature for its treatment of a technical subject in such a rhetorical manner” (77). While Cicero and Manilius are indeed not relevant, S.’s conclusion is seriously inadequate. In terms of stylized (and quirky) writing about technical subjects, Varro stands out as an obvious model — not just the de lingua Latina but the de re rustica. Numerous chapters of Gellius on the technicalities of diction, spelling and semantics come to mind for precisely similar combinations of colloquial, rhetorical, and technical. And in terms of style, my own (very subjective) judgment would compare Romanus’ text to the fragmentary preface of Florus’ Vergilius orator an poeta, a strange and difficult text which uses abstract nouns in a similarly unusual way, and is no more than a century earlier than Romanus’ book. These suggestions are tentative; my only point is that S. should have thought of them himself.
In sum, while S. makes some useful contributions both to the text and to the interpretation of Romanus, his work is uneven and at times perfunctory. There is a good article on the argument of the Introduction in chapter 3, and in some passages he has improved the text. But that is not enough to justify a book that costs, at my rough calculation, 14 times what it would cost to photocopy. Romanus is a text that could use closer study, and if S.’s book arouses interest, that will be a service. But any serious progress will need a broader and more thorough approach than this.