Eliot may not have been wrong in valuing ‘workshop criticism’, or criticism by poets. True, criticism as we know it consists largely of interpretation and evaluation, activities in which writer-critics have no special advantage over critics pure and simple (if the latter description will quite do for Post-Structuralists). But writers have a manifest advantage in criticism that addresses the craft of literature. And this, it seems to me, is a genre of criticism we need much more of, if the art is to thrive. For we now have so many interpretations, and schools of interpretation, and interpretations of schools of interpretation, that many have become unsure what a good poem is – or whether it may not really be bad, or even bourgeois. But how is our need to be supplied? For to write adequately on literary craft calls for more than just any old critic. Only a generalist can hope to put it all together and restore a sense of what poems are really up to, as distinct from what critics have tried to make them.
John Hollander, himself a fine poet, is such a generalist; and Melodious Guile, to my mind the best of his critical books, takes its place – along with Donald Davie’s Articulate Energy and Winifred Nowottny’s The Language Poets Use – among the very few enjoyable and enriching studies of how poetry works. Where Davie discusses syntax and Nowottny diction, Hollander bats the bounding breeze on poets’ ploys (to extend Karen Sparck Jones’s useful term for the way a sentence is employed).
Hollander characterises ploys such as poetical questions and answers, imperatives and invocations, rhymes and reasons (and non-reasons), fetters and garlands, word tasks and word-plays: all in crisp vignettes, with examples that range easily from Job to Jakobson, Hillel and Horace to Dickens and Dickinson, Plato to Wittgenstein, Aeschylus to Ashbery. And he returns to the essay’s origin by adopting georgic digression as his structural principle. Philosophers’ questions lead to questions answered by questions; self-questionings lead on to closure by question, and that, in turn, to the ubi sunt motif and its troping by Spenser. En route, there are many Menippean delights, and much ostranenie, or defamiliarising freshness, in Hollander’s imitatively ‘slant’ procedures; although wanton laterality can make uptake a little difficult at times: ‘Poetry’s own theory of fiction is a book of reflections on transparency.’ The important truth in this called for a less tropic expression, if it was to win ready agreement.
Hollander is fond of a sort of eirenic correctio – ‘I have appeared to suggest that poems talk only to themselves and never to readers’ – which betrays a certain normative purpose. For he almost writes as if defining the sly or self-referring ploys he describes so brilliantly could amount to defining poetry. Can it be that Hollander is in danger of making Emily Dickinson’s ‘Tell the Truth but tell it slant’ a general prescription? (The Achilles heel of poet-critics is that they are crypto-specialists, and know the kind of poetry they want.) Is Waller always slant, or Crabbe, or Whitman? On the last of these Hollander writes convincingly, yet leaves a doubt whether indirection is, for him, quite the main thing.
In some of the best passages, Hollander brings home afresh the sheer extraordinariness and depth of poetic language. However much of a disaster poetry may be as philosophy, poetic questions propound issues of ‘staggering import’. Of course, the answers are likely to be non-answers – questions, perhaps, about the first questions. So Keats follows ‘Where are the songs of spring?’ with ‘Aye, where are they?/Think not of them, thou hast thy music too –’: a ‘revisionary meta-question’ taking the first question to itself. Death hovers throughout the ode, yet Keats turns his persona aside ‘with a strange shudder ... of dread half-warming into erotic excitement’. Perhaps. Autumn, with her ‘hair soft-lifted’ and ‘laden head’, watching ‘the last oozings’, may be half-warmly conceived; and certainly death waits behind the season’s watch. The questions press further here than in ‘Ode on a Grecian Urn’, where they try to evade time, to hold unrealised potentialities. (The love ‘still to be enjoyed’ recalls many 17th-century poems ‘Against Fruition’; ‘Those unheard/Are sweeter’ is not all that far from Cowley’s ‘Thy sweetness is so much within me placed,/That shouldst thou nectar give, ’twould spoil the taste.’) It is extraordinary, when you think of it, how many questions poetry contains. I can well believe that 20 per cent of Shakespeare’s sonnets are interrogative. Not, though, that only 12 per cent of Spenser’s are: I make it 17.
Melodious Guile is partly a grammar of lyric, more wide-ranging and less austerely taxonomic than Christine Brooke-Rose’s A Grammar of Metaphor. Besides questions, Hollander explores literature’s strange imperatives and jussives. ‘Call me Ishmael’, for example, does not so much give information obliquely as instantiate an act of heroic renaming that the reader is to perform. (I wonder if it may not also serve as an epic propositio, taking high mythological ground, and as a self-fulfilling invocatio that adjures the Spirit to address him as Ishmael – that is, as ‘God hears’.) Imperatives are particularly frequent in refrains and closure devices, from which envoys are but a short digression away. Hollander’s preoccupation with the ‘reflexive troping’ of such self-referring devices is noticeable, and prompts the reflection that Melodious Guile may in some sense have turned out a poetics specifically of Post-Modernist poetry.
But this is only half the story. Melodious Guile also offers a philosophy (one of the first) of the conventions of poetry. As such, its main thrust is inevitably against deconstruction, since every formal device, being ‘centred’, counts on the side of resonant redundancy. And indeed Hollander roundly condemns ‘totalitarian scepticism’. But elsewhere he is a little less severe on deconstruction than consistency demands. Perhaps as a good pluralist he feels the need to incorporate it, too. Thus, he is excellent on ‘intertextual answers’ like Frost’s to Keats’s ‘silken ties’, but leaves unquestioned the assumption that every echo is an ‘intertextuality’. And can one legitimise talk of Virgil ‘deconstructing’ Homer, just by slipping in an ‘as it were’? More importantly, the musculature of Melodious Guile seems a little distorted by the effort of showing that the propositions deconstruction subverts never were the point of poetry anyway.
Perhaps in consequence, Hollander tends to follow Philip Sidney rather too closely in identifying poetry with fiction. He even takes literally the Sidneian symbolon that the poet ‘nothing affirms’, playfully adopting in his own criticism the evasive formula ‘whether ... or ... will not admit of easy solution.’ Similarly, he accepts the Renaissance conception of trope as ‘speech wrested from his own natural signification’: but that may not be a tenable idea, now that we have stopped assuming significance to be ‘natural’. Similarly, considering how much of the real language of men in real situations is tropic, it may be a mistake to identify the figurative with the non-illocutionary. Literature can be non-illocutionary even when literal; and many of its most pleasurable indirections have nothing to do with trope. Naming, for example; and catalogues, like the wonderful lists de la Mare is so good at.
Not that Hollander passes up the pleasures of catalogues and word-patterns with their secret plots – as his fine interpretation of Wordsworth’s sonnets on sonnets shows. In ‘Scorn not the Sonnet’, chronological arrangement would have started with Dante: ‘that the chronology is so contorted by the zigzag line of Shakespeare – Petrarch – Tas – so – Camoens – Dante – Spenser – Milton must mean that it is Calliope’s list and that this is indeed poetic, or non-literal, historiography ... Instead of an octave-sestet division – here overridden by a strong enjambment – there is a first group of four poets and a second of three: Dante, Spenser, Milton.’ Evidence for this ‘figure’ the poem is making might have been drawn from Wordsworth’s chiasmic syntax, and his pairing of two Italian poets, to suggest an abba pattern. And I suspect that his ‘private sense of precursorship’, together with the fact that he and Dante were both visionary poets, entitles us to infer the guile of a second quatrain, Dante – Spenser – Milton – (Wordsworth), and so to complete the ghostly octet.
On Wordsworth’s other sonnet-sonnet ‘Nuns fret not’, Hollander’s old-liberal slip shows briefly, when he tries to reassure us ‘that the workers in this sonnet “sit blithe and happy” in no cold or fatuous suppression of the horrible lives of mill-workers.’ Poetry’s pentacle cannot seal out social contexts quite so hermetically; and Wordsworth was perfectly capable of envisaging literally happy workers – as in his ‘Song for the Spinning Wheel’, which is not only full of the happiness of spinning, but of its pleasures: ‘Ply the pleasant labour, ply!’
Often Hollander’s own ploy (the figure, if not exactly the trope, that his own ‘poem’ makes) is to approach some apparently inconsiderable poetic feature from unexpected angles, commenting obliquely but so penetratingly that it comes to seem positively characteristic or central. This mostly succeeds, as in his wonderful anatomy of interspersed long lines, a chapter only a good poet could have written. But occasionally the tactic fails, or needed to be carried further, as with certain of the analyses of texture (an area of weakness in American criticism). Although Hollander is not afraid of linguistic description, in correcting Johnson’s ‘Life of Cowley’ he needed to do more to explain the effect of extension in ‘Till the whole stream which stopped him should be gone’ – a line that depends on such means as rhythm combined with vowel quantity (long in ‘whole stream’). Again, the effect of Thomson’s ‘till all the plume-dark air/And white resounding shore are one wild cry’ is not due only to the prominence of the /ai/ diphthong, nor even (more interestingly) to the fact that ‘the alliterating “one wild” and the assonant “wild cry” contend dubiously for spondaic prominence for an instant, until they fold up into the accented package of “one wild cry”.’ There is also, more simply and directly, a sequence whereby ‘dark’ back vowels contend with front vowels, until at last both merge in the ‘one’ final blurring diphthong of ‘cry’.
Second World and Green World is a selection from Harry Berger’s output during more than two decades. The first book I ever reviewed was his The Allegorical Temper (1957), a rather too adventurous Spenser study. But his subsequent writing has had a not very usual history of improvement beyond all comparison, while changing a little, naturally, with critical fashion (not modishly), until now his essays are among the most deeply considered in the Renaissance field. Their range is remarkable: from theory to close reading, microcosm and mind to More and Vermeer, poetry and prose to theatre, philosophy and visual art. Berger interprets ‘Upon Appleton House’ with perspicuous subtlety, and challenges one to think and think again about Paradise Lost: ‘The very absurdity of the war in Heaven implies the great gap between visual symbol and spiritual referent and shows the need of the new symbolic relations developed in the second half of the poem.’ His philosophical essays, although interesting, read too loosely to impress for their history of ideas. The essays on visual art, on the other hand, are excellent throughout, bringing to bear a precision at times superior to that of the art historians themselves. On Vermeer, in particular, Berger’s accurate descriptions and analyses are more persuasive than Svetlana Alpers’s on the matter of narrative content; and occasionally they even improve on Lawrence Gowing’s.