A Short Historical Latin Grammar: Chapter III: Noun

English Grammar Advanced 19th century W. M. Lindsay

Pages 37–62 of the book.

The I.-Eur. Noun had three Genders, Masc., Fem., and Neuter; three Numbers, Sing., Dual, and Plural; and at least eight cases, Nom., Gen., Dat., Acc., Voc., Abl., Instrumental, and Locative. The Dat., Abl., Instr., and Loc. were not always distinguished in the Plural (cf. Athenae, ‘for A.’, ‘from A.’, ‘with A.’, ‘at A.’).

The three Genders are retained in Latin, but of the Numbers the Dual has disappeared. Traces of it remain in the Numeral duo Nom., duobus Dat., duo, afterwards duos, Acc. (on octo see ch. iv. § 5), and in the Pronoun ambo Nom., ambobus Dat., ambo, afterwards ambos, Acc. (cf. Greek dyo, older dywo, and ampho, both of which are Duals).

Of the Cases the Voc. hardly survives, except in the Singular of the Second Declension (e.g. domine Voc., dominus Nom.), for elsewhere the Nom. is used instead; and even in the Second Declension we find, e.g. puer Nom. used as Voc. in classical Latin, though in the Latin of Plautus’ time the Voc. is puere (on deus, etc., Voc. see § 6). The Instrumental may survive in some Adverbs like modo, cito, though this is doubtful (ch. vii. § 1). The Locative Sing. of the First Declension, ending in -ai (a long diphthong), which passed in time into -ae, was identical in form with the Gen. and Dat. Sing., and came to be looked upon as a Genitive case (e.g. Romae, at Rome, had the same form as Romae, of Rome, and was called a Genitive); in the Second Declension it ended originally in -ei, later -i, and was used as a Genitive as well as a Locative, passing under the name of Genitive (e.g. Corinthi, at Corinth, had also the sense ‘of Corinth’); in the Third Declension it ended originally in -i, later -e, and was used as an Ablative, as well as a Loc., under the name of Ablative (e.g. Carthagine, at Carthage, had also the sense ‘from Carthage’). (On these and other traces of the Locative see §§ 4, 6, 8, 10, 11, 13.)

The I.-Eur. Cases were indicated sometimes by the addition of suffixes, e.g. -s for Nom. Sg. Masc., -m for Acc. Sg., sometimes by a modification of the stem (1), e.g. pater, Nom. Sg. of the stem pater-, ‘a father,’ sometimes by both, e.g. patr-os or pater-es Gen. Sg., patr-om Gen. Pl. of the same stem. In these Genitives we see not merely the addition of the suffix -os or -es (Gen. Sg.) and the suffix -om (Gen. Pl.), but also the modification of the stem from pater- to patr-. This modification of the stem played a great part in the I.-Eur. Declension, but has been effaced in Latin by the natural tendency to make one case like another in everything but the suffix. Thus the Acc. Sg. of pater was originally paterem, but has become patrem on the analogy of patr-is, patr-i, patr-e, as in Homer we find the Gen. Sg. mētros on the analogy of mētera, Acc.

(1) By 'stem' is meant that part to which the case-suffixes are added to form the different cases.

By ‘Heteroclite’ Declension is meant the appropriation of different stems to different cases. An example of I.-Eur. Heteroclite Declension is the declension of certain Neuter Nouns (1), which had an R-stem in the Nom. and Acc. Sg., but an N-stem in the Gen., Dat., etc., and Latin femur Nom., femoris Gen. retains this ancient type. Examples of Latin Heteroclite Declension are the declension of (2) iter, a journey, where the stem iter- is used in the Nom. and Acc. Sg., but the stem itiner- for the other cases, e.g. itineris Gen., itineri Dat. (2); senex, an old man, with the stem senec- (cf. senec-io, an old man, Senec-a, ch. xi. § 11) in the Nom. Sg. and the stem seni- (or sen-) in the other cases (3); supellex, furniture, a Fem. Noun (or rather an Adjective agreeing with res) derived from sup(er)- and lego, I lay, with the stem supelleg- (for super-leg-, ch. xi. § 19) in the Nom. Sg. and the stems supellectil- (for super-lectil-, this lectilis being Verbal Adj. from lego as coctilis, cookable, from coquo) in the other cases. The rarity of I-stems caused the I-stem vis to be declined like the numerous S-stems (ch. xi. § 19) in the Plural, vires, etc., though we have vis, the older Nom., Acc. Pl. even in Lucretius (iii. 265 multae vis; ii. 586 vis multas).

(1) Sanguis, M. may be a relic of a Neuter I-stem with -i in the Nom. and -n- in the other cases. Hence the confusion between its I- and its N-stem. In Lucr. we have sanguen Neut. (i. 837, etc.) and sanguis (for sanguin-s) Nom. (iv. 1050); sangui Abl. in Ennius.

(2) The stem itiner- seems to be a patchwork of the E-stem of the Nom. iter- and the N-stem of the Gen. iten-. So jecinoris, beside jecoris, from jecur. In the older poetry we find that the classical usage has not yet established itself, e. g. ignoti iteris sumus, 'we do not know the road,' Naev.; and even Lucretius has concussos itere (v. 653) auditinerque sequatur (tI. 339).

(3) Priscian quotes from Plautus a Gen. Sg. senicis.

‘Defective’ Nouns are for the most part obsolete Nouns, isolated cases of which have been preserved in adverbial phrases, e.g. formae causa, ‘for form’s sake,’ secus (ch. vii.), forte, by chance (cf. forsitan, ch. viii. § 35), sponte, willingly. Instar, an equivalent (e.g. instar montis equus ‘a horse as large as a hill’, connected with insto, to be steady, in equipoise — of a balance, is perhaps an Inf., instare (cf. exemplare, ch. ii. § 16), used as a Noun.

The reason why certain nouns are masculine, others feminine, and others neuter in Latin is not always to be found. In I.-Eur. the names of fruits were neuter, and so in Latin, e.g. malum, an apple (Gk. μῆλον), while the name of the mother-tree that bears the fruit is feminine in Latin, e.g. malus, an apple-tree (Gk. μηλέα). The names of the months and winds are masculine in Latin, because they are really Adjectives agreeing with mensis or ventus, e.g. Ianuarius (mensis), Februarius (mensis), Martius (mensis), Auster (ventus), Caurus (ventus), Favonius (ventus). The reason why the names of rivers are masculine in Latin seems to be that fluvius (cf. amnis) is masculine, just as they are masculine in Greek because ποταμός, a river, is masculine.

But the proneness of Nouns to take a new Gender by analogy of a Noun which had the same termination or a kindred meaning, or with which they were often joined in speech, made great confusion among the original Genders of Latin. Thus the occasional feminine Gender of dies may be due to the analogy of nox, a word with which it is frequently joined, e.g. dies noxque maesta, dies noxque longa, or to the fact that most Nouns with the termination -ies are feminine.

The Fem. Gender was originally associated with Abstract Nouns, e.g. Lat. optio, choice. But if an Abstract Noun came to be used as a Concrete, it would change its gender; and so optio, in the sense of ‘a centurion’s assistant,’ was masculine. Similarly agricola, literally ‘field-tillage’ (from ager, a field, and colo, I till) became masculine in the sense of ‘a field-tiller’; and this is the reason why A-stems (i.e. First Declension Nouns), which were associated with the Fem. Gender in I.-Eur., are often masculine in Latin. O-stems (i.e. Second Declension Nouns) were similarly associated with the masculine (with Nom. Sg. in -os, classical Lat. -us) and the neuter gender (with Nom., Acc. Sg. in -om, class. Lat. -um); but we have a few feminine O-stems in Latin, e.g. names of trees like malus. In Latin these feminine O-stems show a tendency to pass into the Fourth Declension, e.g. domus, a Second Decl. Noun till Sulla’s time (§ 6). The confusion of masculine and neuter O-stems may be illustrated by the words collum, the neck, which in Plautus is collus, and which was masculine in I.-Eur. (cf. Germ. Hals, ‘the neck,’ Masc.), and uterus, which in Plautus as in I.-Eur. is neuter, uterum. The heteroclite declension of masculine nouns like locus, with neuter plural loca, is usually the result of a primitive practice of using a Fem. Sg. Collective A-stem as a plural of a masculine O-stem. Thus loca, originally a Fem. Sg. meaning ‘a collection of spots,’ ‘a district,’ was used as the plural of locus, ‘a single spot,’ and received the declension of a plural, locorum Gen., locis Dat., etc. A good parallel is supplied by the Greek mēra, meaning ‘a mass of thigh-bones,’ which was originally a Fem. Sg. Collective, but was treated in common use as if it were a plural of mēros, ‘a single thigh-bone.’ The origin of the neut. Pl. in -a from a collective Fem. Sg. explains the use in Greek (and the earliest Sanscrit) of a singular verb with a neuter plural noun, e.g. μἠρα καίεται.

Some nouns are from their nature confined to the Singular or to the Plural Number. Abstract Nouns are naturally Singular, though the Plural is often found in a concrete sense (e.g. opera, work, exertion, operae, workmen; auxilium, help, auxilia, auxiliary troops), occasionally with transference of this concrete sense to the Singular (e.g. accedes opera agro nona Sabino, ‘you will be thrown in as ninth hand on my Sabine farm,’ Hor. Sat. ii. 7. 118). The use of the Plur. in certain words dates from Indo-European times, e.g. tenebrae, and in names of parts of the body, e.g. praecordia (cf. Gk. φρένες), cervices (cervix was a poetic innovation). Often a Noun has a different sense in the Singular and in the Plural, e.g. aedes, Sg. a single room, hence ‘a temple,’ Pl. a house (cf. Homer’s δόμος and δόμοι). Liberi is Plur. of the Adj. liber, ‘the freeborn children of the house,’ as opposed to the vernae.

The Latin First Declension consists of A-stems. These A-stems were, as we saw in the last paragraph, originally feminine, but sometimes became masculine when they changed their meaning from an abstract to a concrete sense. Agricola, from ager and colo, indicated originally the abstract idea of ‘field-tillage,’ and in this sense was feminine, like fuga, flight (Greek φυγή), or any other abstract noun. It became masculine when it passed to the concrete sense of ‘a man who tills the fields,’ by the same transition as is seen in our words ‘relation,’ ‘youth,’ which meant originally ‘relationship,’ ‘youthhood,’ then secondly ‘a related person,’ ‘a youthful person.’ In Greek these masc. concrete A-stems took an -s in the Nom. Sg. like O-stems, e.g. νεανίας, a youth (from a lost νεανία, the period of youth), with the same ending as νέος, young; and in two very early Latin words we seem to have a trace of a similar usage, viz. hosticapas, a capturer of the enemy, and paricidas, or parricidas (1), a murderer.

But if Masc. Ā-stem Nominatives in -as ever existed beside Fem. Ā-stem Nominatives in -a in Latin, they had dropt out of use as early as the time of Plautus, for with him, as with all subsequent writers, the Nom. Sg. ending of masculine and feminine Ā-stems alike is -a. This -a must have been originally long (cf. Gk. χώρᾱ), but had been shortened at a very early period, first in iambic words like fŭga, mŏra (see ch. ii. § 16), then by their analogy in all words. Even in the earliest poetry there is no trace of -ā (2).

The Gen. Sg. ended originally in -ās, an ending which is common in feminine Nouns in the early poetical language (3), and which, owing to the conservative spirit of Roman Law, was retained in the legal phrase pater-familias, lit. ‘head of the household.’ Masculine Ā-stems, following the analogy of O-stems (2 Decl.), took the ending -ī, an ending which they added to the final -ā of their stem, e. g. agricolā-ī, and this ending came to be assumed by feminine Ā-stems too as early as the time of Plautus (cf. § 5). Through the Roman habit of shortening a long vowel before another vowel (ch. ii. § 16), -āī became -ăi, and the two vowels were united into a diphthong -ai, class. ae (ch. x. § 11).

(1) The word is spelt with one r in the ancient spelling, in which a double consonant was never written double (ch. i. § 8).

(2) Ennius’ ‘Et densis aquila pinnis obnixa volabat’ is a case of lengthening in arsi like his ‘Sic expectabat populus atque ora tenebat.’ This lengthening in arsi was an imitation of Homer’s prosody.

(3) Examples are Latonas, escas, Monetas, Liv. Andronicus; Terras, fortunas, Naev.; vias Enn. Even in Virgil the commentator Servius favours the reading auras for aurae in Aen. xi. 801.

The full ending -ăĭ did not quite disappear from use in speech till the time of Terence; and poets like Lucretius and Virgil, who love to insert archaic forms in their lines for the sake of poetic effect, offer many examples of this form, e. g. aulai Aen. iii. 354; aurai, vi. 747; aquai, vii. 464. Notice the dignity which it lends to a famous passage of Lucretius (i. 84–6):

Aulide quo pacto Triviai virginis aram Iphianassai turparunt sanguine foede Ductores Danaum delecti, prima virorum.

To Martial it seemed typical of the uncouth, early Latin poetry (xi. 90. 5):

Attonitusque legis ‘terrai frugiferai, Accius et quicquid Pacuviusque vomunt.

The Dative Sg. ending was originally -āī (a long diphthong; cf. Gk. χώρᾳ). This ending, according to its position in the sentence, would assume the forms -ā or -ai (the ordinary diphthong, with the first element short); and in early inscriptions we find both these forms, e. g. IVNONEI LOUCINA and LOUCINAI, ‘to Juno Lucina,’ PROSEPNAI, ‘to Proserpine’ (1). In time, however, the second form, -ai, was preferred to the other and developed into the classical -ae.

The Locative Sg. ending was likewise -āi, which became -ae at the same time as Gen. -ai (originally -āī) became -ae. Examples of Locatives (called Genitives by the Latin writers on Grammar and their modern imitators) are Romae, at Rome, militiae, at warfare, on the field, and Plautus’ phrase for ‘next door,’ ‘at the next house,’ viz. proxumae viciniae.

(1) This was at first wrongly read PROSEPNAIS, and supposed to be a Genitive. An early Latin ‘Genitive in -ais’ is a fiction.

In the Acc. Sg. ending, -am, the a was originally long (cf. Gk. χώρᾱν), but, like all long vowels before final -m in Latin (ch. ii. § 16), became short.

The Abl. Sg. had originally -ād. This final -d, like every final d after a long vowel (ch. ii. § 9), was dropt in pronunciation before the time of Plautus, though we find it used by Naevius in his Epic:

Noctu Troiad exibant capitibus opertis, ‘cloaking their heads they passed at night-time from Troy.’

The Nom. Pl. in Latin, as in Greek, discarded the original ending -ās for an ending modelled after the Second Declension, viz. the diphthong -ai, written and pronounced in the classical period -ae. But -as remained in dialects of Italy, and some think that a dialectal Nom. Plur. of this type is found in a play of Pomponius, descriptive of life in an Italian country town:

Quot laetitias insperatas modo mi inrepserē in sinum.

The Gen. Pl. ending in I.-Eur. was -āsōm, which by the Latin laws of sound became -ārŏm, classical -ārŭm, e. g. dearum (see ch. x. § 19; ch. ii. §§ 9, 14), by the Greek, first -άων, e. g. Homeric θεάων, then -ῶν, e. g. Att. θεῶν.

The Dat., Abl., Loc., or Instr. Plur. (these cases, as we have seen, are not always to be distinguished in the Plural) had originally the ending -ais (one syllable), which became -eis (ch. ii. § 14), then -īs (ch. x. § 11). In old Latin we find another ending -ābus (cf. the ending of the Third, Fourth, and Fifth Declensions); and in legal language this was kept up in a few nouns for the sake of distinguishing them from cognate O-stems, e. g. filiabus (but filiis Masc.), deabus (but deis Masc.), libertabus (but libertis Masc.), just as we, without the same justification, keep up the old Plural suffix in a few words like ‘oxen,’ but have dropped it in the rest, e. g. ‘shoes’ (earlier ‘shoon’).

The Acc. Plur. had -ās from -āns.

Thus the scheme for the First Declension will be: —

Singular

Case	Ending	Development
Nom.	-ă	became -ā
Gen. (1)	-ăs	retained in familias
Gen. (2)	-āī	became -aī, -ae
Dat.	-āī	became -aī, -ae
Loc.	—	same as Dat.
Acc.	-ăm	became -am
Abl.	-ād	became -ā

Plural

Case	Ending	Development
Nom.	-aī	-ae
Gen.	-āsŏm	became -ārum
Dat. / Abl. / Loc. (1)	-ābus	from -ābhos; retained in filiabus, etc.
Dat. / Abl. / Loc. (2)	-ais	became -eis, then -īs
Acc.	-ās	—

Greek Proper Names sometimes retain their Greek declension in poetry, e. g. Pēlides, Acc. Pēliden, Gen. Pl. Atridum (cf. amphorum, drachmum) (1), though Cicero gives these Masculine names the O-stem Gen. in -ī, e. g. Aristidi, thus repeating the early formation of the Gen. Sing. in -a-ī (see above § 4). And in epitaphs of slaves, freedwomen, and the like, we often find the Greek Gen. Sg. in -ης expressed in Latin characters by -aes, e. g. Faustinaes, Anniaes (Greek η had the long, open vowel-sound of later Latin ae; see ch. ii. § 2).

(1) The poetic caelicolum is a Graecism.

The Latin Second Declension consists of Ō-stems, which are either Masculine (with Nom. Sg. in -ŏs, class. Lat. -ŭs), or Neuter (with Nom., Acc. Sg. in -ŏm, class. Lat. -ŭm). The few Fem. O-stems, e. g. domus, a house (Gk. δόμος Masc.); colus, a distaff (Gk. πόλος Masc.), nurūs, a daughter-in-law (Gk. νυός), and names of trees (§ 2) like mālus, an apple-tree, laurus, a laurel-tree, tend to pass into the Fourth Declension (Nom. Pl. laurīs and laurī, &c.). Pelagus, a Greek loan-word (πέλαγος, -εος) is curiously treated as a Second Declension Noun in Latin, much as words like poema in the Plural, e. g. Dat., Abl. poemātis.

In the Nom. Sg., masculine RO-stems (1) dropt at a very early period the final -ŏs, when a short vowel preceded the r, e. g. vir, not virŏs (virūs), satŭr, not satūros (satūrus), ager (cf. Gk. ἀγρός); but not when a long vowel preceded the r, e. g. sevērus, amārus. Plautus, however, uses socērus for socer and Ennius voltūrus for class. vultur, -ūris Gen., and in classical Latin we have uterus (older uterum, § 2). Compounds of gĕro, fĕro show usually -ger, -fer, but not mōrĭgerus (Plaut.), from mos and gĕro, ‘complying with a man’s humour,’ ‘complaisant.’

(1) Not LO-stems. Lucretius’ famul infumus (iii. 1035) is a direct imitation of Ennius’ famŭl ultĭmus, ‘famulus ultimus.’ Ennius seems to have borrowed the form from his native Oscan.

Other exceptional Adjectives are ferus (but semĭfer) and prŏ-perus (the latter, like pauper, from parŏ). Numĕrus, humĕrus, ērus, pīrus (cf. Junō-pīrus, lit. ‘Juno’s pear,’ like Jū-glans ‘Jove’s acorn’) are not RO-stems, for their r was originally s. RO-stems took in familiar language the ending -ĭs, as well as -ĭŏs (class. -ius), in the Nom. and in the Gen. -ī, in the Acc. -im, in the Voc. -ī. For example, in an early law of Plautus’ time, the Decree of the Senate against Bacchanalian orgies (the S. C. Bacch.), the consuls’ names are written in the ceremonious form, Marcius, Postumius, while the secretaries’ names have the other ending, Clauidi(s), Valeri(s), Minuci(s) (p. 36 n.). (On equos, ecus, equus, &c., see ch. ii. § 10.)

For the Gen. Sg. the Locative was used with the ending -ei (which would become in classical Latin -ī), but in the IO-stems with the ending -ī. This -ī of IO-stems seems to have encroached at an early period on the other. In the Augustan age it became the fashion to use -ī as the ending of the Gen. Sg. of IO-Noun-stems, e. g. praediī, consiliī, Virgiliī, as -ī had been at a previous period used in Adjectives, e. g. patrii sermonis egestas Lucr. Varro, in his edition of the plays of Plautus, found that the plays of another dramatist, Plautius, had been confused with them, owing to the similarity of the Gen. Case, Plauti fabulae.

The Dat. Sg. ending was -ōi (the long diphthong; cf. Gk. οἴκῳ), which, according to its position in the sentence, would assume the forms -ō or -oi (the ordinary diphthong). These two ‘doublets’ are both found in very early Latin (cf. Numasioi ‘Numerio’, p. 33), but the contest between them was decided before the literary period in favour of the first. (Contrast the history of the Dat. Sg. of the First Declension, § 4.)

The Voc. Sg. had -ĕ, an ending which remained unchanged, e. g. dominĕ, though we find the Nom. occasionally, e. g. deus (always), vos O Pompilius sanguis (Hor. A. P. 292). (On the variation of ĕ with ŏ in I.-Eur. declension and conjugation see ch. x. § 12.)

The Abl. Sg. ended originally in -ŏd, class. -ō (ch. x. § 17) e. g. Gnaivod ‘Gnaeo’ on a Scipio epitaph. A byform in -ēd (class. ē) was reserved for Adverbial Ablatives, e. g. facilumed ‘facillime’* on the S. C. Bacch. (see ch. vii. § 4).

The Nom. Plur. Masc. of O-stems borrowed the ending of the Pronoun Declension, viz. -oi (e. g. Gk. *οἶκοι) like οὗτοι), which in Latin became first -ei, then -ī (ch. ii. § 14). The o-diphthong is seen in a very early phrase, pilumnoi poploi (i. e. pilumni populi), ‘the tribes armed with the pilum,’ a description of the Romans in the Carmen Saliare, the ancient hymn of the leaping priests of Mars. In the Latin of Plautus’ time we find an -s added to this ending for Pronouns, e. g. heis or his (class. hī) illeis or illis (class. illī), but not for Nouns; although, on inscriptions of a century later, we find the irregular Nom. Pl. magistreis or magistris for magistri, &c.

The Nom. Acc. Neut. Plur. in -a had originally -ā, which like the -a of the Nom. Sg. of the First Declension (§ 4) had become short before the literary period. (On the origin of the Neut. Plur. in -a from a Fem. Collective Sing., see § 2.)

The proper Gen. Pl. ending -ŏm, which became -ōm (ch. ii. § 16), then -ŭm (ch. ii. § 14) in Latin, was retained in a few Genitives like deum, nummum, triumvirum. But an ending -ŏrŏm (class. -ōrŭm), originating in the Pronoun Declension, where it followed the analogy of the -ārŭm of Ā-stems (§ 4) (e. g. illorum like illarum), spread from Pronouns to Adjectives, where it was found useful in the discrimination of Gender (e. g. bonorum (1) beside bonarum), and ultimately to Nouns (e. g. filiŏrum beside filiarum). In Nouns it did not succeed in fully asserting itself against the proper ending -um till the time of Cicero, who tells us that he yielded to the new fashion in certain words, but not in others.

(1) We have duonorum(m), the old form of bonorum, in an early epitaph of one of the Scipios, and elsewhere duonorum quorom uiro, i. e. ‘bonorum optimum virum,’ where the last word may be Gen. Pl. or Acc. Sg.

He allowed pro deorum fide! or pro deum fide! but only triumvirum, sestertium, nummum, &c. (Orat. 46. 155). One of these Genitives Plural, sestertium (sc. mille), ‘a thousand (of) sesterces,’ came to be regarded as a Neuter Singular and formed a Plural sestertia.

The Dat., Abl., Loc., Instr. Plural ending was -ois (cf. Gk. οἴκοις), which became -eis (ch. ii. § 14), then -īs (ch. x. § 11). The oldest form of the ending appears in the ancient Carmen Saliare, in the word priviclois, i. e. prīvĭcŭlis, with the sense of singulis. The Acc. Pl. ending is -ŏs, from -ŏns.

§ 7. Scheme of the Second Declension.

Singular

Case	Ending	Development
Nom. M.	-ŏs	became -ŭs; dropped in RO-stems, e.g. ager
Nom. N.	-ŏm	became -ŭm
Gen.	-ī	properly the Gen. suffix of IO-stems
Dat.	-ōī	became -ō
Loc.	-ei	became -ī
Acc.	-ŏm	became -ŭm
Voc.	-ĕ	—
Abl.	-ōd	became -ō

Plural

Case	Ending	Development
Nom. M.	-oi	became -ei, then -ī
Nom. N.	-ă	became -ă
Gen. (1)	-ŏm	became -ōm, then -ŭm
Gen. (2)	-ōrum	originally proper to Pronouns, then to Adjectives also
Dat. / Loc.	-ois	became -eis, then -īs
Acc.	-ŏs	from -ŏns
Voc.	-oi	same as Nom.
Abl.	-ois	same as Dat., Loc.

This should properly be divided into two declensions, (1) Consonant-stems, (2) I-stems, but these two are so similar and so intermixed in certain cases, that the Latin writers on Grammar joined them into one. I-stems have a Gen. Pl. in -ium, an Acc. Sg. in -im, and an Abl. Sg. in -ī (earlier -id); Cons. Stems have a Gen. Pl. in -um, an Acc. Sg. in -em, and an ‘Abl.’ (properly Locative) Singular in -ĕ (earlier -ē). But, e. g., vātī-, an I-stem, has vātum Gen. Pl., and most I-stems assumed in time the Cons.-stem Acc. Sg. and ‘Abl.’ Sg., although Neuter I-stems kept the Abl. in -ī for the sake of distinction from their Nominative Case (e. g. mari Abl., mare Nom.). And Consonant-stems sometimes show the I-stem endings. The Gen. Plur. is the case where the distinction between Consonant and I-stems has been best retained. In the Nom. Sing. some I-stems retain their i, e. g. vest-i-s, but it is lost in sors (sortis Plaut.), Arpinas (Arpinatis, Plaut.), imber (stem imbri-), &c. (see ch. ii. § 12).

The Consonant-stems consist of (1) Guttural-stems, e. g. dūx (stem dūc-, the weak form of the root deuc-, ‘to lead’), rēx (stem rēg-, the lengthened form of the root reg-, ‘to stretch,’ ‘to govern’), nix (stem nig-, the weak form of a root meaning ‘to be wet,’ ‘to snow’) [1]. (2) Labial-stems, e. g. princeps, caelēbs [2]. (3) Dental-stems, e. g. pes (stem pēd-), herēs (stem herēd-), anas (stem anāt-), comēs (stem com-it- from cum and eo). The numerous NT-stems are mostly Present Participles, e. g. serpens (stem serpent-), lit. ‘crawling,’ cliēns, lit. ‘hearing,’ ‘obeying,’ rudēns, lit. ‘rattling,’ oriēns (sc. Sol) [3]. (4) S-stems, e. g. cinis (stem cinis-, tellus (stem tellus-), honor, older honos (stem honos-). The r has found its way into the Nom. from the Oblique Cases, where s stood between vowels and so became r by the Latin phonetic law), opus, earlier opos (stem opŏs-, opēs-), Venus, earlier Venos (stem venŏs-, venēs-, originally neuter and meaning ‘glamour,’ whence venēnum for venēs-num, lit. ‘a love-philtre’). (5) N-stems, e. g. homo, -ĭnis, sermo, -ōnis, caro, -nis, with the M-stem hiems, -mis. (6) R-stems, e. g. frater, -tris, conditor, -tōris, fur, fūris. (7) A few L-stems, e. g. sal, sol.

[1] The g of this root was a ‘Labio-velar’ guttural (see ch. x. § 18), like gw (cf. ninguīt). Between vowels a g of this kind became v in Latin (ch. x. § 18); so we have in the Genitive niv-is.

[2] A word of uncertain derivation. The etymology of the Roman Grammarians ‘quasi caelestium vitam ducens’ will not do!

[3] Gens, mens, etc. are I-stems, g(e)n-ti-, men-ti- (ch. xi. § 12).

The Nom. Sg. of Masc. and Fem. Nouns adds -s to the Stem, but in Neuter Nouns the bare Stem is used. Examples are of Masc. and Fem. I-stems, finis, civis, vesti-s, sors (sorti-s Plaut.), pars, and of Masc. and Fem. Cons.-stems, princeps, dux for dūc-s, heres for herēd-s; of Neuter I-stems, mare (originally marĭ, ch. ii. § 14 n.), triste, Neut. of tristis, originally tristi (cf. Gk. ἴδρις, ἴδρις, ἴδρι with Lat. tristis, tristis, triste), of Neuter Cons.-stems, allec, caput, aes (I.-Eur. āyēs-; cf. aēnus for āyēs-nos), crūs, opus, carmen, ver. (On Fem. Noms. in -ēs, see § 13).

The -ĕ of Neut. I-stems was often dropped (cf. ch. ii. § 12), so that animāle (properly the Neuter of the Adj. animalis) became animal with consequent shortening of the ā before final l (ch. ii. § 16); calcarē (sc. ferrum), a spur, lit. ‘the (iron) fastened on the heel (calx)’, became calcar; lacte, the form found in Plautus, became lact and then lac, owing to the difficulty of pronouncing -ct at the end of a word, the same difficulty as we have with a group like -mb in ‘lamb’; corde (Gen. Pl. cordium) in the same way became cord, a form that should perhaps be written in lines of Plautus where the word is scanned as a long syllable, then cor.

Masc. and Fem. R- and N-stems formed their Nom. Sg. in a different way, R-stems in (1) -ōr, e. g. dātor (class. dātor), (2) -ēr, e. g. mātēr (class. mātēr) (cf. Gk. δωτώρ, πατήρ), N-stems in -ō (Gk. -ων), e. g. cārō, flesh, originally ‘a portion of anything,’ then ‘a portion of flesh’ (connected with Gk. κείρω, to cut), hōmō, connected with hūmus, the ground [1]. (On this ā and ō see ch. x. § 12.)

The Gen. Sg. ending of Cons.-stems was (1) -ēs, (2) -ŏs. Of these the former was adopted in Latin, and became in class. Lat. -is (ch. ii. § 14), e. g. reg-is. There are traces of the latter in early spellings on inscriptions like nominus for class. nominis, hominus for class. hominis [2]. I-stems properly took -eis, which would be in class. Lat. -īs, but there is no instance of a Latin I-stem using any other than the Cons.-stem ending, e. g. turris, pars (on an inscr. of the second cent. B. C. we find partus).

The Dat. Sg. ending of Cons.-stems was -ai, which became -ei and in class. Lat. -ī, e. g. regī, and Latin I-stems show the same ending, e. g. turri. (On the -ĕ of jūre dicundo, &c., see ch. i. § 7).

In the Acc. Sg. m was added to the stem, making in Latin Cons.-stems -em (ch. x. § 14), e. g. reg-em, and in I-stems -im, e. g. turrim. But the I-stems came almost all in time to take the Cons.-stem ending, e. g. turrem, partem. When an I-stem Acc. was used as Adverb however the older ending was retained, e. g. partim (but saltem with -tem by analogy of aut-em); and -im is invariable in vim (an I-stem, § 10).

[1] In Old Latin we have a by-form hēmo (cf. nēmo for nē-hēmo) and a by-declension in -onis, -ōni, etc., e. g. Enn.: Volturius in spinis miserum mandebat hēmonem.

[2] On a leaden bullet (glans) used in the siege of Perusia is carved the ‘Trochaic’ line: Antoni calev, peristi Caesaris victoria.

The Abl. Sg. ending of I-stems was -id, which became -ī in the latter part of the third cent. B. C. (ch. x. § 17), e. g. turri. Cons.-stems used their Locative, ending in -ĕ (class. Lat. -ē), instead of an Ablative, e. g. reg-ĕ, Carthagin-ĕ, as the Gk. Loc. was used as Dat., e. g. γεροντ-ι. These two endings are often exchanged, e. g. for I-stems turre, parte, for Cons.-stems majorĭ and in early Latin militĭ Plaut., coutentionid (i. e. ‘contione’) on the S. C. Bacch.

The Loc. Sg. ending of I-stems was -ĕyĭ (Gk. πόλε(y)ι), which in Latin would become -ei, class. -ī, e. g. lucī, in daylight, of Cons.-stems -ĕ (sometimes omitted in I.-Eur. Locatives), class. Lat. -ē. The use of the Cons.-stem Loc. as an Ablative, e. g. Carthaginĕ, from Carthage, had this result, that Cons.-stems sometimes took the I-stem ending for the sake of distinction; e. g. rurī, in the country, might be contrasted with rure, from the country. I-stems often show the Cons.-stem ending, e. g. manĕ, in the morning.

The Nom. Plur. ending of Masc. and Fem. Cons.-stems was properly -ēs, which would be in class. Lat. -ēs (ch. ii. § 14), and of I-stems -ĕyes (Gk. πόλε(y)ες), in Latin -ēs (ch. x. § 13). But there is no trace of a Latin Cons.-stem with any but the I-stem ending [1]. I-stems sometimes show -eis, class. -īs, properly the ending of their Acc. Pl. (see below).

In the Gen. Pl. -ŏm (Lat. -ŏm, then -ŭm, ch. ii. §§ 16, 14) was added to the stem, so that reg-um, turri-um are correctly formed examples of a Cons.-stem and an I-stem. But the usual interchange is found. We have vātum from the I-stem vātĭ-, ferentium Masc. from the NT-stem ferent- (cf. Gk. φερόντ-ων), and the like [2]; though as a rule the formation of the Gen. Pl. is the best test whether a Third Decl. Noun is an I-stem or a Cons.-stem. The distinction was better preserved in this case than in the Acc. and Abl. Sg.

[1] Plautus’ canēs is a mere shortening like vidēs, avē (ch. ii. § 16). [2] Caesar used pantium, Verrius panum; Caesar partium, Pliny partium.

In the Dat., Abl., Instr., Loc. Plur. the I.-Eur. ending was -bhŏs, Lat. -bŏs, class. -ibus. An I-stem Dat. like turri-bus is regular; but the Cons.-stems borrowed the i of the I-stems to form their Dat., e. g. reg-i-bus instead of reg-bus, milit-i-bus instead of milit-bus.

The Acc. Pl. was formed by adding -ns to the stem. Latin Cons.-stems show -ēs from -ens, Lat. I-stems -īs from -ins, e. g. regēs, turris. By the end of the Republic and the beginning of the Empire this distinction came to be lost sight of, and turres, partes, and the like, were allowed instead of turris, partis. (On the occasional use of -īs in the Nom. Pl. by confusion with the Acc., see above.)

Consonant-Stems

Singular

Case	Ending	Development
Nom. M., F. (1)	-s	—
Nom. M., F. (2)	—	Vowel lengthened in R- and N-stems, e.g. datōr (class. datōr), homō
Nom. N.	—	The bare stem
Gen.	-ēs	became -īs
Dat.	-ai	became -ei, then -ī
Loc.	-ĭ	used also as Abl.
Acc.	-ēm	from an original -ĕm
Abl.	-ĭ	(see Loc.)

Plural

| Case | Ending | Development | |---|---|---| | Nom. M., F. | — | the I-stem ending was used | | Nom. N. | -a | — | | Gen. | -ŏm | became -ōm, then -ŭm | | Dat. / Loc. | — | the I-stem ending was used | | Acc. | -ēs | from -ĕns from an original -ens | | Abl. | — | (same as Dat., Loc.) | ---

I-Stems

Singular

Case	Ending	Development
Nom. M., F.	-īs	—
Nom. N.	-ĭ	became -ĕ (dropped in animal, etc.)
Gen.	—	(the consonant-stem ending was used)
Dat.	—	(the consonant-stem ending was used)
Loc.	-ei (from -ĕyi)	became -ĭ
Acc.	-ĭm	—
Abl.	-īd	became -ī

Plural

Case	Ending	Development
Nom. M., F.	-ēs	from -ĕyēs
Nom. N.	-ia	—
Gen.	-ium	—
Dat. / Loc.	-ĭbŏs	from -i-bhŏs; became -ĭbŭs
Acc.	-īs	from -ĭns
Abl.	—	same as Dat., Loc.

Nouns like bōs, Jupiter (better spelt Juppiter) must be treated separately. They are Diphthong-stems bŏu-, Jŏu- (I.-Eur. gʷou-, dyeu-), the latter being the same stem as is found in diēs (§ 13). Bōs is probably not a true Latin word, I.-Eur. gʷ being represented by v, not b, in Latin (ch. x. § 18), but a word of some country dialect. Its Gen. Pl. boum stands for bov-um with loss of v before u (ch. x. § 13); its Dat. Pl. būbus, bōbus (from gʷoubhŏs) shows ō, ū, the long vowels to which the original diphthong ou developed in Latin (ch. x. § 11). Ju-piter (cf. Dieepiter) is, of course, a compound, like Mars pater, its first part showing ū (older ou), the sound to which the original diphthong eu developed in Latin (ch. x. § 11). In early Latin we find the spelling Diovem, Diove, &c. (see ch. x. § 13). On vīs

(Gk. ἴς, ἶ-φι), an I-stem (Acc. vim), and its wrong declension as an S-stem in the Plural, vires, virium, &c., see § 1. Sūs (Gk. ὗς) was originally a Ū-stem, with stem sū- before Consonants (e. g. sū-bus; cf. Gk. ὗ-ν), sŭ(v)- before vowels (e. g. su-is, su-ā, su-em). Navis was an I.-Eur. Diphthong-stem nāu- (Gk. ναῦς), but has passed in Latin, like I.-Eur. U-stem Adjectives, e. g. lēvis (Gk. ἐλαχὺς) (ch. iv. § 1; ch. xi. § 13), into the I-declension.

The Fourth and Fifth Declensions do not preserve their individuality so well as the others. The Fourth, consisting of U-stems, does not always keep itself separate from the Second, while the Fifth is closely connected with the First. In the ordinary Latin of every-day life it is doubtful to what extent of their declension U-stems would show a different treatment from O-stems; and even the most careful writers decline only a few Fifth Declension Nouns, e. g. dies, res, throughout in what may be called a Fifth Declension form. The U-stems of the Fourth Declension are Masculine or Neuter. There are a few Feminines, e. g. domus, laurus, which seem to have been originally O-stems (see § 6); but it is possible that at an earlier period there were more, for there are traces of Abstract U-stems like metus, fear (cf. Ion. Gk. ἀρτύς, a fitting, F., but Lat. artus, a limb, M.), having been Feminine in early Latin, e. g. nec metus ulla tenet, Ennius.

The U-declension being in a manner the property of the educated section of the Roman world, was greatly subjected to the theories of Grammarians. The proper Nom. Sg. Masc. ending -ūs was by some Grammarians pronounced -ŭs, just as some made the -u of Nom. Sg. Neut. long, others short; the I.-Eur. Gen. Sg. ending -eus, in Latin -ous, then -ūs (ch. x. § 11), was by many

discarded for the Gen. Sg. ending of Ū-stems (§ 10), -uis, e. g. senatuis, domuiis (cf. Ter. Haut. 287 eius anuis causa, ‘to oblige that old woman’), an ending which suited the Dat. Sg. ending -uī, older -uei (from I.-Eur. -ēwai). The Emperor Augustus, who like his great predecessor, Julius Caesar, paid great attention to matters of Grammar, and once cashiered an officer for using the vulgarism ipse instead of ipse, always made the Gen. Sg. of domus domos. This domos, if it be anything more than a coinage of Augustus to suit some theory of his own, may be a genuine relic of another I.-Eur. ending -ous, which would be in Latin -ous, then -ōs (ch. x. § 11). But the Second Declension form of the Genitive, used in popular Latin, steadily maintained itself against all these rules and theories of the Grammarians, and Quintilian in the first century A. D. declares senati to be as good a Genitive as senatus. In the Dative, beside -uī, there is an occasional use of -ū, which seems to be the Locative ending (cf. noctū, in the night), from I.-Eur. -eu (ch. x. § 11). (On the use of this Locative in the Second Supine, see ch. vi. § 16.)

In the Nom. Plur. we should expect -uēs, -uās, from I.-Eur. -ēwēs (Gk. πήχε(ϝ)ες), but the Acc. ending is used instead, -ūs, from -ūns. The Gen. Pl. often shows -um (the Second Declension ending) as well as the U-stem ending -ŭum, older -ŭŏm, from -ŭŏm (ch. ii. §§ 14, 16) (I.-Eur. -ĕwŏm, Gk. πήχε(ϝ)ων) e. g. mille passuum or mille passuum, a mile, lit. ‘a thousand (of) paces.’ In the Dat. Pl. -ŭbus became by the Latin laws of sound -ibus (ch. ii. § 14), but the older form was kept up for the sake of distinction in words like arcubus, ‘to bows’ (cf. arcibus, ‘to citadels’).

As regards Neuter U-stems the Latin Grammarians

disputed, as we have seen, whether the -u of the Nom. Sg. was short or long. The Augustan poets certainly show the scansion cornū, &c., though the length of the u is difficult to justify. There was usually a Second Declension by-form in -um, e. g. cornum, Lucr. ii. 388:

Praeterea lumen per cornum transit.

Singular

Case	Ending	Development
Nom. M.	-ŭs	—
Nom. N.	-u	—
Gen. (1)	-ŭs	from -eus
Gen. (2)	-uis	from -ŭwēs, the Ū-stem suffix
Gen. (3)	-ŏ	the O-stem suffix
Dat.	-uī	older -uei, from -ēwai
Loc.	-ŭ	from -eu; used also as Dat.
Acc.	-ŭm	—
Abl.	-ŭd	became -ū

Plural

Case	Ending	Development
Nom. M.	—	see Acc.
Nom. N.	-ua	—
Gen. (1)	-uŭm	from -ēwōm
Gen. (2)	-ŭm	from -ōm, the O-stem suffix
Dat. / Loc.	-ŭbŭs	older -ŭbŏs, later -ǐbŭs from -ŭbhŏs
Acc.	-ŭs	from -ŭns
Abl.	-ŭbŭs	same as Dat., Loc.

This may be called the Ē-declension, for all the Nouns that belong to it end their Nom. Sg. in -ēs and show the long vowel ē in at least some cases. But it is a veritable medley of most widely differing stems. We have (1) Verbal Noun-stems in -ē, e. g. sordē- (cf. sordē-re Inf., sordē-facio, &c.), with Abl.

sordē and Gen. Pl. sordērum (in Plautus, but in late Latin sordium); similarly from facio, faciēs, and from specio, speciēs. (2) Verbal Nouns in -iēs derived from Third Conj. Verbs whose Present has -o, not -io, e. g. rabies from rabo, scabies from scabo, perniciēs from a lost 3 Conj. Verb neco, along with its synonym pernicitiēs. (3) Nouns in -iēs, which are not Verbal Nouns, e. g. temperiēs, derived from tempus, especially Nouns in -tiēs from Adjs., e. g. vastitiēs from vastus, mollitiēs from mollis, segnitiēs from segnis, amicitiās from amicus. (4) diēs, with stem dyēw-, a by-form of dyēw-, meaning ‘sky,’ ‘day.’ The stem dyēw- is the stem of Jovis Gen., Jovi Dat., &c. (§ 10). (5) rēs, with stem rēy-, and spēs, with stem sphey-.

Diēs and rēs are the most consistently declined according to what is called the ‘Fifth Declension,’ e. g. Abl. diē, rē, Gen. Pl. dierum, rērum, Dat. Pl. diēbus, rēbus. Next comes spēs, which however is given by Ennius a Nom. Pl. spērēs, as if it were an S-stem (cf. vires Nom. Pl. of the I-stem vis, § 1). But the Verbal and other Nouns in -iēs belong as much to the First Declension as to the Fifth, for they have by-forms in -ia, e. g. vastitia, temperia, effigia, and the farther back we go in the literature the more does this side of their nature show itself. In Plautus the iē-forms are generally confined to the Nom., Acc. Sg., while he prefers -iae in the Nom. Pl., e. g. intemperiae, and in the Dat. Sg., e. g. materiae, -iā in the Abl. Sg., e. g. barbariā, -iīs in the Abl. Pl., e. g. mollitiis, and so on (but faciē Abl., speciē Abl.). After his time the number of Nouns in -iēs and the number of cases to which this Ē-declension is extended steadily increases, though the Plural is always reserved by good writers for the -ā-forms. All this points to those ĪE-stems having been originally ĪĀ-stems; and as we know that

the presence of i had in certain circumstances the effect of giving ā the sound of ē (open ē) in Latin [1], it seems likely that the greater part of the Fifth Declension Nouns are due to this change of sound. Verbal Nouns in -ēs, on the other hand, have affinity with I-stems (cf. plēbēs and plēbs), and most of them show the Ē-declension to a very limited extent only, e. g. tābē Abl. [2] Requiē Abl. (for requiētē), requiem Acc. (for requietem), may be due to the false analogy of the Nom. Sg. requiēs (for requiet(ēs)).

[1] Jajunus became jejūnus, and Jenuarius was the vulgar form of Januarius (cf. Italian Gennaio) (page 10 n.)

[2] Many so-called Nominatives in -ēs are really Nominatives Plural of I-stems, e. g. ambāgēs, nūbēs (with a by-form nūbēs), saepēs (with a by-form saeps). Some feminine animal-names seem to be dialectal, e. g. fēles, pŭlambes. They take the I-stem declension.

The formation of the Cases is closely modelled on the First Declension, ē being substituted everywhere for ă. The Gen. Sg. has -ēī, which was allowed to keep its long ē when an i preceded, e. g. faciēī, but shortened it in other circumstances, fideī (in early Latin fideī, e. g. plenu’ fideī at the end of a hexameter line of Ennius). This shortening was in accordance with the Latin custom of shortening a long vowel before another vowel (ch. ii. § 16), and in point of fact affected the ending -iēī as well, changing it to -iēī, then to -iī; for the Republican writers have faciī, perniciī, progeniī, &c. Some Grammarians changed this ending -iī to -iē, in order that the Genitive might show the ē-vowel, which was the distinguishing mark of the Fifth Declension; and Julius Caesar, we are told, stamped with his approval forms like speciē and diē. Others preferred the old Gen. in -iēs (like -ās, the old Gen. of the First Declension, § 4); and we hear of a great controversy as to whether diēs or diī or diē was the form used by Virgil in a line of the Georgics (i. 208):

Libra dies somnique pares ubi fecerit horas,

and whether the same poet wrote diī ‘of the day’ or dei ‘of the god’ in Aen. i. 636:

Munerā laetitiamque dei.

The -ēī of fideī, and the like, also appears as -ī, e. g. fami (Lucilius), tribunus plebi, with a by-form in -ē, e. g. fide in Horace (C. iii. 7. 4):

Constantis juvenem fide.

The Genitive of res, spes is as a rule a monosyllable, that of fides a disyllable in Plautus and Terence, writers who follow the conversational language of their time.

The Dat. Sg. has also by-forms, possibly due to the theorizing of Grammarians, possibly genuine relics of an ancient divergence of declension, e.g. faciēī, fideī (the forms generally accepted), faciī, faciē. In Plautus and Terence the treatment of the Dat. Sg. is the same as that of the Genitive. The grammarian Aulus Gellius tells us that purists preferred facie to faciē.

Of the other cases we need mention only the Dat. Abl. Pl. in -ēbus, found in rēbus, diēbus.

Singular

Case	Ending	Development
Nom.	-ēs	—
Gen. (1)	-ēī, -ěī, -ei, -ī	—
Gen. (2)	-ē	changed from -ī
Gen. (3)	-ēs	cf. 1 Decl. -ās
Dat. (1)	-ēī, -ěī, -ei, -ī	—
Dat. (2)	-ē	—
Acc.	-ēm	from -ĕm
Abl.	-ē	from -ēd

Plural

Case	Ending	Development
Nom.	-ēs	—
Gen.	-ērum	cf. 1 Decl. -ārum
Dat. / Abl.	-ēbus	cf. 1 Decl. -ābus
Acc.	-ēs	from -ēns
Abl.	-ēbus	same as Dat.

§11

The Fourth Declension

The Fourth Declension

The Fourth and Fifth Declensions do not preserve their individuality so well as the others. The Fourth, consisting of U-stems, does not always keep itself separate from the Second, while the Fifth is closely connected with the First. In the ordinary Latin of every-day life it is doubtful to what extent of their declension U-stems would show a different treatment from O-stems; and even the most careful writers decline only a few Fifth Declension Nouns, e. g. dies, res, throughout in what may be called a Fifth Declension form. The U-stems of the Fourth Declension are Masculine or Neuter. There are a few Feminines, e. g. domus, laurus, which seem to have been originally O-stems (see § 6); but it is possible that at an earlier period there were more, for there are traces of Abstract U-stems like metus, fear (cf. Ion. Gk. ἀρτύς, a fitting, F., but Lat. artus, a limb, M.), having been Feminine in early Latin, e. g. nec metus ulla tenet, Ennius.

The U-declension being in a manner the property of the educated section of the Roman world, was greatly subjected to the theories of Grammarians. The proper Nom. Sg. Masc. ending -ūs was by some Grammarians pronounced -ŭs, just as some made the -u of Nom. Sg. Neut. long, others short; the I.-Eur. Gen. Sg. ending -eus, in Latin -ous, then -ūs (ch. x. § 11), was by many

discarded for the Gen. Sg. ending of Ū-stems (§ 10), -uis, e. g. senatuis, domuiis (cf. Ter. Haut. 287 eius anuis causa, ‘to oblige that old woman’), an ending which suited the Dat. Sg. ending -uī, older -uei (from I.-Eur. -ēwai). The Emperor Augustus, who like his great predecessor, Julius Caesar, paid great attention to matters of Grammar, and once cashiered an officer for using the vulgarism ipse instead of ipse, always made the Gen. Sg. of domus domos. This domos, if it be anything more than a coinage of Augustus to suit some theory of his own, may be a genuine relic of another I.-Eur. ending -ous, which would be in Latin -ous, then -ōs (ch. x. § 11). But the Second Declension form of the Genitive, used in popular Latin, steadily maintained itself against all these rules and theories of the Grammarians, and Quintilian in the first century A. D. declares senati to be as good a Genitive as senatus. In the Dative, beside -uī, there is an occasional use of -ū, which seems to be the Locative ending (cf. noctū, in the night), from I.-Eur. -eu (ch. x. § 11). (On the use of this Locative in the Second Supine, see ch. vi. § 16.)

In the Nom. Plur. we should expect -uēs, -uās, from I.-Eur. -ēwēs (Gk. πήχε(ϝ)ες), but the Acc. ending is used instead, -ūs, from -ūns. The Gen. Pl. often shows -um (the Second Declension ending) as well as the U-stem ending -ŭum, older -ŭŏm, from -ŭŏm (ch. ii. §§ 14, 16) (I.-Eur. -ĕwŏm, Gk. πήχε(ϝ)ων) e. g. mille passuum or mille passuum, a mile, lit. ‘a thousand (of) paces.’ In the Dat. Pl. -ŭbus became by the Latin laws of sound -ibus (ch. ii. § 14), but the older form was kept up for the sake of distinction in words like arcubus, ‘to bows’ (cf. arcibus, ‘to citadels’).

As regards Neuter U-stems the Latin Grammarians

disputed, as we have seen, whether the -u of the Nom. Sg. was short or long. The Augustan poets certainly show the scansion cornū, &c., though the length of the u is difficult to justify. There was usually a Second Declension by-form in -um, e. g. cornum, Lucr. ii. 388:

Praeterea lumen per cornum transit.

The Fourth and Fifth Declensions do not preserve their individuality so well as the others. The Fourth, consisting of U-stems, does not always keep itself separate from the Second, while the Fifth is closely connected with the First. In the ordinary Latin of every-day life it is doubtful to what extent of their declension U-stems would show a different treatment from O-stems; and even the most careful writers decline only a few Fifth Declension Nouns, e. g. dies, res, throughout in what may be called a Fifth Declension form. The U-stems of the Fourth Declension are Masculine or Neuter. There are a few Feminines, e. g. domus, laurus, which seem to have been originally O-stems (see § 6); but it is possible that at an earlier period there were more, for there are traces of Abstract U-stems like metus, fear (cf. Ion. Gk. ἀρτύς, a fitting, F., but Lat. artus, a limb, M.), having been Feminine in early Latin, e. g. nec metus ulla tenet, Ennius.

The U-declension being in a manner the property of the educated section of the Roman world, was greatly subjected to the theories of Grammarians. The proper Nom. Sg. Masc. ending -ūs was by some Grammarians pronounced -ŭs, just as some made the -u of Nom. Sg. Neut. long, others short; the I.-Eur. Gen. Sg. ending -eus, in Latin -ous, then -ūs (ch. x. § 11), was by many

discarded for the Gen. Sg. ending of Ū-stems (§ 10), -uis, e. g. senatuis, domuiis (cf. Ter. Haut. 287 eius anuis causa, ‘to oblige that old woman’), an ending which suited the Dat. Sg. ending -uī, older -uei (from I.-Eur. -ēwai). The Emperor Augustus, who like his great predecessor, Julius Caesar, paid great attention to matters of Grammar, and once cashiered an officer for using the vulgarism ipse instead of ipse, always made the Gen. Sg. of domus domos. This domos, if it be anything more than a coinage of Augustus to suit some theory of his own, may be a genuine relic of another I.-Eur. ending -ous, which would be in Latin -ous, then -ōs (ch. x. § 11). But the Second Declension form of the Genitive, used in popular Latin, steadily maintained itself against all these rules and theories of the Grammarians, and Quintilian in the first century A. D. declares senati to be as good a Genitive as senatus. In the Dative, beside -uī, there is an occasional use of -ū, which seems to be the Locative ending (cf. noctū, in the night), from I.-Eur. -eu (ch. x. § 11). (On the use of this Locative in the Second Supine, see ch. vi. § 16.)

In the Nom. Plur. we should expect -uēs, -uās, from I.-Eur. -ēwēs (Gk. πήχε(ϝ)ες), but the Acc. ending is used instead, -ūs, from -ūns. The Gen. Pl. often shows -um (the Second Declension ending) as well as the U-stem ending -ŭum, older -ŭŏm, from -ŭŏm (ch. ii. §§ 14, 16) (I.-Eur. -ĕwŏm, Gk. πήχε(ϝ)ων) e. g. mille passuum or mille passuum, a mile, lit. ‘a thousand (of) paces.’ In the Dat. Pl. -ŭbus became by the Latin laws of sound -ibus (ch. ii. § 14), but the older form was kept up for the sake of distinction in words like arcubus, ‘to bows’ (cf. arcibus, ‘to citadels’).

As regards Neuter U-stems the Latin Grammarians

disputed, as we have seen, whether the -u of the Nom. Sg. was short or long. The Augustan poets certainly show the scansion cornū, &c., though the length of the u is difficult to justify. There was usually a Second Declension by-form in -um, e. g. cornum, Lucr. ii. 388:

Praeterea lumen per cornum transit.

§12

The Scheme of the Fourth Declension

The Scheme of the Fourth Declension

Singular

Case	Ending	Development
Nom. M.	-ŭs	—
Nom. N.	-u	—
Gen. (1)	-ŭs	from -eus
Gen. (2)	-uis	from -ŭwēs, the Ū-stem suffix
Gen. (3)	-ŏ	the O-stem suffix
Dat.	-uī	older -uei, from -ēwai
Loc.	-ŭ	from -eu; used also as Dat.
Acc.	-ŭm	—
Abl.	-ŭd	became -ū

Plural

Case	Ending	Development
Nom. M.	—	see Acc.
Nom. N.	-ua	—
Gen. (1)	-uŭm	from -ēwōm
Gen. (2)	-ŭm	from -ōm, the O-stem suffix
Dat. / Loc.	-ŭbŭs	older -ŭbŏs, later -ǐbŭs from -ŭbhŏs
Acc.	-ŭs	from -ŭns
Abl.	-ŭbŭs	same as Dat., Loc.

Singular

Case	Ending	Development
Nom. M.	-ŭs	—
Nom. N.	-u	—
Gen. (1)	-ŭs	from -eus
Gen. (2)	-uis	from -ŭwēs, the Ū-stem suffix
Gen. (3)	-ŏ	the O-stem suffix
Dat.	-uī	older -uei, from -ēwai
Loc.	-ŭ	from -eu; used also as Dat.
Acc.	-ŭm	—
Abl.	-ŭd	became -ū

Plural

Case	Ending	Development
Nom. M.	—	see Acc.
Nom. N.	-ua	—
Gen. (1)	-uŭm	from -ēwōm
Gen. (2)	-ŭm	from -ōm, the O-stem suffix
Dat. / Loc.	-ŭbŭs	older -ŭbŏs, later -ǐbŭs from -ŭbhŏs
Acc.	-ŭs	from -ŭns
Abl.	-ŭbŭs	same as Dat., Loc.

§13

The Fifth Declension

The Fifth Declension 

This may be called the Ē-declension, for all the Nouns that belong to it end their Nom. Sg. in -ēs and show the long vowel ē in at least some cases. But it is a veritable medley of most widely differing stems. We have (1) Verbal Noun-stems in -ē, e. g. sordē- (cf. sordē-re Inf., sordē-facio, &c.), with Abl.

sordē and Gen. Pl. sordērum (in Plautus, but in late Latin sordium); similarly from facio, faciēs, and from specio, speciēs. (2) Verbal Nouns in -iēs derived from Third Conj. Verbs whose Present has -o, not -io, e. g. rabies from rabo, scabies from scabo, perniciēs from a lost 3 Conj. Verb neco, along with its synonym pernicitiēs. (3) Nouns in -iēs, which are not Verbal Nouns, e. g. temperiēs, derived from tempus, especially Nouns in -tiēs from Adjs., e. g. vastitiēs from vastus, mollitiēs from mollis, segnitiēs from segnis, amicitiās from amicus. (4) diēs, with stem dyēw-, a by-form of dyēw-, meaning ‘sky,’ ‘day.’ The stem dyēw- is the stem of Jovis Gen., Jovi Dat., &c. (§ 10). (5) rēs, with stem rēy-, and spēs, with stem sphey-.

Diēs and rēs are the most consistently declined according to what is called the ‘Fifth Declension,’ e. g. Abl. diē, rē, Gen. Pl. dierum, rērum, Dat. Pl. diēbus, rēbus. Next comes spēs, which however is given by Ennius a Nom. Pl. spērēs, as if it were an S-stem (cf. vires Nom. Pl. of the I-stem vis, § 1). But the Verbal and other Nouns in -iēs belong as much to the First Declension as to the Fifth, for they have by-forms in -ia, e. g. vastitia, temperia, effigia, and the farther back we go in the literature the more does this side of their nature show itself. In Plautus the iē-forms are generally confined to the Nom., Acc. Sg., while he prefers -iae in the Nom. Pl., e. g. intemperiae, and in the Dat. Sg., e. g. materiae, -iā in the Abl. Sg., e. g. barbariā, -iīs in the Abl. Pl., e. g. mollitiis, and so on (but faciē Abl., speciē Abl.). After his time the number of Nouns in -iēs and the number of cases to which this Ē-declension is extended steadily increases, though the Plural is always reserved by good writers for the -ā-forms. All this points to those ĪE-stems having been originally ĪĀ-stems; and as we know that

the presence of i had in certain circumstances the effect of giving ā the sound of ē (open ē) in Latin [1], it seems likely that the greater part of the Fifth Declension Nouns are due to this change of sound. Verbal Nouns in -ēs, on the other hand, have affinity with I-stems (cf. plēbēs and plēbs), and most of them show the Ē-declension to a very limited extent only, e. g. tābē Abl. [2] Requiē Abl. (for requiētē), requiem Acc. (for requietem), may be due to the false analogy of the Nom. Sg. requiēs (for requiet(ēs)).

[1] Jajunus became jejūnus, and Jenuarius was the vulgar form of Januarius (cf. Italian Gennaio) (page 10 n.)

[2] Many so-called Nominatives in -ēs are really Nominatives Plural of I-stems, e. g. ambāgēs, nūbēs (with a by-form nūbēs), saepēs (with a by-form saeps). Some feminine animal-names seem to be dialectal, e. g. fēles, pŭlambes. They take the I-stem declension.

The formation of the Cases is closely modelled on the First Declension, ē being substituted everywhere for ă. The Gen. Sg. has -ēī, which was allowed to keep its long ē when an i preceded, e. g. faciēī, but shortened it in other circumstances, fideī (in early Latin fideī, e. g. plenu’ fideī at the end of a hexameter line of Ennius). This shortening was in accordance with the Latin custom of shortening a long vowel before another vowel (ch. ii. § 16), and in point of fact affected the ending -iēī as well, changing it to -iēī, then to -iī; for the Republican writers have faciī, perniciī, progeniī, &c. Some Grammarians changed this ending -iī to -iē, in order that the Genitive might show the ē-vowel, which was the distinguishing mark of the Fifth Declension; and Julius Caesar, we are told, stamped with his approval forms like speciē and diē. Others preferred the old Gen. in -iēs (like -ās, the old Gen. of the First Declension, § 4); and we hear of a great controversy as to whether diēs or diī or diē was the form used by Virgil in a line of the Georgics (i. 208):

Libra dies somnique pares ubi fecerit horas,

and whether the same poet wrote diī ‘of the day’ or dei ‘of the god’ in Aen. i. 636:

Munerā laetitiamque dei.

The -ēī of fideī, and the like, also appears as -ī, e. g. fami (Lucilius), tribunus plebi, with a by-form in -ē, e. g. fide in Horace (C. iii. 7. 4):

Constantis juvenem fide.

The Genitive of res, spes is as a rule a monosyllable, that of fides a disyllable in Plautus and Terence, writers who follow the conversational language of their time.

The Dat. Sg. has also by-forms, possibly due to the theorizing of Grammarians, possibly genuine relics of an ancient divergence of declension, e.g. faciēī, fideī (the forms generally accepted), faciī, faciē. In Plautus and Terence the treatment of the Dat. Sg. is the same as that of the Genitive. The grammarian Aulus Gellius tells us that purists preferred facie to faciē.

Of the other cases we need mention only the Dat. Abl. Pl. in -ēbus, found in rēbus, diēbus.

This may be called the Ē-declension, for all the Nouns that belong to it end their Nom. Sg. in -ēs and show the long vowel ē in at least some cases. But it is a veritable medley of most widely differing stems. We have (1) Verbal Noun-stems in -ē, e. g. sordē- (cf. sordē-re Inf., sordē-facio, &c.), with Abl.

sordē and Gen. Pl. sordērum (in Plautus, but in late Latin sordium); similarly from facio, faciēs, and from specio, speciēs. (2) Verbal Nouns in -iēs derived from Third Conj. Verbs whose Present has -o, not -io, e. g. rabies from rabo, scabies from scabo, perniciēs from a lost 3 Conj. Verb neco, along with its synonym pernicitiēs. (3) Nouns in -iēs, which are not Verbal Nouns, e. g. temperiēs, derived from tempus, especially Nouns in -tiēs from Adjs., e. g. vastitiēs from vastus, mollitiēs from mollis, segnitiēs from segnis, amicitiās from amicus. (4) diēs, with stem dyēw-, a by-form of dyēw-, meaning ‘sky,’ ‘day.’ The stem dyēw- is the stem of Jovis Gen., Jovi Dat., &c. (§ 10). (5) rēs, with stem rēy-, and spēs, with stem sphey-.

Diēs and rēs are the most consistently declined according to what is called the ‘Fifth Declension,’ e. g. Abl. diē, rē, Gen. Pl. dierum, rērum, Dat. Pl. diēbus, rēbus. Next comes spēs, which however is given by Ennius a Nom. Pl. spērēs, as if it were an S-stem (cf. vires Nom. Pl. of the I-stem vis, § 1). But the Verbal and other Nouns in -iēs belong as much to the First Declension as to the Fifth, for they have by-forms in -ia, e. g. vastitia, temperia, effigia, and the farther back we go in the literature the more does this side of their nature show itself. In Plautus the iē-forms are generally confined to the Nom., Acc. Sg., while he prefers -iae in the Nom. Pl., e. g. intemperiae, and in the Dat. Sg., e. g. materiae, -iā in the Abl. Sg., e. g. barbariā, -iīs in the Abl. Pl., e. g. mollitiis, and so on (but faciē Abl., speciē Abl.). After his time the number of Nouns in -iēs and the number of cases to which this Ē-declension is extended steadily increases, though the Plural is always reserved by good writers for the -ā-forms. All this points to those ĪE-stems having been originally ĪĀ-stems; and as we know that

the presence of i had in certain circumstances the effect of giving ā the sound of ē (open ē) in Latin [1], it seems likely that the greater part of the Fifth Declension Nouns are due to this change of sound. Verbal Nouns in -ēs, on the other hand, have affinity with I-stems (cf. plēbēs and plēbs), and most of them show the Ē-declension to a very limited extent only, e. g. tābē Abl. [2] Requiē Abl. (for requiētē), requiem Acc. (for requietem), may be due to the false analogy of the Nom. Sg. requiēs (for requiet(ēs)).

[1] Jajunus became jejūnus, and Jenuarius was the vulgar form of Januarius (cf. Italian Gennaio) (page 10 n.)

[2] Many so-called Nominatives in -ēs are really Nominatives Plural of I-stems, e. g. ambāgēs, nūbēs (with a by-form nūbēs), saepēs (with a by-form saeps). Some feminine animal-names seem to be dialectal, e. g. fēles, pŭlambes. They take the I-stem declension.

The formation of the Cases is closely modelled on the First Declension, ē being substituted everywhere for ă. The Gen. Sg. has -ēī, which was allowed to keep its long ē when an i preceded, e. g. faciēī, but shortened it in other circumstances, fideī (in early Latin fideī, e. g. plenu’ fideī at the end of a hexameter line of Ennius). This shortening was in accordance with the Latin custom of shortening a long vowel before another vowel (ch. ii. § 16), and in point of fact affected the ending -iēī as well, changing it to -iēī, then to -iī; for the Republican writers have faciī, perniciī, progeniī, &c. Some Grammarians changed this ending -iī to -iē, in order that the Genitive might show the ē-vowel, which was the distinguishing mark of the Fifth Declension; and Julius Caesar, we are told, stamped with his approval forms like speciē and diē. Others preferred the old Gen. in -iēs (like -ās, the old Gen. of the First Declension, § 4); and we hear of a great controversy as to whether diēs or diī or diē was the form used by Virgil in a line of the Georgics (i. 208):

Libra dies somnique pares ubi fecerit horas,

and whether the same poet wrote diī ‘of the day’ or dei ‘of the god’ in Aen. i. 636:

Munerā laetitiamque dei.

The -ēī of fideī, and the like, also appears as -ī, e. g. fami (Lucilius), tribunus plebi, with a by-form in -ē, e. g. fide in Horace (C. iii. 7. 4):

Constantis juvenem fide.

The Genitive of res, spes is as a rule a monosyllable, that of fides a disyllable in Plautus and Terence, writers who follow the conversational language of their time.

The Dat. Sg. has also by-forms, possibly due to the theorizing of Grammarians, possibly genuine relics of an ancient divergence of declension, e.g. faciēī, fideī (the forms generally accepted), faciī, faciē. In Plautus and Terence the treatment of the Dat. Sg. is the same as that of the Genitive. The grammarian Aulus Gellius tells us that purists preferred facie to faciē.

Of the other cases we need mention only the Dat. Abl. Pl. in -ēbus, found in rēbus, diēbus.

§14

The Scheme of the Fifth Declension

The Scheme of the Fifth Declension

Singular

Case	Ending	Development
Nom.	-ēs	—
Gen. (1)	-ēī, -ěī, -ei, -ī	—
Gen. (2)	-ē	changed from -ī
Gen. (3)	-ēs	cf. 1 Decl. -ās
Dat. (1)	-ēī, -ěī, -ei, -ī	—
Dat. (2)	-ē	—
Acc.	-ēm	from -ĕm
Abl.	-ē	from -ēd

Plural

Case	Ending	Development
Nom.	-ēs	—
Gen.	-ērum	cf. 1 Decl. -ārum
Dat. / Abl.	-ēbus	cf. 1 Decl. -ābus
Acc.	-ēs	from -ēns
Abl.	-ēbus	same as Dat.

Singular

Case	Ending	Development
Nom.	-ēs	—
Gen. (1)	-ēī, -ěī, -ei, -ī	—
Gen. (2)	-ē	changed from -ī
Gen. (3)	-ēs	cf. 1 Decl. -ās
Dat. (1)	-ēī, -ěī, -ei, -ī	—
Dat. (2)	-ē	—
Acc.	-ēm	from -ĕm
Abl.	-ē	from -ēd

Plural

Case	Ending	Development
Nom.	-ēs	—
Gen.	-ērum	cf. 1 Decl. -ārum
Dat. / Abl.	-ēbus	cf. 1 Decl. -ābus
Acc.	-ēs	from -ēns
Abl.	-ēbus	same as Dat.