Variation of the Stem

106. Variation of the Stem. The phonetic influence of the Ending on the form of the Stem, which plays so large a part in the inflection of athematic tenses, vwas originally no less important in the Nouns. In Sanskrit a Nominal Stem of the consonantal Declension appears in general in at least two forms, a 'strongʼ and a 'weakʼ form; the strong form being used in the Nom. and Ace. Sing. and Dual and the Nom. Pluur., the ςweak form in other Cases. The weak form, agan, may have tςwο degrees, which are then called the 'weak b or 'middleb and the ' weakestʼ form. A few traces of these variations remain in the Greek Declension.

1. In the words of relationship, πατήρ, μήτηρ, 8xc. and in ἀνήρ. Γhus ςwe find Nom. πατήρ, Ace. πατέρ-α, but Gen. πατρ-ός(πατέρ-ος only Od. 11. 500), Dat. πατρ-ί (sixty times in Homer, πατέρ-ι thrice); μήτηρ, 0. μητέρα (only), en. and Dat. μητρ-ός, μητρ-ί, less commonly μητέρ-ος, μητέρ-ι. ἀνήρ uses ἀνερ-and ἀνδρ- (for ἀνρ-) almost promiscuously ; the latter is also seen in the Dat. Pl. ἀνδρά-σι (for ἀνδρσι). The genitive plural δαέρων (Il. 24.769) is scanned as a spondee : it should probably be ςwrtten δαιρ-ῶν, the stem δαιρ- standing to δαήρ (for δᾳξήρ) as ἀνδρ- to ἀνήρ (Ubel, 3. δ. i.293.
2. Ζεύς, for δηηύς (Sanscr.dνdμs) forms the Gen and Dat. from the Stem διb. The original Ace. is Ζῆν, Sanscr. dγdm (ςwith loss of μn): Δία follοςws the analogy of Διός, Διί. Similarly βοῦς, for βωῦς(Sanscr.ράμnw), lea. βοξ-ός, Acc.m Hon. βῶν(Sanscr.ράρα). κύων, Voc. κύον, forms the other Cases from the Stem κῦν-. Cp. Sanscr. 2ναn, Ace. 2νάππ-aρm, Gen. 2αn-aς, 8dc. The Ace. κύν-α (like Δία) follows the analogy of the Gen. and Dat. Similarly, aρήν a aρmb (sαrviving in πολύρρην-ες) forms Gen. ἀρν-όςς (for τν-ός), etc.
3. Adjectives in -πις, Gen. -εντος (Stem -2εντ-), form the Dat. Plur. in -ἐσσι, -εσι, Γo explain this we must first suppose the weak Stem in ξἄτ- (ςwith ς for ἐν, cp. § 31.5 and § 37), which ςwοuld give a Dat. Plur. in -ασσι, -ἄἄσι ; this form then vwas assim- ilated to the other Cases by change of ἄἄ to ε. Α form in -ασι has survived in φρασί[fn]Fοund in Pindar, alsο in an Old Attic inscription given by Joh. Schmidt, K. Z. xxv. p. 38.[/fn] for φρεσί (φρᾶ: φρεν ππκ άτ: εντ). In the same vway δαίμοσι, ποιμέσι, ὅRʼc. are not for δαίμον-σι, ποιμέν-σι, but for δαίμᾶ-σι, aποιμά-σι. The Adverb ἀγκάς has been explained as ἀγκάσ(ι), the trαe Dat. Plur. of ἀγκών.
4. The primitive variation sometimes gives rise to paralle forms of a word : e.g. πτώξ and πτάξ a 2ατe (πτήσσω), which origihnate in the declension πτώξ, Ace. πτῶκ-α, Sea. πτακ-ός. So from πούς and Let. ρθτ, ρed-iw we may infer original πούς (or rather πώς), Ace. πόδα or πῶδα, Gen. πεδ-ός : and so in other cases.[fn]Much, hovwever, remains uncertain in the attempts that have been made to reconstruct the primitive declension of these and smiler vwords. The Sanscrit forms vwοuld furnish a fairlgy complete key, but for two defects: (1) the Sanskrit a may represent either ε or ο, so that (e. g.) padás may be ποδός or πεδός, and similarly ā may be η or ω : and (2) Sanskrit ā often answers to Greek ο, so that (e.g.) pādam may point to either πόδα or πῶδα. See Joh. Schmidt, K, Z. xxv. 23 ff., Brugmann, Grundr. i. 5 311, p. 251.