In the forest of Arden

Frank Kermode introduces the third edition of "the most useful Shakespeare texts" in history

This is third time round for the Arden Shakespeare. The first set was ground out between 1899 and 1944, and the second between 1946 and 1982 - slow going. The present new edition starts off briskly with Titus Andronicus, King Henry V and Antony and Cleopatra (Routledge, £30 each in hardback or £5.99 in paperback), but the pace may slow, for there are tougher problems ahead than any encountered in these inaugural volumes. Still, the new series will, like the old ones, be worth waiting for. Despite inevitable variations in quality, these editions have been for the best part of a century about the most useful Shakespeare texts you could hope to find.

Their history has not been uniformly glorious. The editors enlisted at the outset of the second series, including myself, were told to revise the old versions rather than edit from scratch. The publishers warned us that they were merely doing a service to the scholarly public, that there was no money in the project; they couldn't afford to reset text and commentary and would have to go on using the ancient stereotypes. This meant that if we changed anything - and in the nature of the case there was plenty that needed changing - we had to replace it with exactly the same number of characters as we had deleted. One had visions of printers close to mutiny, chiselling away at the plates and laboriously soldering on the substituted letters.

My own Edwardian predecessor, under no such constraints, had enjoyed an occasional autobiographical flourish - there were notes that began "I well remember when I was a boy in Warwickshire. . ." Obviously these had to go, but I would have to replace them with language probably only marginally less irrelevant, and of exactly the same length. After a year or so of hopeless struggle, it became as clear to the publisher as it always was to the editors that this method could not succeed, so we were set free and encouraged to begin all over again. I admit that there are mildly redundant notes in my volume which somehow survived the change of plan.

Another notion had to be abandoned: the assumption that the series was of interest to very few people. Around 1950, nobody foresaw the age of the academic paperback, but when it arrived the Arden Shakespeare took the form in which most of the world now knows it, and sold in hundreds of thousands, to the great and deserved benefit of the publishers, though not of the unlucky editors, their contracts being harder to alter than stereotypes, and their mess of pottage long since consumed.

Meanwhile, as more plays came out, and the early titles were updated, the editorial material tended to grow more recondite. Shakespearian bibliography was becoming ever more sophisticated; for instance, it was possible to know which individual compositor set which parts of the original texts, and to study his peculiar working habits. Although this knowledge had little effect on the texts themselves, it filled the introductions with all manner of information that would formerly have been thought irrelevant to the "general reader"; but he or she seemed not to mind. And of course critics, as they must, found new ways of talking about the plays.

The later volumes of the postwar series - Philip Brockbank's Coriolanus, for instance - are a rather different proposition from the earliest, but the obsolescence of Shakespeare editions is a fact of life, and the publishers are probably right to have started all over again. Brockbank's volume will doubtless be ripe for replacement by about 2040.

Of the first three volumes of this present series, the liveliest and most ambitious is Jonathan Bates's Titus Andronicus. Having drawn one of the less favoured plays, he clearly decided he could prove it is far better than his benighted predecessors had believed, indeed that it is a masterpiece in its own right. One sceptical forerunner was the 1953 editor of the Arden Titus, J C Maxwell: Bate assures us that he is "taking a completely different approach". To the generation of second-series editors, Maxwell was a formidable figure, an icily brilliant scholar, famous for the terse post-cards in which he informed his contemporaries of their laughable errors. Had he survived, Bate might have had to brace himself for some bleak communications.

However, it is true that he has done it all vigorously and very differently: he is far more interested in performance than Maxwell (who was admittedly working before the famous Peter Brook production of 1955), and he has expended a lot of time and ingenuity on working out the staging. Indeed he invents stage directions as liberally as Dover Wilson did in the old days. But there is nothing in the least old-fashioned about Bate, and altogether his is a brisk, opinionated piece of work, virtuously modern in its technical aspects and in its attention to theatrical possibility, though a shade strident in its critical claims.

John Wilders, the editor of the new Antony and Cleopatra, is a rather older hand, but he too pays much attention to staging, a topic on which he is well informed. Remembering Evans and Tearle, though, and Ashcroft and Redgrave, I was surprised to read that the play "has seldom been performed satisfactorily on the stage". Now that directors have seen that an unadorned stage is required, rather than the Pharos in one scene and the Colosseum in the next, all it needs is some very good actors, and in that respect at least we are reasonably well off.

Antony and Cleopatra, which survives only in the version of the 1623 Folio, is textually fairly straightforward, and Wilders deals conservatively with such problems as there are, T W Craik faces a rather more complex set of them in his edition of King Henry V, for there is an unreliable Quarto as well as a Folio (the Quarto is reproduced in photographic facsimile, a useful innovation). Craik startled me at the outset by going on about when he was a boy in Warrington, but he proceeds to deal genially with the real issues, gently dissenting from the recent Oxford and Cambridge versions, and following what are obviously the instructions of the current General Editors, to attend more closely than before to the history and problems of performance. This is understandable and right; the first Ardens tended to treat the plays as closet dramas, and so did some of the second batch.

The real test of those new editions will be the day-to-day use of their annotations, still fortunately on the same page as the text. Their value can't be judged on a hasty sampling; but on a first inspection they all look sound, sober and useful, which, since they may have to last for 50 years, is just as well.