Pindar, perhaps more than any other ancient poet, seems to demand from his interpreters declarations of their critical premises. In recent years scholars customarily have made initial acknowledgment to the work of E. R. Bundy, as psychoanalysts must to Freud, before they begin to offer their own modifications to and expansions of his fundamental work. Much contemporary scholarship has concentrated on the identification and classification in the odes of the elements whose function Bundy labelled and explained. But useful as this (...) type of analysis has been for exorcising the demon of biographical interpretation, it has, like all orthodoxies, prevented perception of other equally important truths. It constitutes no radical heterodoxy to try to account for the fact that each individual ode, for all its dependence on common conventions of structure and of content, makes a different impression. Nor is it unreasonable to try to explain what makes Pindar's style and approach distinctive.In my own work I have argued, though perhaps not always convincingly, that language as well as structure contributes to an ode's coherence. Scholars trained in America are more willing to assume that repetition of phrase or theme within a poem has significance, and that metaphors can simultaneously bear more than one connotation. The issues at stake have most recently been delineated by Michael Silk, in his discussion of the effect of metaphor in archaic poetry: ‘By “patent”, I mean effects whose existence is not in doubt, though their character may be disputed; by “latent”, those whose effective significance is so tenuous or marginal that one resents the impression of solidity that even mentioning them produces. Such insensitivity is more common than it should be among American classicists, many of whom have also been influenced by the “New Criticism”…’ As illustration of the erroneous American approach Silk cites Cedric Whitman's description of the thematic relation of fires in theIliad.Silk himself avoids the trap Whitman falls into by considering only ‘patent’ metaphors, and these consistentlyout of context, so that there is no necessity to comment on the existence or non-existence of thematic connections among them. But it is possible—at least logically—to frame the question differently, and to ask whether a metaphor cannot have patent and latent associations at the same time. (shrink)