What is the aesthetic status of these interactions? I am tempted to answer: that is it, aesthetic. The answer to a further question, ’whose aesthetic?’, is implicit in my opening argument. The aesthetic must be dynamic, representing ’not a having and a resting, but a growing and a becoming’, as Matthew Arnold phrased the human ideal. It must accord with newly recognised possibilities of literature – of any literature – and equally with those long recognised. Above all, if it hopes to illuminate the particular literature in hand, it must be supported by that literature, must not supplant it.

This book is not an attempt to apply to classical literature the habits of modem literary criticism, but as it might be supposed to be just that, I may as well forestall the supposition at the outset. Despite intermittent efforts in recent times, it is still comparatively rare for practising classicists to attempt such ’applications’ and, if anything, especially rare for Hellenists. But the present work represents something much less fashionable altogether: an attempt by a practising classicist to extend the ’theory’ of an aspect of literature in general in the practical context of the literature of antiquity. But in case the claim should seem unduly immodest, it can be said at once that the ’aspect of literature’ in question is, in itself, a small aspect, although not a trivial one. And by way of glossing the claim, let it be said also that the ’habits of modem literary criticism’ and the theoretical apparatus (if any) that accompanies them are not simply separable from their ’traditional’ counterparts. There is rather, as anyone familiar with the ancestry of modern criticism will know, a developing tradition, complex and many-sided, but continuous, which, in its development, sometimes modifies, sometimes innovates entirely and sometimes reconstitutes, in effect, earlier modes.

Intrusion, the second main type of interaction, has comparatively few distinct varieties. The chief reason for this would seem to be that for once considerations of word order or structuring hardly arise. Another contributory factor is that in general the distinctions between the different forms of imagery have little practical relevance here. In particular, the distinction between implicit imagery (i.e. metaphor) and explicit (e.g. simile) is not of much consequence in most cases – with one exception. The exceptional case has already been treated, albeit briefly. In its most rudimentary form, intrusion is confined to comparison and short simile, where a predictable ’N as V’ (or ’more N than V’) is replaced by ’T as V’ (’more T than V’).

It is convenient to regard neutral-based interaction as, so to speak, the nominative kind and, accordingly, define extra-grammatical interaction with reference to it, as I in fact did, to some extent, in the case of intrusion. In the exposition that follows, this is done without further argument. Whatever minor qualifications must be made, it is clear that on the whole neutral terms operate within or through the grammatical structure of the sentence to which they belong. I say and have said ’grammatical’, although it may appear that what I am ultimately appealing to is not grammar but logic. But whether the structure is better regarded as logical or grammatical, there is the fundamental distinction to be made between interactions that work within it and interactions that work outside and obliviously of it.

It may be useful at this point to recapitulate and/or clarify the precise scope of the two chief factors in the discussions that follow, the Greek corpus and the concept of interaction. First, the corpus. The poetry in my period extends from Archilochus at one end to Aeschylus, Pindar and Bacchylides at the other. It comprises, that is, the whole of archaic and early classical verse with the exception of the two hexametric traditions: Homer with the later hymns and epic fragments, and the Hesiodic corpus together with its distant Presocratic relatives, notably Parmenides and Empedocles. In the case of Homer, the main reason for exclusion is the peculiar difficulty of establishing standard usage for the period (’ which period?’), and the situation of the other early hexametrists is comparably problematic. Their successors are omitted only for the sake of genre consistency.

I begin with those categories that involve neutral terminology. The inherent connection of neutral terms with the ground, the likeness of an image, has been indicated already. Interaction of other kinds does not have this characteristic. It is worth stressing that, in many ways, the more apparent the connection with the ground, the less interesting the interaction. This applies especially to explicit imagery, where, of course, the ground, or part of it, is usually made explicit as a matter of predictable organisation.

All the critical twiddle-twaddle about style and form, all this pseudo-scientific classifying and analysing of books in an imitation-botanical fashion, is mere impertinence and mostly dull jargon.’ The first thing to say is that Lawrence’s protest deserves honest respect. If one had to make an exclusive choice between that version of ’criticism’ which confines itself to the technical and the typical, and a kind that sees as its task assessment of particulars unfettered by reference, even, to types and to any sort of technical consideration: if one must choose, one must choose the latter. Comparative inarticulacy is preferable to a decreative sophistication. And the second thing to say is that we need not make such a choice. Our ability to confront literature fruitfully - to be creative - requires articulacy; and true articulacy requires the direction of the recreative mind. But must articulacy imply classification and analysis?

Inasmuch as all imagery embodies the temporary displacement of the terminology ’at issue’ in favour of ’extraneous’ terminology, all imagery embodies a deviation from the terminological norm, albeit a familiar kind of deviation. Metaphor alone has the distinction of achieving this deviation through a simultaneous departure from the normal usage of the language as a whole. This, as is well known, is precisely what the so-called ’dead’ or ’faded’ or ’linguistic’ metaphor does not do. Tree in a family tree, for instance, is a ’dead metaphor’ and involves no departure from normal usage.