Quintilian's Institutes of Oratory
Previous Chapter

Book 1 - Chapter 5

Next Chapter
Necessity of correctness in speaking and writing § 1. On single words, 2, 3. Choice of words, 4. Barbarisms, 5-10. Barbarisms in poets and other writers, 11-17. Faults in pronunciation, 17, 18. On the aspiration, 19-21. The accents, 22-24. On ending a word with an acute accent, 25-30. Legitimate accentuation, 31-33. On the solecism, 34-37. Different kinds of solecisms, 38-41. No dual number in Latin, 42-44. Solecisms in various parts of speech, 45-51. Figures of speech, 52-54. On foreign words, 55-57. Greek words, 58-64. Compound words, 65-70. Words proper, metaphorical, common, new, 71, 72.

1. SINCE all language has three kinds of excellence, to be correct, perspicuous, and elegant (for to speak with propriety, which is its highest quality, most writers include under elegance), and the same number of faults, which are the opposites of the excellences just mentioned, let the grammarian consider well the rules for correctness which constitute the first part of grammar. 2. These rules are required to be observed, verbis aut singulis aut pluribus, in regard to one or more words. The word verbum I wish to be here understood in a general sense, for it has two significations: the one, which includes all words of which language is composed, as in the verse of Horace,

Verbaque provisam rem non invita sequentur,

"And words, not unwilling, will follow provided matter"; the other, under which is comprehended only one part of speech, as lego, scribo. To avoid this ambiguity, some have preferred the terms voces, dictiones, locutiones. 3. Words, considered singly, are either our own, or foreign, simple or compound, proper or metaphorical, in common use or newly invented.

A word taken singly is more often objectionable than faultless, for however we may express anything with propriety, elegance, and sublimity, none of these qualities arise from anything but the connection and order of the discourse, since we commend single words merely as being well suited to the matter. 4. The only good quality which can be remarked in them is their vocalitas, so to speak, called εὐφωνία (euphony). This depends upon selection, when, of two words which have the same signification and are of equal force, we make choice of the one that has the better sound.

5. First of all, let the offensiveness of barbarisms and solecisms be put away. But as these faults are sometimes excused, either from custom, or authority, or perhaps from their nearness to beauties (for it is often difficult to distinguish faults from figures of speech), let the grammarian, that so uncertain a subject of observation may deceive no one, give his earnest attention to that nice discrimination of which we shall speak more fully in the part where we shall have to treat of figures of speech. 6. Meanwhile, let an offense committed in regard to a single word be called a barbarism.

But some one may stop me with the remark, what is there here worthy of the promise of so great a work? Or who does not know that barbarisms are committed, some in writing, others in speaking? (because what is written incorrectly must also be spoken incorrectly, though he who speaks incorrectly may not necessarily make mistakes in writing). The first sort is caused by addition, curtailment, substitution, or transposition; the second by separation or confusion of syllables, aspiration, or other faults of sound. 7. But though these may be small matters, boys are still to be taught, and we put grammarians in mind of their duty. If any one of them, however, shall not be sufficiently accomplished, but shall have just entered the vestibule of the art, he will have to confine himself within those rules which are published in the little manuals of professors. The more learned will add many other instructions, the very first of which will be this, that we understand barbarisms as being of several kinds. 8. One, with reference to country, such as is committed when a person inserts an African or Spanish term in Latin composition, as when the iron ring with which wheels are bound is called canthus, though Persius uses this as a received word; as when Catullus got the word ploxenum, "a box," on the banks of the Po; and in the speech of Labienus (if it be not rather the speech of Cornelius Gallus), the word casnar, "a parasite," is brought from Gaul against Pollio; as to mastruca, "a shaggy garment," which is a Sardinian word, Cicero has used it purposely in jest. 9. Another kind of barbarism is that which we regard as proceeding from the natural disposition, when he, by whom anything has been uttered insolently or threateningly or cruelly, is said to have spoken like a barbarian. 10. The third kind of barbarism is that of which examples are everywhere abundant and which every one can form for himself, by adding a letter or syllable to any word he pleases, or taking one away, or substituting one for another, or putting one in a place where it is not right for it to be. 11. But some grammarians, to make a show of learning, are accustomed, for the most part, to take examples of these from the poets and find fault with the authors whom they interpret. A boy ought to know, however, that such forms of speech, in writers of poetry, are considered as deserving of excuse or even of praise, and learners must be taught less common instances. 12. Thus Tinca of Placentia (if we believe Hortensius, who finds fault with him) was guilty of two barbarisms in one word, saying precula instead of pergula; first, by the change of a letter, putting c for g, and secondly, by transposition, placing r before the preceding e. But Ennius, when committing a like double fault, by saying Metoeo Fufetioeo, is defended on the ground of poetic licence. 13. In prose, too, there are certain received changes, for Cicero speaks of an army of Canopitae, though the people of the city call it Canobus; and many writers have authorized Trasumennus for Tarsumennus, although there is a transposition in it. Other words suffer similar treatment; for if assentior, "I assent," be thought the proper way of spelling that word, Sisenna has said assentio, and many have followed him on analogy; or, if assentio be deemed the right method, the other form, assentior, is supported by common practice. 14. Yet the prim and dull teacher will suppose that there is either curtailment in the one case or addition in the other. I need hardly add that some forms, which taken singly, are doubtless faulty, are used in composition without blame. 15. For dua, tre, and pondo, are barbarisms of discordant gender; yet the compounds duapondo, "two pounds," and trepondo, "three pounds," have been used by everybody down to our own times, and Messala maintains that they are used with propriety. 16. It may perhaps seem absurd to say that a barbarism, which is incorrectness in a single word, may be committed in number and gender, like a solecism; yet scala, "stairs," and scopa, "a broom," in the singular, and hordea, "barley," and mulsa, "mead," in the plural, as they are attended with no change, withdrawal, or addition of letters, are objectionable only because plurals are expressed in the singular, and singulars in the plural; and those who have used gladia, "swords," have committed a fault in gender. 17. But this point, too, I am satisfied with merely noticing, that I myself may not appear to have added another question to a branch of study already perplexed through the fault of certain obstinate grammarians.

Faults which are committed in speaking require more sagacity in criticizing them, because examples of them cannot be given from writing except when they have occurred in verses, as the division of the diphthong in Europaï, and the irregularity of the opposite kind, which the Greeks call συναίρεσιν (synaeresis) and ἐπισυναλοιφήν (episynaloiphē), and we conflexio, "combination," as in the verse in Publius Varro,

Quum te flagranti dejectum fulmine Phaeton;

18.For if it were prose, it would be possible to enunciate those letters by their proper syllables. Those peculiarities, also, which occur in quantity, whether when a short syllable is made long, as in Ītaliam fato profugus, or when a long one is made short, as in Unĭus ob noxam et furias, you would not remark except in verse; and even in verse they are not to be regarded as faults. 19. Those which are committed in sound are judged only by the ear; as to the aspirate, whether it be added or retrenched in variation from common practice, it may be a question with us whether it be a fault in writing, if h indeed be a letter and not merely a mark, as to which point opinion has often changed with time. 20. The ancients used it very sparingly even before vowels, as they said aedos and ircos, and it was long afterwards withheld from conjunction with consonants, as in Graccas and triumpus. But suddenly an excessive use of it became prevalent, so that choronae, chenturiones, praechones, are still to be seen in certain inscriptions, on which practice there is a well-known epigram of Catullus. 21. Hence there remain, even to our times, vehementer, conprehendere, and mihi. Among the ancient writers, also, especially those of tragedy, we find in old copies mehe for me.

22. Still more difficult is the marking of faults in respect to the tenores, "tones" (which I find called by the old writers tonores, as if, forsooth, the word were derived from the Greeks who call them τόνους (tonos)), or accents, which the Greeks call προσῳδίας (prosodiai) when the acute is put for the grave, or the grave for the acute; as if, in the word Camillus, the first syllable should receive the acute accent; 23. or if the grave is put for the circumflex, as when the first syllable of Cethegus has the acute, for thus the quantity of the middle syllable is altered; or if the circumflex is put for the grave, as when the second syllable is circumflexed in Appi, by contracting which from two syllables into one, and then circumflexing it, people commit two errors. 24. But this happens far more frequently in Greek words, as Atreus, which, when I was young, the most learned old men used to pronounce with an acute on the first syllable, so that the second was necessarily grave, as was also that of Tereus and Nereus. Such have been the rules respecting accents. 25. But I am quite aware that certain learned men, and some grammarians also, teach and speak in such a manner as to terminate a word at times with an acute sound, for the sake of preserving certain distinctions in words, as in circum in these lines,

Quae circum litora, circum
Piscosos scopulos

lest, if they make the second syllable in circum grave, a circus might seem to be meant, not a circuit. 26. Quantum and quale, also, when asking a question, they conclude with a grave accent; when making a comparison, with an acute; this is a practice, however, which they observe almost only in adverbs and pronouns, in other words they follow the old custom. 27. To me it appears to make a difference that in these phrases we join the words, for when I say circum litora, I enunciate the words as one, without making any distinction between them; thus one syllable only, as in a single word, is acute. The same is the case in this hemistich,

Trojae qui primus ab oris.

28. It sometimes happens, too, that the law of the meter alters the accent: as,

Pecudes, pictaeque volucres;

For I shall pronounce volucres with an acute on the middle syllable, because, though it be short by nature, it is long by position, that it may not form an iambus, which a heroic verse does not admit. 29. But these words, taken separately, will not vary from the rule, or if custom shall triumph, the old law of the language will be abolished. The observation of which law is more difficult among the Greeks (because they have several modes of speaking, which they call dialects, and because what is wrong in one is sometimes right in another). But among us, the principle of accentuation is very simple. 30. For in every word the acuted syllable is confined within the number of three syllables, whether those three be the only syllables in the word or the three last; and of these, the acuted syllable is either the next, or next but one, to the last. Of the three syllables of which I am speaking, moreover, the middle one will be long, acute, or circumflex; a short syllable in that position will, of course, have a grave sound and will accordingly acute the one that stands before it, that is, the third from the end. 31. But in every word there is an acute syllable, though never more than one; nor is that one ever the last, and consequently in dissyllables it is the first. Besides, there is never in the same word one syllable circumflexed and another acuted, for the same syllable that is circumflexed is also acuted; neither of the two, therefore, will terminate a Latin word. Those words, however, which consist but of one syllable will be either acuted or circumflexed that there may be no word without an acute.

32. In sounds also occur those faults of utterance and pronunciation, of which specimens cannot be given in writing; the Greeks, who are more happy in inventing names, call them iotacisms, lambdacisms, ἰσχνότητες (ischnotētes), and πλατείασμοι (plateiasmoi); as also κοιλοστομία (koilostomia), when the voice is heard, as it were, in the depths of the throat. 33. There are also certain peculiar and inexpressible sounds, for which we sometimes find fault with whole nations. All the incorrectnesses, then, which we have mentioned above, being removed, there will result that which is called ὀρθοέπαι (orthoepia), that is, a correct and clear utterance of words with an agreeableness of sound; for so may a right pronunciation be termed.

34. All other faults arise out of more words than one; among these faults is the solecism, though about this also there has been controversy. For even those who admit that it lies in the composition of words, yet contend that, because it may be corrected by the amendment of a single word, it is the incorrectness of a word and not a fault in composition; 35. since, whether amarae corticis or medio cortice constitutes a fault in gender (to neither of which do I object, Virgil being the author of both; but let us suppose that one of the two is incorrect), the alteration of one word, in which the fault lay, produces correctness of phraseology, so that we have amari corticis or mediâ cortice. This is a manifest misrepresentation, for neither of the words is wrong, taken separately, but the fault lies in them when put together, and it is a fault therefore of phrase. 36. It is, however, a question of greater sagacity whether a solecism can be committed in a single word, as if a man, calling one person to him, should say venite, or, sending several away from him, should say abi, or discede; or, moreover, when an answer does not agree with the question, as if to a person saying quem vides? you should reply ego. Some also think that the same fault is committed in gesture when one thing is signified by the voice and another by a nod or by the hand. 37. With this opinion I do not altogether agree, nor do I altogether dissent from it, for I allow that a solecism may occur in one word, but not unless there be something having the force of another word to which the incorrect word may be referred. A solecism arises from the union of things by which something is signified or some intention manifested, and, that I may avoid all cavilling, it sometimes occurs in one word, but never in a word by itself.

38. But under how many, and what forms, the solecism occurs, is not sufficiently agreed. Those who speak of it most fully make the nature of it fourfold, like that of the barbarism, so that it may be committed by addition (as nam enim, de susum, in Alexandriam); by retrenchment (as Ambulo viam, Aegypto venio, ne hoc fecit); 39. by transposition, by which the order of words is confused (as, Quoque ego, Enim hoc voluit, Autem non habuit). Whether igitur, placed at the beginning of a phrase, ought to be included may be a matter of dispute, because I see that eminent authors have been of opposite opinions as to the practice, it being common among some, while it is never found in others. 40. These three sorts of irregularity some distinguish from the solecism, and call a fault of addition "a pleonasm," of retrenchment "an ellipsis," of inversion "an anastrophe," and allege that if these fall under the head of solecism, the hyperbaton may be included under the same title. 41. Substitution is without dispute when one thing is put for another; it is an irregularity which we find affecting all the parts of speech, but most frequently the verb, because it has most modifications. Accordingly, under the head of substitution, occur solecisms in gender, tense, persons, moods, (or states or qualities, if any one wish that they should be so called), being six, or, as some will have it, eight in number (since into however many forms you distinguish each of the parts of speech of which mention has just been made, there will be so many sorts of errors liable to be committed), as well as in numbers, of which we have the singular and plural, the Greeks also the dual. 42. There have, indeed, been some who assigned us also a dual, scripsere, legere, a termination which was merely a softening for the sake of avoiding roughness of sound, as, among the old writers, male merere for male mereris. What they call the dual consists in that one sort of termination only, whereas among the Greeks it is found not only through almost the whole system of the verb, but also in nouns, though even so the use of it is very rare. 43. But in no one of our authors is this distinction of ending to be discovered; on the contrary, the phrases, Devenere locos and Conticuere omnes and Consedere duces show us plainly that no one of them refers to two persons only; dixere, too, though Antonius Rufus gives it as an example of the contrary, the crier pronounces concerning more advocates than two. 44. Does not Livy, also, near the beginning of his first book, say, Tenuere arcem Sabini, and a little afterwards, In adversum Romani subiere? But whom shall I follow in preference to Cicero, who, in his Orator, says, "I do not object to scripsere, though I consider scripserunt to be preferable"?

45. In appellative and other nouns, likewise, the solecism shows itself in regard to gender and to number, but especially to case. Whichsoever of those three shall be put in the place of another, the error may be placed under this head, as also incorrectnesses in the use of comparatives and superlatives, as well as cases in which the patronymic is put for the possessive or the contrary. 46. As to a fault committed in regard to quantity, such as magnum peculiolum, there will be some who will think it a solecism, because a diminution is used instead of the integral word, but for my own part, I doubt whether I should not rather call it a misapplication of a word, for it is a departure from the signification. The impropriety of a solecism is not an error as to the sense of a word, but in the junction of words. 47. In respect to the participle, errors are committed in gender and case, as in the noun; in tense, as in the verb; and in number, as in both. The pronoun, also, has gender, number, and case, all of which admit mistakes of this kind. 48. Solecisms are committed, too, and in great numbers, as to parts of speech, but it is not enough merely to remark this generally, lest the pupil should think a solecism committed only where one part of speech is put for another, as a verb where there ought to have been a noun, or an adverb where there ought to have been a pronoun, and the like. 49. For there are some nouns cognate, as they say, that is, of the same kind, in regard to which he who shall use another species than that which he ought to use, will be guilty of no less an error than if he were to use a word of another genus. 50. Thus an and aut are both conjunctions, yet you would be incorrect in asking, hic, aut ille, sit? Ne and non are both adverbs, yet he who should say non feceris for ne feceris would fall into a similar error, since the one is an adverb of denying, the other of forbidding. I will add another example: intro and intus are both adverbs of place, yet eo intus and intro sum are solecisms. 51. The same faults may be committed in regard to the different sorts of pronouns, interjections, and prepositions. The discordant collocation of preceding and following words, also, in a sentence of one clause, is a solecism.

52. There are expressions, however, which have the appearance of solecisms and yet cannot be called faulty, as tragoedia Thyestes, ludi Floralia, and Megalesia, for though these modes of expression have fallen into disuse in later times, there was never any variation from them among the ancients. They shall therefore be called figures, which are more common among the poets, but allowable also to writers and speakers in prose. 53. But a figure will generally have something right for its basis, as I shall show in that part of my work which I just before promised. Yet what is now called a figure will not be free from the fault of solecism, if it be used by any one unknowingly. 54. Of the same sort, though, as I have already said, they have nothing of figure, are names with a feminine termination which males have, and those with a masculine termination which females have. But of the solecism I shall say no more, for I have not undertaken to write a treatise on grammar, though, as grammar met me in my road, I was unwilling to pass it without paying my respects to it.

55. In continuation, that I may follow the course which I prescribed to myself, let me repeat that words are either Latin or foreign. Foreign words, like men, and like many of our institutions, have come to us, I might almost say, from all nations. 56. I say nothing of the Tuscans, Sabines, and Praenestines, for though Lucilius attacks Vectius for using their dialect, as Pollio discovers Patavinity in Livy, I would consider every part of Italy as Roman. 57. Many Gallic words have prevailed among us, as rheda, "a chariot," and petorritum, "a four-wheeled carriage," of which, however, Cicero uses one, and Horace the other. Mappa, "a napkin," too, a term much used in the circus, the Carthaginians claim as theirs; and gurdus, a word which the common people use for foolish, had, I have heard, its origin in Spain. 58. But this division of mine is intended to refer chiefly to the Greek language, for it is from thence that the Roman language is, in a very great degree, derived, and we use even pure Greek words where our own fail, as they also sometimes borrow from us. Hence arises the question, whether it is proper that foreign words should be declined with cases in the same way as our own. 59. If you meet with a grammarian who is a lover of the ancients, he will say that there should be no departure from the Latin method, because, as there is in our language an ablative case, which the Greeks have not, it is by no means becoming for us to use one case of our own and five Greek cases. 60. And he would also praise the merit of those who studied to increase the resources of the Latin language, and asserted that they need not introduce foreign practices; under the influence of which notion they said Castorem, with the middle syllable long, because such was the case with all our nouns whose nominative case ends in the same letters as Castor; and they retained the practice, moreover, of saying Palaemo, Telamo, and Plato (for so Cicero also called him), because they found no Latin word that terminated with the letters -o and -n. 61. Nor did they willingly allow masculine Greek nouns to end in as in the nominative case, and accordingly, we read in Caelius, Pelia Cincinnatus; in Messala, Bene fecit Euthia; in Cicero, Hermagora; so that we need not wonder that the forms Aenea and Anchisa were used by most of the old writers: for, said they, if those words were written as Maecenas, Suffenas, Asprenas, they would end in the genitive case, not with the letter -e, but with the syllable -tis. 62. Hence, to Olympus and tyrannus they gave an acuted middle syllable because our language does not permit the first syllable of a word, if short, to have an acute accent when two long syllables follow. 63. Thus the genitive had the forms Achilli and Ulixi, and many others similar. The modern grammarians have now made it a practice rather to give Greek declensions to Greek nouns, a practice which cannot, however, always be observed. For myself, I prefer following the Latin method, as far as propriety allows, for I would not now say Calypsonem, like Junonem, though Caius Caesar, following the older writers, uses this mode of declining. 64. But custom has prevailed over authority. In other words, which may be declined without impropriety in either way, he who shall prefer to use the Greek form will speak, not indeed like a Roman, but without incurring blame.

65. Simple words are what they are in their first position, that is, in their own nature. Compound words are either formed by subjoining words to prepositions, as innocens (care being taken that there be not two prepositions inconsistent with each other, as imperterritus, otherwise two may be at times joined together, as incompositus, reconditus, and, a word which Cicero uses, subabsurdum); or they coalesce, as it were, from two bodies into one, as maleficus. 66. For to form words out of three constituent parts, I should certainly not grant to our language; though Cicero says that capsis is compounded of cape si vis, and some are found to maintain that Lupercalia also consists of three parts of speech, luere per caprum. 67. As to solitaurilia, it is now believed that it is for suovetaurilia, and such indeed is the sacrifice, as it is described also in Homer. But these words are constructed, not so much of three words, as of parts of three words. Pacuvius, however, appears to have formed compounds, most inelegantly of a preposition and two other words:

Repandirostrum, incurvicervicum pecus,

"The broad-nosed, crook-necked flock of Nereus." 68.Compounds, however, are formed either of two entire Latin words, as superfui, subterfugi (though it is a question whether these are indeed formed of entire words), of an entire and incomplete word, as malevolus; of an incomplete and entire word, as noctivagus; of two incomplete words, as pedissequus; of a Latin and a foreign word, as biclinium; of a foreign and a Latin word, as epitogium and Anticato; or of two foreign words, as epirhedium, for though the preposition epi- is Greek, and rheda Gallic, and though neither the Greek nor the Gaul uses the compound, yet the Romans have formed their word of the two foreign words. 69. Frequently, too, the union causes a change in the prepositions, as abstulit, aufugit, amisit, though the preposition is merely ab-, and coit, the preposition being con-; and so ignavi, erepti, and similar compounds. 70. But the composition of words in general is better suited to the Greeks; with us it is less successful, though I do not think that this results from the nature of the lauguage; but we look with more favor on foreign compounds, and, accordingly, while we admire the Greek κυρταύχενα (kurtauchena), we hardly defend incurvicervicum from derision.

71. Words are proper when they signify that to which they were first applied; metaphorical, when they have one signification by nature, and another in the place in which they are used. Common words we use with greater safety; new ones we do not form without some danger; for if they are well received, they add but little merit to our style, and, if rejected, they turn to jokes against us. 72. Yet we must make attempts; for, as Cicero says, even words which have seemed harsh at first, become softened by use.

As to the onomatopoeia, it is by no means granted to our language; for, if we should venture to produce anything like those justly admired Greek expressions λίγξε βιός (linxe bios), "the bow twanged," and σίζε ὀψθαλμός siz ophthalmos, "the eye hissed," who would endure it? We should not even dare to say balare, "to bleat" or hinnire, "to neigh," unless those words were supported by the sanction of antiquity.

Previous Chapter
Lee Honeycutt (honeycuttlee@gmail.com) Last modified:1/15/07
Next Chapter