(Metam. xiii. 607.) Between two words the vowel is rarely lengthened, except in the arsis of a verse. The last syllable of a word thus remains short, e. g. in Horace at the beginning of an hexameter:-quem mală stultitia aut; or at the end: praemiă scribae. An instance in which the vowel is lengthened by the accession of the arsis occurs in Virgil, Bucol. iv. 51.: Terrasque tractusque maris coelumque profundum.

Qu is not accounted as two consonants, for u is not a true consonant, though we usually pronounce it as such. But j alone is sufficient to make position, because this consonant was pronounced double (in early times it was also written double); e. g. mājor like major, and in like manner in ējus and Trōja. In the compounds of jugum alone it does not lengthen the preceding vowel, as bijugus, quadrijugus, nor does it, according to the rule mentioned above, lengthen the vowel when it begins a new word, and the preceding word ends in a short vowel, as in the hexameter of Virgil (Georg. i. 125.): —Antě Jovem nulli subigebant arva coloni.

Note. The determination of the quantity of a vowel before muta cum liquida within a word has great difficulties, and we must add the following observations. The practice of the different poets varies greatly. Virgil, e. g., is particularly fond of lengthening a vowel by its position before muta cum liquida; and he and the poets in general usually contrive to make the vowel thus lengthened coincide with the arsis in the verse; by the same contrivance he also lengthens the short final syllable of a word, especially the enclitic que, in the second foot of an hexameter, by the muta cum liquida which follow it. We have further to observe particular words which have their vowel short, viz. liber, niger, piger, and ruber; but in their inflections, where the muta cum liquida occurs, the vowel almost always becomes long; coluber, e. g., is short; but colūbrae, colūbris, are long, and migro is made long by the best poets in the hexameter. Other words however are either never lengthened, as arbitror, or very seldom, as locuples. There are, on the other hand, some cases of muta cum liquida, which form a strong position both in Latin and Greek, viz. where the liquid is either l, m, or n, and the mute either b, g, or d. (See Buttmann's Greek Grammar, § 7. 10.) Thus the Latin words publicus, agmen, regnum and ignarus, always have their first syllable long.

It is almost superfluous to repeat here, that we are speaking only of such vowels as are naturally short; for, when the vowel is naturally long, a lengthening by positio debilis is out of the question, and we therefore always say ambulacrum, lavācrum, delūbrum, involūcrum and salubris. When the consonants muta cum liquida belong to different syllables, as in ab-luo, ob-ruo, quam-ob-rem, they make real position.

CHAP. IV.

OF THE ACCENT OF WORDS.

[§ 32.] 1. IT is a general rule that every word has an accent on one particular syllable. This accent is twofold, either the circumflex (4), or the acute ('), for what is called the grave in Greek means only the absence of either accent. Some words have no accent, viz. the enclitics ne, que, ve, ce, which never appear by themselves, but are attached to other words. Prepositions lose their accent when they precede the cases which they govern.

Note. The addition of these enclitics produces a change in the accent of the words to which they are attached, and which thus become compounds. The ancient grammarians have established the rule, that whenever an enclitic has a meaning of its own, the accent is thrown back upon the syllable immediately before the enclitic, and either as the acute (if the vowel of that syllable is short), or as the circumflex (if the vowel is long), as in Musáque (nominat.) hominéque, and Musâque (ablat.) armisque. When, on the other hand, the enclitic has no meaning by itself, and forms only one word with that to which it is attached, the accent varies, as will be shown hereafter. This is the case with que; for in some compounds it either does not possess the meaning of "and" at all, or only very indistinctly. Hence in itáque (and so) the accent belongs to the short penultima, and in ítaque (therefore), in which the meaning of "and" is quite obscured, the pronunciation places the accent upon the antepenultima. In the same manner we have to distinguish between utique (and that) and útique (certainly). By way of exception the same grammarians place the accent on the penultima in utráque and pleráque, on account of the accent of the masculine forms utérque and plerique, although according to the general rule, que not meaning "and," we ought to pronounce útraque and pléraque. They further inform us that we should pronounce néquando and síquando, in order that quando may not be taken for a separate word, and alíquando in order to distinguish it from aliquánto.

[§ 33.] 2. Monosyllables are pronounced with the circumflex, when their vowel is long by nature and not merely by position, as in dôs, mós, flôs, jûs, lûx, spês, fôns and môns; but when the vowel is naturally short, they are pronounced with the acute, although the syllable may be long by position; e. g. árs, párs, fáx, dúx.

Note. Sic (so) the adverb should be pronounced with the circumflex, and sic, which indicates a wish, with the acute; e. g. Síc te, diva potens Cypri, &c. in Horace Comp. Priscian, De XII. Vers. Æn.

3. Words of two syllables have the accent on the first, either as circumflex, when the vowel of that syllable is naturally long, and that of the second one short; or as acute, when the vowel of the first syllable is short and that of the second long; or when the vowel of the first as well as that of the second is long; e. g. Rômă, mūsă, lûce, jûris; but hómo because both syllables are short; déas, because the first is short and the second long; árte, because the first is long only by position; and dóti, for although the vowel of the first is naturally long, yet that of the second is likewise long. The ancient grammarians do not notice those cases where a syllable long by position is at the same time long by the nature of its vowel (see above, § 30.); but it is probable that consul, mônte, dênte, êsse (for edere), asthma and sceptrum, were pronounced in the same manner as lûce.

4. Words of three syllables may have the accent on the antepenultima and penultima; the acute on the antepenultima, when the penultima is short, as in caédere, pérgere, hómines; the accented syllable itself may be long or short. The circumflex is placed on the penultima on the conditions before-mentioned, as in amâsse, Românus; and the acute, when those conditions do not exist, and yet the penultima is long, as in Románis, Metéllus. No word can have the accent further back than the antepenultima, so that we must pronounce Constantinópolis, sollicitudinibus.

Note. Priscian (p. 803. ed. Putsch.) remarks as an exception, that the compounds of facere, which are not formed by means of a preposition, such as calefacit, tepefăcit, and (p. 739.) the contracted genitives in i instead of ii (see § 49.), have the accent on the penultima, even when it is short, as in ingeni, Valeri, so that we must pronounce calefácit, ingéni. He asserts the same with regard to the vocative of proper names in ius, e. g. Virgili, Valeri; while other grammarians (A. Gellius, xiii. 25.) leave to this case its regular accentuation, Virgili and not Virgili.

[§ 34.] 5. Words of two or more syllables never have the accent on the last, and it appears that it was only the grammarians who invented a different mode of accentuation for the purpose of distinguishing words which would otherwise sound alike. They tell us that the words poné (behind) and ergó (on account of) should have the accent on the last syllable, to distinguish them from pône (put) and érgo (therefore). They further accentuate the last syllables of the adverbs circum, docte, raro, primo, solum, and modo, to distinguish them from the cases which have.

the same terminations. The interrogatives quando, qualis, quantus, ubi, and others, are said to have the accent on the first syllable, according to the rule; but when used in the sense of relatives, to have the accent on the last syllable, unless the acute be changed into the grave by reason of their connection with other words which follow. The words ending in as which originally ended in atis, such as optimas, nostras, Arpinas, are said to have the accent on the syllable on which they had it in their complete form, and which is now the last. The same is asserted with regard to the contracted perfects, such as audit for audivit. It is impossible to determine how much of all this was really observed by the ancients, since it is expressly attested by earlier writers, such as Quintilian, that in Latin the accent was never put on the last syllable. But it is certainly wrong to put the grave on the last syllable of all adverbs, as some persons still do, or to use accents for the purpose of indicating the natural length of a vowel, which is better expressed by a horizontal line (-).

[§ 35.] 6. These rules concerning accentuation ought to lead us to accustom ourselves to distinguish accent from quantity; to read, for example, hóminēs and not hōmines, and to distinguish in our pronunciation édo (I eat) from ēdo (I edit), légo (I read) from lego (I despatch), and in like manner fúris (thou ravest), légis (thou readest) and régis (thou rulest) from the genitives fūris, rēgis and legis; further, lévis (light) from levis (smooth), málus (bad) from mâlus (an apple-tree), pálūs, ūdis (a marsh), from pâlus, i (a post), ánus (an old woman) from ânus (πρwêтós), lútum (mud) from lūtum (a dyer's weed), and also lū'teus (dirty or muddy) from luteus (yellow), and populus (the people) from pō'pulus (a poplar). In our own language accent and quantity coincide, but it is very wrong to apply this peculiarity to a language to which it is foreign.

25

THE ACCIDENCE.

CHAP. V.

DIVISION OF WORDS ACCORDING TO THEIR SIGNIFICATION.

[§ 36.] THE words of every language are either nouns, verbs, or particles.

A noun serves to denote an object or a quality of an object, and may accordingly be either a substantive, as domus (a house), a pronoun, as ego (I), or an adjective, as parvus (small). Nouns are declined to indicate their different relations.

A verb expresses an action or condition which is ascribed to a person or a thing, as scribo, ire, dormire, amari. A verb is conjugated in order to indicate the different modes in which an action or condition is ascribed to a person or a thing.

Particles are those parts of speech, which are neither declined nor conjugated, and which are neither nouns nor verbs. They are divided into the following classes. 1) Adverbs express the circumstances of an action or condition, as scribit bene, he writes well; diu dormit, he sleeps long. 2) Prepositions express, either directly or indirectly (§ 295.), the relations of persons or things to one another or to actions and conditions; as, amor meus erga te, my love towards thee; eo ad te, I go to thee. 3) Conjunctions express the connexion between things, actions, or propositions; as, ego et tu; clamavit, sed pater non audivit. 4) Interjections are the expressions of emotion by a single word; as ah, ohe,

vae.

These are the eight parts of speech in Latin; all of them occur in the following hexameter:

Vae tibi ridenti, quia mox post gaudia flebis.