Vague Boundaries: Delimiting grammatical fragments in Charisius

Eruditio Antiqua 3 (2011) : 57-72 VAGUE BOUNDARIES: DELIMITING GRAMMATICAL FRAGMENTS IN CHARISIUS* JAVIER URÍA VARELA RAMÓN GUTIÉRREZ GONZÁLEZ UNIVERSIDAD DE ZARAGOZA UNIVERSITÀ DI BOLOGNA Abstract This paper deals with different problems in the delimitation of grammatical fragments transmitted by the fourth century grammarian Charisius. Cases are analysed in which the various ways of doctrine reporting (indirect style, direct style and quasi indirect style) are involved. Establishing fragment boundaries has implications for both doctrine and terminology assignment as well as for textual criticism. Passages in which several levels of reporting are implied pose especial difficulties. Fragments are analysed which contain the doctrine of grammarians such as Varro, Caesar, Pliny and Verrius Flaccus. Resumen Este artículo analiza los diferentes problemas que plantea la delimitación de fragmentos gramaticales transmitidos por el gramático Carisio (s. IV). Se analizan casos en los que aparecen los diferentes modos de transmisión de doctrina (estilo indirecto, estilo directo y estilo cuasi indirecto). El establecimiento de límites de fragmentos tiene consecuencias tanto para la atribución de doctrina y terminología como para la crítica textual. Los pasajes que contienen varios niveles de transmisión plantean especiales dificultades. Se analizan fragmentos que contienen doctrina de gramáticos como Varrón, César, Plinio y Verrio Flaco. * This paper is part of an ongoing project, directed by Javier Uría, and has benefitted from a grant from the Spanish ‘Ministerio de Ciencia y Tecnología’ (Project FFI2008-05202). www.eruditio-antiqua.mom.fr JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES It is widely acknowledged that Roman literature is in some ways a vast citation: terms such as parallels, loci similes, allusion, imitation or intertextuality have been used, either traditionally or only recently, to refer to one or another form of citation lato sensu. Each literary genre has its own rules regarding this matter, and it is beyond the scope of this paper to deal with each of them1. So, let us focus on grammatical tradition, where we can observe a particular form of citation, namely citation stricto sensu, inasmuch as it expressly recognises the existence of a previous text from which some words or ideas are borrowed. As Robert Kaster has repeatedly emphasised, ancient Latin grammar values tradition and continuity as the mainstay of the grammarian’s prestige. That is why when reading through Keil’s Grammatici Latini one gets an impression of repetition and lack of innovation. In Kaster’s own words (1986, p. 324): ‘A day spent with the daunting bulk of Keil’s Grammatici Latini is enough to confirm this general characteristic [scil. the lack of originality]: belonging for the most part to the third through sixth centuries, the handbooks collected in those volumes appear intent, one after the other, on imposing a deadly uniformity on their material, as the grammarians repeat the same patterns of analysis, teach the same lessons – often in precisely the same words, repeated from older sources [our italics] – and use for illustrations the same examples that had been used for generations’. 1. First and second hand citations in Charisius The ars grammatica of the fourth century uir perfectissimus Flavius Sosipater Charisius was long ago considered a privileged source for previous grammatical doctrine, because he allegedly follows his sources with a high degree of fidelity. This is the impression one gets when reading his preface: Char. Gramm. p. 1.5-7 B.: …artem grammaticam sollertia doctissimorum uirorum politam et a me digestam in libris quinque dono tibi misi2. However, even if we were disposed to accept that Charisius is sincere when stating his fidelity, still it is hard to believe that his arrangement did not result in 1 On this topic, see CONTE – BARCHIESI 1989. 2 On this preface, see SCHENKEVELD 2004, p. 4-5 and URÍA 2006. Eruditio Antiqua 3 (2011) 58 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES some alterations, additions, insertions or abridgements. In fact, we know that for example in chapter I 15 (the so-called Anonymus de extremitatibus), Charisius inserted small portions which he borrowed from chapter I 17 (De analogia by Julius Romanus3). Even more importantly, Charisius uses many second-hand citations, in which the fidelity of the transmitter is even more doubtful. All this, together with the awareness of disagreements in the collection of fragments, invites a systematic review of those fragments. Indeed, as a result of the lack of a systematic study of the way in which later grammarians quote or abstract their predecessors4, it is possible to detect in current critical editions a number of inaccuracies in the text as well as in the attribution and extension of grammatical fragments. It is to the latter that we will pay attention in this paper, by focusing on the problem of fragment boundaries. 2. A typology of reporting A simple collatio of the main collections of grammatical fragments (those by Funaioli 1907, Goetz–Schoell 1910, Mazzarino 1955 and Della Casa 1969) and the editions of Keil (1857) and Barwick (1925) provides evidence that consensus is far from full regarding where a fragment begins or ends. Even where agreement exists, the accepted solution can often be challenged. In total, in Charisius we have detected more than eighty quotations whose boundaries can be disputed; roughly two thirds of them include direct quotations (oratio recta), whereas the rest are either indirect quotations (oratio obliqua) or what is called ‘modalisation en discours second’5 (alias oratio quasi obliqua), namely utterances marked with an insert of the type ut ait/dicit. These three main types pose distinct problems, as the expected loyalty to the source from each way of reporting is different. At the least faithful, most disloyal extreme we have indirect quotations, which were long ago described as referentially opaque, inasmuch as they do not always allow the so-called de dicto interpretation, but only a de re interpretation, as in the classical example Oedipus 3 On Romanus, see SCHENKEVELD 2004, p. 29-53, and P. L. SCHMIDT in SALLMANN 2000, p. 269-271 (§439.1); on the Anonymus de extremitatibus, see P. L. SCHMIDT in SALLMANN 2000, p. 272 (§439.2). 4 This is one of the main objectives of the above mentioned project. 5 A distinction must be made between this type and the so-called ‘modalisation autonymique’, which implies ‘îlots textuels’ (parts of direct speech inside a reported speech). See AUTHIERREVUZ 1992, p. 39, ROSIER 2008, p. 35 and TOUMARLA 2000, p. 156-163. Eruditio Antiqua 3 (2011) 59 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES said that his mother was beautiful6. This implies that indirect quotations may suffer from interference from the reporter’s speech and the quoted speech7. 3. Whose term is this? Assigning terminology in indirect quotations This is directly applicable to the reporting of grammatical doctrine, as illustrated in the following passage, where the term ablatiuus seems to be assigned to Varro: Char. Gramm. p. 132.8-13 B.: Leontion et Chrysion et Phanion ex neutris Graecis feminina [neutra] fecere et Plautus quod dixit haec ‘Phronesium’ [truc. 12 etc.] et Caecilius ‘Leontium’ [frg. 286 Ribbeck = 271 Guardì]. Varroni [frg. 258 Funaioli = 39 Goetz–Schoell] autem placet talia nomina datiuo tantum casu et ablatiuo declinari, de ceteris uero sic efferi ut nominatiuo. Indeed, Varro might have known the term datiuus, but it is certain that he did not know the term ablatiuus, as confirmed by Diomedes8, Gramm. I 302.4-5: ablatiuum Graeci non habent. hunc tamen Varro sextum, interdum Latinum appellat. Accordingly, caution is also needed to ascribe the term ablatiuus to Varro’s contemporary Julius Caesar, since the relevant passage is reported in indirect style. Moreover, three levels of reporting can be detected in that passage: Caesar (reported by) Pliny (reported by) Romanus (reported by) Charisius: Char. Gramm. p. 170.13-18 B.: Iubare. Plinius [frg. 76 Mazzarino = 38 Della Casa] ait inter cetera etiam istud C. Caesarem [frg. 24a Funaioli = 26b Klotz] dedisse praeceptum, quod neutra nomina ar nominatiuo clausa per i datiuum ablatiuumque singulares ostendant; iubar tamen ab hac regula dissidere. nam ut huic iubari dicimus, ab hoc iubare dicendum est, ut huic farri et ab hoc farre. 6 De dicto interpretation: *My mother is beautiful (Oedipus could not have said this, for he did not know that Jocasta was his mother). De re interpretation: Jocasta is beautiful / My wife is beautiful / The mother of my children is beautiful. 7 On this matter see COULMAS 1986, p. 3-4. However, the use of reported speech does not exclude the so-called ‘îlots textuels’ (see n. 5). 8 Cf. SCHAD 2007, p. 3: ‘Varro seems not to have known this term [scil. ablatiuus] since he uses sextus and Latinus for the ablative case (…). The earliest direct attestation is Quint. 1.4.26’. CALBOLI (1972, p. 105) proposes that Pliny is the first author where the term is found; however, in the relevant passage (CHARISIVS Gramm. p. 154.11 B.) Pliny is only reported in oratio quasi obliqua (ut ait Secundus). On the term ‘dativ’ and its history, see DE MAURO 1965. Eruditio Antiqua 3 (2011) 60 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES An interpretation de re seems to be preferable to an interpretation de dicto. In passing, you can observe that the part of the text in bold belongs to Romanus, the underlined part to Pliny, and the double underlined part to Caesar. Finally, the two parts underlined and in bold can be either Romanus’ or Pliny’s responsibility (this is in fact under debate, for Mazzarino and Della Casa choose different solutions9). That the terms datiuus and ablatiuus depend on Pliny rather than on Caesar seems to be confirmed by some parallel passages: Char. Gramm. p. 154.5-12 B.: Aedile, ab hoc aedile, non aedili (...) quod cum ratione dictum esse monstrabis, ut ait Secundus sermonis dubii libro VI [frg. 78 Mazzarino = 17 Della Casa], quod nomina quaecumque genetiuo singulari is syllaba finiuntur, exceptis his quae similiter faciunt nominatiuo, oportet ablatiuo singulari e littera terminari, a prudente, ab homine. Char. Gramm. p. 156.12-15 B.: Aplustre. omnium nominum quae sunt neutri generis et in e terminantur ait Plinius [frg. 74 Mazzarino = 21 Della Casa] Caesarem [frg. 23 Funaioli = 25a Klotz] scisse eosdem esse ablatiuos quales sunt datiui singulares. Char. Gramm. p. 156.16-18 B.: Ar litteris nomina neutralia terminata item non minus ait Caesar [frg. 24 Funaioli = 26a Klotz], quia datiuo et ablatiuo pari iure funguntur, ut idem Plinius [frg. 75 Mazzarino = 37 Della Casa] scribit. item Nipperdey : idem N pari N : an per i legendum? It is especially these final words, ut idem Plinius scribit, which make it very likely that what we have in these passages is Pliny’s own rephrasing of Caesar’s doctrine. The use of an inserted ut-clause supports this view, because oratio quasi obliqua tends to imply an identification of the points of view of the reporter (Romanus) and the quoted character (Pliny). Also, if a uerbum dicendi is used, this means that the reporter of Pliny intends to report not just content (de re) but also form (de dicto)10. Moreover, if our proposal of reading per i instead of pari is accepted, then the idea that we are reading Pliny’s words is reinforced, given that similar uses of per are often found in Plinian contexts11. It is true that all this does not exclude the possibility that Pliny himself may be conveying Caesar’s words as well as his ideas, but this is less probable. Had Romanus meant to indicate that Pliny was reflecting Caesar’s words, he would have probably used a different 9 MAZZARINO 1955, frg. 76 p. 295 (de re); DELLA CASA 1969, frg. 38, p. 125 (de dicto, cf. p. 234-235). 10 Cf. DELLA CASA 1969, p. 234: ‘La tesi del Mazzarino [1948b, p. 67], secondo cui si rispecchiano le precise parole di Plinio, pare si possa sufficientemente documentare e sottoscrivere’. 11 See MAZZARINO 1948a, p. 208. Eruditio Antiqua 3 (2011) 61 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES phrasing: funguntur should rather have been fungantur, so that it could be clear that it is Caesar’s reasoning that Pliny was reporting. Anyway, the paragraph is not easy at all, for item is only a conjecture (N gives idem) and quia could be read both as a completive or as a causal conjunction depending on the punctuation: either 1) ‘Caesar states that neuter nouns ending in -ar also take the same form in the dative and the ablative, as Pliny writes’, or 2) ‘Neuter nouns ending in -ar. Caesar makes the same statement as well, because in the dative and the ablative they take the same form, as Pliny writes’. 4. Where should we place quotations marks? Problems with direct quotations As for direct quotations, they represent, as we have previously stated, two thirds of the quotations with ‘vague boundaries’ in Charisius. Of course, this is partly a consequence of the lack of any written sign for quotation marks. It is exceptional that our manuscripts use that kind of sign, and obviously there is no trace of them in the Neapolitanus Latinus transmitting Charisius ars12. Consequently, we entirely depend on other criteria to establish the limits of a citation. Among them, the most important is the use of the verb inquit, which is always coherently used in inserted position13, even if this has sometimes been overlooked by editors, as in the following passage: Char. Gramm. p. 145.18-23 B.: Velius Longus de hac regula dixit in V ea parte ‘ergo alacris cunctosque putans excedere palma’ [Verg. Aen. 5.380]. Romanus libro de analogia ita inquit, ‘Alacris, licet consuetudo, ut ait Plinius libro VI dubii sermonis [frg. 24 Mazzarino = 49 Della Casa], alacer dicat, ut equester ordo…’. In this passage, the adverb ita is not used by Charisius to introduce Romanus’ citation, as Barwick’s puctuation seems to imply (as does Keil’s), but must rather be ascribed to Romanus himself, and consequently included in the quotation. This change implies either that Romanus also includes a citation from Virgil, or that he refers to Velius Longus’ doctrine; in both cases, we must translate something like ‘It takes that form, alacris, even if the use says alacer, as Pliny says…’. What caused the editors’ mistake is that ita is sometimes used to introduce direct quotations, for example with refert, loquitur. This ambiguous use of ita poses some problems of interpretation in the passage below, where, in our view, the reverse solution should be given: 12 A useful introduction to this topic can be found in MCGURK 1961, with further bibliography. 13 Regarding the use of inquam in Charisius, see PÉREZ ALONSO – GUTIÉRREZ GONZÁLEZ 2009, p. 132-134. Eruditio Antiqua 3 (2011) 62 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES Char. Gramm. p. 134.5-8 B.: Φυλακτήριον quod Graeci appellant, amuletum Latine dicimus. nam et Varro diuinarum XIII [frg. 129 Funaioli = 86 Cardauns] ita dixit, siue a molendo, id est infringendo uim mali, siue ab emulatione. Following Barwick’s punctuation we should interpret either that Varro used the word amuletum or that he glossed the Greek word φυλακτήριον with the Latin word amuletum. However, in this case it is more likely that ita refers not to the preceding, but to the following sentence and that it introduces a literal citation (a proper fragment!) from Varro, in which he explained the etymology of amuletum. In short, the sentence siue... siue... should be put into quotation marks14. Several factors support this view: there are parallels of siue a... siue a... in Varro15, which also makes frequent use of id est16 and of the gerund17 to explain an etymology from a verb. We can wonder why Keil and Barwick did not detect the literal quotation; first, it is true that the Anonymus de extremitatibus (the author of chapter 15 of book I) makes little use of direct quotations (in fact, most of the direct quotations in that chapter are Charisius’ own additions from Iulius Romanus), but among the few literal quotations we have found, there are two which belong to Varro (p. 132.2 B. [frg. 48 Funaioli = 105 Goetz–Schoell] and p. 139.15 B. [frg. 226 Funaioli]). 5. Fragment boundaries, textual criticism and doctrine assignment: a case study The adverb ita shares its referential ambiguity with sic. We can illustrate this with a very controversial passage, in which there are doubts about whether a de dicto or a de re interpretation applies; moreover, this passage shows that fixing citation boundaries is often interwoven with textual matters: Char. Gramm. p. 246.3-14 B.: Collatiua sunt aduerbia. Varro [frg. 50 Funaioli = 48 Goetz–Schoell] sic ait in III περὶ χαρακτήρων, propius proxime. in his extra consuetudinem communem frequenter perfectis uti solet Plautus, ut in Aulularia [668] ‘ea subleuit os mihi penisssume’ et in Mostellaria [656] <‘ita mea consilia perturbat paenissime’. et in Cistellaria> 14 This also appears to be CARDAUNS’ view (cf. 1976, I, p. 62), although in his commentary on this fragment (1976, II, p. 182) he does not discuss or justify its length. 15 For instance, VARRO Ling. 7.44 id tutulus appellatus … siue ab eo quod id tuendi causa capilli fiebat, siue ab eo quod altissimum in urbe quod est, arcs, tutissimum uocatur. 16 VARRO Ling. 5.33 dictus peregrinus a pergendo, id est a progrediendo; 5.37 seges ab satu, id est semine; 5.39 iacta, id est proiecta; al. 17 VARRO Ling. 5.44 uelabrum a uehendo; 5.91 tubicines a tuba et canendo; al. Eruditio Antiqua 3 (2011) 63 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES [63] ‘quid faciam? in latebras condas pectori penitissimo’ et in Curculione [122] ‘salue oculissime homo’. sed num oculissime βραχέως legendum? collatiua C, Nettleship : collatia N : conlata Keil ait Keil : esse N proprius N in his N : iis C Whether sic refers to collatiua sunt aduerbia or to propius proxime or even further can be discussed. As Usener (1867, p. 247) puts it: ‘Wo sollen denn die Varronischen Worte, die der Grammatiker durch ein sic ait einleitet, gesucht werden (…)? Etwa in propius proxime oder in der folgenden Zusammenstellung über adverbiale Superlative oder gar in dem vorausgehenden Ausspruch “es komt Komparation vor bei Adverbien”?’. Funaioli (1907, p. 207) ascribes to Varro just the words propius proxime, whereas Goetz and Schoell (1910, p. 201 and 295), following Usener, openly criticise that proposal and extend Varro’s citation to line 13 (at the end of the quotation from Plautus’ Curculio). But this view heavily depends on the reading ait which is only Keil’s conjecture, the manuscript reading being esse. In our opinion, it is very likely that the text once read sic esse ait18 so that sic might refer to collatiua sunt and after propius proxime an additional extended citation begins: ‘There are comparative adverbs: in book III of On paradigms Varro <says> this is the case of propius proxime’. The text that follows might be a literal quotation from Varro19: communis consuetudo is found also in Varro Ling. 5.6, 5.8, 6.82, 9.9, 9.114, 10.16, 10.74 and 10.76; moreover, Varro’s interest in and admiration for Plautus is well-known20. However the suspicion arises that there could be an intermediate source that incorporates Varro’s doctrine as well as some of his words, and there is also the possibility that, provided that sic refers to the preceding words, the words after propius proxime have nothing to do with Varro, being rather an additional remark triggered by Varro’s comment on propius proxime. In order to evaluate all these possibilities, we have first to make sure we understand the text, which is not clear enough as it stands and needs, in our opinion, some amendment. We propose to read it as follows: iis extra consuetudinem communem frequenter <pro> perfectis uti solet Plautus. iis C : in his N pro suppleuimus ‘These adverbs are often used by Plautus, contrary to the common use, instead of the positive ones’. 18 See also LINDEMANN 1840, p. 111, n. 17: ‘Praeterea deest in seqq. aliquid ad explendam sententiam, velut Varro sic esse refert cett. Vel: Varro sic posuit in tertio. Quomodo vero reconcinnanda sit oratio, difficile est dicere’. 19 See ThLL X.1, 1378, 31: ‘e Varrone, ut uid.’. 20 Let us just mention the famous passage from QUINTILIAN Inst. 10.1.99: licet Varro Musas, Aeli Stilonis sententia, Plautino dicat sermone locuturas fuisse si Latine loqui uellent… See further LEHMANN (2002, p. 49-56) on Varro as ‘exégète de Plaute’. Eruditio Antiqua 3 (2011) 64 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES The reading iis is provided by the excerpta Cauchii, and pro could easily have been lost by haplology21. The text makes much better sense with these changes, because it gives an early example of a remark on the absolute use of the comparative and the superlative (i.e. its use for the corresponding positive22 adverb or adjective), a remark that is found elsewhere in Latin grammatical texts23. The other controversial matter is about the source of this piece of doctrine. The context seems ‘Palaemonian’, since the piece is at the end of a section which is introduced (p. 241, 20 B.) with Aliis de aduerbiis longius disserere ita placuit and contains the example Palaemon docet24. In spite of this evidence, Barwick excludes Varro’s citation25 from Palaemon’s section26, arguing that a man who called Varro porcus (Suet. Gramm. 23.4) could hardly have used him in his own work27. Also, Barwick supports his view by pointing out that the parallel passages from Diomedes (Gramm. I 408.4-728) and Dositheus (Gramm. VII 412.28-413.4 Keil = 44.3-8 Bonnet) state that penes and penitus do not have a comparative or superlative, whereas in Varro’s citation a superlative of penitus is used. Accordingly, he argues that Varro’s citation must have been added by a grammarian who wanted to criticise the statement on penes and penitus lacking a superlative. In our opinion though, Barwick’s explanation of the passage is not entirely right: firstly, the above-mentioned passages from Diomedes and Dositheus are not at all strict parallels of this one29, but rather of Char. Gramm. p. 148.6-8 B.: sunt comparatiua aduerbia ex prototypis. uelut est prope aduerbium; ex hoc fit 21 See a similar confusion in PORPHYRIO Hor. epist.: 2.1.71 pro <per>fecto. 22 For perfectus with the sense of ‘positive (grade)’ see SCHAD 2007, p. 297 and ThLL X 1, 1378, 27-34. 23 See CHARISIVS Gramm. p. 354.14-17 B.: aduerbiis, cum relatiuum pro absoluto <ponitur>, ut ‘saepius…’ pro saepe; QUINTILIAN Inst. 9.3.19: utimur uulgo et comparatiuis pro absolutis, ut cum se quis infirmiorem esse dicet. 24 See KEIL 1857, p. XLIX and BARWICK 1922, p. 26, n. 2. 25 So we are again facing a problem of fragment boundaries. 26 The same criterion seems to be applied by MAZZARINO (1955, p. 78-83), but no justification is offered. See also LINDEMANN (1840, p. 111, n. 17): ‘Haec et sequentia sine dubio ab hoc loco sunt aliena’. 27 BARWICK 1922, p. 120: ‘Man wird nicht glauben, daß Pal. einen Mann, der ihm zuwider war und den er porcus (Sueton. de gramm. 23) gescholten hat, in seiner ars zitierte’. 28 Here penes penitus should be read instead of Keil’s penitus penitissimus (cf. BARWICK 1922, p. 120, n. 1). 29 Barwick was probably deceived by the fact that Diomedes’ passage is in exactly the same position as Charisius’ passage (after the paragraph on supine datives translatui and so on). But this is not unusual in Diomedes. Eruditio Antiqua 3 (2011) 65 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES comparatiuum propius proxime: item intus interius intime, et siqua alia30. Secondly, the doctrine in p. 246.3 ff. B. is consistent with the one in p. 148.6-8, for in both passages the grammarian deals with adverbial comparative/superlative forms which are derived from ‘primitive forms’; only Charisius p. 246.3 B. preserves the very interesting additional remark on the special use of some of those derived forms instead of the primitive ones. Admittedly, the Palaemonian piece is not free from later additions: since it is clear that Palaemon did not use Pliny (Barwick 1922, p. 119 n. 2), the paragraph at p. 242.19-22 B. (Plinius Secundus inter aduerbia qualitatis posuit…) must be seen as an addition; also, the example Beryti at p. 243.27 B. must have been inserted into the text after Probus (Barwick 1922, p. 190; Mazzarino 1955, p. 83)31. We must also pay attention to the fact that from p. 245.1 B. the proper classification of adverbs leaves room to quaestiones, that is to say to controversial issues (mostly related to anomaly) arising in grammatical description. This procedure is typical of chapter 15 of book I, where Caper’s influence is very marked. Moreover, those controversial issues seem to show an increasing difficulty: first (p. 245.1-4 B.), an adverb (numquam) of disputed semantic ascription; then, adverbs which share traits and have links (consortia32) with other parts of speech (prepositions and nouns); after that, adverbs in comparative or superlative form which lack a strict positive counterpart; finally, and most interestingly, remarks on the use of those forms instead of the positive forms. The latter comment fits better in a commentary than in a handbook, and the same is true of the critical remark on oculissime, which is actually best understood in the classroom context (Uría 2005, p. 104-105). Accordingly, we feel inclined to ascribe the doctrine that follows propius proxime not to Varro33, but rather to a tradition of philological commentaries on 30 Anyway, Charisius seems to understand and state this doctrine in a different way. While Diomedes and Dositheus refer to adverbs that lack a base-form, that is to say, that are not derived from an adjective, Charisius’ wording implies that some adverbial comparatives come directly from an ‘original word’, unlike most comparative adverbs, which are seen as coming either from the corresponding comparative adjective, or from a derived adverb (these ones being not original, but derived words): an adverb such as breuiter is itself derived from breuis so that the adverb breuius is seen either as coming from breuior or from breuiter. It seems that it is Diomedes’ and Dositheus’ version that is the one that implies some degree of alteration or even misinterpretation. 31 From this and similar evidence BARWICK (1922, p. 191) suggested that references to the usage of the ueteres were added to the Palaemonian edition soon after his death, Flavius Caper being the most likely responsible for those additions (Caper used both Probus and Plinius). See also URÍA 2005, p. 103-105. 32 Cf. CHARISIVS Gramm. p. 247.4-5 B. (e Romano). 33 The technical use of perfectus advises against an ascription to Varro, since, when he refers to comparison of adjectives in De lingua Latina (8.17, 52, 75-78 and 9.72), he does it without an established terminology. Eruditio Antiqua 3 (2011) 66 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES Plautus (Probus? Sisenna?34), which probably reached Charisius (or his source) through Flavius Caper35. 6. Who says what? Problems in second-hand quotations Let us now analyse some other passages. As we have already shown, Barwick’s edition contains not a few inconsistencies regarding the use of quotation marks in literal citations. Some of them have been noted and amended by the editors of the grammarians quoted by Charisius, as in the following case: Char. Gramm. p. 157.3-9 B. (= Iulius Romanus, De analogia): Aenigmatis Varro de utilitate sermonis III [frg. 5 Funaioli = 53 Goetz–Schoell]. Ait enim Plinius [frg. 55 Mazzarino = 83 Della Casa] ‘quamquam ab hoc poemate his poematibus facere debeat, tamen consuetudini et suauitati aurium censet summam esse tribuendam, ut in Aceste et in Anchise Maroni diximus [cf. p. 85, 11-19 B.] placitum; et quia Graeca nomina non debent Latinis <nomi>nibus alligari’. It is quite surprising that, even if Barwick expressly notes that diximus is referring to Char. Gramm. p. 85.11-19 B., he nevertheless considers that Pliny’s quotation goes far beyond it, and reaches alligari. A similar solution had been offered by Keil, who was followed by Beck (1894, p. 20.12-17) in his edition of Plinius’ Dubius sermo (see also Goetz–Schoell 1910, p. 202-203 frg. 53 and Funaioli 1907, p. 186 frg. 5). However, modern editors of Pliny’s grammatical treatise have been more careful in placing the limits of the fragment, for they unanimously consider that it finishes with tribuendam, the rest belonging to the reporter, namely Iulius Romanus (the fragment is in chapter 17 of book I, the socalled De analogia by Iulius Romanus). Even so, they give no reason for this choice, perhaps because they consider it obvious. Mazzarino (1955, p. 279 frg. 55 in app.) just notes that ‘uerba ut in Aceste — alligari aliena habenda Pliniano contextui sunt mihi uisa’; the reason for this, as we have already said, cannot be anything else than the fact that the rejected words contain a reference to Char. Gramm. p. 85.11-19 B. (and not to a fragment of Plinius’ work, as one might expect). Since the referred passage belongs to a different chapter, namely the socalled Anonymus de extremitatibus (chapter 15 of book I), one should assume that 34 Both of these authors commented on Plautus and show an interest in adverbs: for Probus, see CHARISIVS Gramm. p. 257.27-31 (efflictim) and p. 274.22-24 B. (parcissime); cf. also URÍA 2005, p. 104; for Sisenna, see CHARISIVS Gramm. p. 285.20-21 and 24-26 B. and cf. PERUTELLI 2004, p. 60. 35 Caper is known to have used Probus, as witnessed by CHARISIVS Gramm. p. 150.31-32 B.: Fl. tamen Caper Allecto monoptoton esse Valerium Probum putare ait. He also made extensive use of Pliny’s Dubius sermo, and Pliny used Sisenna. Eruditio Antiqua 3 (2011) 67 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES either the internal reference is an addition by Charisius himself, or that it belongs to Romanus, but refers to a part (or some lemmata) of his work that Charisius did not copy. The second possibility seems more likely, in our view. Finally let us examine a more complex example of direct quotation, which can also illustrate the different levels of citation in Charisius’ work, as well as the difficulties for determining a hierarchy among them. In this case, Barwick, who follows Beck (1894, p. 11.5-9), may be right against the modern editors of Pliny: Char. Gramm. p. 160.15-19 B.: Diligente. ‘Verrius Flaccus [frg. 21 Funaioli] – inquit Plinius [frg. 88 Mazzarino = 26 Della Casa] – eorum nominum quae -ns finiuntur casu nominatiuo ablatiuus in -e dirigendus est’. Itaque Caesar epistularum ad Ciceronem ‘neque – inquit – pro cauto ac diligente se castris continuit’ (frg. 2 Klotz = 48 Cugusi). (The position of the quotations marks is the same in Keil’s edition.) Even if the use of quotation marks in this passage is not very clear, in so far as Barwick quotes Beck, we can conclude that for both of them the words comprised between Verrius Flaccus and continuit were taken from the Dubius sermo. Nevertheless, the modern editors of Pliny unanimously state that the citation finishes with dirigendus est, although they hold differing views about its beginning. Mazzarino places a lacuna after Verrius Flaccus and attributes to Pliny just from eorum to dirigendus est, leaving aside the reference to Verrius Flacus. This proposal is quite problematic, and Mazzarino tries to justify it by assuming that the lemma was cut down, for it allegedly shows a doctrine which is opposite to Pliny’s doctrine36. That Mazzarino is wrong was soon noticed by Della Casa, who pointed out that Pliny has always claimed that cognomina end in -e in the ablative37. To judge from her translation of the passage38, Della Casa assumes that Pliny is reporting Verrius’ doctrine, although in her introduction (1969: 42) she implies that there is no firm evidence to establish that Pliny used Verrius. Indeed, the syntax of the fragment is difficult, and this has caused many different punctuations. It is worth mentioning the solution by Funaioli (1907, 36 MAZZARINO (1955, p. 301 [frg. 88 in app]): ‘lemma decurtatum puto. Nam si, ut est uerisimilius, frg. 87 [= Charisius Gramm. p. 159,17-22 B.] rectam Plinii doctrina praebet, intellegere non possumus cur diligens quoque, quod cognomen potest esse, (…) ablatiuo sing. -e claudi debeat’. Cf. DETLEFSEN 1867, p. 701. 37 DELLA CASA 1969, p. 225 ‘Non si capisce con molta chiarezza perché il Mazzarino dica: “intellegere non possumus (…)”. Anzi, Plinio ha sempre sostenuto la uscita in -e dei cognomina (cf. fr. 18 Agile [Char. Gramm. p. 154.18-20 B]; fr. 23 Auxiliare [Char. Gramm. p. 155.14-17 B]; fr. 25 Constante [Char. Gramm. p. 159.17-22 B]; ecc.); anche in questo caso il suo pensiero è che diligens, sia come participio che come cognomen, deve avere come ablativo diligente’. 38 DELLA CASA 1969, p. 225: ‘diligente. (Dice) Verrio Flacco – (secondo quanto) afferma Plinio – che l’ablativo di quei nomi che al nominativo terminano in ns debe terminare in e’. Eruditio Antiqua 3 (2011) 68 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES p. 519 ‘Diligente Verrius Flaccus’ inquit Plinius). As for terminology, the special use of dirigere is found in Pliny, whereas the gerundive is often employed by Verrius39. Accordingly, if neither syntax nor terminology provide a satisfactory conclusion, the best solution might well be to regard the passage as an example of what is called ‘fusion of points of view’, that is to say when the reporter accepts the ideas of the quoted authority, thus resulting in a sort of ‘blended doctrine’. There is yet another possibility: if we extend the quotation to also include Caesar’s words, then both Caesar and Verrius could be used, not properly as grammatical authorities whose doctrine is reported, but rather as plain examples of the use of the word. We can find some support for this in two other passages from Charisius, where Caesar and Verrius are also put together, this time apparently as grammatical authorities: Char. Gramm. p. 114.2-4 B. (Anon. de extremitat.): Panis autem genetiuum pluralem Caesar de analogia II [frg. 8 Funaioli = 9 Klotz] ‘panium’ dixit, sed Verrius [frg. 19 Funaioli] ‘panum’ sine i. Char. Gramm. p. 178.30-178.1 B. (Iulius Romanus = Plin. Dub. serm. p. 25.10-12 Beck): Panium Caesar de analogia libro II [frg. 8 Funaioli = 9 Klotz] dici debere ait. Sed Verrius [frg. 19 Funaioli] contra. Nam i detracta panum ait dici debere. Also, a close look at similar uses of itaque for bringing in additional examples (see ThlL VII.2, 530.65 f. ‘in exemplo afferendo’) favours the attribution of Caesar’s passage to Pliny. Consequently, if we propose that the passage from Pliny’s Dubius sermo contained both fragments, we can take it as a collection of material dealing with the abl. sg. of diligente, and consisting of two juxtaposed citations, something like: Diligente. Verrius Flaccus: ‘eorum nominum quae -ns finiuntur casu nominatiuo ablatiuus in -e dirigendus est’. Itaque Caesar epistularum ad Ciceronem: ‘neque pro cauto ac diligente se castris continuit’. 39 Actually, according to MAZZARINO (1948a, p. 208), dirigere as used at p. 160,15 B. can be seen as a technical term of Pliny. For dirigere in Pliny see CHARISIVS Gramm. p. 167.7-11 B. (= Plin. Dub. serm. frg. 54 Mazzarino = 82 Della Casa) glossemata ut toreumata enthymemata noemata schemata poemata et his similia omnia Varronis [frg. 256 Funaioli = 52 Goetz–Schoell] regula – inquit Plinius – datiuo et ablatiuo plurali in -bus derigit, quia singularis ablatiuus e littera finiatur. For Verrius frequent use of the gerundive, see FESTVS p. 218.2-3 L. Obsidionem potius dicendum esse, quam obsidium; p. 316.33-34 L. <Ru>ctare <non ructari dicendum est> (cf. Paul. Fest. p. 317.13); p. 356.35-36 L. Ruri esse, non rure dicendum, testis est Terentius in Phormione; CHARISIVS Gramm. p. 139.17-19 B. (= Verrius frg. 17 Funaioli.) *qui necesse sit. Hic masculine dicendus est, ut Verrius ait, quoniam neutra in i et us non exeunt; p. 124.11-13 B. (= Verrius frg. 8 Funaioli) Manibias per duo i dicendum, quia sunt a manibus, ut putat Verrius <dictae>. Eruditio Antiqua 3 (2011) 69 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES This passage bears witness to the difficulties of disentangling sources and doctrines in Charisius and proves that a satisfactory solution cannot always be reached. 7. Conclusion To summarise, we hope we have illustrated the different implications that fragment boundaries have, not just for the ascription of both grammatical doctrine and terminology, but also for textual improvement. Additionally, we think it has been established that any attempt at (re)-editing grammatical fragments should be based on a good acquaintance with the primary sources as well as on a full understanding of their mechanisms for reporting the preceding doctrine. REFERENCES AUTHIER-REVUZ J. 1992, ‘Repères dans le champ du discours rapporté’, L’information grammaticale 55, p. 38-42. BARWICK K. 1925, Flauii Sosipatri Charisii Artis Grammaticae libri V, Leipzig. BECK J. W. 1894, Plinius secundus. Librorum dubii sermonis VIII reliquiae, Leipzig. BONNET G. 2005, Dosithée. Grammaire latine, Paris. CALBOLI G. 1972, La linguistica moderna e il latino, Bologna. CARDAUNS B. 1976, M. Terentius Varro. Antiquitates rerum diuinarum, vol. I-II, Wiesbaden. CONTE G. B. – BARCHIESI A. 1989, ‘Imitazione e arte allusiva. Modi e funzioni dell’intertestualità’, in Lo spazio Letterario di Roma antica, vol. I, G. Cavallo – P. Fedeli – A. Giardina (edd.), Roma 1989, pp. 81-114. COULMAS F. 1986, ‘Reported speech: Some general issues’, in F. Coulmas (ed.), Direct and Indirect Speech, Berlin-New York-Amsterdam, p. 1-28. CUGUSI P. 1979, Epistolographi Latini minores, vol. II.1, Torino. Eruditio Antiqua 3 (2011) 70 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES DE MAURO T. 1965, ‘Il nome del dativo e la teoria dei casi greci’, Atti della Accademia Nazionale dei Lincei. Rendiconti. Classe di scienze morali, storiche e filologiche s. VIII, 20, p. 1-61 (repr. in Opuscula collecta, Roma 2005). DELLA CASA A. 1969, Il Dubius sermo di Plinio, Genova. DETLEFSEN D. 1867, ‘Zur Flexionslehre des ältern Plinius’, in Symbola philologorum Bonnensium in honorem Friderici Ritschelii collecta, II, p. 697-714. FUNAIOLI G. 1907, Grammaticae Romanae fragmenta, Leipzig. GOETZ G. – SCHOELL F. 1910, M. Terenti Varronis De lingua Latina quae supersunt, Leipzig. GUARDÌ T. 1974, Cecilio Stazio. I frammenti, Palermo. HERZOG R. – SCHMIDT P. L. (éds.) 1993, Nouvelle histoire de la littérature latine. V. Restauration et renouveau. La littérature latine de 284 à 374 après J.-C., Turnhout. KASTER R. A. 1986, ‘Islands in the Stream: The Grammarians of Late Antiquity’, Historiographia Linguistica 13, p.323-343. KEIL H. 1857-1880. Grammatici Latini, I-VIII, Leipzig. KLOTZ A. 1927, C. Iuli Caesaris Commentarii, vol. III, Leipzig. LEHMANN A. 2002, Varron critique littéraire. Regard sur les poètes latins archaïques, Bruxelles. LINDEMANN F. 1840, Corpus grammaticorum Latinorum ueterum, t. IV., Leipzig. MAZZARINO A. 1948a, ‘Una nuova pagina di Plinio il Vecchio’, Maia 1, p. 201222. — 1948b, ‘Due note filologiche. II. Carisio, 175,12 Barw.’, Maia 1, p. 65-67. — 19552, Grammaticae Romanae fragmenta aetatis Caesareae. Volumen primum, Torino. MCGURK P. 1961, ‘Citation marks in early manuscripts’, Scriptorium 15, p. 3-13. PÉREZ ALONSO M. A. – GUTIÉRREZ GONZÁLEZ R. 2009, ‘Charisiana III (Char. Gramm. p. 153.30 y p. 162.19 Barwick)’, Exemplaria Classica 13, p. 129142. Eruditio Antiqua 3 (2011) 71 JAVIER URÍA VARELA – RAMÓN GUTIÉRREZ GONZÁLEZ VAGUE BOUNDARIES PERUTELLI A. 2004, Prolegomeni a Sisenna, Pisa. RIBBECK O. 18732, Comicorum Romanorum praeter Plautum et Terentium fragmenta, Leipzig. ROSIER L. 2008, Le discours rapporté en français, Paris. SALLMANN K. (éd.) 2000, Nouvelle histoire de la littérature latine. IV. L’âge de transition. De la littérature romaine a la littérature chrétienne de 117 à 284 après J.-C., Turnhout. SCHAD S. 2007, A Lexicon of Latin Grammatical Terminology, Rome. SCHENKEVELD D. M. 2004, A Rhetorical Grammar. C. Iulius Romanus, Introduction to the Liber de Aduerbio, Leiden-Boston. TOUMARLA U. 2000, La citation: mode d’emploi. Sur le fonctionnement discursif du discours rapporté direct, Helsinki. URÍA J. 2005, ‘What can we learn from place-names in Charisius’ Ars grammatica?’, in The Origins of European Scholarship. The Cyprus Millenium International Conference, I. Taifacos (ed.), Stuttgart, p. 95-107. — 2006, ‘Consideraciones sobre el prefacio del Arte gramática de Carisio’, Studium. Revista de Humanidades 12, p. 113-125. USENER H. 1867, ‘Antwort an Herrn Dr. Moritz Crain’, Jahrbücher für Classische Philologie 95, p. 247-250. © Eruditio Antiqua 2011 ISSN 2105-0791 www.eruditio-antiqua.mom.fr [email protected] Image : © Kunsthistorisches Museum,Vienna Eruditio Antiqua 3 (2011) 72