114. The form of the verrb stem in primitive nouns is liable to the same variations as in the Tenses (§ 38). it will be seen that these variations are connected with the accent; but this part of the subect will be best treated separately (§ 115).
The chief primary suffixes are as follows.
-o, feminine -ᾱ, -η: the verb stem taking three forms
- The weak form.
ἰαχή (ϝι-ϝαχ-ή) cry
ἵ-στο-ς (στα-) web
- The O-form.
τόκ-ο-ς (τεκ-) offspring
ἀρωγ-ό-ς (ἀρήγω) helper
σπονδ-ή (σπένδ-ω) libation
- Attic reduplication.
The radical vowel appears as ω.
-ι: as τρόφ-ι (τρέφ-ω) thick, τρόπ-ι-ς keel of a ship, φρόν-ι-ς understanding (with the verb stem in the O-form).
-ι̯ᾰ: seldom with stems of clearly verbal meaning.
φύζα (φυγι̯α) flight
σχίζα (σχίδι̯α) chip
More often with roots used as nouns
and as a feminine suffix in adjectives (infra).
The Greek -ι̯α takes the place of -ῑ, the original declension of which is lost in Greek; see Brugmann, Grundr. ii. 109, p. 313 ; Joh. Schmidt, Pluralb. p. 42.
-υ: with two forms of declension.
- Genitive -ε-ος: with the weak stem; chiefly in masculine and neuter adjectives.
ταρφ-ύ-ς (τρέφ-ω) thick
εὐρύς (for ἐ-ϝρυ-, root ϝερ-).
But ἡδύ-ς has the strong stem, and ὠκύ-ς the O-form.
Feminine -ειὰ (for -εϝ-ι̯α), -πᾰ, as ἡδεῖα, ὠκέα.
- Genitive -υ-ος: in substantives (chiefly feminine).
νέκ-υ-ς (masculine) cοrpse
γῆρυ-ς νοίce, cry
As to the declension of nοuns in -ις, genitive -ιος, and -υς, genitive -υος, see § 94.
-εσ: with the strong form of the stem.
βένθ-ος depth (cp. βαθ-ύ-ς)
θέρ-ος warmth, summer
Feminine -ειᾰ (for -εσ-ι̯α), as ἠριγένεια.
The O-form of the stem is found in ὄχ-ος chariot (cp. the perfect ὄκωχα, § 26.5); the weak form in θάλ-ος blossom (but cp. νεο-θηλ-ής), κάρτος(alsο κράτος), θάρσος(cp. έ2ερσ-ίτηςἈλι-θέρσ-ης), ἄχ·ος grief The forms πάθ-ος, βάθ-ος are not Homeric.
Note however that in Homer the substantive is θάρσος (for which θράσος occurs only once, Il. 14.416), the adjective always θρασύς; so that a distinction of quantity is kept up in place of the original distinction between a *θέρσος and θρασύς. On θέρσος as the original Greek form see Osthoff, M. U. ii. 49.
ῑ and ῡ appear in these stems as in the present tense (§ 29).
-ασ: as δέμ-ας "build".
The stem is in the strong form; indeed the stem vowel is always, except in γῆρας old age, κῶας fleece, and οὗδας floor; cp. γέρας, δέπας, κέρας, κνέφας, κρέας, κτέρας, πέρας, σέβας, σέλας, σκέπας, σφέλας, τέρας. Also *ἔρας (ἐραννός for ἐρασ-νός) and *γέλας (ἐ-γέλασ-σα).
-εν, -ᾰν, -ον, -ων.
τέρ-ην, genitive -εν-ος (τείρω) sοft
πέπ-ον (vocative) tender one
ἀγκ-ών,genitive -ῶν-ος elbow
Feminine -αινα (-αν-ι̯α), in λέαινα, imitated by way of sarcasm in θέ-αινα (Il. 8.5).
-ντ, -οντ: in participles, and in a few substantives.
δράκ-ων a serpent (literally the "staring" animal, δέρκ-ομαι)
-αντ: notably in compounds, as ἀκάμας, ἀδάμας, πολύτλας.
-ᾰν: in τάλας, μέλας. Perhaps originally stems in -αντ, which have followed the analogy οf -εν, -ον (Meyer, G. G. p. 354).
-ερ, -ωρ, ᾰρ.
ἀήρ (ἀϝ-ήρ) air
αἰθ-ήρ (αἴθ-ω) bright sky
δα-ήρ husband's brother (leνir)
μάκ-αρ great (Il. 11.68)
-ορ: in the Homeric ἄορ sword, ἦτορ breast is perhaps only the Aeοlic form of -αρ (-r̥). As to the nominative and accusative neuter forms in -ωρ see § 114*.8.d.
-ιο, -ι̯o is very rare in Greek as a primary suffix. Brugmann gives ἐρείπ-ια rains and (post-Hοmeric) ἅγ-ιος, στύγ-ιος, σφάγ-ιον, πάγ-ιος. We may add ταμ-ίη dispenser, πεν-ίη poverty; also δῖος (διϝ-ι̯ο-ς) bright, πεζός (πεδ-) on fοοt, κραδ-ίη (κῆρ for κηρ-δ) heart, in which the stem is a root noun.
The word ἀ-οσση-τήp helper presupposes a stem ὁσσο- for σοκ-ι̯ο-, answering to Latin sοc-iu-s (seq-, Greek ἑπ-).
In ἄλλος (alius), μέσσος (medius), δεξιός the suffix appears to give the force of a comparative; see Brugmann, Grundr., ii. § 3, p. 125.
-ιοσ, -ι̯οσ, -ισ: the comparative suffix, as πλέω (πλε-ι̯οσ-α) πλεῖστος (πλε-ισ-τος) : see § 114*.7.
-ϝο: κεινός (κεν-ϝός) empty, οὖλος (ὁλ-ϝος) whole, λαι-ός lae-νus, ὀρθός ard-uus.
-ϝωσ, -ϝοτ, feminine -υιᾰ: in the perfect participle, and in the nouns ὄργ-υια fathom, ἅρπ-υια stοrm-wind, ἄγ-υια street.
-ϝᾰρ: as πῖαρ (for πί-ϝαρ) fαtness, ὄνειαρ (ὀνη-ϝαρ?) help, εἶδαρ (ἐδ-ϝαρ) fοοd, εἶλαρ shelter, etc.; -ϝερ in πίειρα, feminine of πίων fat. The ancient grammarians noticed that the stem before -ᾰρ is long (Herοdian ii.769 ed. Lentz).
-μο with the Ο-form.
πότ-μο-ς (πετ-) fall
κορ-μό-ς (κείρω) a trunk
ὅλ-μο-ς (ϝελ-) a rolling stone
ῥωχ-μός (ῥηγ-) gully
-μι: in φῆμι-ς report, δύνα-μι-ς power.
-μῖν: in ῥηγ-μῖν beach on which the waνes break, dative ὑσμῖν-ι fight also nominative ὑσμίνη.
-μεν, -μον, -μων.
πυθ-μήν (genitive -μέν-ος) base
δειμῶν (-μον-ος) fearing
θη-μῶν-α (accusative) a heap.
Also the infinitives in -μεν-αι (dative) and -μὲν (locative); see § 84.
-μᾰτ: as δεῖ-μα, genitive -ματ-ος, fear, ὄνομα name, etc.
Of these suffixes -μον and -μᾰτ go with the strong form of the stem, -μεν with the weak form.
With -ο, -η are formed -μενο (in participles), and -μνο, -μνη, as βέλεμνο-ν a dart, λίμνη a marsh; -μνᾰ (-μν-ι̯ᾰ), in μέρι-μνα care.
-μαp, -μωρ: as τέκ-μαρ and τέκ-μωρ a deνice; -μεpο, in ἵ-μερο-ς desire.
-νεσ: τέμε-νος enclοsure, ἴχ-νος imprint, γλῆνος jewel
-νυ: θρῆ-νυ-ς a fοοt-stοοl.
-ρο, -λο: generally with the weak stem.
Also with an auxiliary ᾰ.
-ρι: in ἴδ-ρι-ς knοwίng, ἄκ-ρι-ς mountain top.
-ρυ, -λυ: δάκ-ρυ tear, θῆ-λυ-ς female (θῆ-σθαι).
-τ: θής θη-τ-ός, νύξ νυκ-τ-ός; but chiefly in compounds, as προ-βλής, ἀγνώς.
-ετ, -ητ: accusative ἀργ-έτ-α white (Il. 21.127), also ἀργῆτα (Il. 8.133), dative ἀργέτι and ἀργῆτι (Il. 11.818), κέλ-ης, λέβ-ης.
-το: found with stems
- In the O-form.
κοῖ-το-ς, κοί-τη (κεῖ-μαι) lair
νόσ-το-ς going, return (νέομαι for νεσ-ο-μαι)
οἶ-τος (εἶ-μι) cοurse, fοrtune
βροντή (βρέμω) thunder
- In the weak form.
-τί, σι: generally with the weak stem, as φά-τι-ς saying, πίσ-τι-ς (for πιθ-τις) trust, τί-σι-ς vengeance, δόσις, βόσις, βρῶσις, γένεσις, νέμεσις, ἄνυσις, ἄροσις.
-σιη: as κλισίη a tent, ὑπο-σχε-σίη promise.
-τίνη in δω-τίνη (from δῶ-τις) gift.
κλῑ-τύ-ς a slope
This suffix is especially common in Homer: ἀγορητύς, ἀλαωτύς, βοητύς, γραπτύς, ἐλεητύς, κιθαριστύς, ἀκοντιστύς, ὀαριστύς, ὀρχηστύς, ὀτρυντύς, ῥυστακτύς, τανυστύς.
-τερ: in πατήρ, μήτηρ, θυγάτηρ, εἰνά-τερ-εςς, γαστήρ, ἀστήρ.
-τηρ, -τορ, τωρ: as δο-τῆρ-α and δώ-τορ-α (acc.) giνer
βοτῆρ-ες and βώτορες herdsman
ἐπ-ακτήρ "driνer," huntsman
μήσ-τωρα (μήδ-ομαι) advisor
also of things, with a touch of personification, κρητήρ, ζωστήρ, λαμπτήρ. Feminine -τειρα (-τερ-i̯ᾰ), as δμή-τειρα subduer.
τρ-ο, as ἰη-τρός healer, ἄρο-τρο-ν plough, σκῆπ-τρον, λέκτρον.
-δ, -ῐδ, -ᾰδ; as accusative ἐλπ-ίδ-α hope; λευκ-άδ-α white.
κέλα-δο-ς noise (κέλ-ομαι)
The suffixes -θ-ρο, θ-λο, -θ-μο are produced by combining the verbal suffix or root-determinant -θ (§ 45) with -ρο, -λο, -μο. Thus ὄλε-θ-ρος, γενέ-θ-λη, στα-θ-μός presuppose the verbs *ολέ-θω, *γενέ-θω, *στά-θω (cp. ἐϋ-σταθ-ής, also στῆ-θος) formed like πλή-θω, φλεγέ-θω, μινύ-θω, etc. Practically, however, they are single primary suffixes.
-θμο is especially common in Homer, cp. ἀρ- θμός, ἀρι-θμός, κηλη-θμός, ἑλκη-θμός, ὁρχη-θμός, κνυζη-θμός: λύ-θρον, ῥέε-θρα, μέλπη-θρα, μέλα-θρον, βέρε-θρον. Cp. also -θμα in ἴ-θμα-τα going.
114*. Variation of Suffixes.
- Primary suffixes were originally liable to variation of the kind already noticed (§ 106). From the Sanskrit declension, in which the variation is preserved with singular fidelity, it appears that a suffix in general has three different forms or degrees of quantity, called by Sankcrit grammarians the strong, the middle, and the weakest form. Just as in the declension of dyas, Gr. Ζεύς, we find
(1) dyāu- in the nominative
(2) dyău- in the locative dyăv-i (Latin Jŏvi for di̯ĕν-i)
(3) dĭν- or diu- in other cases, so in dā-tǎ "giνer" we have (1) -tār- in the accusative dā-tǎr-am, (2) -τar- in the locative dā-tár-i, and (3) -tr- in the dative dā-tr-é, instrumental dā-tr-ǎ.
Similarly we have the series -ār, -ăr, -r; -mān, -măn, -mn; -νān, -νăn, -νn ; -ān, -ăn, -n, etc., the rule being that the first or strong form contains a long vowel, which in the second is short, and in the third disappears altogether.
In the combinations -νa, -i̯a the a is lost and the semivowel becomes a vowel, thus giving -u, -i
- In Greek we find the same suffixes as in Sanskrit, with the further distinction that the vowel may be η or ω, ε or ο. Thus we may have
-τωp, -τοp, τηp, -τερ, -τρ
-μων, -μον, -μην, -μεν, -μν (-μᾰ, -μᾰν
-ωσ, -οσ, -εσ
-ϝωσ, -ϝοσ, -ϝεσ, -υσ
-ι̯ωσ, -ι̯οσ, ι̯εσ, -ισ
and so in other cases. Sometimes both sets of forms occur with the same root; as δώ-τωρ, δώτορ-ος and δο-τήρ, δοτῆρ-ος.
The interchange of ο and ε in the suffix -ο (as φίλο-ς, vocative φίλε) belongs to this head.
The three forms of a suffix are hardly ever to be seen in the Greek declension, one of them being usually taken as the stem of all the oblique cases. Thus the strong form is generalized in μήσ-τωρ, -τωρ-ος, the second in δώ-τωρ, -τορ-ος, to the exclusion of the original *μηστρ-ός, *δωτρ-ός, etc. The a weakest form, however, often appears in derivatives.
ποιμήν, ποιμέν-ος, ποίμν-η
δείμων, δείμον-ος, δειμαίνω (for -μᾰν-ι̯ω, -μν-ι̯ω)
θεράπων, feminine θεράπν-η, also θεράπαινα (for -πν-ι̯α)
τέκ-μωρ, τεκμαίρομαι (for τεκμᾰρ-ι̯ο-μαι), etc.
Cp. Latin cαr-ō(n), genitive car-n-is.
- The relation of the forms -ων (-μων, -ϝων), -ωρ (-τωρ), etc., to -ην, -μην, -ϝην, -ηp, -τηρ, etc., has been the subject of much controversy. It is generally agreed that the difference is not original, but arises in each case by differentiation from a single form. Probably it is due to shifting of accent, the suffixes with η being generally accented, while those with ω are found in barytone words. Thus we have the pairs δοτήρ and δώτωρ, ῥητήρ and ῥήτωρ, βοτῆρες and βώτορες, πατήρ but φρᾱ́τωρ, also Latin sor-ōr (Sanskrit sνásā). In composition, too, the loss of accent is regularly accompanied by the change from η, ε to ω, ο.
φρήρ, ἄφρων, etc.
Many exceptions, however, remain unexplained.
- The Nouns of Relationship (the group πατήρ etc.) with one or twο similarly inflected words (ἀστήρ, γαστήρ) are distinguished from the Nouns of the Agent in -τηρ (-τωρ) by the use of the shorter form -τερ in the accusative: πατέρ-α, Sanskrit pitár-am, but δοτῆρ-α, Sanskrit dātā́r-am. Similarly among stems in -n ἄρσην, ἄρσεν-α answer to Sanskrit νŕ̥sh-a, νŕ̥shan-am (instead of -ān-am). This peculiarity has been explained as the result of an original difference of quantity. That is to say, the form pitar (Greek πατερ-) has been taken to be the strong stem, because it is the stem of the accusative. If so, the η of the nominative has to be explained as due to the analogy of the -ηρ of δοτήρ, etc. But this view cannot well be reconciled with the fact that the stem pitar- occurs not only in the accusative pitáram but also in the locative pitár-i. The locative is a case which regularly takes the middle stem; cp. dātā́r-am, locative dātár-i, áçmān-am, locative áçman-i. Hence we must recognlse a group of stems in -r and -n fοrming the accusative with the middle form. Thus the original declension wοuld be (e. g.)
Strong form: Nominative πατήρ
Middle form: Accustaive πα-τέρ-α, Locative πα-τέρ-ι, Vocative πά-τερ
Weakest form, Genitive πα-τρ-ός
The cause of this difference in the treatment of the accusative has still to be found.1
- The stems in -ant, -mant, -νant, (Greek -οντ, etc.) interchange with shorter forms in -at, mat, -vat, Greek -ᾰτ, -μᾰτ, -ϝᾰτ. In Greek the suffix -οντ is used to form the participle present, as φέροντ-α. The chief trace of -ᾰτ is the Doric ἔασσα (ἐσ-ᾰτ-ι̯α) for ἐοῦσα. The forms -μᾰτ, -ϝᾰτ are found in the neuters, such as δεί-ματος, πείρατος, (περ-ϝᾰτ-ος), etc. So in Latin nōmen, nōminis, for nō-mn-is (Sanskrit nā-mn-as).
On the other hand some stems in -ν take -ντ in the oblique cases: λέων, λέοντ-ος, but feminine λέαινα (for λε-ϝν-ι̯α, cp. Latin leō, leōn-is); θεράπων, -οντος, but θεράπ-ν-η; πρόφρων, feminine πρόφρασσα for προφρα-τι̯ᾰ. Cp. § 107.2.
- The suffix of the perfect active participle presents anomalies, both in Sanskrit and Greek, which are not yet satisfactorily explained. The Sanskrit -vāṁs, -νas, -us and Greek -ϝωσ, -ϝοτ, -ῠσ (in -υια for -ῠσ-ι̯ᾰ) seem to represent the original gradation; but the τ of the masculine and neuter oblique cases is peculiar to Greek, as the nasal to Sanskrit. If we suppose a primitive declension (e. g.) ϝιδ-ϝώς, accusative ϝιδ-ϝῶσ-α, genitive ϝιδ-ύσος, etc., this might become accusative ϝιδ-όσ-α, genitive ϝιδ-ϝόσ-ος, etc. (by the same leveling which we have in δώ-τωρ, accusative δώ-τορ-α, genitive δώ-τορ-ος), then accusative Fιδ-ό-α, genitive ϝιδ-ϝό-ος etc. At this stage the endings -ότ-ος, -ότ-α etc., may have been introduced through analogy—perhaps οf the present participle. However this may be, this is one of several instances in nominal declension of τ creeping in to form a stem for the oblique cases.
- Α suffix which originally was closely parallel to the ϝώς of the perfect is to be seen in the -ίων or -ι̯ων of the comparative; Sanskrit -yāṁs, -yas, (-is), Greek -ιων, -ιον, -ισ (in -ισ-τος). Here the ν, in spite of the Sanskrit nasal, is as difficult to explain as the τ of the perfect. However the older endings -ο-α, -ο-ες (for -οσ-α, -οσ-ες) are preserved in the accusative singular masculine and nominative and accusative plural neuter (ἀμείνω for ἀμειν-οσ-α), and the nominative plural (ἀμείνους, etc.). In the Latin -iō̆r, -iōr-is, etc. there is no trace of a nasal. We may compare the variation in αἰών, κυκεών (§ 107 ad fin.).2
- Heteroclite forms occur when different suffixes are brought into a single declension. In particular
(a) Suffixes ending in -ν interchange with suffixes in -ρ. Thus we find πίων, genitive πίον-ος fat, but feminine πίειρα (πῑ-ϝερ-ι̯ᾰ) and the neuter substantive πῖαρ fatness. Also χειμών, but χειμέρ-ιος. (Cp. the Latin femur, femin-is, and jec-ur, jecin-or-is, which is for an older jecin-is.)
(b) Similarly along with ἠώς we have ἠέρ-ιος at dawn, and the adverb ἦρι (Sanskrit ushâs and ushâr).
(c) Final τ is introduced in the suffix; as in ἥπα-τ-ος (for ἡπν-τ-ος, cp. the Sanskrit yakṛt, genitive yakn-as, and the other neuters in -ᾰρ, -ωρ, genitive -ᾰτ-ος, as πεῖραρ, -ᾰτος (for περ-ϝᾰρ, -ϝν-τ-ος); also in neuters in -μᾰ, genitive -μᾰτ-ος (for -μν-τ-ος).
(d) It is probable that the neuters in -ωρ—ὕδωρ, ἕλωρ, πέλωρ, ἐέλδωρ, τέκμωρ, νύκτωρ (accusative used adverbially)—were originally collective or abstract nouns (Joh. Schmidt, Pluralb. p. 193). On this view ὕδωρ waters (German gewässer) is properly a different word from the stem *ὑδα or *ὑδαρ which we infer from the oblique cases. τέκμωρ is originally a collective or abstract from τέκμαρ; similarly ἕλωρ, ἐέλδωρ, πέλωρ, νύκτωρ (cp. νυκτερ-ίς), which only occur in the nominative accusative, are nouns formed like χειμών (χεῖμα), αἰδώς (αἰδεσ- in αἰδέομαι, ἀν-αιδής), γέλως (γελασ- in γελάω), etc. When ὕδωρ, etc., were brought into use as nominatives answering to neuter oblique cases, they naturally followed these in respect of gender. Cp. § 110 (ad fin.).
- 1. Cοllitz in Bezz. Beitr. x. 37 ff.
- 2. The suffixes of the perfect active participle and the comparative have lately been the subject of much controversy, see Brugmannn, K. Z. xxiv. 79 ff., Grundr. §§ 135, 136, pp 403, 417; Joh. Schmidt, K. Z. xxvi. 341 ff., 378 ff., Pluralb. p. 157; Cοllitz, Bezz. Beitr. x. 25, 63. The chief difficulty lies in the nasal of the Sanskrit strong cases. Such a gradation as -vōns (or -vēns), -ves, -us, or -iōns, -iοs (or -ies), -is, is unexampled. Joh. Schmidt takes the nasalized forms (Sanskrit -vāms,-iāms) as his point of departure, but has been unable to explain -vas, -i̯as, -us, -is to the satisfaction of other scholars. Those who assume a primitive -vōs, -i̯ōs have hitherto been equally unsuccessful in accounting for Sanskrit -vāms, -iāms and Greek -ι̯ωv. The explanation of the τ of -στ-ος, etc., is also difficult, but there it is at least certain that it is of secondary origin. It is to be noted that the traces of -ι̯οσ in the comparative are confined to strοng cases, as accusative singular -οσ-α, nominative plural -οσ-εs. Hence the genitive ι̯οv-οs, dative -ι̯ον-ι, etc., perhaps did not take the place of middle forms -ιοσ-οs, -ιοσ-ι, but of the primitive weak forms (-ισ-ος, -ισ-ι?).