INTRODUCTION.

which have determined the virtue of that word, the happy effect is spoiled by an emerging incongruity. For when you pass into imaginative literature, no coquettes are so capricious as words, so easily spoiled in more than one sense, their humors requiring the patient study of a lover.

Nor is the mere sound of a word insignificant. In poetry of all ages, effects frequently depend on similar sounds which represent quite different meanings, as in Pindar’s ἄλλοισι δ' ἅλικες ἄλλοι Original: "alloisin d' halikes alloi", in Homer’s ὠδίνων ὀδύνησι Original: "odinon odynesi", ἀθύρματα θυμῷ Original: "athyrmata thymo", or in Rossetti’s

This is carried further, the poet, as it were, drawing attention to it, when Viola says in Twelfth Night:

The effect of these lines depends on the assonance of the names. Now to the Greeks, similarity in sound meant far more than to modern ears, for they (except a few rationalists) regarded language as a divine invention; and of this view it was a corollary that behind a likeness in sound lay some hidden likeness in fact. And this theory, in combination with a belief in omens, suggested special significances in proper names; onoma ornis, a name is an omen (bird). References to such significances, common to all Greek poets, are a notable feature in Pindar, occurring in almost every hymn¹. And this was recognized by Greek critics. In a note which probably comes from Didymus we read the words: εἴωθε δὲ ὁ Πίνδαρος ταῖς ὁμωνυμίαις ἐπαναπαύεσθαι ἔθει ἰδίῳ Original: "eiothe de ho Pindaros tais homonymiais epanapauesthai ethei idio"—Pindar is accustomed to rest upon homonyms in his own peculiar way. (Scholia on Nemean II. 11). There is a good example in the Second Pythian Ode. Rhadamanthys is there introduced for the sake of his name, interpreted as 'easily learning', and contrasted with the ape who also 'learns in a way':—

Unless we recognize this intention, we shall have to think that Pindar, introducing Rhadamanthys without a motive, had forgotten his cunning.

It is obvious that in many cases, where it would have been improper to mention names, unmistakable allusions could easily be made by

¹ Instances will be found in most of the Odes in this volume. It is needless to cite here the familiar instances of paronomasia from Homer, Aeschylus, etc. The derivations of Iamus and Aias in Pindar are well known.