Techniques such as these are closely related to the non-illusionistic nature of the Elizabethan stage, in which the mechanics of production were frequently visible. Many scenes take place nowhere in particular. Awareness of place was conveyed through dialogue and action rather than through scenery; location could change even within a scene (as, for example, in 2 Henry IV, where movement of the dying King’s bed across the stage establishes the scene as ‘some other chamber’: 4.3.132). Sometimes Shakespeare uses conflicting reactions to an imagined place as a kind of shorthand guide to character: to the idealistic Gonzalo, Prospero’s island is lush, lusty, and green; to the cynical Antonio, ‘tawny’ (The Tempest, 2.1.57-9): such an effect would have been dulled by scenery which proved one or the other right.

In some ways, the changes in Shakespeare’s practice as his art develops favour naturalism. Thus verse becomes freer, metaphor predominates over simile, rhyme and other formalistic elements are reduced, the proportion of prose over verse increases to the middle of his career (but then decreases again), some of his most psychologically complex character portrayals—Coriolanus, Cleopatra—come late. But his drama remains rooted in the conventions of a rhetorical, non-scenic (though not unspectacular) theatre: the supernatural looms largest in his later plays—Macbeth, Pericles, Cymbeline, and The Tempest. The Tempest draws no less self-consciously on the neo-classical conventions of five-act structure than The Comedy of Errors, and Prospero’s narration to Miranda (1.2) is as blatant a piece of dramatic exposition as Egeon’s tale in the opening scene of the earlier play. Heroines of the late romances—Marina (in Pericles), Perdita (in The Winter’s Tale), and Miranda (in The Tempest)—are portrayed with less concern for psychological realism than those of the romantic comedies—Viola (in Twelfth Night) and Rosalind (in As You Like It)—and the revelation to Leontes at the end of The Winter’s Tale is both more improbable and more moving than the similar revelation made to Egeon at the end of The Comedy of Errors.

The theatre of Shakespeare’s time was his most valuable collaborator. Its simplicity was one of its strengths. The actors of his company were the best in their kind. His audiences may not have been learned, or sophisticated, by modern standards; according to some accounts, they could be unruly; but they conferred popularity upon plays which for emotional power, range, and variety, for grandeur of conception and subtlety of execution, are among the most demanding, as well as the most entertaining, ever written. If we value Shakespeare’s plays, we must also think well of the theatrical circumstances that permitted, and encouraged, his genius to flourish.

10. A performance of Julius Caesar in progress at the reconstructed Globe Theatre on Bankside, London

The Early Printing of Shakespeare’s Plays

For all its literary distinction, drama in Shakespeare’s time was an art of performance; many plays of the period never got into print: they were published by being acted. It is lucky for us that, so far as we know, all Shakespeare’s finished plays except the collaborative Cardenio reached print. None of his plays that were printed in his time survives in even a fragment of his own handwriting; the only literary manuscript plausibly ascribed to him is a section of Sir Thomas More, a play not printed until the nineteenth century. In this edition of the Complete Works we replace the fragment previously offered with a newly edited text of the whole play. The only works of Shakespeare that he himself seems to have cared about putting into print are the narrative poems Venus and Adonis and The Rape of Lucrece. A major reason for this is Shakespeare’s exceptionally close involvement with the acting company for which he wrote. There was no effective dramatic copyright; acting companies commonly bought plays from their authors—as a resident playwright, Shakespeare was probably expected to write about two each year—and for whatever reason, companies generally preferred that their plays should not get into print except when they needed to raise money by selling them to publishers. Nevertheless, by one means and another, and in one form and another, about half of Shakespeare’s plays were printed singly in his lifetime, almost all of them in the flimsy, paperback format of a quarto—a book made from sheets of paper that had been folded twice, and normally costing sixpence. Some of the plays were pirated: printed, that is, in unauthorized editions, from texts that seem to have been put together from memory by actors or even, perhaps, by spectators, perhaps primarily to create scripts for other companies, perhaps purely for publication. Some of the plays were printed in shorter editions in some of which the text is variously adapted, paraphrased, and garbled. It has been argued that they have been subject to ‘memorial reconstruction’—put together by actors—though the extent to which the reporters’ defective memories gave rise to the peculiar features of these texts has been much disputed. They make up an unstable grouping often referred to as the ‘bad’ quartos: bad not because they were, necessarily, badly printed or surreptitiously published, but because the quality of the text is different from, and by and large inferior to, that of the alternative versions. The term has become contentious, as it places texts of different character and possible provenance under a single pejorative label. The earliest texts of The First Part of the Contention and Richard Duke of York (usually known by the titles under which they were printed in the First Folio—2 Henry VI and 3 Henry VI) appeared in 1594 and 1595 respectively; they may well be reported texts, but the reports seem to have been based on an earlier version of the plays as Shakespeare wrote them. Also in 1594 appeared The Taming of A Shrew, perhaps better described as an imitation of Shakespeare’s The Taming of the Shrew (the titles may have been regarded as interchangeable) than as a detailed reconstruction of it. The 1597 edition of Richard III is perhaps the best of these heavily variant texts; it seems to be closely based on a performance version. The text of Romeo and Juliet, as printed in the same year, may have been put together by a few actors exploiting a popular success, though some critics regard this too as an acceptable theatre version of the play. The 1600 quarto of Henry V may present a text made for a smaller company of actors than that for which it had been written; here again theatrical adaptation seems to have played a major part in its evolution. The Merry Wives of Windsor of 1602 seems to derive largely from the memory of the actor who played the Host of the Garter Inn—perhaps a hired man no longer employed by Shakespeare’s company. Worst reported of all is the 1603 Hamlet, which also appears to derive from the memory of one or more actors in minor roles. Last printed of the ‘bad’ quartos is Pericles, of 1609, where there are strong indications of memorial reporting even in the absence of a fuller text with which to compare it.

The worst of these reported texts have many faults. Frequently they garble the verse and prose of the original—To be or not to be; ay, there’s the point‘, says the 1603 Hamlet; usually they abbreviate—the 1603 Hamlet has about 2,200 lines, compared to the 3,800 of the good quarto; occasionally they include lines from plays by other authors (especially Marlowe); sometimes they include passages clearly cobbled together and in distinctly un-Shakespearian language to supply gaps in the reporter’s memory. For all this, they are not without value in helping us to judge how Shakespeare’s plays were originally performed. As the fuller texts may have been too long for performance, the shorter versions potentially offer an indication as to how the plays were reshaped for the Shakespearian stage. Their directions may give us more information about early performances than is available in other texts: for instance, the reported text of Hamlet has the direction ‘Enter Ofelia playing on a lute, and her hair down, singing‘—far more vivid than the good Quarto’s ‘Enter Ophelia’, or even the Folio’s ‘Enter Ophelia distracted’. Because these are post-performance texts, they may preserve, in the midst of corruption, authentically Shakespearian changes made to the play after it was first written and not recorded elsewhere. A particularly interesting case is The Taming of the Shrew: the play as printed in the Folio, in what is clearly, in general, the more authentic text, abandons early in its action the framework device which makes the story of Katherine and Petruccio a play within the play; the quarto continues this framework through the play, and provides an amusing little episode rounding it off. These passages may derive from ones written by Shakespeare but not printed in the Folio: we print them as Additional Passages at the end of the play. In general, we draw more liberally than most previous editors on the reported texts, in the belief that they can help us to come closer than before to the plays as they were acted by Shakespeare’s company as well as by others. However, we are mindful that, no matter what their various origins, these texts constitute distinct versions of the plays.