In these notes it will be argued that our text of Cicero's Philippics may be improved in nearly a score passages by printing manuscript readings that editors have repeatedly spurned.Footnote ¹ If my detailed arguments are accepted, this article will have also a wider import, serving as a reminder that when a manuscript tradition is bifid both branches should be taken seriously, even if one branch is in general more corrupt than the other, and showing how modern technology may be harnassed as a powerful aid in the study of Cicero's idiom.

If late hybrids are discounted, the dozens of surviving manuscripts of the Philippics may be divided into two classes. In one stands V = Vatican, Archivio S. Pietro H 25, s. ix, in more or less glorious isolation.Footnote ² The other, whose earliest member dates from s. x, is known as D (for decurtati); the siglum may serve also for the lost parent of the class. All readings of D to be discussed in this article can be ascribed without difficulty to this parent, and the identity and individual readings of its best descendants need not concern us.Footnote ³ Shared corruptions prove that V and D derive from a common lost ancestor.Footnote ⁴ V stops at 13.10, and before that point there are passages for which one or other of V or D is defective, but for most of 1.1–13.10 the broad outline of the method that editors should adopt is uncontroversial: they should print either V's reading or D's or, in those passages where their common ancestor was corrupt, a conjecture.

Of the general characteristics of V and D there has been agreement for long among editors. V, although containing much nonsense, is an ‘honest’ MS, in that its text contains very few readings that derive from scribal attempts to remove inherited corruption;Footnote ⁵ D, by contrast, was ‘dishonest’, often presenting readings that are the result of late antique or mediaeval scribes' attempts to correct or improve the text.Footnote ⁶ Lest my defence of some spurned readings of D should suggest naivety, I cite here three simple examples of D's dishonesty. At 1.21 editors print de ui et maiestatis damnati, maiestatis being Halm's conjecture for maiestates in V and de maiestate in D. Although D's reading makes sense, it offers an expression (de maiestate) not attested in Cicero. By contrast, Halm's conjecture offers well paralleled standard idiom (cf. 1.23 legibus … quae iubent ei, qui de ui, itemque ei, qui maiestatis damnatus sit, aqua et igni interdici, to look no further) and accounts for the readings of both V and D: maiestates is a slight corruption which must have stood also in the archetype; de maiestate is D's faulty attempt to correct maiestates. At 1.34 editors print Vtinam, M. Antoni, auum tuum meminisses! de quo tamen audisti multa ex me eaque saepissime: eaque is the easy and stylish conjecture of Faernus; V has aquae; and D omits the word (presumably finding aquae in its parent but being unable to make sense of it).Footnote ⁷ At 2.21, where C. wonders what people would have thought if Antony had killed Clodius on the occasion of his chasing him: Quidnam homines putarent, si tum occisus esset, cum tu illum in foro inspectante populo Romano gladio insecutus es negotiumque transegisses, nisi se ille in scalas tabernae librariae coniecisset … For gladio (V, Schol. Bob. ad Mil. 40 [p.123 Stangl]) D had gladio stricto; this makes good sense but could be right only if V and the Scholiast are united in coincidental omission, or if the Scholiast took his text from an ancestor of V. It is easier to believe that D made a banalizing conjecture.

Although V and D each present many corruptions, their common ancestor seems to have presented a text in good condition, and therefore, when both are present, modern editors have been able to print relatively few conjectures. Usually it is obvious which of V or D is right, and editors agree on the majority of choices that have to be made; and, despite the fact that D has been subject to interpolation, all, including Giuseppina Magnaldi, rightly adopt a very substantial number of readings from it. However, there remains a significant minority of passages in which the readings of both V and D make good sense, and for many of these it is hard to produce decisive arguments in favour of one or the other. In consequence there is some variation in the readings which editors adopt from each of V and D. Shackleton Bailey was slightly more hospitable than his predecessors to the readings of D. More recently Magnaldi, whose knowledge of the textual problems and MSS of the Philippics and the bibliography pertaining to them is more profound than that of any previous scholar, has argued that D is far more dishonest than editors have realized, that this dishonesty stems from the attempts of late antique grammatici to emend the text, and that in consequence many more readings of V should be adopted.Footnote ⁸ A full analysis of Magnaldi's contentions would involve much ad feminam argument and would do little more than restore a previously accepted consensus; it suffices here to say that she overstates her case and exalts V to an undeserved pinnacle in which it often seems to be improbably immune to the corruptions that beset even the most honest of MSS.Footnote ⁹ It is to be hoped that subsequent editors will restore many of the readings of D that she has ejected.

In this article I shall champion readings of D which have been unjustifiably rejected by all or many recent editors, and not just Magnaldi:Footnote ¹⁰ in some of the passages to be discussed V presents readings of unacceptable Latinity; in others, although V offers Latin that is in itself acceptable, D offers better Ciceronian idiom; and in others D offers readings that are for various reasons attractive and unlikely to have been produced by corruption. For some of the passages to be discussed I adduce evidence that it has been easy to find only in recent years when computers have been able to search through Latin literature. Indeed, most of my parallels have been gleaned from such searches: any merit which this article may have in establishing a few new facts about Ciceronian idiom comes not from hard work in reading Latin in a search for parallels but from being open-minded towards the possibility that D may have been right where it has generally been thought wrong and from being willing to embrace modern technology to establish its correctness.

1.26–7. Cicero says that some of his fears relate to the future. He requests Dolabella, once his son-in-law, not to be angry with him:

In the first crux, if D and not V transmits what C. wrote, then ego in this position in the sentence must carry some emphasis. Often an emphatic pronoun takes its emphasis from an explicit or implicit contrast. Here the implicit contrast is made clear by uobis and uos in the following two sentences: Cicero champions the future of the Republic (pro re publica dicenti, in which dicenti looks back to dico), which Dolabella and Antony threaten; he advises against laws, which it is still open to Dolabella and Antony not to force on the people.

V's coupling atque haec is (it need hardly be said) found scores of times in C., but D's reading is paralleled at Div. Caec. 33 atque ego haec quae in medio posita sunt commemoro; sunt alia magis occulta furta, quae ille, ut istius credo animos atque impetus retardaret, benignissime cum quaestore suo communicauit and Q Fr. 1.2.9 atque ego haec tam esse quam audio non puto. In addition, he has atque ego … dico at 2Verr. 3.79 atque ego nunc eum frumenti numerum et eas pecunias publice decumanis ab Herbitensibus datas esse dico and Sest. 8 atque ego de Antonio nihil dico praeter unum, numquam illum illo summo timore ac periculo ciuitatis neque communem metum omnium nec propriam non nullorum de ipso suspicionem aut infitiando tollere aut dissimulando sedare uoluisse; he uses ego with dico over a score of times in the speeches; and he has the juxtaposition ego haec in a dozen other passages.Footnote ¹¹

A champion of V may point out truly that ego is not entirely necessary here, but the same is true of Sest. 8; there is no good reason for D to have interpolated;Footnote ¹² and since D's reading corresponds to C.'s style it should be printed.

In the second crux there is nothing obviously wrong with V's reading, but D's is supported by passages in which C. uses enim after noui and cognoui. Particularly comparable to ours are the following in which he states that he knows someone's character: Att. 2.24.1 ac ne sis perturbatus; noui enim te et non ignoro quam sit amor omnis sollicitus atque anxius; Fam. 1.9.22 His ego consiliis, si te praesentem habuissem, ut opinio mea fert, essem usus isdem; noui enim temperantiam et moderationem naturae tuae, noui animum cum mihi amicissimum …; 5.11.1 Grata tibi mea esse officia non miror; cognoui enim te gratissimum omnium; Lig. 12 ‘Ego uero istud non postulo’ inquies. ita me hercule existimo, Tubero. noui enim te, noui patrem …; Sen. 1 licet enim mihi uersibus eisdem adfari te, Attice, quibus adfatur Flamininum ‘ille uir haud magna cum re, sed plenus fidei’; quamquam certo scio non ut Flamininum ‘sollicitari te Tite sic noctesque diesque’, noui enim moderationem animi tui et aequitatem; note too Sen. Dial. 12.14.2 Prior pars mihi leuiter perstringenda est; noui enim animum tuum nihil in suis praeter ipsos amantem. Less close but still parallel are Att. 5.5.1, Fam. 13.1.4, 14.5.1. In late antiquity Ambrose and Augustine often used (cog)noui enim. That D had interpolated a late antique manner of speech, or that it had interpolated enim to clarify the parenthesis,Footnote ¹³ is less likely than that it faithfully preserved a characteristic manner of Ciceronian utterance. If so, V's omission of enim is of a kind that is extremely common in MSS of all Latin authors.Footnote ¹⁴

In the third crux Reeve (n. 1), 30 has pointed out that D's reading, adopted by e.g. Shackleton Bailey and Ramsey, ought once again to be right: it gives a clausula consisting of a cretic + trochee, whereas V offers an unwelcome triple trochee.

2.19. C. contrasts his own use of force against Catiline with Antony's surrounding of the senate:

D's word order may be supported by the following arguments. First, in a sentence in which dementius shows that the primary contrast is between ipse and alteri (there is a secondary contrast between perniciosa and salutaria), it is preferable to place ipse rather than rei publicae in the emphatic position after cum. Second, cum is followed by a part of ipse on about 179 occasions in the Ciceronian corpus (with the collocation cum ipse itself accounting for about 80 instances, including Phil. 2.42).

Analogies for the word order given by D may be found at 2Verr. 5.106 fit clamor et admiratio populi tantam esse in homine impudentiam atque audaciam ut aut aliis causam calamitatis attribueret quae omnis propter auaritiam ipsius accidisset, aut cum ipse praedonum socius putaretur, aliis proditionis crimen inferret; Prov. cons. 27 sum Cn. Pompei uirtutem et animi magnitudinem admiratus, quod cum ipse ceteris omnibus esset omni honore antelatus, ampliorem honorem alteri tribuebat quam ipse erat consecutus; Fam. 5.12.8 accedit etiam ut minor sit fides, minor auctoritas, multi denique reprehendant et dicant uerecundiores esse praecones ludorum gymnicorum, qui, cum ceteris coronas imposuerint uictoribus eorumque nomina magna uoce pronuntiarint, cum ipsi ante ludorum missionem corona donentur, alium praeconem adhibeant, ne sua uoce se ipsi uictores esse praedicent; Lig. 7 apud quem igitur hoc dico? nempe apud eum … qui me cum ipse imperator in toto imperio populi Romani unus esset, esse alterum passus est.

Unless future editors can adduce good parallels or analogies for V's reading, they should adopt that of D.

2.26. Cicero states that, because of their family history, the two Bruti would hardly have needed his advice and urging:

At first glance V's hi igitur his seems self-evidently right: to emphasize the inappropriateness of the Bruti seeking advice from himself, C. repeats forms of hic in polyptoton. However, when one asks how common is the collocation hi igitur in extant Latin and then finds that it is unattested elsewhere before Christian texts of the third century a.d. onwards, and when one finds that the same is true for hae igitur,Footnote ¹⁵ that ii igitur is first attested even later,Footnote ¹⁶ that eae igitur is not attested at all, one concludes that the awkwardness of pronouncing such collocations led to their avoidance in classical Latin and that V's reading imports a gross solecism. By contrast, D's reading is quite acceptable: his igitur is found hundreds of times in classical Latin and more than a score of times in C. (first at Inv. rhet. 2.158), and on close inspection one sees that there is no need to emphasize the Bruti as well as their ancestors. If D's reading was corrupted to hi[s] igitur, then maioribus would need qualification and progressive corruption could easily have occurred.

I am less inclined to disagree with recent editors in their rejection of D's orti. If V is right to omit it, then his maioribus is to be explained as combining elements of both ablative of description and ablative of attendant circumstances; if D was right to have it, then his maioribus is a separative ablative used to signify origin. The best argument in defence of D's reading is that maioribus ortus may be paralleled at Off. 1.116, Hor. Sat. 1.5.55, 1.6.10, Sen. Ben. 4.30.4, Stat. Silv. 1.2.71, Tac. Ann. 12.37.1, 14.35.1, Mar. Vict. Rhet. 1.24 (Halm, Rhet. min. p. 216.36). However, the hexametric rhythm that orti gives to the close of the colon tells against it (at Off. 1.116 C. used the inoffensive order orti maioribus), and it could be a banalizing intrusion: Magnaldi, Parola 192 compares the presence of dici at 2.19, accepisti at 2.74 and uolumus at 12.12, in all of which she argues that D has added a verb;Footnote ¹⁷ we may also compare the addition of stricto at 2.21 (discussed above).

2.75. Pompey's children fight against Caesar in Spain in order to recover their country:

Faernus' conjecture is a very slight adaptation of V's reading, but it is D's reading that is probably right. That di penates is standard Latin needs no illustration. That di could be qualified by both penates and patrii in asyndeton (as with D's reading) is shown by 2Verr. 4.17 Tametsi lex est de pecuniis repetundis, ille se negat pecuniam repetere quam ereptam non tanto opere desiderat: sacra se maiorum suorum repetere abs te dicit, deos penatis te patrios reposcit. Very closely analogous are the following passages, all of which exhibit a connecting particle rather than asyndeton and some of which have a different word order: Sull. 86 Quam ob rem uos, di patrii ac penates, qui huic urbi atque huic rei publicae praesidetis …; Har. resp. 37 haruspices haec loquuntur an patrii penatesque di?; Sest. 45 uos, penates patriique dei; Nep. Them. 7.4 hos postquam Athenas peruenisse ratus est, ad magistratum senatumque Lacedaemoniorum adiit et apud eos liberrime professus est: Atheniensis suo consilio, quod communi iure gentium facere possent, deos publicos suosque patrios ac penates, quo facilius ab hoste possent defendere, muris saepsisse neque in eo quod inutile esset Graeciae fecisse (understand deos with suos); and Liv. 1.47.4 di te penates patriique et patris imago et domus regia et in domo regale solium et nomen Tarquinium creat uocatque regem.

So common is the expression di patrii that it is hard to see that D would have felt any need to interpolate. Magnaldi, Parola 161 wonders whether D's reading reflects a school exercise in which the question ‘What were the di patrii called?’ was asked;Footnote ¹⁸ but it is not clear why D should have been concerned with such a question (let alone with recording the answer), and it is much easier to argue that V has an omission.

3.22. C. has been listing some remarks of Antony which, he claims, display his oratorical incompetence. He then comments sarcastically on the reward of land given by Antony to his teacher of rhetoric:

Faernus' deletion of sit was urged by Cobet (‘Vitiosum est asyndeton sed sit expungendum’),Footnote ¹⁹ has been accepted by Boulanger–Wuilleumier, Fedeli, Shackleton Bailey and Magnaldi, and is probably right, since it makes the second part of the sentence follow more logically on the first. D's publico populi Romani has been adopted by no editor and appears to involve two difficulties: first, the expression seems tautologous; second, the presence of a further expression in the genitive (campi Leontini) straight after populi Romani seems inelegant.

The first difficulty is illusory. In formal writing publicus is often reinforced by populi Romani. The expression res publica populi Romani is found at Phil. 2.94 and often elsewhere, but note also 2Verr. 4.88 publicum populi Romani signum; Deiot. 37 publicis populi Romani litteris monumentisque; Lex agr. (CIL 1².585 = M.H. Crawford, Roman Statutes [London, 1996], no. 2) 89 uectigalia publica populi Romani; Liv. 1.32.6 ego sum publicus nuntius populi Romani; 8.9.4 pontifex publicus populi Romani; 10.7.5 sacra publica populi Romani; 23.17.2 bonaque eorum iussit publica populi Romani esse; 26.47.2 opifices ad duo milia hominum erant; eos publicos fore populi Romani edixit; 38.45.7 istud publicum populi Romani bellum et non tuum priuatum latrocinium. This reinforcement is commonly found, as in our passage, in the context of ager publicus: 2Verr. 3.13 perpaucae Siciliae ciuitates sunt bello a maioribus nostris subactae; quarum ager cum esset publicus populi Romani factus, tamen illis est redditus; Lex agr. 1 quei ager poplicus populi Romanei … fuit; 4, 5, 15 (also in many supplements and in many other passages coupled with locus [e.g. 13 quei ager locus publicus populi Romanei … fuit]); Liv. 3.71.7 agrum de quo ambigitur finium Coriolanorum fuisse captisque Coriolis iure belli publicum populi Romani factum; 3.72.6, 28.46.5, 31.4.2, 40.38.3, 44.16.7. Note also Cic. Leg. agr. 2.38 quicquid igitur sit extra Italiam quod publicum populi Romani factum sit L. Sulla Q. Pompeio consulibus aut postea, id decemuiros iubet uendere; 2.39 primum hoc quaero, ecqui tandem locus usquam sit quem non possint decemuiri dicere publicum populi Romani esse factum; 2.40 regnum Bithyniae, quod certe publicum est populi Romani factum. Magnaldi, Parola 161 dismisses populi Romani as the answer to a grammarian's question (‘To whom does the ager publicus belong?’); we have seen that it is standard Latin idiom.Footnote ²⁰

The second difficulty is real but eased if one takes campi Leontini with duo milia iugerum, as at C.'s earlier reference to this land (Phil. 2.43 duo milia iugerum campi Leontini Sex. Clodio rhetori adsignasti). If C. did write D's reading he would have intended possideat in agro publico populi Romani to make a colon, and campi Leontini probably to make a colon of its own but perhaps to be taken in one breath with duo milia iugerum immunia.

Although D shows evidence elsewhere of being an interpolated text, there is no obvious reason why a late antique or early mediaeval scribe should either have thought that publico needed explication or been familar with the expression ager publicus populi Romani; it is much easier to believe that its reading is true.

3.25. Cicero explains his commendation of the tribunes for providing the senate with an armed guard:

Here the reign of V's reading has not been uninterrupted: Fedeli prints his own conjecture, which includes an important element of D's reading, and he has been followed by Shackleton Bailey; Hine championed D's reading,Footnote ²¹ and he has been followed by Manuwald and by Manuwald and Ramsey in their revision of Shackleton Bailey. Hine, Manuwald and Ramsey are probably right; but Magnaldi, Parola 144–5 and in her edition has returned to V, and the arguments in favour of D therefore need to be set out more precisely.

First, carere metu et periculo makes excellent sense. Although ‘How are we able to lack fear and danger’ sounds stilted to the modern English ear (which prefers the more positive ‘feel safe’), mention of fear suits the context, and the Romans were happy enough to use expressions of this kind: Mil. 5 Quid enim nobis duobus iudices laboriosius, quid magis sollicitum, magis exercitum dici aut fingi potest, qui spe amplissimorum praemiorum ad rem publicam adducti, metu crudelissimorum suppliciorum carere non possumus?; Ov. Ars am. 3.92; Sen. Phaedr. 243; Quint. Decl. min. 271.12.Footnote ²² Second, the coupling of metus and periculum is common both generally in Latin and in C. in particular (e.g. 2Verr. 1.39, 5.14, 5.80, 5.137; Clu. 104; Leg. agr. 2.45; Mur. 52; Sest. 35; Div. 2.58; Phil. 7.27). Third, D's reading is to be preferred to V's because metu et periculo provides a better balance to cupiditate et audacia than just periculo alone: as Dr Berry kindly points out, metus (fear of the loss of one's property) is produced by cupiditas (the greed of other people), whereas periculum (physical danger) is produced by audacia (reckless violence).Footnote ²³ Fourth, it is more likely that metu et was omitted in error by V or an ancestor than that it was deliberately interpolated into D. Magnaldi thinks that ‘metu et periculo è una tipica locuzione ciceroniana qualcuno può averla memorizzata e annotata qui’; Footnote ²⁴ but this kind of argument, on which Magnaldi has often to fall back, reads like special pleading in the face of standard Ciceronian idiom, and why should anyone have felt that periculo needed further explication?

Hine stated that, compared to Fedeli's conjecture, ‘D's word order gives a much commoner rhythm’. It offers a double cretic at the end of the colon, Fedeli's conjecture the less favoured cretic + iambus.Footnote ²⁵ Fedeli's conjecture has the advantage of accounting for the readings of both V and D but the disadvantage of postulating corruptions in both; since D's reading is quite acceptable, it is safer to print it and postulate a double corruption in V.

9.4. C. discusses the placing of statues on the rostra:

V's reading imputes an expression to C. that is certainly extraordinary and probably arrogant. Roman politeness demanded a reference to the collective memory of the audience or readership and not a limiting first person singular. Indeed, the expression ad meam memoriam seems to be found nowhere else in ancient Latin, while ad memoriam meam is found only three times, and then in late writers, and in the quite different context of bringing something to one's mind.Footnote ²⁶ By contrast, D's ad nostram memoriam is attested in a score of passages, including three from C.: Inv. rhet. 2.1 is (i.e. Zeuxis) et ceteras conplures tabulas pinxit, quarum nonnulla pars usque ad nostram memoriam propter fani religionem remansit; Leg. Man. 54 Rhodiorum, quorum usque ad nostram memoriam disciplina naualis et gloria remansit; ibid. populus Romanus, cuius usque ad nostram memoriam nomen inuictum in naualibus pugnis permanserit; Leg. 2.56 Eodemque ritu in eo sepulcro quod haud procul a Fontis ara est regem nostrum Numam conditum accepimus gentemque Corneliam usque ad memoriam nostram hac sepultura scimus esse usam; Sall. Iug. 114.2 illique et inde usque ad nostram memoriam Romani sic habuere, alia omnia uirtuti suae prona esse, cum Gallis pro salute, non pro gloria certare; Nep. Them. 10.3 Huius ad nostram memoriam monumenta manserunt duo: sepulcrum prope oppidum, in quo est sepultus, statua in foro Magnesiae; Liv. 6.38.13 et quod usque ad memoriam nostram tribuniciis consularibusque certatum uiribus est, dictaturae semper altius fastigium fuit; Plin. HN 30.13, 35.19; Tac. Germ. 42.2; Prol. ad Paul. epist. (PL 68 col. 414.2), Mar. Vict. Rhet. 1.3 (twice), 2.1 (Halm, Rhet. min. pp. 166.33–4, 167.39–40, 258.23); Censorinus, DN 20.11; Aur. Vict. Caes. 37.5; SHA 17.24.6, Sulp. Sev. Chron. praef. 1; Eutrop. praef. 1.6.1; and Rufin. Euseb. Hist. eccl. 7 capit. (p. 633.16 Schwartz–Mommsen).

The best defence of meam would be that C. was by now so old that he might wish to stress what he, but few others, could remember. Such a defence would hardly counterbalance the lack of parallels. Nor does the use of mea with memoria in other contexts and cases offer any real support for V's reading: the contexts of De or. 2.122, Fam. 2.1.2, 5.8.2, Amic. 104, Att. 13.31.4 and Phil. 13.11 are entirely different; at Sest. 36 and Phil. 7.6 C. does use mea memoria, where he could perhaps have used nostra memoria, but there is an appropriate emphasis on his personal recollection (the same could be said of Varro, Rust. 2.1.14); the best support would be Ad Brut. 1.15.10 nullum enim bellum ciuile fuit in nostra re publica omnium quae memoria mea fuerunt, in quo bello non, utracumque pars uicisset, tamen aliqua forma esset futura rei publicae: hoc bello uictores quam rem publicam simus habituri non facile adfirmarim, uictis certe nulla umquam erit, but again C.'s personal recollection is emphasized, and the personal nature of the letter makes the context different from a speech or treatise. The Elder Pliny regularly uses nostra aetas (in various cases) but at HN 12.57 varies it with mea aetate and at Epist. 2.14.3 the Younger Pliny writes At hercule ante memoriam meam (ita maiores natu solent dicere) ne nobilissimis quidem adulescentibus locus erat nisi aliquo consulari producente (sc. in the centumviral court; the parenthesis refers not to Pliny's expression but explains how he knows what happened before his own adolescence), but these passages hardly compensate for the lack of parallels for ad meam memoriam. C. and others qualify memoria with nostra so often that illustration is not needed.

Since words are often corrupted into synonyms or near synonyms, the corruption in V is easily explained.Footnote ²⁷

9.10–11. C. praises Ser. Sulpicius Rufus:

This problem is more complex than it may appear. At first glance V's ita ea quae seems attractive, since it allows such a translation as: ‘For indeed he was as expert in dispensing justice as he was on points of law. Accordingly such matters which took their starting point from the statutes or matters of civil law he used always to judge by the criteria of ease of management and equity …’. However, those living in the digital age are able to discover that ita ea quae is a collocation not found in Latin before Augustine and that even the simpler ita ea is attested before Augustine only at Vitr. 6.5.1 and Gai. Inst. 3.179. This discovery makes it hard to believe that V's reading is right.

D's itaque quae may seem a little harder, since itaque (‘therefore’) may seem too emphatic. Nevertheless, it may be defended as an example of itaque's being used to illustrate a general proposition:Footnote ²⁸ the general proposition is introduced in the sentence beginning nec enim, and the sentence introduced by itaque gives two illustrations of this. And D's reading is much easier to parallel. The following passage is precisely parallel to ours, in that itaque comes first in its sentence, quae is neuter nominative plural, and the antecedent ea (on which it depends) is omitted but would have been in the accusative: Vitr. 9.7.7 itaque quae nobis tradita sunt et a quibus sint inuenta, dicam. Others are very similar, but the omitted ea would have been in the nominative: Off. 1.143 sed potest eadem esse prudentiae definitio de qua principio diximus; hoc autem loco de moderatione et temperantia et harum similibus uirtutibus quaerimus. itaque quae erant prudentiae propria suo loco dicta sunt; quae autem harum uirtutum de quibus iam diu loquimur, quae pertinent ad uerecundiam et ad eorum approbationem quibuscum uiuimus, nunc dicenda sunt; Varro, Rust. 3.9.9 itaque quae ante aut post nata sunt et etiam prima eo tempore, non supponenda (but animals and not abstract concepts are being discussed). Others differ from the previous category only in that quae is not nominative neuter plural: Varro, Rust. 2.2.3 itaque quae id non habent, maiores nostri apicas appellabant ac reiciebant; Vitr. 2.10.2; Sen. Ep. 108.23. In the following passages the antecedent ea is found later in the sentence: Luc. 99 itaque quae contra sensus contraque perspicuitatem dicantur ea pertinere ad superiorem diuisionem, contra posteriorem nihil dici oportere; Off. 2.17 itaque quae in rebus inanimis quaeque in usu et tractatione beluarum fiunt utiliter ad hominum uitam, artibus ea tribuuntur operosis, hominum autem studia ad amplificationem nostrarum rerum prompta ac parata uirorum praestantium sapientia et uirtute excitantur; and Sen. Ep. 86.8 itaque quae concursum et admirationem habuerant cum dedicarentur, ea (Reynolds: et uel haec codd.) in antiquorum numerum reiciuntur, cum aliquid noui luxuria commenta est, quo ipsa se obrueret. At Liv. 5.10.10 ea is again found but quae is neuter accusative plural. These facts about usage suggest that, despite its slight awkwardness, D's reading should be adopted.

9.12. Cicero has been discussing the honours that should be given to Ser. Sulpicius Rufus and now mentions the devotion of his son Servius:

Manuwald, p. 1079, explains Cicero's thought at the beginning of the passage thus:

And then she explains the sentence which has the MS variants: ‘Cicero describes the situation just outlined from the opposite perspective, by turning from the effects of the son's grief on the father to the son himself: it contributes to his own reputation if he is seen to have paid his father the honour he deserved.’ All this is just. She does not discuss the reading of D but explains the text of V thus:

If V's text is to be defended, this is probably the best defence one can make. However, even if etiam is not to be taken with et quidem, it is troubling that Nat. D. 2.36 is autem est gradus in quo rerum omnium natura ponitur; quae quoniam talis est ut et praesit omnibus et eam nulla res possit inpedire, necesse est intellegentem esse mundum et quidem etiam sapientem is one of only two passages in which V's collocation et quidem etiam is attested before the Middle Ages. The other is Zeno of Verona, Tract. 2.7 (CCSL 22, p. 171) At ego non pertimescam, sermonis publici quae de me fabuletur inuidia; non enim nuptias condemno, sed nuptiis meliora praepono, et quidem etiam apostolo hortante sic Paulo: ‘dico autem innuptis et uiduis: bonum est illis, si sic perseuerauerint sicut ego; si autem non fuerint continentes, nubant: melius est enim nubere quam uri’. Were et quidem etiam sound in our passage it would have to introduce a new sentence, which is as much contrasting as additional (Manuwald's ‘from the opposite perspective’ is somewhat at odds with her ‘a reinforcement of the preceding one’); but in the other two passages there is no contrast, merely an emphatic addition.

D's equidem etiam is entirely idiomatic and is to be preferred. Kühner–Stegmann 1.805 have this to say about equidem: ‘Es unterscheidet sich von quidem nicht bloß durch die stärkere Bedeutung, sondern auch dadurch, daß es auf einen ganzen Gedanken, nicht, wie quidem, auf einen einzelnen Begriff bezogen wird’. Here, as we have seen, we have a new thought. After remarking that in several authors it is confined almost entirely to the first person, they go on to say: ‘Es wird also am häufigsten in solchen Versicherungen gebraucht, welche sich auf den Redenden beziehen, daher wird es sehr oft mit den Verben puto, credo, arbitror, duco, spero, scio, uideo, mihi uideor verbunden.’ And whereas Cicero never uses quidem to link two sentences in which the first person present arbitror is the main verb of the second,Footnote ²⁹equidem is found eight times with arbitror, with some link to the previous sentence always implicit. A good example is Tusc. 5.121 sed quoniam mane est eundum, has quinque dierum disputationes memoria comprehendamus. equidem me etiam conscripturum arbitror – ubi enim melius uti possumus hoc, cuicuimodi est, otio? –, ad Brutumque nostrum hos libros alteros quinque mittemus, a quo non modo inpulsi sumus ad philosophiae scriptiones, uerum etiam lacessiti.

When etiam appears in close proximity to equidem, it usually goes not with equidem but with a word in the sentence that needs emphasizing, normally placed after it (as conscripturum is in the example above). In striking contrast to the collocution et quidem etiam, Cicero has the juxtaposition equidem etiam often, especially in his later writing: De or. 2.362; Sen. 26; Amic. 101 Vicissim autem senes in adulescentium caritate acquiescimus, ut in uestra, ut in Q. Tuberonis; equidem etiam admodum adulescentis P. Rutili, A. Vergini familiaritate delector (this illustrates in brief compass the appearance of a new thought in the second sentence); Div. 1.80; Fin. 4.13, 5.2; Off. 1.37; Att. 14.11.1, 15.16a; Fam. 15.4.14 (perhaps also Att. 5.10.3; Fam. 4.7.4). From other authors note Plin. Ep. 6.4.3; August. Ep. 266.2; and Prisc. Inst. 12.18, Gramm. Lat. 2, p. 587.20 Keil (in which last the verb is not first person).

11.12. Cicero refers to the gifts made by Antony to Decidius Saxa and Cafo:

That V's reading is possible is at once obvious: there are many instances in Latin of contentus esse not dependent on the verb posse (for just one from Cicero, see Att. 6.1.3). However, throughout his career Cicero liked to couple contentus esse and posse; see Quinct. 12; 2Verr. 1.3, 1.40; Clu. 149; Leg. Man. 14 nam ceterarum prouinciarum uectigalia, Quirites, tanta sunt ut eis ad ipsas prouincias tutandas uix contenti esse possimus, Asia uero tam opima est ac fertilis ut et ubertate agrorum et uarietate fructuum et magnitudine pastionis et multitudine earum rerum quae exportentur facile omnibus terris antecellat; De or. 1.249, 3.146 Tum Caesar ‘unum’ inquit ‘me ex tuo sermone maxime, Crasse, commouit, quod eum negasti, qui non cito quid didicisset, umquam omnino posse perdiscere; ut mihi non sit difficile periclitari et aut statim percipere ista, quae tu uerbis ad caelum extulisti, aut, si non potuerim, tempus non perdere, cum tamen his nostris possim esse contentus’; Orat. 174; Deiot. 29; Att. 12.19.1, 12.25.1; Fin. 1.39 nam si ea sola uoluptas esset quae quasi titillaret sensus, ut ita dicam, et ad eos cum suauitate afflueret et illaberetur, nec manus esse contenta posset nec ulla pars uacuitate doloris sine iucundo motu uoluptatis, and Luc. 54. As the examples quoted in full make clear, he uses this expression even in the subjunctive mood, even though the potentiality inherent in the subjunctive might sometimes seem to make the addition of posse redundant.Footnote ³⁰

One reason why C. was willing to use posse in this way even in the subjunctive is that it sometimes gave him a better clausula (here a molossus + trochee + spondee as opposed to five heavy syllables);Footnote ³¹ the same principle underlies his use of e.g. esse uideatur, even in passages where the uideatur is largely redundant. However, D's reading is not only rhythmically preferable but also more likely to have been corrupted to essent than vice versa, both psychologically and on palaeographical grounds (essepossent > essent, with saltus oculi from one sse to the other).Footnote ³²

For a parallel corruption see Liv. 27.27.5, where the codex Puteaneus rightly reads euadere possent; ‘Spirensian’ witnesses, wrongly followed by some recent editors, have euaderent.Footnote ³³

11.13. Amidst general talk of the indebtedness of Antony's henchmen, C. makes fun of the obscure P. Decius:

Shackleton Bailey, in his textual appendix (p. 388), has argued that D's clarissimi should be adopted: ‘The sarcasm depends on the superlative, Cicero's standard way of referring to particularly distinguished senators. He does not so use uir clarus.’ Shackleton Bailey does not cite parallels, but from many possibilities, see 11.10 and Cat. 4.13.Footnote ³⁴ However, Magnaldi, Parola 252–3, in part following Fedeli, points to the use of uir clarus at 9.4, 9.11 and 13.41 and writes ‘Sebbene clari sia qui [s. 13.411] rinforzato da magni, ciò non sembra sufficiente per stampare in 11.13 il superlativo clarissimi, nato probabilmente, come in 1.3 clara per praeclara, da un'esercitazione scolastica sul grado degli aggettivi.’ Her argument has two weaknesses. First, the variants clarus and clarissimus are very likely to have a graphic origin and there is no need to invoke a scholastic exercise. Second, here (as elsewhere) she forgets that D's reading is equipollent with that of V on the stemma; although the demonstrable presence of some interpolation in D should make editors cautious, its readings do not have to be considered only when V is obviously corrupt. Here clarissimus gives both C.'s standard idiom and is rhetorically more pointed.

D's uide is rightly spurned by editors. Its suorum has suffered the same fate, and Magnaldi, Parola 166, noting that in a few other passages D seems to interpolate pronouns and possessive adjectives, specifically describes it as a characteristic banalization of this group. However, C.'s joke about Decius' devoting himself to debt depends upon knowledge of the self-sacrificial devotions made by the Decii Mures in the fourth and third centuries. It is highly improbable either that this Decius was related to these famous men or that C.'s contemporaries imagined that he was. Adoption of D's reading both sharpens C.'s irony on this point and allows a more pointed contrast with the three instances of alienus in the passage, esp. the first (in which aere alieno has a double reference: to Decius' debt and to someone else's money). The argument that D has made a banalizing insertion may be countered by the observation that suorum could easily have been omitted by V or an ancestor through homoioteleuton (maiorvm suorvm).

A decisive resolution of the final crux (in which V reads exercere) is not possible. I observe merely that D's reading, adopted only by Wuilleumier, is just conceivable if indebtedness is regarded as a kind of bondage (note Liv. 7.19.5 sorte ipsa obruebantur inopes nexumque inibant: C. could perhaps be reversing this use of inire), but the corruption in V would then be a little surprising; that V's reading is impossible; that Halm's conjecture is a characteristically intelligent attempt to make sense of V's reading, which, however, is unsupported by precise parallels; that Müller's conjecture is likewise unsupported by precise parallels and is further from the paradosis; and that exsoluere (perhaps not conjectured before) + direct object + aere alieno has precise parallels at Liv. 6.14.11 and 37.54.8.Footnote ³⁵

13.5. Cicero says that peace with Antony and his followers is impossible:

The sentence for which variants are here reported is well translated by Shackleton Bailey, who adopts V's reading: ‘If he were a foreign enemy, it would indeed be difficult after such doings, but it would be possible somehow or other.’ If one follows D, one could replace ‘indeed’ with ‘still indeed’.

Although V's reading gives good sense, D's is the lectio difficilior and the less likely to have been produced by banalization.Footnote ³⁶ More importantly, it is well paralleled. The best parallels are three further passages in which C. uses the construction si … tamen id ipsum: cf. Planc. 5 nam si tantummodo mihi necesse esset contra Laterensem dicere, tamen id ipsum esset in tanto usu nostro tantaque amicitia molestum; Tull. 5 nunc cum coactus dicam, si quid forte dicam, tamen id ipsum uerecunde modi<ce>que faciam, tantum ut, quoniam sibi me non esse inimicum potuit priore actione Fabius iudicare, nunc M. Tullio fidelem certumque amicum esse cognoscat; Att. 12.44.4 tu mihi, etiam si nihil erit quod scribas, quod fore ita uideo, tamen id ipsum scribas uelim te nihil habuisse quod scriberes, dum modo ne his uerbis. Analogous are those passages in which C. uses the construction si … tamen id but with no reinforcing ipsum: 2Verr. 2.155 qui si te publice laudarent, tamen id more potius suo quam merito tuo facere uiderentur; Clu. 126; Brut. 111; Fin. 1.2, 3.21 (note too Plaut. Mil. 747). Other passages in which tamen id ipsum is found are worth citing, since they show C.'s willingness to use this collocation: 2Verr. 4.36; Planc. 31, De or. 3.138 quid Pericles? de cuius dicendi copia sic accepimus, ut, cum contra uoluntatem Atheniensium loqueretur pro salute patriae seuerius, tamen id ipsum, quod ille contra populares homines diceret, populare omnibus et iucundum uideretur; Nat. D. 2.23 (note too that C. uses hoc tamen ipsum at Att. 11.5.2). However, I have ignored such passages as Att. 1.13.1, where the collocation is found but tamen is to be taken closely with a preceding neque or sed or uerum.

Sense should always be paramount in solving textual problems, but the evidence of an author's usage is often decisive when two readings make sense. In this article I have tried to show that new facts about Ciceronian usage may still be discovered, and that modern technology removes much of the pain involved in such discovery.