We have two papyri of Terence, both found in Egypt, and both relating to the same play, the Andria, first performed on stage in 166 bce. Before presenting them, I would like to thank Maria Chiara Scappaticcio and Alessandro Garcea for having drawn my attention to these papyri and provided me with the documents, in particular the PLATINUM data, which enabled me to study them.
1.1 The Papyri and Their Context of Production and Use
1.1.1 The Papyri
The first is a Vienna papyrus, P.Vindob. inv. L 103, dated to the fourth century. It contains 46 verses, taken from vv. 489–582 of Act III (sc. 2, 3, 4).Footnote 1 The second, partly written in Alexandrian capital letters, comes from Oxyrhynchus, P.Oxy. XXIV 2401. A little later (end of the fourth or beginning of the fifth century), it presents 116 verses extracted from vv. 602–88 of Act III (sc. 4, 5) and vv. 924–79A of Act V (sc. 4, 5, 6).Footnote 2 These two papyri are fragmentary, discontinuous, and very mutilated, especially the first one. The delicate editing work that has been carried out in recent years by Roberto Danese (Reference Danese1989) for the first, and for the second, by Gabriel Nocchi Macedo (Reference Nocchi Macedo2018a) and Maria Chiara Scappaticcio (Reference Scappaticcio2018c), in the framework of the PLATINUM project, is to be commended.
The harvest is meagre: 162 verses preserved out of 981, that is to say only 16.5 per cent of the Andria, and 2.6 per cent of the whole of Terence’s theatrical corpus. But their early date makes them invaluable, since we have here two original and authentic editions from late Latin, predating the Carolingian and medieval traditions. To these must be added a parchment from a later century (late fifth/early sixth), the famous Vaticanus Latinus 3226, known as Bembinus, glossed, which offers 188 verses, extracted from vv. 787–981 of Acts IV (sc. 4, 5) and V (sc. 3, 4, 5, 6), 50 of which overlap with those of the Oxyrhynchus papyrus (vv. 924–50, 957–79). This privileged situation of ancient editions is explained by the important role played by Terence in the grammatical tradition. It is also related to the development of grammar in the fourth century, both in the Roman West and in the Greek–speaking East.
1.1.2 The Grammatical Boom of the Fourth Century
We know the role of ‘circularity’ played by the classical authors in the conception of Latin grammar, whose function is both exegetical, exegetice (enarrare, interpretari, to explain the texts) and descriptive, prescriptive, horistice (recte scribendi loquendique ratio, praecepta demonstrare, to state the rules of correct expression):Footnote 3 the grammars are based on the texts, from which they borrow numerous quotations, but they are also used to interpret these same texts …
It is also known that the notion of ‘grammar’ in antiquity is to be taken in a very broad sense. It encompasses the artes (grammars proper), orthographic and metrical treatises, and various thematic opuscules, gathered in the corpus of Keil’s Grammatici Latini. But it also has ‘auxiliaries’, which exploit the same theoretical, methodological, and terminological principles, in particular the commentaries, which accompanied the editions, as well as lexica, collections of glosses, and differentiae, to which can be added the connections it has with rhetoric, and various remarks by the ancient scholars that can be found here and there.
Our two papyri are part of a grammatical and exegetical tradition that is already several centuries old:Footnote 4
First century bce: Varro, Cicero;
First century ce: Palaemon, Quintilian, Pliny, Probus, Asconius;
Second century: Caper, Terentianus Maurus, Velius Longus, Festus, Acron, Gellius;
Third century: Sacerdos, Iulius Romanus, Porphyrion, Arruntius Celsus;
Fourth century: Donatus, Artes (and commentators of the fifth century: Sergius, Pompeius, Cledonius, Explan. in Don.), Commentum Terenti; Servius, In Donati Artem, Vergili commentarii; [Probi], De catholicis, Instituta artium, De nomine; Arusianus Messius, Exempla elocutionum; Nonius Marcellus, De compendiosa doctrina; Charisius, Ars; Diomedes, Ars; Evanthius, De fabula, De comœdia excerpta; Dositheus, Grammatica; Jerome;
Fifth century: Audax, Cledonius, Consentius, Asper, [Probi] Appendix, Rufinus;
Sixth century: Priscian, Eutyches, Cassiodorus.
As early as the first century bce, it provided commentators on Terence, such as Probus. This grammatical tradition really took off in the fourth century. In the West, Donatus and Servius were both grammarians and commentators, the former of Terence, the latter of Virgil, but Servius’ commentary contains many references to Terence (illud Terentianum). In the eastern part of the empire, especially in Constantinople, grammars were produced more directly for Greek-speakers, such as that of Charisius. They include numerous references to the Greek language. It is in this context of practice of the Latin language by Greek-speakers that the two papyri must be placed.
At the beginning of the third century, the Edict of Caracalla, the Constitutio Antoniniana (212), was promulgated, granting Roman citizenship to all free inhabitants of the Roman Empire and thus opening the doors to careers in the Roman administration, which required knowledge of the Latin language and culture. This led to the development of works that set out the specific features of the Latin language as opposed to those of the Greek language: De Latinitate, Proprietates Latinorum, Exempla elocutionum. Book 5 of the Grammar of Charisius includes a chapter De idiomatibus. This examination of the specific features of each of the two languages naturally opens the way to bilingual works, such as Dositheus’ Grammar, and to what can be called a ‘comparative grammar’ of Latin and Greek, as in Macrobius’ opuscule on the differences and similarities (differentiis vel societatibus) between Greek and Latin verbs, or the catalogue of book 18 of Priscian’s Grammar:
Pliny (first century), Dubius sermo;
Caper (second century), De Latinitate;
Charisius, Ars 5 (fourth century), De idiomatibus, De differentiis;
Arusianus Messius (fourth century), Exempla elocutionum;
Priscian, Ars 17, gramm. 3.107.4*, Proprietates Latinorum; 18 (catalogue);
Macrobius, De verborum Graeci et Latini differentiis vel societatibus excerpta.
This Hellenic-speaking context is found in both papyri. They contain lexical glosses in Greek, which have been well studied,Footnote 5 in relation to the monolingual and bilingual lexica known elsewhere.
1.2 Terence: Ancient Grammatical Education and Culture
1.2.1 A Key Reference and Founder of the Grammatical Tradition
Terence has long been an essential and founding reference in the grammatical tradition and in teaching practice.Footnote 6 He owes this to the universal value of his humanism, to the clarity and classicism of his language, but also to the rhetorical and legal dimension of the speeches made by some of his characters, which can be formative for future executives of the Roman administration. It is perhaps in this sense, more than in the strict learning of the Latin language, that the educational purpose of these two papyri should be considered.
Terence was exploited and quoted as early as the first century bce by Varro and Cicero, who presents sixty-seven quotations. But it was in the fourth century that he was institutionally recognised as one of the four reference authors (the ‘quadriga of Messius’),Footnote 7 along with Virgil, Cicero, and Sallust, in the work of Arusianus Messius on the case constructions of Latin verbs. After the world of the theatre, Terence had a second life in grammatical teaching, to the point of becoming an example of grammar, in the company of his three companions, as in the typical example of the vocative, Terenti, scribe, ‘Terence, write’.Footnote 8
Consultation of the Grammatici Latini database, put online by A. Garcea and V. Lomanto, already gives an idea of the importance of Terence in the grammatical tradition. He is the subject of 1,078 citations, which places him in fourth place after Virgil (6,280), Horace (1,269), and Cicero (1,183), and before Sallust (592) and Plautus (492). Of these 1,078 quotations, 293 concern the whole of the Andria, which is the most quoted play.
1.2.2 Citations of Terence, Andria
Sixteen verses from the papyrological corpus (in its attested parts) are explicitly quoted in the Grammatici Latini corpus. Most of them come from Arusianus Messius and Priscian, and they mainly concern problems of syntax, of casual rection of Latin verbs, paralleled, in Priscian, with the corresponding Greek verbs, thus of credo/πιστεύω. But there are also questions of morphology, such as the archaisms hoc(c)ine and the genitive nulli (for nullius), or the diathetic variation of alterco/altercor:
| 497 | credon | Prisc. gramm. III 347.14–15 (credo // πιστεύω + acc. / dat.) |
| 546 | in re(m) est | Arvs. 58.4–5 (Di Stefano Reference Di Stefano2011) |
| 608 | nullius / nulli | Char. 142.8–10 (Barwick Reference Barwick19642); Prisc. gramm. II 227.14–16, III 7.12–14 |
| 614 | me(i) | Prisc. gramm. III 16.8–9. Cf. III 189.4 |
| 625 | hoc(c)ine | Prisc. gramm. II 592.24 (uetustissimi, in antiquissimis codicibus) |
| 639 | expos[t]ulem | Arvs. 38.1–2 (Di Stefano Reference Di Stefano2011) (cum eo); Prisc. gramm. III 288.11–12 (ἐξαιτούμενος + acc. / gen.) |
| 639 | [adea]mne / adeon | (adire ad) Arvs.11.16 (Di Stefano Reference Di Stefano2011); Prisc. gramm. III 353.16–17; Prisc. gramm. III 353.16–17 (προσέρχονται + dat. / acc. / πρὸς + acc.) |
| 648 | lactasses | Eutych. gramm. V 483.28 |
| 653 | altercasti / altercatus est | Prisc. gramm. II 392.24–393.1 |
| 932 | cuiam | Prisc. gramm. III 179.16–17 |
| 933 | arrige aures | (uitium, cacenphaton) Diom. gramm. I 451.5; Don. mai. 658.7–8 (Holtz Reference Holtz1981); Serg. gramm. V 264.4–5 (metaphora); Pomp. gramm. V 293.17; Ivl. Tol. ars 187.5 (Maestre Yenes Reference Maestre Yenes1973) |
| 939 | dignus (+ abl. / acc.) | [Don. mai.: uaria auctoritate]; Pomp. gramm. V 173.10–12, 188.6–7; Cledon. gramm. V 45.28 |
| 940 | scrupulus (m. / -um, n.) | [Prob.] De nomine 70.1 (Passalacqua Reference Passalacqua1984) |
| 940–1 | dignus odio / odium | Pomp. gramm. V 173.6, 188.6–7 |
| 961 | intercesserit | Arvs. 60.2–3 (Di Stefano Reference Di Stefano2011) (intercedit + acc. / dat.); Prisc. gramm. III 107.16–17* (tibi) |
| 964 | gauisurum gaudia | (gaudeo + acc. / abl.) Arvs. 42.4 (Di Stefano Reference Di Stefano2011) |
These are only explicit mentions of Terence. The implied harvest is much wider, but more difficult to identify.
1.3 An Ancient Editorial Tradition
The papyri of Terence, when compared with other witnesses of the grammatical tradition, constitute a beautiful ‘playground’ for appreciating the variations in editorial practice, the variae lectiones already numerous in Late Antiquity.
1.3.1 P.Vindob. Inv. L 103 and P.Oxy. XXIV 2401
The two papyri are hardly comparable, given that they do not contain the same sequences of texts. But some words, especially names of characters, are found in both. The papyri both favour the full form consiliis, also given by Donatus, whereas other manuscripts and modern editors retain, for metrical reasons, the contracted form consilis.
Consiliis: P.Vindob. inv. L 103, v. 576; P.Oxy. XXIV 2401, vv. 616, 650. Don. / consilis: codd., edd.
The two papyri differ, however, on the spelling of the final sequence of Davom/Davum:
Davo[m]: P.Vindob. inv. L 103, v. 579 / Davum: P.Oxy. XXIV 2401, v. 96.
Don. Ter. Andr. 173.4 et 580.1 Davus, ut recte scribatur, Davos scribendum est, ‘Davus, if you want to write it correctly, should be written Davos.’Footnote 9
They also differ on the form of the vocative of Chremes, in -ēs or -ē, which raises the problem of the integration of Greek words in Latin declensions. The question is a textbook case in the Latin grammatical tradition, as in Charisius and Priscian, and it has its source precisely in the theatre of Terence:
Voc. [Chrem]es: P.Vindob. inv. L 103, v. 574 / Chreme: P.Oxy. XXIV 2401, vv. 930, 946; codd., edd.
Char. 85.22–8 (Barwick Reference Barwick19642) Chremes et Laches apud comicos … varie declinantur […] Nullum … nomen in Latina lingua desinit in flexum -ē, ‘Chremes and Laches show variations in declension in comics […] No noun in Latin has an inflectional ending in -ē.’
Prisc. (gramm. 2.287.24–289.4) Est tamen quando in -ēs productam terminantium Graecorum vocatiuus in -ē longam exit secundum Graecos […] ut Chremē, ‘It happens, however, that the vocative of Greek words with a long -ēs ending, ends in -ē long, as in Greek […], thus in Chremē’ [there follow examples from Ter. Andr. (outside the papyrological corpus)].
The vocative in -ē is incongruous in Latin, but it has the advantage of being well marked as opposed to the nominative Chremēs.
1.3.2 P.Oxy. XXIV 2401 and Donatus, Andria, ad loc. (Legitur et)
It is also interesting to compare the lessons of the papyri with those given by Donatus’ commentary. Donatus points out the variants he was able to find in the editions he had at his disposal (legitur et, ‘one reads also’), such as, in v. 627, the choice of the singulars gaudeat and comparet, instead of the plurals gaudeant and comparent:
P.Oxy. XXIV 2401, v. 627 gaudeant … comparent / Don. Ter. Andr. 627.2: legitur et gaudeat et comparet.
In v. 653, too, he notes that of the periphrastic perfect altercatus es instead of the active form altercasti, which gives rise in Priscian to a debate on the uses attested in the vetustissimi and in common usage (consuetudo), as regards verbal voice:
P.Oxy. XXIV 2401, v. 653 [alte]rcasti
Don. Ter. Andr. 653.2 legitur et altercatus es. Non enim alterco dicimus, ‘There is also the variant altercatus es. One does not say alterco.’
Prisc. (gramm. 2.392.6–393.1) plurima inveniuntur apud vetustissimos quae contra consuetudinem vel activam pro passiva, vel passivam pro activa habent terminationem, ut […] alterco … pro altercor, unde Terentius in Andria [653] ‘Scio, cum patre altercasti dudum’, ‘In archaic authors, we find many verbs which, contrary to usage, have an active ending instead of passive, or passive instead of active, such as alterco instead of altercor, as in Terence in the Andria [v. 653]: “I know that you had a quarrel with your father just now.”’
In v. 650, instead of confecit (‘cause’), attested in the Oxyrhynchus papyrus and the Calliopian manuscripts (Σ), Donatus mentions a variant conflavit (‘stir up’), retained by modern editors:Footnote 10
P.Oxy. XXIV 2401, v. 650 quantasque hic suis consiliis confecit sollicitudines,
‘all the anxieties his advice has caused me’
confecit: pap., MSS Σ, Don. / conflavit: edd.
Don. Ter. Andr. 650.3 confecit: legitur et conflavit
Don. Ter. Andr. 167.3 confore: confieri, perfici, unde confectum negotium dicitur vel confecta res, quae ad plenum perficiuntur, ‘confore means confieri, “to be achieved”, perfici, “to be completed”, and confectum negotium or confecta res, “a done deal”, for that which has been brought to a full completion (perficiuntur).’
The more expressive conflavit variant was probably influenced by alternative expressions such as invidiam conflare, ‘to stir up jealousy’, and iudicium conflare, ‘to stir up legal action’. They are well attested in Cicero (Verr. 2.2.73 and 116, for example), and already integrated in the Latin grammatical tradition. Charisius, in his chapter 5, De Latinitate (406.1 Barwick Reference Barwick19642), cites the expressions invidiam movit and iudicium conflavit, glossed by δικαστήριον συνεκρότησεν in a list of idiomatic expressions of judicial eloquence appended to Dositheus’ bilingual grammar (gramm. 7.426.13–14).
Finally, the question of the alternative ending (alter exitus), obviously apocryphal, which characterises the Andria, should be mentioned. Donatus specifies that it does not appear in the good manuscripts (exemplaribus bonis):
Don. Ter. Andr. 978.1 Hi versus, usque ad illum ‘gnatam tibi meam Philumenam uxorem’, negantur Terentii esse, adeo ut in plurimis exemplaribus bonis non inferantur, ‘These verses, until gnatam tibi meam Philumenam uxorem, are not by Terence; indeed, they do not appear in most good manuscripts.’
The archetype of the Oxyrhynchus papyrus was clearly not one of these, since it contains three verse beginnings of an alter exitus, different from the one known elsewhere.Footnote 11
1.3.3 The Different ‘Hands’ (m1, m2, m3): Corrections and Glosses
Both papyri, and especially the Oxyrhynchus papyrus, contain corrections made by the original copyist or by a later reader, as well as supralinear Greek glosses. These modifications of the text also allow for interesting exegetical observations which can be related to the grammatical tradition. Some corrections remedy simple carelessness, such as the deletion of the a in v[a]ecordia (P.Oxy. XXIV 2401, v. 626), due to a homophony between the prefix vē- and the interjection vae, or the addition of the c in fa`c´tis (P.Oxy. XXIV 2401, v. 666), initially influenced by fatis (from fatum, ‘fate’): it may have been generated by the semantic context, which is that of a divine curse (At tibi di dignum fa`c´tis exitium duint!). Some corrections are obviously glosses in an earlier edition, which were then inserted into the text. This is undoubtedly the case, in P.Oxy. XXIV 2401, of the substitution of the usual forms of genitive nullius (nulli{i}ius, v. 608) and dative cuiquam ({cu}quoiquam, v. 626) for the archaic forms nulli and quoiquam, widely referred to by grammarians:
P.Oxy. XXIV 2401, v. 608 tam nulli{i}ius consilii sum
nulli (m1) in nullius (m2) per adiectione (supra lineam) emendatum ab altera manu
nulli, pap.1, codd, edd. / nullius, pap.2, Don.
Don. Ter. Andr. 608.3 nulli nos nullius et solius, recte autem veteres nullus, nulli et solus, soli declinabant, ‘We say nullius and solius, but it was correct for the ancients to decline nullus, nulli and solus, soli.’Footnote 12
We can also quote the substitution of the more explicit and contemporary prefixed verb auferet for the simple verb feret (au{f}feret, v. 610), which testifies to the extension of prefixed forms in late Latin. Donatus and the lexicographersFootnote 13 make a distinction (differentia) between the two forms:
P.Oxy. XXIV 2401, v. 610 au{f}feret
feret (m1) in auferet (m2) per adiectione (supra lineam) emendatum ab altera manu
ferret pap.1 / auferet pap.2, codd., edd.
Don. Ter. Andr. 28.4 auferte: bene ‘auferte’; auferimus enim ea quae cum fastidio cernimus, ferimus ea quae cum honore tractamus, ‘auferte (“carry away”) [is used] rightly; for one carries away (auferimus) what one sees with displeasure, and carries (ferimus) what one treats with respect.’
This is probably also the case for the insulting abstract noun scelus (neuter), ‘crime’, redeveloped into the adjective scelestus (masculine), ‘criminal’, to account for the gender syllepsis, which calls for the use of the masculine relative pronoun qui instead of the neuter quod, as Donatus points out:
P.Oxy. XXIV 2401, v. 607 ubi illic est scel`est´us, qui, me, perdidit‧Footnote 14
scelus (m1) in scelestus (m2) per adiectione (supra lineam) emendatum ab altera manu
scelus pap1, codd., edd. / scelestus pap2
Don. Ter. Andr. 607.4 scelus qui ad sensum, non ad verba redegit; et est figura σύλλημψις. Nam quia scelus ‘scelestus’ intellegitur, modo qui subiecit, non quod, ‘“This criminal who”: he has conformed to the meaning, not to the words; and this is the figure of syllepsis. Indeed, since scelus [neuter] must be understood in the sense of scelestus [masculine], he has used qui instead of quod.’
The Greek glosses have already been well studied (see above, n. 5). One of them, however, deserves attention: the equivalence established in v. 648 between the Latin lactasses and the Greek ἠπάτησας:
P. Oxy. XXIV 2401, v. 648 [n]isi me lactasses, amantem et falsa spē produceres (cf. also v. 912 lactas), ‘if you had not taken advantage of my love to cajole me and entertain false hopes’.
ἠπάτησας supra lineam = lactasses, a tertia manu glossatumFootnote 15
It is never made clear in the Latin sources that there is a homonymy between the verbs lacto1, ‘to suckle’ (denominative of lac, lactis, glossed as θηλάζω), and lacto2, ‘to seduce, deceive’, frequentative of lacio (derived from lax, lacis), ‘to lure into one’s nets’, so the distinction remains implicit.Footnote 16 Donatus (Andr. 648.1–2 and 912.1–2) does identify the meaning of ‘to deceive, to seduce’, and makes the connection with lacio, delecto, and oblecto. But the grammatical tradition, limited to a morphological perspective, seems to retain only the frequentative derived from lac, confusing the two verbs.Footnote 17 The Greek gloss by ἠπάτησας (from the verb ἀπατάω, ‘to seduce, to deceive’, denominative of ἀπάτη, ‘deception’) disambiguates the context, and finds a contemporary echo of the papyrus, in a Christian context, in the synonymic equivalences cited by Jerome, in his biblical commentary on the Book of Hosea, for the expression quasi columba seducta (‘like a seduced dove’, easy to catch):Footnote 18
Hier. in Os. 7.11 p. 878c ab Aquila et Symmacho θελγομένη uel ἀπατωμένη dicitur, id est lactata siue decepta, ‘in Aquila and Symmachus [Lat. seducta] corresponds to [Gr.] thelgomenē [“charmed”] or apatōmenē [“deceived”], i.e. lactata [“seduced”] or decepta [“deceived”]’.
1.4 Written Text and Oralisation
1.4.1 Scriptio Continua: Distinctio and Ordo
The text of the papyri being written in scriptio continua, the major problems posed by their reading are, on the one hand, those of segmentation into distinctive units (littera, syllaba, partes orationis) and their succession (ordo), for a good understanding of the text, and on the other hand, the question of the oralisation of this text with a view to lectio. It is above all a question of avoiding the pitfalls linked to homophony, against which grammarians and orthographers warn. As a result of the assimilation of the voiced stop consonants in an imploding position in front of a voiced consonant, the coordinating conjunction atque is very frequently confused in manuscripts and inscriptions with the sequence adque, consisting of the preposition ad and the enclitic -que).Footnote 19 The Oxyrhynchus papyrus presents in three places the spelling adque instead of atque, as the context requires:
P.Oxy. XXIV 2401, vv. 607, 614, 627 (and 977?) adque pap. / atque legendum
Velius Longus apud Cassiod. gramm. VII 154.17 atque coniunctio composita per t … scribenda, ‘The compound conjunction atque must be written with a t.’
The diacritical marks assigned to this word identification are, as always, randomly distributed and functionally versatile. The few horizontal lines that appear above vowels or syllables in P.Oxy. XXIV 2401 may correspond to quantity signs: they are actually drawn above long vowels. But why have these forms been marked, and not the others? They seem to us to have in common that they draw attention to typical forms and facts of language which are the subject of traditional developments by grammarians, such as spē (v. 648) and uxōrem (v. 976), given as models of declensions:
Char. 69.12 (Barwick Reference Barwick19642) si producta ē littera finiatur ablativus, i accepta faciat genetivum … spē, spei, ‘If the ablative has a long -ē ending, it forms its genitive by adding an -i: spē, spei [“hope”]’;
Ars Bob. 18.29 (De Nonno Reference De Nonno1982) In -or […] uxor, soror, arbor […] genetivo o litteram producunt, velut … uxor, uxōris, ‘Nouns in -or, uxor [“wife”], soror [“sister”], arbor [“tree”], have a long ō in the genitive, such as uxor, uxōris.’
They also flag up the inflectional ambiguity of the forms mē (v. 609) and aurēs (v. 939), or, in impru]dēns (v. 643), warn against the spelling confusion that can result from the homophony of [ēs] endings, corresponding to graphic endings in -ēs (Herculēs) or -ē(n)s (impudens), consecutive to the fall of the implosive consonant n before s:
[Prob.] Appendix 20.19 (Asperti and Passalacqua Reference Asperti and Passalacqua2014) Hercules non Herculens, ‘It should be written Hercules, non Herculens.’
A major difficulty in the process of deciphering the text and segmenting the utterance into distinctive units is the analysis into words and the identification of their boundaries. Acute accent and diastole can be used to individualise words, playing a demarcation function. In the Oxyrhynchus papyrus, the graphic accents placed on forms such as the verb géssero (v. 641), the interjection ó in ó, scelus‧ (v. 665), or the participle complácita (v. 645) do correspond to the tonic accents of these words, but they are very few in number. The compound forms are particularly prone to variation: multis modis hesitates between the nominal syntagm and the univerbed adverb multimodis (as omnibus modis/omnimodis); post, hac are separated by a diastole in the papyrus, but univerbed in the grammatical tradition, while the inscriptions attest both forms, post hac and posthac (‘from now on, henceforth’):Footnote 20
P.Oxy. XXIV 2401, v. 939 mul]tis mod[i]s: pap., Σ / multimodis: A, edd. [multi’modis]
P.Oxy. XXIV 2401, v. 611 post, hac,
Diom. gramm. 1.406.1 adverbiis addi praepositiones … lectum inuenimus … posthac, antehac, ‘As prepositions preceding adverbs, we found in the texts posthac, antehac.’
Dosith. 40.70 (Bonnet Reference Bonnet2005) postea, post, posthac (glossed in Greek as μετὰ ταῦτα).
The diastole within the word hodie is surprising, but it may serve to draw attention to the classical pronunciation of the word, [ho-di-e] or [ho-dye], with a voiced dental consonant [d], as opposed to the late assibilated pronunciation [(d)z], which the grammarians point out ([ho-zie]), and which is confirmed by inscriptions.Footnote 21
P.Oxy. XXIV 2401, v. 620 hod,ie (hod,i,e)Footnote 22
Prisc. gramm. 2.24.5–6 saepe d scribentes Latini hanc exprimunt sono [scil. Z], ut meridies, hodie, ‘It is common for Latins to use the spelling d to express the sound of z, as in meridies [“noon”], hodie [“today”] [pronounced meriz(i)es and hoz(i)e]’.
The high point (punctus) often corresponds, in both papyri, to the end of a sentence, verse, or line, and to semantic sequences, but this is not always the case. Donatus is much more explicit when he suggests the introduction of strong (distinguere) or weak (subdistinguere) pauses, and the effects of meaning they induce, specifying moreover the nature of the intonative curves – assertive, interrogative, or exclamatory – that accompany these breathing pauses in the oralisation of the text.Footnote 23 When one is confronted with the materiality of the reading of papyri (and manuscripts in general), one understands even better the usefulness of the commentaries that quickly accompanied the editions of ancient texts, in helping to understand them …
P.Oxy. XXIV 2401, v. 926 Itane vero o[b]turbat‧ (‘that’s how he interrupts me’).
Don. Ter. Andr. 926.1 Si subdistinguit, accipe ‘insterstrepit’; sin distinguit, ‘evertit’ intellegas, ‘If it is a weak punctuation, the meaning is “he interrupts me”; but if it is a strong punctuation, the meaning is “he makes trouble”’ – 926.2 Potest itane uero subdistingui, sic cum comminatione inferri obturbat, ‘You can put a weak punctuation after itane uero [“ah yes really”]; you must then pronounce obturbat in a threatening tone.’
1.4.2 Occultation of Melody and Rhythm
From the end of the Republic, Terence’s plays were no longer performed onstage. They quickly became reference texts, studied in the school curriculum. Their musical dimension, which required, as Donatus points out, a composer, an instrumentalist, and a singing actor, has left no trace in the manuscript tradition.Footnote 24 The division between spoken and sung parts (diverbia and cantica), indicated in some manuscripts by the abbreviations DV, C, and MMC,Footnote 25 does not exist in our papyri, and colometry is seriously corrupted, especially in the Oxyrhynchus papyrus.Footnote 26 The ends of verses do not necessarily coincide with the ends of lines, and long verses are not distinguished from short verses by ekthesis and eisthesis. In a word, Terence’s verses are treated as prose (ad imaginem prosae orationis).Footnote 27
1.4.3 Occultation of Prosody
This loss of metrics leads to an occultation of prosody and of all the phonetic processes of vowel reduction and compression (vocalium collisio) which characterise the flow of the spoken language, and which are studied by grammarians and rhetors (hiatus, aphaeresis, apocope, apostrophe, synizesis) and by commentators.Footnote 28 In Donatus’ commentary, we do not find the criterion of metrical constraint (metri causa, metro cogente), often mentioned by Servius in his commentary on Virgil, for the choice of variants. In order to avoid any difficulty of understanding, the two papyri write the words in their entirety, inserting diastoles between the forms in hiatus:
P.Vindob. inv. L 103, v. 546 si, in, re, est, / rem est, Don.; remst, codd. edd.
and systematically avoiding apheresis of the verb ‘to be’:
P.Oxy. XXIV 2401 v. 643 inventa est. v. 647 falsus es … visum est,
where modern editions generalise the contracted forms inventast, falsu’s, visumst, which also poses the problem of restoring the missing parts.Footnote 29 For the interrogative-exclamatory particle -ne, often reduced to the final nasal -n, the situation is more complex; unlike in the rest of the tradition, it is, in the two papyri, not even attested in many places:
P.Vindob. inv. L 103, v. 492 itane (itane / itan, codd., Don.;Footnote 30 itan, edd.) – 581 audin (codd., edd.)
v. 497 credo tibi (credon, codd., edd.)
P.Oxy. XXIV 2401, v. 625 hocine est (hoccinest, codd., edd.) – 616 uiden me (codd., edd.)
612 negabo uelle (negabon, codd., edd.) – 943 ego huius (egon A / egone Σ)
647 non tibi (nonne, codd., edd.) – 629 id est (idnest, codd., edd.)
1.5 Standard and Standardisation: Use and Authority
It is thus a standardised, banalised text of Terence, for didactic use, that the two papyri give us. The metrical and prosodic dimension is occulted. The developed, full forms, such as ipsa est, consiliis, peri`i´, mihi, nihil, are privileged, as opposed to the reduced, short forms (ipsast, consilis, peri, mi, nil, codd., edd.) of the spoken language, and one may wonder what was happening to them during the transition to oralisation that the exercise of lectio entails. The archaic spellings and forms are also, and especially in the Oxyrhynchus papyrus, replaced by forms more in line with the classical norm, such as cuius (541) and cuiquam (626), replacing quoius and quoiquam; Davus (-um), tuus (-um), salvus (-vos,-vom, codd., edd.); verser (649) instead of vorser; impudentissima (634) instead of -issuma; acc. pl. aures (933), non in -is, or the genitive nulli (codd., edd.) corrected (or explained?) to nullius:
P.Oxy. XXIV 2401, v. 608 tam nulli{i}us consilii sum
Nulli m1, nullius m2, per adiectione (supra lineam) emendatum
Prisc. gramm. 2.226.16–19 vetustissimi solebant omnium in -ius terminantium genetivum et in -i dativum etiam, in -i genetivum et in -o dativum in genere masculino et neutro, in feminino vero secundum primam declinationem in -ae diphtongum proferre, ‘All words ending in the genitive with -ius and in the dative with -i [in the three genders], the archaic authors declined them in the masculine and neuter genders with a genitive in -i and a dative in -o, and in the feminine with a diphthong in -ae, on the model of the first declension.’
Terence is classified by grammarians among the vetustissimi (archaic authors), and his language is not so much studied for itself, as for the deviations it presents from the norms established later, which define usage, and which feed linguistic debates. The Oxyrhynchus papyrus seems to erase this archaic dimension, in order to make the text more in line with the classical standard taught to a contemporary audience.
The question of consonantal assimilation or non-assimilation of prefixes is a recurrent spelling problem in ancient and modern editions, especially for prefixed verbs. The Oxyrhynchus papyrus transmits in v. 656 an assimilated form apparabantur, whereas it is a non-assimilated form that it presents for adplicat in v. 924. Precisely for this verb, Servius distinguishes the assimilated ancient spelling, with a geminate consonant (applico), from the common usage (praesentem usum) with etymological spelling (adplico); but he also reports a ‘euphonic’ form aplico:
P.Oxy. XXIV 2401, v. 924 adplicat
Serv. Aen. 1.616 adplicat secundum praesentem usum, per d prima syllaba scribitur; secundum antiquam orthographiam, quae praepositionum ultimam litteram in vicinam mutabat, per p; secundum euphoniam, per a tantum, ‘In contemporary usage, it is written with a d in the first syllable [adplicat]; in the old spelling, which equated the last letter of the preposition [scil. prefix] with the next, it was written with a p [applicat]; for reasons of euphony, it is only written with a [aplicat].’
This form aplicat is echoed in the Latin–derived legal Greek ἀπλικιτάριος, glossed as adplico, in a papyrus (P.Cair.Masp. III 67287) that presents a bilingual lexicon of names of taxes and professions.
Linguistic reality is much more complex than the rules of analogy and usage would have us believe. This is what the pseudo-Probus notes about the gender ambiguity of the word scrupulus, which is neuter in usage (consuetudo) but masculine in the ‘good’ authors, Cicero and Terence (erudita, auctoritatem).
P.Oxy. XXIV 2401, v. 940 [at miFootnote 31 unus scru]pulus etiam restat
[Prob.] De nomine 69.22–70.1 (Passalacqua Reference Passalacqua1984) scrupulus hic an et hoc scrupulum? Consuetudo quidem hoc scrupulum, erudita uero hunc scrupulum dixit. Cicero, Pro Sexto Roscio ‘hunc … scrupulum’ […]. Post cuius auctoritatem nemo dubitauerit hunc scrupulum dicere et quod in eius modi quaestiunculis nihil potest analogia. Nam Terentius: ‘at mihi unus scrupulus etiam restat’, ‘Should one say scrupulus [masc.] or scrupulum [neut.]? In usage it is a neuter, scrupulum, but scholarly tradition has used it in the masculine, scrupulus [there follows a quotation from Cic. S. Rosc. 6]. Given the authority he [scil. Cicero] enjoys, no one questioned that it should be scrupulus [masc.] and that in such trifles the analogy did not apply. Indeed Terence [Andr. 940] says scrupulus.’
The analogical rule is that any diminutive retains the gender of the word from which it is derived. Scrupulus (‘little stone’) being derived from scrupus (‘stone’), as Donatus reminds us (Andr. 940.2 scrupulus a scrupo, lapide), it is therefore, like it, masculine, and the neuter is therefore not correct.
The two papyri of Vienna and Oxyrhynchus have only provided us with 162 verses from the whole of Terence’s theatrical output; but this modest papyrological corpus proves to be rich in information, on Terence’s text as well as on the context in which it was used, and it has its rightful place in the history of the grammatical tradition and in the history of the Latin language, just as it does in the exegetical tradition.
