---
title: "The Orator’s Education · Institutio oratoria"
author: "Quintilian"
translator: "Harold Edgeworth Butler"
languages: [la, en]
shelf: "Classical Latin"
section: 40 of 53
url: https://scriptura.si/works/quintilian-institutio-oratoria/40/
cite: "Quintilian, tr. Harold Edgeworth Butler, The Orator’s Education · Institutio oratoria, section 40. Scriptura (scriptura.si), https://scriptura.si/works/quintilian-institutio-oratoria/40/"
---

# The Orator’s Education · Institutio oratoria

*Quintilian, tr. Harold Edgeworth Butler*

Twelve books on training a good man skilled in speaking, from the nursery to the forum.

> Latin: ed. Harold Edgeworth Butler, from the Perseus Digital Library (canonical-latinLit, phi1002.phi001, CC BY-SA 4.0).

> English: Harold Edgeworth Butler, from the Perseus Digital Library (phi1002.phi001.perseus-eng2). Notes are not included.

> Rows follow the references the two texts share (books, chapters and sections), cut where both reach the end of a sentence.

## 40. Book IX · chapter 4 · 4.80–4.147

Source: https://scriptura.si/works/quintilian-institutio-oratoria/40/ · cite a paragraph as https://scriptura.si/works/quintilian-institutio-oratoria/40/#p-N

[¶0] 4.80. He does not, however, conceal the fact that some regard these as rhythms rather than feet: and they are right in so doing, since whatever is longer than three syllables involves more than one foot. Since then there are four feet which consist of two syllables, and eight composed of three, I shall call them by the following names: two long syllables make a spondee; the pyrrhic or pariambus, as some call it, is composed of two shorts; the iambus of a short followed by a long; its opposite, that is a long followed by a short, is a choreus, for I prefer that term to the name of trochee which is given it by others.

> Latin: 4.80. tamen ipse dissimulat, quibusdam numeros videri non pedes; neque immerito; quidquid est enim supra tris syllabas, id est ex pluribus pedibus. cum constent quattuor pedes binis, octo ternis, spondeum longis duabus, pyrrhichium, quem alii pariambum vocant, brevibus, iambum brevi longaque, huic contrarium e longa et brevi choreum, non ut alii trochaeum nominemus.

[¶1] 4.81. Of trisyllabic feet the dactyl consists of a long followed by two shorts, while its opposite, which has the same time-length, is called an anapaest. A short between two longs makes an amphimacer, although it is more often called a cretic, while a long between two shorts produces its opposite, the amphibruachys. Two long syllables following a short make a bacchius, whereas, if the long syllables come first the foot is called a palimbacchius.

> Latin: 4.81. iis vero, qui ternas syllabas habent, dactylum longa duabusque brevibus, huic temporibus parem, sed retro actum appellari constat anapaeston. inter longas brevis faciet amphimacrum, sed frequentius eius nomen est creticus; longa inter breves, amphibrachyn huic contrarium.

[¶2] 4.82. Three shorts make a trochee, although those who give that name to the choreus call it a tribrach: three longs make a molossus. Every one of these feet is employed in prose, but those which take a greater time to utter and derive a certain stability from the length of their syllables produce a weightier style, short syllables being best adapted for a nimble and rapid style. Both types are useful in their proper place: for weight and slowness are rightly condemned in passages where speed is required, as are jerkiness and excessive speed in passages which call for weight.

> Latin: 4.82. longis brevem sequentibus bacchius, totidem longis brevem praecedentibus palimbacchius erit. breves trochaeum, quem tribrachyn dici volunt, qui choreo trochaei nomen imponunt; totidem longae molosson efficient. pedum nullus non ut orationem venit, sed quo quique sunt temporibus pleniores longisque syllabis magis stabiles, hoc graviorem faciunt orationem, breves celerem ac mobilem. locis utile; nam et illud, ubi opus est velocitate, tardum et segne, et hoc, ubi pondus exigitur, praeceps ac resultans merito damnetur.

[¶3] 4.84. It may also be important to remark that there are degrees of length in long syllables and of shortness in short. Consequently, although syllables may be thought never to involve more than two time-beats or less than one, and although for that reason in metre all shorts and all longs are regarded as equal to other shorts and longs, they none the less possess some undefinable and secret quality, which makes some seem longer and others shorter than the normal. Verse, on the other hand, has its own peculiar features, and consequently some syllables may be either long or short.

> Latin: 4.84. ut hoc quoque aliquid fortasse momenti, quod et longis longiores et brevibus sunt breviores syllabae, ut, quamvis neque plus duobus temporibus neque uno minus habere videantur, ideoque ut metris omnes breves longaeque inter se ipsae sint pares, lateat tamen nescio quid quod supersit aut desit. versuum propria condicio est, ideoque ut his quaedam etiam communes.

[¶4] 4.85. Indeed, since strict law allows a vowel to be long or short, as the case may be, when it stands alone, no less than when one or more consonants precede it, there can be no doubt, when it comes to the measuring of feet, that a short syllable, followed by another which is either long or short, but is preceded by two consonants, is lengthened, as for example in the phrase agrestem tenui musam.

> Latin: 4.85. vero quia patitur aeque brevem esse vel longam vocalem, cum est sola, quam cum eam consonantes una pluresve praecedunt, certe ut dimensione pedum syllaba, quae est brevis, insequente alia vel longa vel brevi, quae tamen duas primas consonantes habeat, fit longa, ut
tenui musam:
nam A brevis, gres brevis, faciet tamen longam priorem;

[¶5] 4.86. For both a and gres are short, but the latter lengthens the former, thereby transferring to it something of its own time-length. But how can it do this, unless it possesses greater length than is the portion of the shortest syllables, to which it would itself belong if the consonants st were removed? As it is, it lends one time-length to the preceding syllable, and subtracts one from that which follows. Thus two syllables which are naturally short have their time-value doubled by position.

> Latin: 4.86. dat igitur illi aliquid ex suo tempore. modo, nisi habet plus quam quae brevissima, qualis ipsa esset detractis consonantibus? unum tempus accommodat priori et unum accipit a sequente; ita duae natura breves positione sunt temporum quattuor.

[¶6] 4.87. I am, however, surprised that scholars of the highest learning should have held the view that some feet should be specially selected and others condemned for the purposes of prose, as if there were any foot which must not inevitably be found in prose. Ephorus may express a preference for the paean (which was discovered by Thrasymachus and approved by Aristotle) and for the dactyl also, on the ground that both these feet provide a happy mixture of long and short; and may avoid the spondee and the trochee, condemning the one as too slow and the other as too rapid; Aristotle may regard the heroic foot, which is another name for the dactyl, as too dignified and the iambus as too commonplace, and may damn the trochee as too hasty and dub it the cancan; Theodectes and Theophrastus may agree with him, and a later critic, Dionysius of Halicarnassus, may adopt a similar view;

> Latin: 4.87. autem ut hac opinione doctissimos homines fuisse, ut alios pedes ita eligerent aliosque damnarent, quasi ullus esset, quem non sit necesse ut oratione deprehendi. igitur paeana sequatur Ephorus, inventum a Thrasymacho, probatum ab Aristotele, dactylumque ut temperatos brevibus ac longis, fugiat spondeum et trochaeum, alterius tarditate nimia, alterius celeritate damnata, herous, qui est idem dactylus, Aristoteli amplior, iambus humilior videatur, trochaeum ut nimis currentem damnet eique cordacis nomen imponat, eademque dicant Theodectes ac Theophrastus, similia post eos Halicarnasseus Dionysius:

[¶7] 4.89. but for all they say, these feet will force themselves upon them against their will, and it will not always be possible for them to employ the dactyl or their beloved paean, which they select for special praise because it so rarely forms part of a verse rhythm. It is not, however, the words which cause some feet to be of more common occurrence than others; for the words cannot be increased or diminished in bulk, nor yet can they, like the notes in music, be made short or long at will; everything depends on transposition and arrangement.

> Latin: 4.89. irrumpent etiam ad invitos, nec semper illis heroo aut paeane suo, quem, quia versum raro facit, maxime laudant, uti licebit. sint tamen allis alii crebriores, non verba facient, quae neque augeri nec minui nec sicuti toni modulatione produci aut corripi possint, sed transmutatio et collocatio.

[¶8] 4.90. For a large proportion of feet are formed by the connexion or separation of words, which is the reason why several different verses can be made out of the same words: for example, I remember that a poet of no small distinction writing the following line:
Astra tenet caelum, mare classes, area messem,
a line which, if the order of the words be reversed, becomes a Sotadean; again, the following Sotadean, if reversed, reads as as an iambic trimeter:
caput exeruit mobile pinus repelita.
Feet therefore should be mixed, while care must be taken that the majority are of a pleasing character, and that the inferior feet are lost in the surrounding crowd of their superior kindred.

> Latin: 4.90. enim ex commissuris eorum vel divisione fiunt pedes; quo fit ut iisdem verbis alii atque alii versus fiant, ut memini quendam non ignobilem poetam talem exarasse:
tenet caelum, mare classes, area messem.
hic retrorsum fit sotadeus; itemque e sotadeo retro trimetros:
exeruit mobile pinus repetita.

[¶9] 4.91. The nature of letters and syllables cannot be changed, but their adaptability to each other is a consideration of no small importance. Long syllables, as I have said, carry the greater dignity and weight, while short syllables create an impression of speed: if the latter are intermixed with a few long syllables, their gait will be a run, but a gallop if they are continuous.

> Latin: 4.91. miscendi ergo sunt, curandumque, ut sint plures qui placent, et circumfusi bonis deteriores lateant. vero ut litteris syllabisque natura mutatur, sed refert, quae cum quaque optime coeat. igitur auctoritatis, ut dixi, et ponderis habent longae, celeritatis breves; quae si miscentur quibusdam longis, currunt, si continuantur, exultant.

[¶10] 4.92. When a short syllable is followed by a long the effect is one of vigorous ascent, while a long followed by a short produces a gentler impression and suggests descent. It is therefore best to begin with long syllables, though at times it may be correct to begin with short, as in the phrase novum crimen: a gentler effect is created, if we commence with two shorts, as in the phrase animadverti iudices: but this opening, which comes from the pro Cluentio, is perfectly correct, since that speech begins with something similar to partition, which requires speed.

> Latin: 4.92. , quae ex brevibus ad longas insurgunt; leniores, quae a longis ut breves descendunt. incipitur a longis, recte aliquando a brevibus, ut crimen; lenius a duabus, ut indices; sed hoc pro Cluentio recte, quia initium eius partitioni simile est, quae celeritate gaudet.

[¶11] 4.93. Similarly the conclusion of a sentence is stronger when long syllables preponderate, but it may also be formed of short syllables, although the quantity of the final syllable is regarded as indifferent. I am aware that a concluding short syllable is usually regarded as equivalent to a long, because the time-length which it lacks appears to be supplied from that which follows. But when I consult my own ears I find that it makes a great difference whether the final syllable is really long or only treated as the equivalent of a long. For there is not the same fullness of rhythm in diccre incipieniem timere as there is in ausus est confiteri.

> Latin: 4.93. quoque e longis firmissima est; sed venit et ut breves, quamvis habeatur indifferens ultima. enim ego ignoro, ut fine pro longa accipi brevem, quia videtur aliquid vacantis temporis ex eo, quod insequitur, accedere; aures tamen consulens meas, intelligo multum referre, verene longa sit, quae cludit, an pro longa. enim tam plenum est incipientem timere, quam illud est confiteri.

[¶12] 4.94. But if it makes no difference whether the final syllable be long or short, the concluding feet in these two instances must be identical: and yet somehow or other one gives the impression of sitting down and the other of a simple halt. This fact has led some critics to allow three timebeats for a final long syllable, adding the extra time-length which a short syllable derives from its position at the end of a sentence to the long syllable as well. And it not merely makes a difference with what foot a sentence ends, but the penultimate foot is also of importance.

> Latin: 4.94. si nihil refert, brevis an longa sit ultima, idem pes erit; verum nescio quo modo sedebit hoc, illud subsistet. moti quidam longae ultimae tria tempora dederunt, ut illud tempus, quod brevis ex loco accipit, huic quoque accederet. solum refert, quis pes claudat, sed claudentem quis antecedat. autem neque plus tribus, iique, si non ternas syllabas habebunt, repetendi erunt (absit enim poetica observatio), neque minus duobus;

[¶13] 4.95. It is not, however, necessary to go back further than three feet, and only that if the feet contain less than three syllables, for we must avoid the exactitude of verse: on the other hand, we must not go back less than two: otherwise we shall be dealing with a foot and not with rhythm. But in this connexion the dichoreus may be regarded as one foot, if indeed a foot consisting of two chorei can be considered as a single foot. The same is true of the paean composed of the choreus and a pyrrhic, a foot which is regarded as specially suitable to the beginning of a sentence, or of the other paean, formed of three shorts followed by a long, to which the conclusion is specially dedicated. It is of these two forms that writers on rhythm generally speak. Some, however, call all feet containing three short syllables and a long by the name of paean, irrespective of the position of the long syllable, and merely taking into account the total number of time-lengths that it contains.

> Latin: 4.95. alioqui pes erit, non numerus. tamen vel unus esse dichoreus, si unus est, qui constat e duobus choreis. paean, qui est ex choreo et pyrrhichio, quem aptum initiis putant, vel contra, qui est e tribus brevibus et longa, cui clausulam adsignant; de quibus fere duobus scriptores huius artis loquantur. omnes, ut quocunque sit loco longa, temporum quod ad rationem pertinet,paeanas appellant.

[¶14] 4.97. The dochmiac, again, which consists of a bacchius and an iambus, or of an iambus and a cretic, forms a solid and severe conclusion. The spondee, so frequently employed in this position by Demosthenes, is used with varying effect. It is most impressive when preceded by a cretic, as in the following instance: De qua ego nihil dicam, nisi depellendi criminis causa. Again there is a point, of the importance of which I spoke above, namely that it makes a considerable difference whether two feet are contained in a single word or whether they are both detached. Thus criminis causa makes a strong and archipiratae a weak ending, while tile weakness is still further increased if the first foot be a tribrach, as for instance in words like facilitates or temeritates.

> Latin: 4.97. et dochmius, qui fit ex bacchio et iambo vel ex iambo et cretico, stabilis ut clausulis et severus. quoque, quo plurimum est Demosthenes usus, non eodem modo semper se habet. praecedet eum creticus, ut ut hoc, qua ego nihil dicam, nisi depellendi criminis causa. nihil est, quod supra dixi multum referre, unone verbo sint duo pedes comprehensi an uterque liber. Sic enim fit forte casa; molle, mollius, si tribrachys praecedat, facilitates, temeritates.

[¶15] 4.98. For the mere fact that words are separated from each other involves an imperceptible length of time: for instance, the spondee forming the middle foot of a pentameter must consist of the last syllable of one word and the first of another, otherwise the verse is no verse at all. It is permissible, though less satisfactory, for the spondee to be preceded by an anapaest: e.g. muliere non solum nobili, verum etiami nota.

> Latin: 4.98. enim quoddam ipsa divisione verborum latens tempus, ut ut pentametri medio spondeo, qui nisi alterius verbi fine, alterius initio constat, versum non efficit., etiamsi minus bene, praeponi anapaestos: non solum nobili, verum etiam nota. anapaestus et creticus, iambus quoque, qui est utroque syllaba minor (praecedet enim tres longas brevis), sed et spondeus iambo recte praeponitur: [iisdem ut ] armis fui.

[¶16] 4.99. Andit may also, in addition to the anapaest and cretic, be preceded by the iambus, which is a syllable less in length than both of them, thus making one short syllable precede three long. But it is also perfectly correct to place a spondee before an iambus, as in armis fui, or it may be preceded by a bacchius instead of a spondee, e.g. in armis fui, thereby making the last foot a dochmiac. From this it follows that the molossus also is adapted for use in the conclusion provided that it be preceded by a short syllable, though it does not matter to what foot the latter belongs: e.g. illud scimus, ubicunque sunt, esse pro nobis.

> Latin: 4.99. spondeus, et bacchius, sic enim fiet ultimus dochmius: armis fui. iis quae supra probavi apparet molosson quoque clausulae convenire, dum habeat ex quocunque pede ante se brevem: scimus ubicunque sunt, esse pro nobis.

[¶17] 4.101. The effect of the spondee is less weighty, if it be preceded by a palimbacchius and pyrrhic, as in iudicii Iuniani. Still worse is the rhythm when the spondee is preceded by a paean, as in Brute, dubitavi, although this phrase may, if we prefer, be regarded as consisting of a dactyl and a bacchius. As a rule, endings composed of two spondees, a termination which causes comment even in a verse, are to be deprecated, unless the phrase is composed of three separate members, as in cur de perfiigis nostris copias comparat is contra nos? where we have a word of two syllables preceded and followed by a monosyllable.

> Latin: 4.101. gravis erit spondeus, praecedentibus palimbacchio et pyrrhichio, ut Iuniani, et adhuc peius priore paeane, ut Brute, dubitavi; nisi potius hoc esse volumus dactylum et bacchium. Duo spondei non fere se iungi patiuntur, quae ut versu quoque notabilis clausula est, nisi cum id fieri potest ex tribus quasi membris: de perfugis nostris copias comparat is contra nos? una syllaba, duabus, una.

[¶18] 4.102. Even the dactyl ought not to precede a final spondee, since we condemn verse-endings at the period's close. The bacchius is employed at the conclusion, sometimes in conjunction with itself as in venenum timerss while it is also effective when a choreus and spondee are placed before it as in ut venenum timeres. Its opposite, the palimbacchius, is also employed as a conclusion (unless, of course, we insist that the last syllable of a sentence is always long), and is best preceded by a molossus, as in civis Romanus sum, or by a bacchius, as in quod hie potest, nos possemus.

> Latin: 4.102. dactylus quidem spondeo bene praeponitur, quia finem versus damnamus ut fine orationis. et claudit et sibi iungitur: timeres; vel choreum et spondeum ante se amat: venenum timeres. quoque qui est, cludet, nisi si ultimam syllabam longam esse volumus, optimeque habebit ante se molosson: Romanus sum; aut bacchium, hic potest, nos possemus.

[¶19] 4.103. It would, however, be truer to say that in such cases the conclusion consists of a choreus preceded by a spondee, for the rhythm is concentrated in nos possemus and Romanus sum. The dichoreus, which is the repetition of one and the same foot, may also form the conclusion, and was much beloved by the Asiatic school: Cicero illustrates it by Patris dictum sapiens temeritasfilii comprobavit.

> Latin: 4.103. verius erit claudere choreum praecedente spondeo, nam hic potius est numerus, possemus et Romanus sum. et dichoreus, id est idem pes sibi ipse iungetur, quo Asiani sunt usi plurimum; cuius exemplum Cicero ponit, dictum sapiens temeritas filii comprobauit.

[¶20] 4.104. The choreus may also be preceded by a pyrrhic, as in omnes prope cives virtute, gloria, digitiate superabat. The dactyl also may come at the close, unless indeed it be held that, when it forms the final foot, it is transformed into a cretic: e.g. muliercula nixus in litore. The effect will be good if it is preceded by a cretic or an iambus, but unsatisfactory if it is preceded by a spondee, and worse still if by a choreus. The amphibrachys may close the cadence, as in Q. Ligarium in Africa fuisse, although in that case some will prefer to call it a bacchius. The trochee is one of the less good endings, if any final syllable is to be regarded as short, as it undoubtedly must be. Otherwise how can we end with the dichoreus, so dear to many orators? Of course, if it be insisted that the final syllable is long, the trochee becomes an anapaest.

> Latin: 4.104. ante se choreus et pyrrhichium: prope cives virtute, gloria, dignitate superabat. et dactylus, nisi eum observatio ultimae creticum facit: nixus ut litore. ante se bene creticum et iambum, spondeum male, peius choreum. amphibrachys: Q. Ligarium ut Africa fuisse, si non eum malumus esse bacchium. optimus est trochaeus, si ulla est ultima brevis, quod certe sit necesse est; alioqui quomodo claudet, qui placet plerisque, dichoreus? observatione ex trochaeo fit anapaestus.

[¶21] 4.106. If preceded by a long syllable, the trochee becomes a paean, as is the case with phrases such as sipotero, or dirit hoc Cicero, or obstat invidia. But this form of paean is specially allotted to the beginnings of sentences. The pyrrhic may close a sentence if preceded by a choreus, thereby forming a paean. But all these feet which end in short syllables will lack the stability required for the cadence, and should as a rule only be employed in cases where speed is required and there is no marked pause at the ends of the sentences.

> Latin: 4.106. trochaeus praecedente longa fit paean, quale est Si potero et hoc Cicero, obstat inuidia. hum initiis dederunt. et pyrrhichius choreo praecedente, nam sic paean est. omnes hi, qui ut breves excidunt, minus erunt stabiles, nec alibi fere satis apti, quam ubi cursus orationis exigitur et clausulis non intersistitur. et initiis optimus:

[¶22] 4.107. The cretic is excellent, both at the beginning (e.g. quod precatus a diis immortalibus sum ) and at the close (e.g. in conspeclu populi Romani vomere postridie ). The last example makes it clear what a good effect is produced when it is preceded by an anapaest or by that form of paean which is regarded as best suited to the end of a sentence. But the cretic may be preceded by a cretic, as in servare quam plurinos. It is better thus than when it is preceded by a chores, as in quis non turpe duceret? assuming that we treat the final short syllable as long. However, for the sake of argument, let us substitute duceres for duceret.

> Latin: 4.107. precatus a diis immortalibus sum, et clausulis: conspectu populi Romani vomere postridie. vero, quam bene eum praecedant vel anapaestos vel ille, qui videtur fini aptior, paean. et se ipse sequitur: quam plurimos. melius quam choreo praecedente, non turpe duceret? si ultima brevis pro longa sit; sed fingamus sic, turpe duceres.

[¶23] 4.108. Here, however, we get the rest of which I spoke: for we make a short pause between the last word and the last but one, thus slightly lengthening the final syllable of turpe; otherwise quis non turpe duceret? will give us a jerky rhythm resembling the end of an iambic trimeter. So, too, if you pronounce ore excipere liceret without a pause, you will reproduce the rhythm of a licentious metre, whereas if triply punctuated and thus provided with what are practically three separate beginnings, the phrase is full of dignity.

> Latin: 4.108. hic est illud inane, quod dixi: paulum enim morae damus inter ultimum ac proximum verbum et turpe illud intervallo quodam producimus; alioqui sit exultantissimum et trimetri finis: non turpe duceret? illud excipere liceret, si iungas, lascivi carminis est; sed interpunctis quibusdam et tribus quasi initiis fit plenum auctoritatis.

[¶24] 4.109. In specifying the feet above-mentioned, I do not mean to lay it down as an absolute law that no others can be used, but merely wish to indicate the usual practice and the principles that are best suited for present needs. I may add that two consecutive anapaests should be avoided, since they form the conclusion of a pentameter or reproduce the rhythm of the anapaestic metre, as in the passage, nam ubi libido dominatur, innocentiae leve praesidinun est, where elision makes the last two syllables sound as one.

> Latin: 4.109. ego, cum praecedentes pedes posui, legem dedi ne alii essent, sed quid fere accideret et quid ut praesentia videretur optimum, ostendi. quidem optime est sibi iunctus anapaestos, ut qui sit pentametri finis, vel rhythmos qui nomen ab eo traxit: ubi libido dominatur, innocentiae leve praesidium est; nam synaloephe facit, ut duae ultimae syllabae pro una sonent.

[¶25] 4.110. The anapaest should preferably be preceded by a spondee or a bacchius, as, for instance, if you alter the order of words in the passage just quoted to leve innocentiae praesidium est. Personally, although I know that in this I am in disagreement with great writers, I am not attracted by the paean consisting of three shorts followed by a long: for it is no more than an anapaest with the addition of another short syllable (e.g. facilitas, agilitas ). Why it should have been so popular, I cannot see, unless it be that those who gave it their approval were students of the language of common life rather than of oratory. It is preferably preceded by short syllables, such as are provided by the pyrrhic or the choreus (e.g.

> Latin: 4.110. fiet praecedente spondeo vel bacchio, ut si idem mutes leave innocentiae praesidium est. me capit (ut a magnis viris dissentiam) paean, qui est ex tribus brevibus et longa. est et ipse una plus brevi anapaestos facilitas et agilitas. ita placuerit is, non video, nisi quod illum fere probaverunt, quibus loquendi magis quam orandi studium fuit.

[¶26] 4.111. mea facilitas, nostra facilitas); on the other hand, if it be preceded by a spondee, we have the conclusion of an iambic trimeter, as indeed we have in the paean considered alone. The opposite form of paean is deservedly commended as an opening: for the first syllable gives it stability and the next three speed. None the less I think that there are other feet which are better suited for this purpose than even this paean.

> Latin: 4.111. et ante se brevibus gaudet pyrrhichio vel choreo, mea facilitas, nostra facilitas; ac praecedente spondeo tamen plane finis est trimetri, cum sit per se quoque. contrarius principiis merito laudatur, nam et primam stabilem et tres celeres habet. hoc quoque meliores alios puto.

[¶27] 4.112. My purpose in discussing this topic at length is not to lead the orator to enfeeble his style by pedantic measurement of feet and weighing of syllables: for oratory should possess a vigorous flow, and such solicitude is worthy only of a wretched pedant, absorbed in trivial detail:

> Latin: 4.112. vero hic locus non ideo tractatur a nobis, ut oratio, quae ferri debet ac fluere, dimetiendis pedibus ac perpendendis syllabis consenescat; nam id cum miseri, tum ut minimis occupati est, neque enim, qui se totum ut hac cura consumpserit, potioribus vacabit, si quidem relicto rerum pondere ac nitore contempto tesserulas, ut ait Lucilius, struet et vermiculate inter se lexis committet.

[¶28] 4.113. since the man who exhausts himself by such painful diligence will have no time for more important considerations; for he will disregard the weight of his subject matter, despise true beauty of style and, as Lucilius says, will construct a tesselated pavement of phrases nicely dovetailed together in intricate patterns. The inevitable result will be that his passions will cool and his energy be wasted, just as our dandies destroy their horses' capacity for speed by training them to shorten their paces. Prose-structure, of course, existed before rhythms were discovered in it, just as poetry was originally the outcome of a natural impulse and was created by the instinctive feeling of the ear for quantity and the observation of time and rhythm, while the discovery of feet came later. Consequently assiduous practice in writing will be sufficient to enable us to produce similar rhythmical effects when speaking extempore.

> Latin: 4.113. ergo refrigeretur sic calor et impetus pereat, ut equorum cursum delicati minutis passibus frangunt? vero numeri non sint ut compositione deprehensi, sicut poema nemo dubitaverit impetu quodam initio fusum et aurium mensura et similiter decurrentium spatiorum observatione esse generatum, mox ut eo repertos pedes. igitur ut hoc nos componet multa scribendi exercitatio, ut ex tempore etiam similia fundamus.

[¶29] 4.115. Further it is not so important for us to consider the actual feet as the general rhythmical effect of the period, just as the poet in writing a verse considers the metre as a whole, and does not concentrate his attention on the six or five individual feet that constitute the verse. For poetry originated before the laws which govern it, a fact which explains Ennius' statement that Fauns and prophets sang.

> Latin: 4.115. vero tam sint intuendi pedes quam universa comprehensio, ut versum facientes totum illum decursum non sex vel quinque partes, ex quibus constat versus, aspiciunt. enim carmen ortum est quam observatio carminis, ideoque illud Fauni vatesque canebant.

[¶30] 4.116. Therefore rhythmical structure will hold the same place in prose that is held by versification in poetry. The best judge as to rhythm is the ear, which appreciates fullness of rhythm or feels the lack of it, is offended by harshness, soothed by smooth and excited by impetuous movement, and approves stability, while it detects limping measures and rejects those that are excessive and extravagant. It is for this reason that those who have received a thorough training understand the theory of artistic structure, while even the untrained derive pleasure from it.

> Latin: 4.116. quem ut poemate locum habet versificatio, eum ut oratione compositio. autem de illa iudicant aures, quae plena sentiunt et parum expleta desiderant et fragosis offenduntur et levibus mulcentur et contortis excitantur et stabilia probant, clauda deprehendunt, redundantia ac nimia fastidiunt. docti rationem componendi intelligunt, etiam indocti voluptatem.

[¶31] 4.117. There are some points, it is true, which are beyond the power of art to inculcate. For example if the case, tense or mood with which we have begun, produces a harsh rhythm, it must be changed. But is it possible to lay down any definite rule as to what the change of case, tense or mood should be? It is often possible to help out the rhythm when it is in difficulties by introducing variety through the agency of a figure. But what is this figure to be? A figure of speech or a figure of thought? Can we give any general ruling on the subject? In such cases opportunism is our only salvation, and we must be guided by consideration of the special circumstances.

> Latin: 4.117. vero tradi arte non possunt. est casus, si durius is, quo coeperamus, feratur., ut quem transeamus ex quo, praecipi potest? laboranti compositioni variata saepe succurrit.? cum orationis, tum etiam sententiae? praescriptum eius rei ullum est? utendum et cum re praesenti deliberandum est.

[¶32] 4.118. Further with regard to the time-lengths, which are of such importance where rhythm is concerned, what standard is there by which they can be regulated save that of the ear? Why do some sentences produce a full rhythmical effect, although the words which they contain are few, whereas others containing a greater number are abrupt and short in rhythm? Why again in periods do we get an impression of incompleteness, despite the fact that the sense is complete?

> Latin: 4.118. vero spatia ipsa, quae ut hac quidem parte plurimum valent, quod possunt nisi aurium habere iudicium? alia paucioribus verbis satis plena vel nimium, alia pluribus brevia et abscisa sunt? ut circumductionibus, etiam cum sensus finitus est, aliquid tamen loci vacare videatur?

[¶33] 4.119. Consider the following example: neminem vestrum ignorare arbitror, iudices, hunc per hosce dies sermonem vulgi atque hanc opinionem populi Romani fiisse. Why is hosce preferable to hos, although the latter presents no harshness? I am not sure that I can give the reason, but none the less I feel that hosce is better. Why is it not enough to say sermonem vulgifuisse, which would have satisfied the bare demands of rhythm? I cannot tell, and yet my ear tells me that the rhythm would have lacked fullness without the reduplication of the phrase.

> Latin: 4.119. vestrum ignorare arbitror, iudices, hunc per hosce dies sermonem vulgi atque hanc opinionem populi Romani fuisse. hosce potius quam hos? enim erat asperum. fortasse non reddam, sentiam tamen esse melius. non satis sit, sermonem vulgi fuisse, (compositio enim patiebatur) ignorabo; sed ut audio hoc, animus accipit plenum sine hac geminatione non esse.

[¶34] 4.120. The answer is that in such cases we must rely on feeling. It is possible to have an inadequate understanding of what it is precisely that makes for severity or charm, but yet to produce the required effect better by taking nature for our guide in place of art: none the less there will always be some principle of art underlying the promptings of nature.

> Latin: 4.120. sensum igitur referenda sunt. si qui non satis forte, quid seuerum, quid iucundum sit, intelligent, facient quidem natura duce melius quam arte; sed naturae ipsi ars inerit. prorsus oratoris, scire ubi quoque genere compositionis sit utendum.

[¶35] 4.121. It is, however, the special duty of the orator to realise when to employ the different kinds of rhythm. There are two points which call for consideration if he is to do this with success. The one is concerned with feet, the other with the general rhythm of the period which is produced by their combination. I will deal with the latter first. We speak of commata, cola and periods.

> Latin: 4.121. duplex observatio est: altera, quae ad pedes refertur; altera, quae ad comprehensiones, quae efficiuntur ex pedibus. de his prius.

[¶36] 4.122. A comma, in my opinion, may be defined as the expression of a thought lacking rhythmical completeness; on the other hand, most writers regard it merely as a portion of the colon. As an example I may cite the following from Cicero: Domus tibi deerat? at habebas: pecunia superabat? at egebas. But a comma may also consist of a single word, as in the following instance where diximnus is a comma: Diximus, testes dare volumus.

> Latin: 4.122. igitur esse incisa, membra, circuitus. (quantum mea fert opinio) erit sensus non expleto numero conclusus, plerisque pars membri. est enim, quo Cicero utitur: tibi deerat? at habebas. superabat? at egebas. autem etiam singulis verbis incisa;, testes dare volumus; incisum est diximus.

[¶37] 4.123. A colon, on the other hand, is the expression of a thought which is rhythmically complete, but is meaningless if detached from the whole body of the sentence. For example O callidos homines is complete in itself, but is useless if removed from the rest of the sentence, as the hand, foot or head if separated from the body. He goes on, O rein excogitatam. At what point do the members begin to form a body? Only when the conclusion is added: quem, quaeso, nostrum fefellit, id vos ita esse facturos? a sentence which Cicero regards as unusually concise. Thus as a rule commata and cola are fragmentary and require a conclusion.

> Latin: 4.123. autem est sensus numeris conclusus, sed a toto corpore abruptus et per se nihil efficiens. O callidos homines perfectum est, sed remotum a ceteris vim non habet, ut per se manus et pes et caput: et O rem excogitatam. ergo incipit corpus esse? cum venit extrema conclusio:, quaeso, nostrum fefellit, id vos ita esse facturos? quam Cicero brevissimam putat. fere incisa et membra mutila sunt et conclusionem utique desiderant.

[¶38] 4.124. The period is given a number of different names by Cicero, who calls it ambitus, circuitus, comprehensio, continuatio and circumscriptio. It has two forms. The one is simple, and consists of one thought expressed in a number of words, duly rounded to a close. The other consists of commata and cola, comprising a number of different thoughts: for example, aderat ianitor carceris, carnriex praetoris and the rest.

> Latin: 4.124. plurima nomina dat Cicero, ambitum, circuitum, comprehensionem, continuationem, circumscriptionem. eius duo sunt, alterum simplex, cum sensus unus longiore ambitu circumducitur, alterum, quod constat membris et incisis, quae plures sensus habent: ianitor carceris, carnifex praetoris, reliqua. periodus membra minimum duo.

[¶39] 4.125. The period must have at least two cola. The average number would appear to be four, but it often contains even more. According to Cicero, its length should be restricted to the equivalent of four senarii or to the compass of a single breath. It is further essential that it should complete the thought which it expresses. It must be clear and intelligible and must not be too long to be carried in the memory. A colon, if too long, makes the sentence drag, while on the other hand, if it be too short it gives an impression of instability.

> Latin: 4.125. numerus videntur quattuor, sed recipit frequenter et plura. eius a Cicerone aut quattuor senariis versibus aut ipsius spiritus modo terminatur. debet ut sensum concludat; sit aperta, ut intelligi possit, non immodica, ut memoria contineri. longius iusto tardum; brevius instabile est.

[¶40] 4.126. Wherever it is essential to speak with force, energy and pugnacity, we shall make free use of commata and cola, since this is most effective, and our rhythmical structure must be so closely conformed to our matter, that violent themes should be expressed in violent rhythms to enable the audience to share the horror felt by the speaker.

> Latin: 4.126. acriter erit, instanter, pugnaciter dicendum,membratim caesimque dicemus, nam hoc ut oratione plurimum valet; adeoque rebus accommodanda compositio, ut asperis asperos etiam numeros adhiberi oporteat et cum dicente aeque audientem inhorrescere.

[¶41] 4.127. On the other hand we shall employ cola by preference when narrating facts, or relax the texture of our periods by considerable pauses and looser connexions, always excepting those passages in which narration is designed for decorative effect and not merely for the instruction of the audience, as for example the passage in the Verrines where Cicero tells the story of the Rape of Proserpine: for in such cases a smooth and flowing texture is required.

> Latin: 4.127. plerumque narrabimus, aut ipsas periodos maioribus intervallis et velut laxioribus nodis resolvemus, exceptis quae non docendi gratia, sed ornandi narrantur, ut ut Verrem Proserpinae raptus. enim lenis et fluens contextus decet.

[¶42] 4.128. The full periodic style is well adapted to the exordium of important cases, where the theme requires the orator to express anxiety, admiration or pity: the same is true of commonplaces and all kinds of amplification. But it should be severe when we are prosecuting and expansive in panegyric. It is also most effective in the peroration.

> Latin: 4.128. apta prooemiis maiorum causarum, ubi sollicitudine, commendatione, miseratione res eget, item communibus locis et ut omni amplificatione; sed poscitur tum austera, si accuses, tum fusa, si laudes. et ut epilogis pollet.

[¶43] 4.129. But we must only employ this form of rhythmical structure in its full development, when the judge has not merely got a grasp of the matter, but has been charmed by our style, surrendered himself to the pleader and is ready to be led whither we will, by the delight which he experiences. History does not so much demand full, rounded rhythms as a certain continuity of motion and connexion of style. For all its cola are closely linked together, while the fluidity of its style gives it great variety of movement; we may compare its motion to that of men, who link hands to steady their steps, and lend each other mutual support.

> Latin: 4.129. autem hoc adhibendum est, quod sit amplius compositionis genus, cum iudex non solum rem tenet, sed etiam captus est oratione et se credit actori et voluptate iam ducitur. non tam finitos numeros quam orbem quendam contextumque desiderat. omnia eius membra connexa sunt et, quoniam lubrica est, hac atque illac fluit, ut homines, qui manibus invicem apprehensis gradum firmant, continent et continentur.

[¶44] 4.130. The demonstrative type of oratory requires freer and more expansive rhythms, while forensic and deliberative oratory will vary the arrangement of their words in conformity with the variety of their themes. I must now turn to discuss the first of the two points which I mentioned above. No one will deny that some portions of our speech require a gentle flow of language, while others demand speed, sublimity, pugnacity, ornateness or simplicity, as the case may be, or that long syllables are best adapted to express dignity, sublimity and ornateness.

> Latin: 4.130. genus omne fusiores habet liberioresque numeros; iudiciale et contionale, ut materia varium est, sic etiam ipsa collocatione verborum. iam nobis pars ex duabus, quas modo fecimus, secunda tractanda est. quis dubitat alia lenius, alia concitatius, alia sublimius, alia pugnacius, alia ornatius, alia gracilius esse dicenda;

[¶45] 4.131. That is to say, while the gentler form of utterance requires length of vowel sounds, sublime and ornate language demands sonority as well. On the other hand, passages of an opposite character, such as those in which we argue, distinguish, jest or use language approximating to colloquial speech, are better served by short syllables.

> Latin: 4.131. gravibus, sublimibus, ornatis longas magis syllabas convenire? ita ut lenia spatium, sublimia et ornata claritatem quoque vocalium poscant; his contraria magis gaudere brevibus, argumenta, partitiones, iocos et quidquid est sermoni magis simile.

[¶46] 4.132. Consequently in the exordium we shall vary our structure to suit the thought. For I cannot agree with Celsus, when he would impose a single stereotyped form upon the exordium and asserts that the best example of the structure required for this purpose is to be found in Asinius: e. g., si, Caesar, ex omnibus mortalibus, qui sunt ac fuerunt, posset huic causae disceptator legi, non quisquam te potius optandus nobis fuit.

> Latin: 4.132. componemus prooemium varie atque ut sensus eius postulabit. enim accesserim Celso, qui unam quandam huic parti formam dedit, et optimam compositionem esse prooemii, ut est apud Asinium, dixit,, Caesar, ex omnibus mortalibus, qui sunt ac fuerunt, posset huic causae disceptator legi, non quisquam te potius optandus nobis fuit.

[¶47] 4.133. I do not for a moment deny that the structure of this passage is excellent, but I refuse to admit that the form of rhythmical structure which it exemplifies should be forced on all exordia. For there are various ways in which the judge's mind may be prepared for what is to come: at times we appeal for pity, at others take up a modest attitude, while we may assume an air of energy or dignity, flatter our audience, attempt to alter their opinions and exhort them to give us their best attention, according as the situation may demand. And as all these methods are different by nature, so each requires a different rhythmical treatment. Did Cicero employ similar rhythms in his exordia to the pro Milone, the pro Cluentio and the pro Ligario?

> Latin: 4.133. quia negem hoc bene esse compositum, sed quia legem hanc esse componendi ut omnibus principiis recusem. iudicis animus varie praeparatur: tum miserabiles esse volumus, tum modesti tum acres, tum graves, tum blandi, tum flectere, tum ad diligentiam hortari. ut sunt diversa natura, ita dissimilem componendi quoque rationem desiderant. similibus Cicero usus est numeris ut exordio pro Milone, pro Cluentio, pro Ligario?

[¶48] 4.134. The statement of fact as a rule requires slower and what I may be allowed to call more modest feet; and the different kinds of feet should, as far as possible, be intermixed. For while the style of this portion of our speech is generally marked by restraint of language, there are occasions when it is called upon to soar to greater heights, although on the other hand its aim will at all times be to instruct the audience and impress the facts upon their minds, a task which must not be carried out in a hurry. Indeed my personal opinion is that the statement of fact should be composed of long cola and short periods. Arguments, inasmuch as they are characterised by energy and speed, will employ the feet best adapted to these qualities.

> Latin: 4.134. fere tardiores atque, ut sic dixerim, modestiores desiderat pedes ex omnibus maxime mixtos. et verbis, ut saepius pressa est, ita interim insurgit; sed docere et infigere animis res semper cupit, quod minime festinantium opus est. mihi videtur tota narratio constare longioribus membris, brevioribus periodis.

[¶49] 4.135. They will not however acquire rapidity at the expense of force by employing trochees, but will rather make use of those feet which consist of a mixture of long and short syllables, though the long should not outnumber the short. Lofty passages, which employ long and sonorous vowels, are specially well served by the amplitude of the dactyl and the paean, feet which, although they contain a majority of short syllables, are yet not deficient in time-length.

> Latin: 4.135. acria et citata pedibus quoque ad hanc naturam commodatis utentur, non tamen ita ut trochaeis quoque celeria quidem, sed sine viribus sint, verum iis, qui sint brevibus longisque mixti, non tamen plures longas quam breves habent.

[¶50] 4.136. On the other hand, where violence is required, the requisite energy will be best secured by the employment of the iambus, not merely because that foot contains but two syllables, with the result that its beat is more frequent, making it unsuited to gentle language, but also because every foot gives the effect of an ascent, as they climb and swell from short to long, a fact which renders them superior to the choreus, which sinks from long to short.

> Latin: 4.136. sublimia spatiosas clarasque voces habentia amant amplitudinem dactyli quoque ac paeanis, etiamsi maiore ex parte syllabis brevibus, temporibus tamen satis pleni. contra iambis maxime concitantur, non solum quod sunt e duabus modo syllabis eoque frequentiorem quasi pulsum habent, quae res lenitati contraria est, sed etiam quod omnibus pedibus insurgunt et e brevibus ut longas nituntur et crescunt, ideoque meliores choreis, qui ab longis ut breves cadunt.

[¶51] 4.137. Subdued passages, such as occur in the peroration, also require slow syllables, which must, however, be less sonorous. Celsus insists that there is a special form of rhythmical structure which produces a particularly stately effect: I do not know to what he refers and, if I did, should not teach it, since it must inevitably be slow and flat, that is to say unless this quality is derived from the words and thoughts expressed. If it is to be sought for its own sake, independent of such considerations, I cannot sufficiently condemn it. But, to bring this discussion to a close, I would remark that our rhythm must be designed to suit our delivery.

> Latin: 4.137. , qualia ut epilogis sunt, lentas et ipsa, sed minus exclamantes exigunt. esse Celsus aliquam et superiorem compositionem, quam equidem si scirem, non docerem; sed sit necesse est tarda et supina, verum nisi ex verbis atque sententiis. se si id quaeritur, satis odiosa esse non poterit.

[¶52] 4.138. Is not our tone subdued as a rule in the exordium, except of course in cases of accusation where we have to rouse the judge or fill him with indignation, full and clear in the statement of fact, in argument impetuous and rapid not merely in our language, but in our motions as well, expansive and fluent in commonplaces and descriptions and, as a rule, submissive and downcast in the peroration?

> Latin: 4.138. , ut semel finiam, sic fere componendum quomodo pronuntiandum erit. non ut prooemiis plerumque summissi, (nisi cum ut accusatione concitandus est iudex aut aliqua indignatione complendus) ut narratione pleni atque expressi, ut argumentis citati atque ipso etiam motu celeres sumus, ut locis ac descriptionibus fusi ac fluentes, ut epilogis plerumque deiecti et infracti?

[¶53] 4.139. But the motions of the body also have their own appropriate rhythms, while the musical theory of rhythm determines the value of metrical feet no less for dancing than for tunes. Again, do we not adapt our voice and gesture to the nature of the themes on which we are speaking? There is, therefore, all the less reason for wonder that the same is true of the feet employed in prose, since it is natural that what is sublime should have a stately stride, that what is gentle should seem to be led along, that what is violent should seem to run and what is tender to flow.

> Latin: 4.139. corporis quoque motui sunt sua quaedam tempora et ad signandos pedes non minus saltationi quam modulationibus adhibetur musica ratio numerorum.? non vox et gestus accommodatur naturae ipsarum, de quibus dicimus, rerum? minus id mirere ut pedibus orationis, cum debeant sublimia ingredi, lenia duci, acria currere, delicata fluere. tragoediae, ubi necesse est, adfectamus etiam tumorem ex spondeis atque iambis quibus maxime continetur:
impero Argis, sceptra mi liquit Pelops.

[¶54] 4.140. Consequently, where necessary, we must borrow the pompous effect produced by the spondees and iambi which compose the greater portion of the rhythms of tragedy, as in the line,
En, impero Argis, sceptra mi liquit Pelops.
But the comic senarius, styled trochaic, contains a number of pyrrhics and trochees, which others call tribrachs, but loses in dignity what it gains in speed, as for example in the line,
quid igiturfaciam?

> Latin: 4.140. At ille comicus aeque senarius, quem trochaicum vocant, pluribus trochaeis, qui trochaei ab aliis dicuntur, pyrrhichiisque decurrit;

[¶55] 4.141. non earn, ne nunc quidem?
Violent and abusive language, on the other hand, even in verse, as I have said, employs the iambic for its attack: e.g.,
Quis hoc potest videre, quis potest pati,
nisi impudicus et vorax et aleo?
As a general rule, however, if the choice were forced upon me, I should prefer my rhythm to be harsh and violent rather than nerveless and effeminate, as it is in so many writers, more especially in our own day, when it trips along in wanton measures that suggest the accompaniment of castanets.

> Latin: 4.141. sed quantum accipit celeritatis, tantum gravitatis amittit:
igitur faciam? non eam ne nunc quidem
Aspera vero et maledica, ut dixi, etiam ut carmine iambis grassantur:
hoc potest videre, quis potest pati,
impudicus et vorax et aleo?
In universum autem, si sit necesse, duram potius atque asperam compositionem malim esse quam effeminatam et enervem, qualis apud multos et cotidie magis lascivissimis syntonorum modis saltat.

[¶56] 4.142. Nor will any rhythm ever be so admirable that it ought to be continued with the same recurrence of feet.

> Latin: 4.142. ne tam bona quidem ulla erit, ut debeat esse continua et ut eosdem semper pedes ire.

[¶57] 4.143. For we shall really be indulging in a species of versification if we seek to lay down one law for all varieties of speech: further, to do so would lay us open to the charge of the most obvious affectation, a fault of which we should avoid even the smallest suspicion, while we should also weary and cloy our audience by the resulting monotony; the sweeter the rhythm, the sooner the orator who is detected in a studied adherence to its employment, will cease to carry conviction or to stir the passions and emotions. The judge will refuse to believe him or to allow him to excite his compassion or his anger, if he thinks that he has leisure for this species of refinement.

> Latin: 4.143. et versificandi genus est unam legem omnibus sermonibus dare; et id cum manifestae adfectationis est (cuius rei maxime cavenda suspicio est), tum etiam taedium ex similitudine ac satietatem creat; quoque est dulcius, magis perdit amittit que et fidem et adfectus motusque omnes, qui est ut hac cura deprehensus. potest ei credere aut propter eum dolere et irasci iudex, cui putat hoc vacare.

[¶58] 4.144. It will therefore be desirable from time to time that in certain passages the rhythm should be deliberately dissolved: this is a task of no small difficulty, if the appearance of effort is to be avoided. In so doing we must not come to the assistance of the rhythm by introducing bhyperhata of extravagant length, for fear that we should betray the purpose of our action: and we should certainly never in our search for smoothness abandon for another any word that is apt and appropriate to our theme.

> Latin: 4.144. interim quaedam quasi solvenda de industria sunt; et quidem illa maximi laboris, ne laborata videantur. neque longioribus, quam oportet, hyperbatis compositioni serviamus ne, quae eius rei gratia fecerimus, propter eam fecisse videamur; et certe nullum aptum atque idoneum verbum permutemus gratia levitatis.

[¶59] 4.145. As a matter of fact no word will be so intractable as to baffle all our attempts to find it a suitable position; but it must be remembered that when we avoid such words, we do so not to enhance the charm of our rhythm, but to evade a difficulty. I am not, however, surprised that Latin writers have paid more attention to rhythmical structure than the Athenians, since Latin words possess less correctness and charm.

> Latin: 4.145. enim ullum erit tam difficile, quod non commode inseri possit, nisi quod ut evitandis eiusmodi verbis non decorem compositionis quaerimus, sed facilitatem. tamen mirabor Latinos magis indulsisse compositioni quam Atticos, cum minus ut verbis habeant severitatis et gratiae;

[¶60] 4.146. Nor again do I account it a fault in Cicero that, in this respect, he diverged to some extent from the practice of Demosthenes. However, my final book will explain the nature of the difference between our language and that of Greece. But I must bring this book to a conclusion without more delay, since it has already exceeded the limits designed for it. To sum up then, artistic structure must be decorous, pleasing and varied.

> Latin: 4.146. nec vitium duxerim, si Cicero a Demosthene paulum ut hac parte descivit. Sed quae sit differentia nostri Graecique sermonis, explicabit summus liber. (nam finem imponere egresso destinatum modum volumini festino) debet esse honesta, iucunda, varia.

[¶61] 4.147. It consists of three parts, order, connexion and rhythm. The method of its achievement lies in addition, subtraction and alteration of words. Its practice will depend upon the nature of our theme. The care which it demands is great, but, still, less than that demanded by expression and thought. Above all it is necessary to conceal the care expended upon it so that our rhythms may seem to possess a spontaneous flow, not to have been the result of elaborate search or compulsion.

> Latin: 4.147. tres partes: ordo, coniunctio, numerus. ut adiectione, detractione, mutatione; usus pro natura rerum, quas dicimus: cura ita magna, ut sentiendi atque eloquendi prior sit; dissimulatio curae praecipua, ut numeri sponte fluxisse, non arcessiti et coacti esse videantur.