Skip to main content

A recent paper by Joshua T. Katz puts forth a new explanation of the morphology of the earliest forms of the active pluperfect in Greek. Building on work by Jay Jasanoff, Katz explains the curious alphathematic endings 1sg. -εα, 2sg. -εαϛ, 3sg. -ει (< -εε) as resting ultimately on changes that arose when Proto-Indo-European secondary desinences were added to dental-final perfect stems, as in (ἐ)πεποίθεα, (ἐ)πεποίθεαϛ, (ἐ)πεποίθει ⟵ *(e-)bhe-bhóidh-ṃ, -s, -t, to the root *bheidh- ‘trust’. Along the way, he accounts also for the endings of the thematic pluperfect (e.g., 3sg. ἄνωγε(ν)) and of the anomalous pluperfect of the verb ‘know’ in Homer (e.g., 3sg. ᾔδη).

The principal difficulty with Katz’s explanation — and arguably an advantage of the leading competing proposal, that of Nils Berg — lies in the fact that Homeric 3sg. forms in -ει have a predilection for verse-final position (cf., e.g., Il. 16.171 ἐπεποίθει), where regular contraction from a proximate preform *-ese is not what anyone would expect. Just how unexpected this is can be seen from the tortured rhetoric of Katz’s very long footnote on the matter. The goal of the present paper is to rescue the essential idea of the pluperfect as the “imperfect of the perfect” by showing how Katz’s proposal can be modified so as to eliminate this contraction. As a bonus, a non-ad hoc reason will be provided for the long augment in the stem ἠϝειδ-.