Читать книгу The Latin Dual and Poetic Diction: Studies in Numbers and Figures - Andrew J. Bell - Страница 17

Table of Contents

In introducing the cases of the dual Brugmann writes: As regards formation the dual seems to have been a singular of which the essential formative elements originally expressed the quality of being coupled or paired (Vergl. Gr. Il². 2. 194). In discussing the Irish da fer (=two men) he notices the use of the nom. sing. masc. fer as a dual, as likewise of the masc. sing. tene, and the neut. sings. dliged, tech, ainm, ascribing the use to a likeness in form that developed there. We noticed Sommer’s belief that genu and cornu are old duals that have become singulars, and that he bases the change, not on form, but on the notion of pair in ‘the two knees’, or in ‘the two horns’, making meaning, not form, the cause of the transfer. The use of Castor, or of Pollux, to name both the Twins, seems based not on the form of the word, but on the idea of pair in the Gemini, because of which the name of the one at once suggests to the reader or hearer the name of the other. Meaning seems more important than form in producing this change: if in a dual you emphasize the idea of ‘two’ or division, it will pass to the plural; if you dwell on the notion of pair, it may pass to the singular.

This change from the dual to the singular, though it escaped the notice of the Roman grammarians, seems to me certain for Latin and probable for Greek, though to a less extent. Just as in Latin we find nasus or nasum in the singular and nares in the plural, while in Skt. we have only nāsā in the dual, in Greek we have ῥίς, ἡ, and ῥίν, τό, as well as ῥῖνες for the nose. We read in Horace: gaude quod spectant oculi te mille loquentem (Ep. 1. 6. 19), not, however, of five hundred spectators, for oculus here evidently means a spectator, i.e. a pair of eyes. Usually we have the plural, as in: qui siccis oculis monstra natantia (vidit) (Od. 1. 3. 18); but what of: quisquis ingentes oculo irretorto special acervos (Od. 2. 2. 23-4) or: solus mullisne coheres, veloci percurre oculo (Sal. 2. 5. 54-5)? We read in Horace: non islic obliquo oculo mea commoda quisquam limat (Ep. 1. 14. 37-8), but in Ovid: altera, si memini, limis subrisit ocellis (Am. 3. 1. 33). In: cum tibi sol tepidus plures admoverit aures (Hor. Ep. 1. 20. 19) I feel that the singular of plures aures is una auris (=auditor unus); and I read in Martial: aurem non ego tertiam timerem (12. 24. 10) ‘I should not fear a third pair of ears’. So in Virgil: simul hoc dicens attollit in aegrum se femur (Aen. 10. 856-7) femur must be for utrumque femur; for his wound was imo inguine; (hasta) ima sedit inguine (imo) v.785-6. More interesting is:

tum pendere poenas

Cecropidae iussi, miserum, septena quotannis

Corpora natorum. (Aen. 6. 20-2.)

Heyne asks: Cum pueri septem septemque puellae mitterentur, quidni alterutrum tantum poni potuit? But he has no answer. Conington’s answer is very close to the mark: ‘The story mentioned seven youths and seven maidens, but Virgil has chosen only to name the former’. But we have here, too, a dual passing to a singular, each corpus natorum consisting of a youth and a maid.

Virgil uses the plural of geminus as an adjective to designate well-matched pairs, as in: gemini ... inmensis orbibus angues (Aen. 2. 203-4), geminae columbae (6.190), geminae belli portae (7.607), geminae somni portae (6.893), geminae quercus (9.681), huc geminas nunc flecte acies (6.788), geminae slant vertice cristae (6.779). The last two examples are noteworthy; in the last the crests are on the same head, and when pairs are thus connected we commonly have the singular, not the plural, as in: gemina teguntur lumina nocte (Catull. 51. 11), gemino demittunt bracchia muro turriti scopuli (Aen. 3. 535), geminum pugnae proponit honorem (5.365), gemina super arbore sidunt (6.203). So we may regard: geminique sub ubere nati (5.285) as short for: gemini (gemino) sub ubere nati, an abbreviation of four to three, from which hypallage takes its origin; cf. geminos huic ubera circum ludere pendentes pueros (8.631-2). So, too, with geminum (geminae) pugnae proponit honorem. In: solem geminum et duplices ... Thebas (Aen. 4. 470) we have, probably, the exchange of number for the sake of variety. That foris in the sing. and fores in the pl. are both used for the pair of doors is plain from: ad geminae limina prima foris (Ov. Her. 12. 150) and: frustra clavis inest foribus (Tib. 1. 6. 34). Along with θύραι ‘the pair of doors’, which Brugmann thinks an old dual, Homer uses θύρη for ‘the door’. Probably foras is the acc. pl. and foris the abl. pl. of an old *fora, corresponding with this; and foris is a new nom. sing. formed on the analogy of fores, the old plural of fora. (Cf. manus ‘good’, and manes ‘the kindly ones’.)

By Homer’s time duals like θύραι had become plurals; in their stead had begun the development of duals in ᾶ for α-stems following the analogy of duals like ἵππω. We see this beginning in Homer, not for feminines, but for masculines like Ἀτρείδᾶ; but the development is soon extended to feminines. It was an irregular and partial development; for the use of the dual in Greek was passing away. We find side by side ταῖν χεροῖν and τοῖν χεροῖν, τὰ κόρα and τὼ στήλα. Attraction played its part, as we see from ἀμφοῖν τοῖν χεροῖν, but ταύταιν ταῖν ἀδελφαῖν. The form ταῖν seems to have developed for the article before τά; Plato and the Orators have ταῖν ὁδοῖν, but τὼ ὁδώ, Sophocles and Aristophanes write τὰ κόρα, but τὼ χεῖρε. One of the results of the struggle thus prolonged of the dual against the prevailing plural seems to have been that, when the dual disappeared, it was almost always the plural that took its place. There is no trace of such a revival of the dual in Latin; it probably disappeared there more readily and speedily; and so more of its uses passed to the singular.

Our most interesting example of this use of the singular for the dual occurs in Horace’s first Canidia epode, where the boy, whom the witches are about to kill to obtain an unguent from his marrow, after a vain appeal to their feelings as women, in a second speech threatens them with the vengeance of gods and men. Its opening words have proved a crux for all editors:

Venena magnum fas nefasque non valent

Convertere humanam vicem (Epod. 5. 87-8).

To this text, given by all manuscripts of any value, Bentley adds the note: Durissimus locus, neque, quocumque modo vertas et excutias, sententiam commodam praebens. Scio equidem ut conati sint explicare veteres novique interpretes; sed si quid video laterem lavarunt.... Multa quidem ipse nequidquam tentavi, quae piget hic referre; tamen, ne omnino asymbolus veniam, dicam aliquid, quod etsi ne mihi quidem placeat, Rutgersiano haud deterius fore credo. Ergo ecce tibi correctionem, qualiscumque est:

Venena magica fas nefasque non valent,

Non vertere humanas vices.

Of our novi interpretes Kiessling reads:

Venena maga non fas nefasque, non valent

Convertere humanam vicem.

In maga non he follows an improvement on Bentley’s magica suggested by Moritz Haupt, and he calls the magnum of the manuscript meaningless (sinnlos); but he does not venture to change the number of humanam vicem.

Turning to the veteres we read in Acron: Venena magnum f. n. n. v. c. h.v. Lex enim humana habet malis poenam, bonis praemia pollicenda, et ideo furens puer dicit, haec eas carminibus mutare non posse. Porphyrio gives us: Ven. magnum fas nefasque. Magnum fas venena sunt, si hostibus dantur, magnum nefas, si amicis. Non valent conv. hum. vicem. Sic sensus est: Quamvis venena multum possint, non tamen valent merita in contrarium vertere, ut liberentur poena, aut mala mereantur. Vices autem appellantur poenae, quae in scelerosis admissis regerentur.

It is evident that both accept the text cited by Bentley as that of all good manuscripts, and it might seem significant that both give in full magnum, the word rejected by Bentley and found meaningless by Kiessling. There are two things of the greatest importance to note in Porphyrio’s second scholium: (1)in his explanation he substitutes vices for the vicem of his text, (2)in arranging Horace’s words he joins non valent with convertere. We have in the distich non valent in the first verse, and convertere in the second, giving us a not very difficult example of distribution; we must fill out each of these, understanding for each convertere non valent. So we have: venena magnum fas nefasque convertere non valent; convertere non valent humanam vicem (=vices, Porph.). Acron’s note gives us the true meaning of humanam vicem: it is a singular where prose would use a plural; a singular for a dual, the praemia and poenae of the lex humana, which here balance the fas nefasque of the gods. It is put in the singular to vary the diction of the couplet, just as in: et solem geminum et duplices se ostendere Thebas (Aen. 4. 470) solem geminum is put in the singular to give variety to the verse.

Why do Bentley and Kiessling feel magnum to be inadmissible, in face of such clear testimony from Scholia and manuscripts? Venena ait magica, says Bentley, id enim epitheton necessarium hic videtur; quippe venena per se et absolute posita non possunt rem magicam denotare. If venena were here used in its proper and absolute sense of ‘drugs’, one might readily supply magica from the theme of the whole poem, which is of the maga Canidia; and Horace might well deem its expression here unnecessary. But Bentley is hopelessly prosaic; usually in poetry words of importance are not used ‘per se et absolute’; and Acron tells us implicitly that venena is here used for carmina ‘spells’. Magnum is ‘sinnlos’, thinks Kiessling. True, it is not here applied to a material object, the size of which can be determined by the modius or the decempeda. But magnus is at times applied to animus; we have them connected in magnanimus ‘high-souled’. From the humanam of the second verse we might have expected divinum in the first, but divinum is already implied in fas as opposed to lex. Convertere means ‘to change wholly’, or ‘to reverse entirely’. In balancing magnum fas nefasque by humanam vicem, the poet has chosen for variety to use the singular, not the plural, to express the dual idea of the rewards and penalties of human law. We shall translate then: Your spells have not the power wholly to reverse the right and wrong of high heaven; they have not the power to reverse even the rewards and punishments of men.

It seems strange at first sight that Brugmann is willing to accept cornus (old cornuus) as for an older gen. dual cornuos corresponding to the Skt. gen.-loc. dual in -os, but is unwilling to accept cornu as the old nom. dual. A similar irregularity is noted, however, for Polish by Delbrück; while reçe, the old nom. dual, ‘the two hands’, has become a plural, reku, its gen.-loc., has passed to the singular (Vergl. Synt. I, p. 145). From examples like: iam cornu petat et pedibus qui spargat harenam (Buc. 3. 87 et Aen. 9. 629), cornu ferit ille (Buc. 9. 25), qui vexat nascenti robora cornu (Juv. 12.9) it is plain that the singular cornu in Latin stands at times for a pair of horns. The plural is far more common; as is genua in: tarda trementi genua labant (Aen. 5. 431-2) but we read: nuda genu nodoque sinus collecta fluentes (1.320), and: impressoque genu nitens terrae adplicat ipsum (12. 303). In:

quamquam tardata sagitta

Interdum genua impediunt cursumque recusant (12. 746-7),

Wagner relying on inferior manuscripts changed tardata to tardante, and he is followed by Forbiger and Ladewig. Only one of Aeneas’s knees is affected by the arrow, and interdum seems to show that it is this one knee which at times impedes him in the race. Ribbeck has retained tardata, as all the manuscripts on which he relies give this reading, which is clearly the more difficult; and this is the reading of Conington and Sidgwick, who do not explain Wagner’s difficulty in translation, however. Probably we have here a use of genua for genu, the opposite of genu for genua.

We read: et gemina auratus taurino cornua voltu (Geo. 4. 371). We have also geminae nares (Geo. 4. 300), geminas aures (Culex 150), geminos lacertos (Moretum 21), geminas acies (Aen. 6. 788), showing that gemini is felt to be the numeral appropriate to pairs. While neither duo nor ambo is joined with the singular, we noticed solem geminum (Aen. 4. 470) ‘twin suns’, gemino muro (3.535) ‘twin walls’, geminum honorem (5.365) ‘a pair of prizes’, gemino ab ovo (Hor. A.P. 147) ‘from the twin eggs’, and geminae foris (Ov. Her. 12. 150) ‘a pair of doors’. It is reasonable to place here: geminus Pollux (Hor. Od. 3. 29. 64) and geminus Castor (Ov. A.A. 1. 746). If the Latin dual passes into the singular at times, it is only to be expected that the numeral for two appropriate to the dual shall also appear as a singular.

The dual is used as the singular at times, in return for the uses of the singular for the dual. I have already noticed the use of Atridas (Aen. 1. 458), and of tardata sagitta genua (Aen. 12. 746-7). Of course in Latin it is the plural used for the dual that is in question.

In poetry a part of the body is often expressed by the plural, where we should expect the singular. We read: coniugis ille suae complexus colla lacertis (Ov. Met. 1. 734); why colla? Many parts of the body occur in pairs; e.g. oculi, aures, genae, manus, pedes, genua; and after analogy of such parts it became a poetic fashion to use the plural even for such parts as were not paired, like cervix, collum, mentum, dorsum. Often such uses are plainly to vary the diction, as in: nunc animis opus, Aenea, nunc pectore firmo (Aen. 6. 261), with which compare: violenta pectora Turni edocet (10. 151-2) and: his animum arrecti dictis (1.579). Here animis is the mind of Aeneas, but animum the minds of Aeneas and Achates; while both pectore and pectora are of a single person. In: manus iuvenem ... post terga revinctum (Aen. 2. 57) terga of a single person seems to follow the analogy of pectora, a paired part. So in Homer we read:

τὸν . . . ὁ μὲν ἰῷ

βεβλήκει τελαμῶνα περὶ στήθεσσι φαεινόν (Il. 12. 400-1),

and:

ἥ τ’ ἀνὰ νῶτα θέούσα διαμπερὲς αὐχέν’ ἱκάνει (Il. 13. 547),

where νῶτα seems to follow the analogy of στήθεα in like manner.

Still more noteworthy is the use of the dual in Sanskrit in the double elliptical form Mitrā Varunā, the two Mitras, the two Varunas, for Mitra and Varuna; or in pitarau-mātarau, the two fathers, the two mothers, for the father and the mother. We find the same idiom in Mithra Ahura in the Avesta (Reichelt, Av. Synt. p. 222). In: lugete, O Veneres Cupidinesque (Catull. 3.1) Schwyzer (Ig. F. 14, p. 28) thinks that the plurals Veneres Cupidinesque are for old duals, and mean merely Venus and Cupid.

The development of this construction seems interesting, and can be traced with the help of Anglo-Saxon and old Irish. In Skt. the dual Mitrā is used for Mitra and Varuna, just as in Greek Ἀτρείδα for Agamemnon and Menelaus (Il. 1. 16), or in Latin Castores for Castor and Pollux. But at times uses like Mitrā ‘the two Mitras’ did not convey clearly enough the name of the second member of the pair, and so the second name would be added, at first in the singular. So we read: ā yad ruhāva Varuṇaś ca navam (R.V. 7. 88.3) ‘when we two and Varuna get on board the ship’, with the meaning ‘when I and Varuna get on board’, &c. We have the same syntax in the Anglo-Saxon phrase: wit Scilling ‘we two Scilling’ for ‘I and Scilling’. Zimmer gives us from Irish (K.Z. 32, p. 152 ff.): doronsat sid ocus Fergal ‘they made peace and Fergal’ for ‘he and Fergal made peace’; a syntax reflected in the Latin life of Fintan: in illo autem die ante vesperum venit Fintanus ad consilium, et salutaverunt se in vicem et Lasserianus. But instead of advancing to two duals as in Sanskrit, or to two plurals as in Latin, Irish changes the verb back to the singular, influenced, Zimmer thinks, by an impersonal use of the verb; and we have a form which Zimmer translates thus: She came from the East in the shape of two swans and her maid.

In Greek Wackernagel cites from Homer:

ἐς δ’ ἐνόησ’ Αἴαντε δύω, πολέμον ἀκορήτω

ἑσταότας, Τεῦκρόν τε νέον κλισίηθεν ἰόντα (Il. 12. 335-6),

where he thought Αἴαντε δύω . . . Τεῦκρόν τε parallel to the āvām Varuṇaś ca of the example cited from the Rig-Veda, where he amplifies āvām from the -va in ruhāva. And this view seems confirmed by:

τὶν δ’ ἐν Ἰσθμῷ διπλόα θάλλοισ’ ἀρετά,

Φυλακίδα, κεῖται, Νεμέα δὲ καὶ ἀμφοῖν

Πυθέᾳ τε, παγκρατίου (Pind. Isth. (5). 17 ff),

‘further, Phylacides, a double crown of glory is at Isthmus stored and at Nemea both for thee and for Pytheas, a pancratiast’s crown’ (Myer’s transl.). The Scholiast explains ἀμφοῖν Πυθέᾳ τε by ἀμφοτέροις ὑμῖν, σοί τε καὶ τῷ Πυθέᾳ, making it parallel with Αἴαντε ... Τεῦκρόν τε. But the Greek never advanced to Αἴαντε Τεύκρω τε as did the Sanskrit to Mitrā Varunā.

It seems clear that this advance in Sanskrit was due to an assimilation of the second to the first in number. So in Latin from an older Veneres Cupidoque, where Veneres primarily meant Venus and Cupid, we have Veneres Cupidinesque (also in Catull. 13. 12 and Martial, 9. 11. 9 and 11. 13.6). But the assimilation of the first to the second was also possible, as we find it developing in Irish from: they and Fergal, to: she ... and her maid. This is what takes place in Greek, giving us the Schema Alemanicum, as we see it in:

ἧχι ῥοὰς Σιμόεις συμβάλλετον ἠδὲ Σκάμανδρος (Il. 5. 774),

where συμβάλλετον becomes intelligible only when we think of its primary subject as Σιμόεντε, ‘the Simois and Skamander’, which, after Σκάμανδρος was added, passed back to Σιμόεις under the influence of its form and meaning.

Of course it would be absurd to think of the Latin construction here as directly influenced by the Skt. syntax. Gauthiot, in the paper Du Nombre Duel already cited, shows that this curious doubling of the dual is not uncommon in Finnish; and from its development in languages so independent of each other we may conclude that it is a natural and easy development of the dual wherever found. Other peculiarities of the dual, such as its passage to the singular, may be even more easily assumed as probable developments of the dual number wherever it occurs. (See AppendixA.)

Delbrück holds that the primary use of the dual is for natural pairs, like ὄσσε or χεῖρε. But that this use, though very old, is older than, or even as old as, that for twins, as in the Skt. Aśvinau, is opposed to much that we see in the development of number and gender. In the next chapter we shall see reasons for believing that the plural was used primarily of persons, and was extended to things later. The use of gemini, as the numeral appropriate to the dual in Latin, and accompanying it in its passage to the singular as well as to the plural, tends to confirm this. The sensation, at times of joy, but oftener of horror, called forth by a birth of twins among savages of to-day, may give us some idea of its importance among primitive men. In Russian the use of the dual is extended from dva, ‘two’, and oba, ‘both’, to trī, ‘three’, and chetyre, ‘four’ (Figgis, Russ. Gram., p. 91). The extension to ‘four’ seems the same as that we cited from Il. 8. 185-6; but its use with three seems easiest to understand when we think of the evident association of gemini with tergemini. Delbrück notes that Slovenian, and Upper and Lower Serbian have kept the same construction and almost to the same extent to which it existed in Old Bulgarian; he thinks it an effect of analogy (Vergl. Synt. I, p. 144-5).