107. Heteroclite Nouns. This term is applicable to nouns that employ distinct stems. The chief variations are
- Between the vowel declension (stems in -ο and -ᾱ, -η) and the corresponding consonantal forms.
δίπτυχο-ς; accusative δίπτυχ-α
ἐρίηρο-ς; plural ἐρίηρ-ες, ἐρίηρ-ας
(ἀνδράποδο-ν pοst-Homeric); dative plural ἀνδραπόδ-εσσι
ἀλκή ; dative ἀλκ-ί
ὑσμίνη; dative ὑσμῖν-ι
ἰωκή; accusative. ἰῶκ-α
Ἀΐδη-ς, genitive Ἀΐδα-ο; also Ἄϊδ-ος, dative Ἄϊδ-ι
φυλάκους (or φυλακούς, as Aristarchus accented the word);
also φύλακ-ας, dative plural φυλάκ-εσσιὄσσε, dative plural ὄσσοισι (Hes. Sc. 426)
πολλό-ς and πολύ-ς are both declined throughout: so δάκρυο-ν and δάκρυ.
- With forms in -τ or -ᾰτ.
γόνυ, genitive γουνός (for γονϝ-ός), plural γοῦν-α, γούν-ων, γούν-εσσι; also γούνατ-ος, etc.
δόρυ, genitive δουρός (for δορϝ-ος), etc.; δούρατ-ος, etc.
ὄνειρο-ς; plural ὀνείρατ-α
πρόσωπο-ν; plural προσώπατ-α, dative προσώπασι
Hence the form ὦπα (εἰς ὦπα ἰδέσθαι, κατʼ ἔν-ωπα ἰδών) may be a neuter singular: cp. Aeοlic ὄππατα eyes[fn]The old explanation of ὄππα from ὁπ-μα, by "progressive assimilation," seems to be groundless.[/fn].
οὖς; genitive οὔατ-ος, dative plural οὔασι and ὠσί
ἦμαρ (cp. ἡμέρ-α); ἤματ-ος, etc. (cp. ἠμάτ-ιος)
So πεῖραρ (πείρατ-α), ἦπαρ, οὖθαρ, εἶδαρ, ὄνειαρ, φρεῖαρ, κτέαρ, ἄλειφαρ, στέαρ.
ὕδωρ, ὕδατ-ος. See § 114*.8.d.
χάρις, accusative χάριν (cp. χαρίεις); plural χάριτ-ες, etc.
μέλι (μείλινος, μελι-ηδέα); μέλιτ-ος, etc.
χρώς, χρο-ός, χρο-ί, χρό-α; also χρωτ-ός (Il. 10.575) and χρῶτα (Od. 18.172, 179)
We should add the whole class of nouns in -μα, genitive -ματ-ος: since the -μᾰ of the nominative and accusative is not for -μᾰτ, but answers to the Latin -men, genitive -min-is.
- Between -ασ- and -εσ.
τέρας, τέραα, τερά-ων, τερά-εσσι; but τείρεα (in the sense of "stars," Ιl. 18. 485)
οὖδας, οὔδε-ος, etc.
So κῶας, κώε-α, κτέρας, κτέρε-α (and New Ionic γέρεα, etc.; Attic βρέτους, κνέφους)
This variation doubtless arose from the Ionic change of ᾰο, ᾰω into εο, εω. Thus the ε first appeared in the genitive, giving (e. g.)
sing. τέρας, τέρεος, τέραι
plur. τέραα, τερέαν, τέρασι or τερά-εσσιThen ε was extended to other cases, and on the other hand α was sometimes restored, as in τεράων, κρεάων. See § 106.4 and Joh. Schmidt, Pluralb. p. 325.
- Comparatives in -ων (genitive -ον-ος) sometimes form cases as if by contraction with a stem in -οσ.
ἀμείνω (for ἀμείνοσ-α, ἀμείνο-α)
πλείους (for πλείοσ-ες)
- Other variations are
ἡνίοχος; accusative ἡνιοχῆ-α, nominative plural ἡνιοχῆ-ες
Aἰθίοπ-ες, etc., but accusative Aἰθιοπῆ-ας
Aντιφάτη-ς, accusative Ἀντιφατῆ-α
Ἄρης, vocative Ἆρες; genitive Ἄρη-ος and Ἄρε-ος, etc.; accusative Ἄρηα and once Ἄρη-ν (Il. 5.909)
ζαής, accusative ζαῆ-ν (Od. 12.313): see § 97.
λᾶα-ς, accusative λᾶα-ν; genitive λᾶ-ος, dative λᾶ-ϊ, dual λᾶε, plural λᾶ-ες, λά-ων, λά-εσσι
The latter forms are doubtless by hyphaeresis (§ 105.4) for λάα-ος, etc.
γρῆϋς, dative γρηΐ, as if from a monosyllabic γρηῦς
μέγα (but μεγṇ, Cp. mag-nus), masculine μέγα-ς, μέγα-ν; the other cases from the derivative stem μεγα-λο-
Three apparently distinct stems are used in υἱός sοn.
(1) υἱό-ς, vocative υἱέ; the forms υἱοῦ, υἱῷ, υἱοῖσι are very rare in Homer.
(2) (υἱυ-), accusative υἱέ-α, genitive υἱέ-ος, dative υἱέ-ϊ, plural υἱέ-ες, υἱέ-ας: and from these by hyphaeresis-
(3) αccustaive υἷ-α, genitive υἷ-ος, dative υἷ-ι, dual υἷ-ε, plural υἷ-ες, υἷ-ας, υἱά-σιcp. γρηΰς, λᾶας.
The form υἱάσι (instead of υἱύ-σι) follows the type πατράσι, etc.
The neuter κάρη head forms
(1) genitive καρήατ-ος, κάρητ-ος, dative καρήατ-ι, κάρητ-ι;
(2) genitive κράατ-ος, dative κράατ-ι, plural κράατ-α(ᾱα);
(3) accusative singular κρᾶτ-α (Od. 8.92). genitive κρᾱτ-ός, dative κρᾱτ-ί, plural genitive κρᾱ́των, dative κρᾱσί. The dative singular form κράτεσφι (Il. 10.156) is quite anomalous.[fn]We might add the stem κρη-, in κατὰ κρῆθεν down from the head, cp. κρή-δεμνον, κρή-νη. The relations of these forms have hardly yet been satisfactorily cleared up: see especiallγ Jοh. Schmidt, Pluralb. p. 363 ff. It is highly probable that κέρας is originally the same word, so that the original declension, answering to Sanskrit çiras, çīrshnás, etc., was κέρας, genitive κρᾱ(σ)νός and κρᾱ́(σ)-ατος (like γόνυ, genitive γονF-ὁς and γόνF-ατος, etc.). The form κάρη must have been originally a derivative, introduced to mean head when κέρας had come to be limited to the sense of horn, From it again καρήατος, etc. were obtained by analοgy.[/fn]The declension of ἔρωs, γέλως and ἱδρώς in Homer is open to some doubt ; it is clear however that the stems in -τ are post-Homeric.
Nominative ἔροs occurs in Il. 14.315, accusative ἔpοv in the phrase ἔξ ἔρον ἕντο put away desire, dative ἔρῳ in Od. 18.212; nominative ἔρως is read in Il. 3.442, 14.294, but the meter allows ἔροs in both places. ἔρωτ-α occurs first in H. Merc. 449.
Nominative γέλως occurs in Il. 1.599; Od. 8.326, 343, 344: in the twο last passages (in the Song of Demodocus) the meter is rather against γέλος. The dative γέλῳ occurs in Od. 18.100 (most MSS γέλω); the accusative γέλον or γέλω in Od. 18.350, 20.346 (MSS. γέλων, γέλον, and γέλω). Thus the wοrd may be either γέλο-ς (genitive -ου) or γέλωs, accusative γέλω (for γέλω-α or γέλο-α) : cp. αἰδῶ for αἰδόα, The stem γελοσ- appears in γελοῖος, cp. αἰδοῖος, ἠοῖος.
From ἱδρώς we have accusative ἱδρῶ; but this must be read ἱδρόα in one place (Il. 10.574 ἱδρῶ πολλόν at the end of the line), and always may be so read. The dative is ἱδρῷ (Il. 17.385, 745), possibly to be written ἰδροῖ. Hence ἱδρώς is probably like χρώς.
Twο other case forms of this type are ἰχῶ (Il. 5.416), accusative of ἰχώρ, and κυκειῶ (Il.) or κυκεῶ (Od.), accusative of κυκεών. Cp. also αἰῶ (Aesch. fr. 413), accusative of αἰών.
The history of all these instances is very similar. The original stem ended with a spirant (commonly σ), the loss of which in the oblique cases caused hiatus (-οοs, -οϊ,-οα, etc.); then these forms were replaced by adopting stems in -τ and -v. Cp. § 114.6-8.
108. Heteroclite Pronouns. The following points remain to be noticed.
- The stems ἐμε (με) and ἑε, ἑ do not form a nominative singular.
It is evident that the original nominative coalesced at a very early period with the stem of the verb, becoming the ending -μι ; just as the French je has ceased to be used except in a fixed place before the verb, so that it is hardly a separate word.
In the plural also the nominative was not originally formed from the same stems as the oblique cases. Both ἄμμες, ὔμμε-ς and ἡμέ-ες, ὑμέ-ες are comparatively late, and due to the analogy of the nominal declension (Meyer, G. G. p. 388).
- The interrogative and indefinite τίς is declined from three stems
(1) τι, giving neuter. τί (for τίδ), also the plural neuter traceable in ἅσσα (for ἅ τι̯α). The indefinite ἄσσα occurs in Od. 19.218 ὁπποῖ᾽ἄσσα, where it would be better to write ὁπποῖά᾽σσα (for τι̯α).
(2) τε-, giving genitive τέο, τεῦ (cp. ἐμέο, ετc.), dative τέῳ, τῷ (Il. 16.227, H. Apοl. 170).
genitive τέων (ε̅ω̅), dative in ὁ-τέοισι (ε̅ο̅ι̅, Il. 15.491).
(3) τιν-, giving accusative τίν-α, dative (very rarely) τίν-ι, plural nominative τίνες (only in the Od.).
In the compound ὅσ-τις the first part is sometimes declined as ὅς, ἥ, ὅ, sometimes undeclined, giving ὅ-τις, ὅ-τευ, etc. The neuter plural is once ὅ-τιν-α (Il. 22.450), usuually ἅσσα.
In the forms with ττ, ππ (as ὅττι, ὅππως) we have to recognise the original neuter ὅδ (Sanskrit yad). Thus ὅδ τι becomes ὅτ τι (not ὅστι, since τι is a distinct word, not a suffix). In ὅττεο, which occurs in the Odyssey (1.124, 17.121, 22.377), ὁδ- is indeclinable (cp. ὅ-τις), and so in ὅππως, ὁππόσος, ὁπποῖος, etc. For the assimilation we may compare κὰδ δέ, κὰπ πεδίον, etc. (for κὰτ δέ, κὰτ πεδίον).
- The article is declined from two stems
ὁ-, feminine -ᾱ̔, which gives ὁ, ἡ, οἱ, αἱ: perhaps also ὥς thus, if it is distinct from the relatival ὡς as.
το-, feminine τᾱ-, which gives the other cases, and second forms of the nominative τοί, ταί: also the adverb τώς thus.
The compound ὅ-δε uses the stem ὅ- for the forms ὅ-δε, ἥ-δε, οἵ-δε, αἵ-δε, and the adverb ὧ-δε, The second part is sometimes declined in the dative plural, τοίσ-δεσσιν or τοίσ-δεσιν (Il. 10.462 and Od.). The -δὲ is enclitic: hence the accent, ἥ-δε, not ἧδε. Strictly, therefore, it should be written ὅ δε, ἥ δε, etc.
The forms ἐμαυτόν, σεαυτόν, etc. are post-Homeric. The earliest instance of a compound of this kind is the word ἑαυτῇ, in Hes. Th. 216.