98. 3). Fronto has: tantus usus studiorum bonarumque artium communicandi (Ep. ad Aur. 1. 24); and Gellius: verborumque fingendi et novandi studium (4.15. 1) and: causarumque orandi cupiens (5. 10. 5). We read in Caesar: neque sui colligendi hostibus facultatem relinquunt (B. G. 3. 6) and: venerunt . . . sui purgandi causa (ib. 4. 13), and in Livy vestri adhortandi causa (21.41. 1), in which examples the gerundive agrees with the form, and not with the meaning of the pronoun. So the gerundive becomes in a measure indifferent to number, and from: sui colligendi facultatem hostibus reliquit we should get: hostium colligendi facultatem reliquit eorum duci. : That this is the true explanation of the idiom is confirmed by examples where it is not number, but gender, that is in question. We read in Plautus: quia tui videndi copiast (Truc. 370), where tui is feminine, but videndi agrees with it in form rather than meaning. Hence the gerundive becomes indifferent to gender at times, and we have lucis das tuendi copiam (Pl. Capt. 1008), eius (= uxoris) videndi cupidus (Ter. Hec. 372), where Donatus takes no notice of the irregularity, copia placandi sit modo parva tui (fem.) (Ov. Her. 20. 74), and: crescit enim adsidue spectandi cura puellae (Prop. 3. 21. 3). Let us turn from this rare and complicated construction to a common and simple one, the agreement of the verb with two subjects connected by κal or et. The answer seems easy to the grammarian : the verb must be in the plural. When Kipling ventured to write: 'The tumult and the shouting dies', he was at once assailed by our pedants. But we read in Homer: Τρώων δὲ κλαγγή τε καὶ ἄσπετος ὦρτο κυδοιμὸς a passage which Kipling may have had in mind. In Greek related pairs of words commonly take a singular verb, as we have already noticed in dealing with the Schema Pindaricum. So in: or in: or: or: ὅθι νητὸς χρυσὸς καὶ χαλκὸς ἔκειτο ἐσθής τ' ἐν χηλοῖσιν ἅλις τ ̓ εὐῶδες ἔλαιον (Od. 2. 338-9), πάντα τύχη καὶ μοῖρα, Περίκλεις, ἀνδρὶ δίδωσιν (Archil. Fr. 16, Β.), ἀλλ ̓ εὐγενὴς μὲν ὁ κτανών τε χὠ θανών (Soph. Phil. 336), εἰ δ ̓ Ορφέως μοι γλῶσσα καὶ μέλος παρήν (Eur. Alc. 357), or: ἔδοξε γὰρ δὴ ἡμῖν ἡ πολιτικὴ καὶ ἡ βασιλικὴ τέχνη ἡ αὐτὴ εἶναι (Plat. Euthyd. 291 c), or : οὗ δὴ καὶ ἐκφανὴς ἐγένετο ἡ τῆς πόλεως ῥώμη καὶ ἀρετή (id. Menex. 243 c), or: ἄνεμος καὶ χειμών διεκώλυσεν αὐτούς (Xen. Hell. 1. 6. 35). It is not that the pair of words describe the same object, nor that the objects they designate are not often opposed; it is that they constitute a correlated pair, such as will naturally be expressed by a dual; and hence, like the dual, they are at times constructed with a singular verb. When we turn to Latin this becomes even plainer. 'Close union often amounts to unity', says Gildersleeve. But in Latin close union very often gives us plurality, and ideas mutually opposed are constructed with a singular verb as well as with a plural. In archaic and classical Latin as a rule with a closely related pair of words the verb is in the singular, as in: tua fama et gnatae vita in dubium veniet (Ter. Ad. 340), (fabulae) novae novum intervenit vitium et calamitas (id. Hec. 2), cum tempus necessitasque postulat (Cic. Off. 1. 81. 23), societas hominum et communitas evertatur necesse est (ib. 3. 22. 5), ut summus furor atque amentia consequatur (Rosc. Am. 66. 24); but in: (fortuna), quam nemo ab inconstantia et temeritate seiunget, quae dignae certe non sunt deo (N. D. 3. 61. 24) we have the plural. This is especially the case when the members of a pair are contrasted, as in ius et iniuria natura diiudicantur (Leg. 1. 44. 16) or: nam vita et mors iura naturae sunt (Sall. Hist. 2. 50. 5, Kr.). But we have both religio et fides anteponatur amicitiae (Off. 3. 46. 10) and: ni virtus fidesque vostra spectata mihi forent (Sall. Cat. 20. 2). While Cicero and Caesar as a rule have the singular with ideas so closely related, Livy has the plural very often, and Sallust and Tacitus favour the plural. So we have: caedes ac tumultus erat in castris (Liv. 10. 20. 10), but: quod passim eos simul pavor terrorque distulerant (id. 6. 42. 8) and: ubi ira et aegritudo permixta sunt (Sall. Jug. 68. 1). We read: si pax veniaque ab dis impetrata esset (Liv. 1. 31. 7), but: postquam pax et concordia speciosis et inritis nominibus iactata sunt (Tac. Hist. 2. 20. 3); and we have the same pairs now with the singular, now with the plural, in: senatus populusque Romanus intellegit (Cic. Fam. 5. 8. 2), and: si antidea senatus populusque iusserit (Liv. 22. 10. 6); but: cum senatus populusque Romanus pacem comprobaverint (Liv. 37. 45. 14), and: auctor essem senatui populoque Romano, ut eam vos habere sinerent (Liv. 36. 32. 5). And in Livy : tempus et locus convenit (1. 24. 2); but in Tacitus: ubi locus veneficii tempusque composita sunt (Ann. 4. 10. 2). And turning to names of persons: Palatium Romulus, Remus Aventinum ad inaugurandum templa capiunt (Liv. 1. 6. 4), but: Tros Tyriusque mihi nullo discrimine agetur (Aen. 1. 574). We have here to do with pairs that are potential duals, and so are constructed now with a verb in the plural, now in the singular. The opposite will also be found: when two subjects are disjoined by oure or й, by neque or aut, while they will usually take a singular, they will be constructed with the plural at times. So we have in Homer: οὐδ ̓ ἐδύναντο οὔθ ̓ ὁ τὸν ἐξελάσαι καὶ ἐνιπρῆσαι πυρὶ νῆα and in a fragment of Bacchylides: θνατοῖσι δ ̓ οὐκ αὐθαίρετοι οὔτ ̓ ὄλβος οὔτ ̓ ἄκαμπτος "Αρης οὔτε πάμφθερσις στάσις (Fr. 36, B.), and in Euripides : καί μ ̓ οὔθ ̓ ὁ Πλούτωνος κύων οὔθ ̓ οὑπὶ κώπῃ ψυχοπομπὸς ἂν Χάρων ἔσχον (Alc. 360-2); and with ἤ in: ὅταν ἀδελφὸς ἢ ἀδελφή τῳ yévшvтaι Kaλoí (Plat. Leg. 838 A). So in Latin: haec si neque ego neque tu fecimus (Ter. Ad. 103), erant enim quibus nec senatus gloriari neque princeps possent (Plin. Pan. 75); and with aut in: si quid Socrates aut Aristippus contra morem consuetudinemque civilem fecerint locutive sint (Cic. Off. 1. 148. 41), or: quaerere puerum aut puellam qui supponantur mihi (Pl. Truc. 403-4). Seu ... seu with a plural verb is recorded only in: et me seu naturalis sollicitudo seu fides sedula, non ad diligentiam modo, verum ad amorem quoque commissae rei instigent (Frontin. de Aquaed. Praef.), and tam. . . quam in: ut proprium ius tam res publica quam privata haberent (ib. 128). When we have more than two subjects connected by et or κaí, in Latin the verb usually agrees with the last of its subjects, in Greek with the nearest, as in: aetas et forma et super omnia Romanum nomen te ferociorem facit (Liv. 31. 18. 3), ζῶντι τῷ δικαίῳ . . . ἆθλά τε καὶ μισθοὶ καὶ δῶρα γίγνεται (Plat. Rep. 614 4). But in : ἡμεῖς δή, ἐγὼ kai καὶ Στράτιος καὶ Στρατοκλῆς παρεσκευάζοντο ἅπαντες λαγχάνειν (Isaeus, 11. 15) the agreement presents difficulty. ... Probably we have ys the third plural dependent on the pair Stratius and Stratocles. What is the construction with plurals joined with duo? With masculine and neuter plurals the dual may be used as in αὐτῷ γίγνεσθον υἱεῖς δύο (Isae. 8. 1o), and : δύο πυρὸς σώματα εἰς ἓν σvvíoτaσlov cidos dépos (Plat. Tim. 56 E). With masculines the plural is the rule, as in: τῆς δ ̓ ἀρχῆς αὐτῷ λοιποὶ δύο μῆνες ἦσαν (Antiphon 6. 42), while with neuters the singular is usual, as in: dúo δέ μοι τῆς κατηγορίας εἴδη λέλειπται (Aeschin. 1. 116). At times, however, we have neuters plural taking a singular and a plural verb in successive clauses, as in : καὶ δὴ δοῦρα σέσηπε νεῶν καὶ σπάρτα λέλυνται (ΙΙ. 2. 135). Very curious is the union of a singular with a plural verb in: τῷ δ ̓ οὗ τι γυνὴ καὶ νήπια τέκνα οἴκαδε νοστήσαντι παρίσταται οὐδὲ γάνυνται (Οd. 12. 42-3). We may have here a use of a singular and a plural with νήπια τέκνα, but more probable seems the use of both singular and plural with the dual union : γυνὴ καὶ νήπια τέκνα. XVII THE SCHEMA ALCMANICUM, AND FURTHER We have already spoken (pp. 64 ff.) of the development of the Schema Alcmanicum from a form of the extended elliptical dual, which was perfected in the Skt. Mitrā Varunā, and becomes a plural in Veneres Cupidinesque. We noted in Irish: doronsat sid ocus Fergal 'they made peace and Fergal', for 'he and Fergal made peace'. In the Latin life of Fintan we read: venit Fintanus . . . et salutaverunt se in vicem et Lasserianus. If in the second clause we supply Fintanus from the first, we have: Fintanus salutaverunt se in vicem et Lasserianus, a construction exactly parallel to several examples of the Schema Alcmanicum. In the development of the extended elliptical dual we seem to have three stages: (1) the dual Mitra for Mitra and Varuna, (2) the addition for clearness of Varuna to this dual, (3) the assimilation of this Varuna to the preceding Mitrā in Mitrā Varunā 'the two Mitras, the two Varunas'. In the examples from AngloSaxon: wit Scilling 'we two Scilling' for 'I and Scilling', and in that cited from Irish we seem to have the second stage. The assimilation to the third stage in Greek seems to have been, not progressive as in Sanskrit, but retrogressive. Take the example we have from Alcman: Κάστωρ τε πώλων ὠκέων δματῆρες, ἱππόται σοφοί, xaì Hoλudeúkŋs kvôpós (Fr. 9); this was probably before the retrogressive assimilation: Κάστορες τε πώλων ὠκέων δματῆρες κτλ., which makes the Suarêpes intelligible. But Káσropes was assimilated in number to the following Πολυδεύκης. In: Τυδεὺς μάχην συνῆψε Πολυνείκης θα ἅμα (Eur. Suppl. 144) and: Λᾶσός ποτ ̓ ἀντεδίδασκε καὶ Σιμωνίδης (Ar. Vesp. 1410) the verb has also been assimilated to the singular, leaving now nothing of the Schema but its peculiar order of words. Of the examples we have of this curious figure we have already cited: ἦχι ῥοὰς Σιμόεις συμβάλλετον ἠδὲ Σκάμανδρος (ΙΙ. 5. 774), where we still see the dual, the primary number of the figure. This dual has become plural in: |