14. This term has been employed to explain a number of forms in which a short vowel is lost before a liquid, and the corresponding long vowel follows the two consonants thus brought together.
- ξυμ-βλή-την (βᾰλ-, βέλ-ος)[fn]Middle βλῆ-το was struck.[/fn]
met
- ἔ-τλη (τάλα-ς)
endured
- πλῆ-το (πέλα-ς)
drew near
- πλῆ-το (Sanskrit par-)
was filled
- ἔ-στρω-το (στορε-)
was scattered
- κλη-τός (καλ-έω, κέλ-ομαι)
called
- κασί-γνη-τος (γεν-)
kinsman
- μέ-μνη-μαι (μεν-)
- δμη-τός (δᾰμᾰ-)
tamed
But this long vowel—ᾱ, η, or ω—is clearly of the same nature as the η of σχή-σω (σεχ-), ἐνι-σπή-σω (σεπ-), πε-πτη-ώς (πετ-, πί-πτ-ω), ἄημι (root αυ in αὔρα), or the ω of πέ-πτω-κα (πετ-), ἔ-γνω-ν (root gan), ζω-ός (root gi, hence Greek ζη- and ζω-, for γι̯-η, γι̯-ω). In these and many similar cases "metathesis" is out of the question. Moreover we find several stems of the same character with the long vowel ῡ as ῥῦ-σθαι to shield (ϝρῡ-), ῥῡ-τός drawn (ϝερῠ-, ϝρῡ-), τρῡ́-ω (cp. τρ-η-, root tar). Hence it is probable that the long vowel is of the nature of a suffix, by which a new verbal stem is formed from the primitive stem or "root". This vowel usually does not vary with the personal endings, but is long in all forms of the tense. It cannot be an accident, however, that the same stems appear also as disyllables with a short final vowel: τᾰλ-ᾰ, πελ-ᾰ, στορ-ε, καλ-ε (in καλέ-σαι), γεν-ε (in γένε-σις), δᾰμ-ᾰ, πετ-ᾰ, ϝερ-ῠ in ἐρύ-σαι, and many others. What then is the relation between these forms and the monosyllabic τλ-η, πλ-η, στρ-ω, κλ-η, γν-η, δμ-η, πτ-η?Apparently the difference is ultimately one of accent. The same disyllable would become τάλ-α or τλ-ή as the stress fell upon the first or the second syllables.[fn]Joh. Schmidt, K. Z. xxiii. 277; Brugmann, M. U. i. 1-68; Fröhde. B. B. ix. 119. The whole subject, as Brugmann has recently warned us (Grundriss, ii. § 8, n.1), is full of uncertainty, and it is possible that forms such as pelē- represent the "root" or primitive word, from which not only plē- (πλη-, Latin plē-nus) and pele-, but also pel- (Sanskrit pi-par-ti) and pl- (πί-πλᾰ-μεν), are derived. We are dealing here, not with the derivation of Greek, etc., from Indo-European—where the comparison of other languages, such as Sanskrit, may give us help—but with the formation of Indo-European itself, to which the comparative method is ex hypothesi inapplicable.[/fn]