[In the essay below, originally published in 1991, Whitworth studies the romantic elements of The Comedy of Errors, urging that the play be recognized as romance in its form and in much of its substance. Whitworth focuses in particular on the structure, content, and language of the framing tale of Egeon of Syracuse.]

What in the world can/should/does an editor do to the text of a Shakespeare play?¹ We are reminded by a growing host of performance critics but also, and more significantly, by textual scholars and editors, that play texts are both potential, to be realized in performance, rather than ends in themselves, and, as things in themselves, unstable. We are enjoined to privilege those early texts of Shakespeare—where there exist more than one—which appear to embody his theatrical practice or that of his colleagues, rather than those which represent his first thoughts or a scribe's transcription, and, generally, to have the play in mind as we edit the text.² What then is the role of the textual editor vis à vis a Shakespeare play? How can whatever he does make any real difference? He works perforce only with the printed signs of Quartos and Folios. The director and actors of the play may start from the text he or another editor prepares, but they can and do deviate from it, cut it, rewrite it, rearrange it, re-edit it at will, even throw it away, as the actor Richard McCabe—or was it Puck?—did with a copy of the New Penguin edition of A Midsummer Night's Dream in the RSC's 1989 production. Some directors go back to the original printed texts, circumventing all subsequent editions. But the vaunted indefiniteness of the dramatic text is not for the editor. He must make choices and fix them in print, however he may equivocate and canvass alternatives in his notes. He must fix that which must remain unfixed, fluid, open, ambiguous, always at the mercy, inspired or banal, of producers.³ But if that very production, or the totality of all productions, past, present, and future, is the essence of the play, is he not in some quixotic, perverse way engaged in denying that essence? Is not his enterprise both subsidiary and paradoxical, at the same time both prior to and parasitic upon the living business of the theatre? He works for years to produce a printed text, agonizing over accidentals, solving and resolving cruxes, emending, guessing, inventing, admitting defeat, and delivers his text to be printed, bound, and read. But reading is not the activity for which that play was written. Performing is. And his text, over which he sweated for so long, will never be performed as he edited it. It will be reproduced, transformed, literally, in the sweat of rehearsal and performance: Brook's Dream, Hall's Hamlet, Nunn's Macbeth, Warner's Titus, not his, the editor's. The edition used as script is rarely even mentioned in the theatre programme.⁴ “A poem should not mean but be,” said MacLeish. An editor's text cannot mean. Can it be? Certainly there can be no play on the editor's carefully constructed page. The editor of a novel or a poem or a treatise has no such anxiety. His text goes to its ultimate consumer, the reader, as he has prepared it, unmediated in its essentials. The play editor's text never goes to its ultimate consumer, the play-goer, as he has prepared it, especially in its essentials.

The editor's quest for his text is, from one perspective, a romantic one. He encounters obstacles, fights giants, dodges between Scylla and Charybdis (who shall remain nameless), comes upon ancient strongholds, smoking ruins, signs of skirmishes, dry bones, dust, the landscape of the editorial history of his text. When it is a Shakespeare play, that landscape is vast, the terrain complicated, with traps for the unwary and the challenges of predecessors, the illustrious and the silly, the flamboyant and the hapless, with many of whom he must pause to do battle. But the quest itself is absurd: he can never succeed. He will think he has it, but like Sir Calidore's Blatant Beast, it will escape again, never be finally pinned down, penned in, tamed. What is worse, all productions of his play will repeatedly release it, in virtually infinite mutations, all of them the play, none of them his play. (Even if he convinces himself that he has got the text, he may have lost, or slain, the play.)

To what or whom is the editor of a dramatic text responsible? To the author? He, in Shakespeare's case, is long dead, and besides, we do not, cannot, know what exactly he wrote or wanted. He is, in a crucial sense, irrelevant. To the reader? To the director, to the actor, all of whom are also readers, but who read to different ends than do the student, the teacher, the literary critic, the “mere” reader? And those to whom the ultimate, infinite re-creations of the play belong, theatre practitioners, amateur and professional, great and small, have no responsibility to any editor's text, They may rely upon a single edition, which they will still cut and rearrange as it suits them, or they may work from several, picking and choosing, referring to earlier printed texts, even preferring less reliable texts to more reliable ones.⁵ Their responsibilities are various: to that chimera, “the play,” in the objectivity of which they sometimes display a touching faith, to the audience, society, the box office, sponsors.

The two activities, reading a printed text and seeing/hearing a performance, are, obviously, radically different.⁶ The one thing you cannot do as you read the book is really see and hear a performance (imagination is something else); at a performance, you see, hear, even smell sometimes (as in the 1980 RSC production of As You Like It when the most delicious aroma of roast chicken wafted over the stalls as the banished duke and his men sat down to their feast in Act II, scene vii), but you cannot read the printed text of the play, the book simultaneously. To the reader, the medium is print; to the spectator-auditor, the media are many, but print is not normally one of them. The editor provides for the one activity, the director, actors, designer, composer, and others for the other. The editor may hope fervently that his edition is adopted by some director and that his cherished readings will be spoken and thus achieve a fleeting immortality. But he knows that even if that happens, his text, what he edited, will be only part of the play. He cannot provide for those other essential parts, including the way each phrase and line of his text is delivered. The actor gives meaning to words, can indeed give different meanings to the same words, meanings which the editor cannot entertain or even imagine.⁷ His text is thus at an even further remove from final or definitive meaning. Whether blueprint or skeleton (or some other, always inadequate, metaphor), his text will be the merest starting-point for the performance, that text which was the goal of his long arduous quest. Not only is it destined to be rehandled in the very act of being realized, it is doomed, even as allegedly fixed, permanent, printed, preserved artifact, to be superseded. The beast breaks out as it is being apprehended.⁸

On the other hand, is it really as hopeless as all that? The writing of a play is an act of literary composition. The editor deals with a literary artifact, written, printed, as any literary work is. Written words are the medium, as they are for a novel or a poem. We cannot edit what is not there, that third dimension: the performance. (Nor should the editor, however much he may yearn to do so, stage the play on the page; to do so is to limit the text's potentiality.) We can only edit that text which is different in its structure and layout—speech prefixes, stage directions, act and scene divisions, and so on, rather than authorial voices, paragraphs, quotation marks, chapters, and so on—but similar in its medium: words. The editor of a dramatic text will always have divided, irreconcilable loyalties: to the written text he works from, the material cause (in Aristotelian terms) of what he aims to make, and to the performed play, the final cause.

These and other related questions have occurred to me, a moderately experienced editor of Renaissance dramatic texts but a relative novice as a Shakespeare editor, as I have worked on an edition of The Comedy of Errors for the Oxford Shakespeare. They are, some of them, simple matters, but they are also, I am convinced, fundamental matters. They are theoretical, or rather philosophical, questions about the nature of the editing enterprise where dramatic texts are concerned, and about the nature of those texts themselves. I do not wish to belabour the obvious, nor to over-dramatize harmless drudgery, but I do wish to pose such questions, even the more naive-sounding ones, and to worry, and encourage my fellow-editors to worry, about what we do. This concern has not sprung ex nihilo, but neither is there yet any articulated theory of dramatic textual editing that addresses these and related questions. Greg's famous theory of copy-text and the guidelines for editors that were derived from it went unchallenged for a surprisingly long time, such were Greg's stature and authority.⁹ But even as editors, the great majority of whom have not considered themselves textual theorists per se, have worked, they have recognized the limitations and contradictions in Greg's impressive rationale. Its shortcomings are more clearly seen as the peculiar extra-literary nature of early play-texts, the differences between playwriting and authoring, and the primacy of performance over mere reading emerge and are articulated. I am one of that majority, just an editor, not a bibliographer or textual critic tout court, and the questions and sceptical reflections which have arisen in the course of the workaday business of editing a Shakespeare play-text (scepticism has not yet induced paralysis) have led, not to a new theory, but to a preoccupation with the peculiarity of that business.

With these queries pending, or looming, I propose to engage in a sort of quizzical intermittent trialogue with recent Shakespearean textual theory, specifically as enunciated by the editors of the Oxford Complete Works and related publications, and by other textual critics, all on one side, and “my” play/text, The Comedy of Errors, on the other.¹⁰ On this side too is some limited experience of testing editorial solutions in the arena of performance when I had the opportunity to advise the director Phyllida Lloyd on textual matters as she prepared and rehearsed a production of The Comedy of Errors for the Bristol Old Vic.¹¹ Like Antipholus of Syracuse, I am “smothered in Errors.” I can only hope that my efforts do not prove me, as he claims to be in the second half of that line, “feeble, shallow, weak” (II.ii.35).

The Comedy of Errors is not a particularly difficult text, compared with others in the Shakespeare canon. It does not raise the “two-text” issue since the Folio of 1623 contains the only early version of the play, and it is not radically corrupt, incomplete, or otherwise maimed. It has its own peculiarities, to some of which I shall return later: a small handful of cruxes, some confusion over characters' names, an occasional vagueness in stage directions, as well as the usual misprints, verse-prose transpositions, unmetrical lines, lacunae, and the like. It is a uniquely Shakespearean amalgam of disparate genres, romance and farce, an early comedy that has more in common with Twelfth Night and Pericles than with the other plays more nearly contemporary to it, and little of the comedy of young love so prominent in most of Shakespeare's first ten comedies.

I want to discuss here some of the problems and puzzles, critical and editorial, it poses, within the framework of the theoretical and procedural issues adumbrated above. The romance and the farce of editing Shakespeare's plays will, I hope, be both evoked and illustrated in this context, the particular surprises and misprisings, double-takes and double thinking involved in doing the mundane job of editing this text framed, as it were, by the larger questions, as the immediate, hectic business of the farcical comedy is framed and overarched by the mythical, romance motifs of erring, losing, seeking, and finding. The harmless, hopeful drudge sets forth, in the giant shadows of his predecessors, equipped with the tools of the trade, instinct, and some notion of what the achieved thing should be like. He should not deceive himself that he has a perfect idea of the thing-in-itself. Romance versus farce; editing texts for readers versus performing plays for theatre audiences; the editor's need to choose, to set something and not something else down in print, but with space for glosses, collations, explanations versus the performer's need to say one thing and not another, despite the “openness” of the text, with no place for verbal glosses or commentary, but virtually unlimited scope for glossing by gesture, expression, inflection; telling, in narrative and in introductions and commentaries on the page versus showing, in performance, with sets, costumes, and music on the stage; diegesis versus mimesis—these complementary, often contradictory sets of conventions, requirements, and procedures seem to me to be reflected in the play, The Comedy of Errors.

Romance is essentially a narrative genre, not a dramatic one. I take it that there are fundamental differences between those two modes, which are more or less identical to Plato's diegesis and mimesis.¹² Drama, as theatre, occurs in the present, is immediate, visual as well as aural; narrative is usually in the past tense, is mediated by a narrator who may or may not be the “author,” and nowadays is experienced silently and privately, by reading, though it used more commonly to be experienced aurally and with others. Romance depends upon discursive passages of description, scene-setting, and mood-making, and upon the omniscient narrator's mediation, guidance, information, suspense-building, reassurance, and so on. The time scale is, or can be, vast: “Once upon a time, long ago” is not the dramatist's opening gambit, but it is the essence of the romancer's. Consider Shakespeare's various ploys to overcome that initial obstacle. He uses prologues and epilogues, choruses, frames (themselves either narrative or dramatic), lumps of narrative within the plays (for example, Prospero, Orlando, the Third Gentleman in The Winter's Tale, or Egeon at the beginning of Errors), even a real historical poet, Gower, who appears in Pericles to tell the story that we are unlikely to credit without his assurances as to its authenticity. Only after forty lines, with references to his book, does Gower hand us over to “the judgement of [our] eye”; not content, he presents dumbshows, refers often to his sources, and in all, appears seven times throughout the play, speaking some 300 lines, including an epilogue, all in order to mediate the romance narrative to a theatre audience, to turn telling into showing. Shakespeare's practice in many of his plays amounts to an inversion, or turning inside out, of Plato's and Aristotle's “mixed mode” (a narrative with some direct dialogue): he writes drama, with a lot of narrative, external and internal, to account for that which is beyond the dramatist's and his audience's reach. An Elizabethan paradigm of this kind of mode-switching is George Peele's marvellous little fantasy, The Old Wife's Tale, which announces itself as “tale” but is a play, but a play in which a tale being told turns into a play being performed for an audience which includes the tale-teller herself and her auditors, a play in which several characters tell romance-like tales of travel, hardship, and enchantment.¹³ In the vogue of performance studies, we must acknowledge that there are limits to the dramatist's art, even as we claim that his written text is not fully realized until it is performed. Shakespeare's and others' metadrama is a recognition of those limits; it is also a challenge to them, pushing out the circumscribing walls of the wooden O's and concrete caverns in which the performance is confined. There are things the dramatist cannot do, qua dramatist, when it comes to story-telling, that the romance narrator can do. But in the theatre the theatrical naturally prevails.

To come to cases: in the theatre, will comedy and its vigorous stepchild, farce, where they are present, inevitably overwhelm, if not subvert, romance? Does romance have a chance where farce pops in its zany face (or arse)? Is The Comedy of Errors, romance in its form and in much of its matter, doomed to live on stage in a single dimension, that of farce? Is the romance to be left to readers only, while theatre audiences get farce(d): two works living under one title? Farce is a viewerly, spectator-friendly genre; romance is a readerly, imagination-friendly one. Romance requires imagination, farce leaves nothing to it. As readers of romance, we have to create our own Arcadias and Faerylands, Illyrias, Bohemias, and Ardens (or Ardennes if we are reading the Oxford Shakespeare). Romance is not visual; it is, “to speak metaphorically, a speaking picture.” Farce has to be seen to be (dis)believed.¹⁴ Is it wrong for directors, designers, and actors to take the farce and let the romance go (or worse, send it up)? Or do the improbabilities of romance pushed further, treated comically on stage, necessarily become farcical? That is, is the difference I am talking about one of degree and not of kind? But if so, why isn't Cymbeline, of all the outrageously improbable plays, called a farce, or Twelfth Night? What is the point of Shakespeare's having encased his “farce”—if that is what it is—in a romance, a story which, on its own, has all the sentiment, pathos, and wonder of Pericles? Will we even agree that The Comedy of Errors, in its larger dimension, is romance? For we are told and have been told for a long time that the play is a farce, and theatre practitioners have, it seems, always treated it so.

In 1819, Frederick Reynolds turned the play into an operatic farce. He added songs from other Shakespeare plays, with musical settings by various composers, including Mozart. A reviewer for the European Magazine found Reynolds's enormities just the remedy for a silly, incredible play:

It was attended by the most crowded house since the beginning of the season, and the audience were throughout in a unanimous temper to applaud. … No illusion of the stage can give probability to the perpetual mutations of four persons, paired in such perfect similitude that the servant mistakes his master, and the master his servant; the wife her husband, and the husband his wife. All this so strongly contradicts common experience, that it repels us even in description; but on the stage, with the necessary dissimilarity of countenance, voice, manner, and movement, that occurs between the actors, however disguised by dress, the improbability becomes almost offensive.¹⁵

The anonymous reviewer is carried away by his own rhetoric: no husband mistakes his wife, because there are no twin women.¹⁶ Even the farce, let alone what was left of the romance, failed to work for this dyspeptic critic, but whether audiences liked the play or not, farce it was and farce it remained. C. E. Flower, in his preface to the play in the Memorial Theatre acting edition, makes much of the text of the “Comedy, or as we should now call it Farce” being fully restored in the 1881 Stratford production.

Many of us would concur in Dr. Johnson's opinion that “Shakespeare's plays are not in the rigorous and critical sense either tragedies or comedies, but compositions of a distinct kind,” accurately reflecting “the real state of sublunary nature” with its “chaos of mingled purposes and casualties.”¹⁷ It is well known that Shakespeare drew upon Plautus' comedy, The Menaechmi, about twin brothers from Syracuse, accidentally separated in childhood, one of whom journeys in search of the other. It is also clear that in Act III, scene i, he had in mind the first scene of Plautus' Amphitruo, in which Jupiter and Mercury impersonate Amphitruo and Sosia, master and servant.¹⁸ But the ingredient that Plautus did not provide was the plot which frames, overarches, and ultimately subsumes the comedy of twins mistook, servants beaten, masters maddened, and merchants thwarted. That plot, which sets The Comedy of Errors in another mode altogether, belongs to a different tradition, one which also went back to antiquity, which Shakespeare knew well and turned to again and again, in which he seems to have been more at home than ever he was in the strictly-structured, rule-governed school of classical comedy. Dr. Johnson, keeping to the Folio's tripartite division of the plays into comedies, histories, and tragedies, opined of Shakespeare that “in comedy he seems to repose, or to luxuriate, as in a mode of thinking congenial to his nature … His tragedy seems to be skill, his comedy to be instinct.”¹⁹ I would refine Johnson's distinction and suggest that it was particularly romance and the dramatic genre that approximates it that were most congenial, even instinctive, to Shakespeare throughout his career. He manifests a peculiar fondness for romance, for old, hoary, much-told tales of wonder and wandering, of storms, shipwrecks, pirates, mistaken identity, oracles and mysteries, treachery and betrayal, bravery and devotion, of parents and children, husbands and wives, brothers and sisters torn asunder and tossed by Fate, but brought together at last against all odds. All of his comedies and tragicomedies contain such elements, some are essentially of that kind.

Just such a tale, of course, is that of old Egeon of Syracuse, his wife, and twin sons and twin servants, shipwrecked, separated, rescued, lost, finally reunited after years of despair and searching, but not before further trials, danger, and anguish. The Comedy of Errors begins with it and ends with it, and its dominant moods and motives run right through the farcical comedy, tempering it and transforming it into a new kind of whole which cannot, without distortion, even denaturing, be described or performed simply as “farce.” That long discursive opening scene, in which the actors, especially the one who plays Egeon, must grip the audience's attention and imagination with pure tale-telling, holds the cruel promise of execution for the sad, worn-out old man.²⁰ The comedy which follows must be coloured by it. Johnson may have had in mind that scene, among others, when he complained of the tediousness of Shakespeare's passages of narration:

He affects a wearisome train of circumlocution and tells the incident imperfectly in many words which might have been more plainly delivered in few. Narration in dramatic poetry is naturally tedious, as it is unanimated and inactive and obstructs the progress of the action; it should therefore always be rapid and enlivened by frequent interruption. Shakespeare found it an encumbrance. …²¹

Well, maybe, but he repeatedly brought it upon himself by his choice of romance material. The language of Egeon's narrative is stylized, formulaic, the language of romance: “Once upon a time, long ago …” is the mode; and Egeon's story begins a long time ago, at the beginning of his life, in fact: “In Syracusa was I born. …”

Within a few lines of the start of the next scene, we see that the two plots, Plautine comedy and Hellenistic romance, are related, that the promise of doom will not be kept because the elements necessary to avert disaster and to bring about the happy dénouement begin immediately to assemble. Scene ii begins, in sharp contrast to the deliberate narrative tempo of Scene i, dramatically in mid-conversation, in mid-sentence with a friendly local merchant warning the newly-arrived Syracusans of their danger as proscribed foreigners in Ephesus, and pointing the warning with news of another Syracusan who is to be executed that very afternoon. The “wearisome train of circumlocution,” the narrative, ends, and the brisk, immediate action of drama begins: in medias res takes over from “Once upon a time. …” The frequent reminders of the time of day in the play—in eight of its eleven scenes—keep Egeon and his impending fate constantly in mind while he is absent from the stage, in tension with the expectation raised by the romance conventions of the first scene that it will be averted. Shakespeare's observation of the unity of time, here as nowhere else before The Tempest, heightens that effect. The theme was rendered visually by the clock in Theodor Komisarjevsky's 1938 Stratford production: its hands moved as the hours ticked away, and sometimes ran to catch up. Reference to Egeon (not by name, of course) in the second scene links his plot to that of Antipholus and Dromio of Syracuse: father, son, and servant, unknown to each other, are in the same place at the same time, aliens all three, despairing seekers for each other and the rest of their family.

Furthermore, the two stories are linked immediately and explicitly by references in both to money. Egeon desperately needs money to save his life; barely half-a-dozen lines after he goes off, “hopeless and helpless,” to seek it, Antipholus receives back from the merchant the money he had held in safe-keeping, the very sum, we are soon told (I.ii.81), that Egeon requires. And the place where all of them are, the alien town of Ephesus, is established as one where money counts, and where the making of profit has priority over the taking of pleasure: the merchant excuses himself from accompanying Antipholus on a sight-seeing tour because he has an appointment with “certain merchants / Of whom [he] hope[s] to make much benefit” (I.ii.24-5).²² Shakespeare's changing of Plautus' Epidamnus to Ephesus was no doubt suggested by the primary source for his Egeon plot, the famous story of Apollonius of Tyre in Gower's Confessio Amantis. It was in Ephesus, where his long-lost wife had been restored to life after shipwreck and become a priestess in the temple of Diana, that Apollonius was reunited with her at last. Egeon finds his long-lost Emilia in Ephesus. Thus, years before he dramatized the story in its entirety in Pericles, he seized upon it as the unlikely frame-plot for his most classical comedy. The alteration of Diana's temple to a Christian priory and Diana's priestess to a Christian Abbess are probably due to the prominence of Ephesus and its affairs in St. Paul's New Testament writings, in Acts and the Epistle to the Ephesians, although Gower supplied a hint by referring to Apollonius' wife as an “Abbess.” From Paul, Shakespeare would certainly have known about the reputation of Ephesus for strange goings-on, with evil spirits, sorcerers, exorcists, and others who practised “curious arts,” as well as its artisans and merchants. Perhaps Demetrius the silversmith in Acts XIX who makes idols for the devotees of Diana suggested Angelo the goldsmith who purveys trinkets for the servants of Venus. Plautus' Epidamnus survives, however, in no fewer than seven references in the play.²³ In his choice of Ephesus, whatever the origin of that choice may have been, Shakespeare gave himself both the strangeness, the menace, and the surreal atmosphere of the typical romance setting, and the urban, mercantile, domestic scene of Roman comedy. The very setting embodies the two primary modes that he fused in this play.

Egeon's story of separation at sea introduces that central motif, and sea imagery recurs frequently. Metamorphosis and loss of identity are introduced in the second scene, expressed in the sea image in Antipholus' first soliloquy:

I to the world am like a drop of water
That in the ocean seeks another drop,
Who, failing there to find his fellow forth,
Unseen, inquisitive, confounds himself,
So I, to find a mother and a brother,
In quest of them unhappy, lose myself.

(I.ii.35-40)²⁴

Adriana uses the same image later when, ironically, she is pleading with this Antipholus, the wrong one, not to tear himself away from her (II.ii.128-32). Transformation, dissolution, loss of oneself—this related group of states and their various images form a major theme, or super-motif. In Act II, Adriana wonders if age is diminishing her beauty, causing her husband to seek his pleasure in the company of other women; in Act V, when his own son does not recognize him, Egeon exclaims that grief and time must have altered him beyond recognition (both characters use the rare word defeatures, its only two occurrences in Shakespeare). Time's ravages, added reminders of the immediate, real time that is passing in the play's day, and related to the transformation and dissolution motif, become the subject of two comic exchanges (II.ii.;IV.ii.). Metamorphosis is mentioned repeatedly, sometimes humorously, sometimes fearfully. The workaday city of Ephesus itself is curiously animate: its very buildings are called “Centaur,” “Phoenix,” “Tiger,” and “Porcupine.”

Enchantment continues to work upon Antipholus: someone hands him a gold chain, Dromio brings him a bag of gold. Convinced they are bewitched, he calls upon divine aid—“Some blessed power deliver us from hence”—whereupon a courtesan appears (IV.iii.44) (in Adrian Noble's 1983 RSC production, she rose spectacularly from beneath the floor, scantily clad in red and black). Antipholus and Dromio behold not a heavenly rescuer, but Satan herself. Divine aid will come, and in female form, when the Abbess appears and gives them sanctuary, but not just yet. (Parenthetically, we may notice that Antipholus of Syracuse falls under the spells, as he believes, of a series of enchantresses: Adriana in II.ii., Luciana in III.ii., the courtesan in IV.iii., finally the Abbess in V.i.—one enchantress per act, a neat distribution—with, as prelude, the soliloquy in I.ii. in which he voices his fears of sorcerers, witches, and the like. This underlines his vulnerability and impressionability, and is reminiscent of the case of a famous hero of chivalric romance, Sir Percival, one of the Grail knights of Arthurian legend, whose experiences with women during his quest, including his mother, his sister, and the fiend in female guise several times, similarly underline his susceptibility to error and his innocence. In contrast, Antipholus of Ephesus is always accompanied by men only—his servant, friends, business acquaintances, creditors, the officer who arrests him—until IV.iv., the conjuring scene, when at last he is surrounded by women, who insist that he is mad; his brother thinks himself mad, the victim of witches. Another way in which Shakespeare differentiates the brothers, making the Syracusan the romance protagonist while the Ephesian retains the role of the thwarted and irate husband of domestic comedy, is by giving the former no fewer than six soliloquies and asides, totalling fifty lines, while his brother has none.)

A new order, that of genuine divine authority, not Dr. Pinch's sham, intervenes in the person of the Abbess. Her claim to be able to heal the supposedly mad Antipholus and Dromio is the claim of a power superior to those of mere magic, sorcery, even the devil. Now the Duke returns, leading old Egeon to execution. At its height of frenzy, the comical-farcical action, which, as we have seen, is far from being only that, is interrupted by the resumption of the tragicomic one. But its progress is halted too, literally, physically, by the prostrate Adriana, imploring the Duke to intercede with the Abbess and get her husband restored to her. Farce impedes romance temporarily. At this moment, the two plots merge, under the auspices, as it were, of both spiritual and temporal authority, both benign, the Abbess and the Duke. Romance resumes, and subsumes farce. To be sure, the unravelling will take some 300 more lines and there will be further supposes and surprises, even pathos, along the way.

Time, which Dromio claimed had gone back an hour, has now gone back years, to when the family was whole, before the events narrated by Egeon a few hours earlier took place. The boys were infants then, new-born. It is, fittingly, the Abbess, the holy mother, who gives explicit utterance to the metaphor of rebirth, describing this moment as one of nativity, repeating the word (if the Folio is right) for emphasis:

Thirty-three years have I but gone in travail
Of you, my sons, and till this present hour
My heavy burden ne'er deliverèd.
The Duke, my husband, and my children both,
And you the calendars of their nativity,
Go to a gossip's feast, and joy with me.
After so long grief, such [nativity].

(V.i.403-9)

Both Pericles and Cymbeline use similar language when they are reunited with children whom they had believed long dead.²⁵ The imminent death with which the play began is transfigured into birth: then we met with things dying, now with things, as it were, new-born. That same death-dealing Duke becomes the life-giving lord: “It shall not need. Thy father hath his life.” Patron already to one Antipholus, the Duke becomes godfather to both at their re-christening. Even the little coda, with its comic business between the pairs of twins, not yet entirely free from error though beyond its more baleful effects, ends on that note. The Dromios resolve that since they do not know which is elder, but came into the world “like brother and brother,” they will now go hand in hand, not one before another, a visual image of recognition and reunion, the joining, not the confounding, of water drops, and a verbal reminder of birth and rebirth. Komisarjevsky's clock should by now have been running furiously backward, whirling away the years, for in the biggest and best of the comedy's errors, Time has indeed gone back, all the way from death to birth, from the intense dramatic final moment to the expansive narrative “Once upon a time,” from the end of the play to the beginning of the story. But that, essentially, is what happens in romance.

The editor in his quest may face anything from minor, uncontentious emendations to hopeless cruxes, from mere commas and full stops to be distributed judiciously, to gaping blanks where text should be, and to heaps of text where less, or none, should be. He will be grateful, in the present case, for the relative brevity and relative cleanness of the Folio text of The Comedy of Errors, and that there is only the Folio text to contend with, no two- or three-headed monsters. Errors is the fifth play in the Folio, following The Tempest, The Two Gentlemen of Verona, The Merry Wives of Windsor, and Measure for Measure, all of which were set by the compositors from transcripts made by Ralph Crane. The four plays which follow Errors—Much Ado About Nothing, Love's Labour's Lost, A Midsummer Night's Dream, The Merchant of Venice—are all reprints of Quartos.²⁶Errors stands alone among the first nine plays in the Folio in having apparently been set from Shakespeare's foul papers, a genesis it shares with only seven others in the volume. This orthodox view, held by Chambers, McKerrow, Greg, and nearly everyone since, of the nature of the printer's copy for Errors has recently been challenged by Paul Werstine.²⁷ His argument that authorial foul papers might have been used in the theatre, and thus that the standard “foul papers versus prompt copy” dichotomy may not be so rigid after all, is rebutted by Wells and Taylor.²⁸ Some of the confusions in the text are of the sort usually attributed to unperfected authorial copy: descriptive or narrative stage directions, imprecise distinctions between characters, uncertain or alternate names for characters, missing or imprecise entrances and exits, and so on. All of these require editorial emendation but not all are necessarily problematic for performers.

Take the names, for example. No editor is likely to hesitate before emending “Iuliana” and “Iulia” in the stage direction and first speech prefix at III.ii. The character in question is clearly Luciana, Adriana's sister, who has already appeared and been named at II.i.3. But the misnamings occur in column a of gathering H4, probably the first column of this play set by Compositor C; meanwhile, C's partner, Compositor D, was getting it right seven times—Luc.—in column b of the same page (this is not to suggest that they were necessarily setting the page simultaneously, side by side). Surely Shakespeare, writing his play seriatim and not by formes, did not forget his character's name between the end of II.ii, where she has the last line, and the beginning of III.ii. Did he on the spur of the moment decide to change her name to “Juliana” to avoid confusion with Luce, who had just made her one and only appearance barely eighty lines before, then revert to “Luciana,” the aberration preserved in the foul papers? Compositor D was in no doubt, nor was B, the third Errors compositor, and C himself conformed subsequently, though he vacillated between Luc. and Luci. on three other pages. Such speculation need not trouble even an editor intent on establishing a consistent text; he emends whether it was Shakespeare the composer or C the compositor who erred. A director may never know if he has not seen the Folio text or an apparatus that records such things.

But Luce is a different, more substantial matter. This character, later identified as the kitchen-maid, appears in III.i. and engages in a slanging match with Dromio and Antipholus of Ephesus, who vainly seek entry to their own house, while their twins are inside, enjoying their usurped places. In the Folio, Luce is named once in a stage direction, seven times in speech prefixes, and three times in the dialogue. This is her only scene, though in many productions she returns in the general mêlées in IV.iv. and V.i. (usually taking the small part of the messenger in V.i., which justifies her presence on stage), and to be reunited with the right Dromio as her master and mistress are reunited. John Dover Wilson, in his first Cambridge Shakespeare edition of the play (1922), listed the character as “Luce, or Nell” and he has been followed by many editors since. A few imply that there are two characters, one of whom does not appear. Only the new Oxford edition goes so far as to change the character's name to “Nell,” and to give only that in Dramatis Personae, stage directions, speech prefixes, and dialogue. Luce is expunged.

The basis for this emendation is Dromio of Syracuse's reference to the woman in question as “Nell” in his comic set-piece duologue with Antipholus of Syracuse in III.ii. There is no reason to identify Luce, who appears in III.i., with Nell the kitchen-maid described as globular in shape by Dromio in III.ii. But the person who stood behind the door and “reviled” Antipholus of Ephesus, that is, Luce, is identified as the kitchen-maid at IV.iv.75-6. So did Shakespeare begin with a maid called “Luce,” then change her name to “Nell,” perhaps to avoid confusion with Luciana, or did he originally plan a second maidservant for the second Dromio? If the latter, then he changed his mind before the end where Dromio of Syracuse gratefully relinquishes any claim in the kitchen-maid to his brother (V.i.417-19). In any case, as the text stands in the Folio, the name “Nell” occurs only once, in a set-piece, where Dromio is desperately inventing witty replies to straight-man Antipholus' questions. The scene, with Dromio's (geo)graphic anatomizing of spherical Nell, is the comical counterpart in prose to Antipholus' lyrical wooing of Luciana in rhyming verse in the immediately preceding seventy lines of the same scene; we witness the first wooing, Dromio reports the other, with grotesque embellishment. “Nell” is an ad hoc invention by Dromio; it allows him (and Shakespeare) their harmless necessary pun on “ell”: “What's her name?” “Nell, sir. But her name and three-quarters—that's an ell and three-quarters—will not measure her from hip to hip” (III.ii.110-13). He even spells it out to be sure we get it, as bad punsters usually do. The actor might pause momentarily before replying “Nell, sir,” as if inventing the name and lining up his gross pun on the spot. This single occurrence of the name in this highly artificial context hardly warrants changing “Luce” to “Nell” eleven times.²⁹ No editor, to my knowledge, has bothered to mention the name “Dowsabell,” let alone proposed that Luce be called that. Yet this same Dromio calls this same woman “Dowsabell” at IV.i.110.

Nevertheless, a director may choose to call the character “Nell,” as Phyllida Lloyd did in her 1989 Bristol Old Vic production. She read and was persuaded by the Oxford editors' argument, liked the jingle of “If thy name be called Nell, Nell thou hast answered him well” (III.i.53), and thought the name suited the actress playing the role. My arguments for retaining “Luce” did not prevail; the director simply chose to do otherwise as she was free to do, citing a recent major edition of the play, and no one could protest that the text was grievously violated or directorial whim irresponsibly indulged. The Oxford editors print “Nell,” I would print “Luce,” and both of us will have done as we did for good reasons. Who will be right? What are the criteria for deciding that question? Shakespeare's intentions? Whatever they were, we cannot recover them, and they may anyway have been one thing at one time, another at another. The beast has not been slain or caught, greasy Nell is forever loose.

Other names and other creatures are less troublesome. The place-name “Epidamium” occurs seven times in the Folio, three on pages set by Compositor C, four on pages set by B. Though minim error is a distinct possibility, the agreement of two compositors makes it more likely that they set what they were reading in their copy. But there was no such place.³⁰ May an editor rectify Shakespeare's errors? Some would not: the Riverside Shakespeare and David Bevington, in both his last revision of Hardin Craig's edition (1980) and his new Bantam (1988), read “Epidamium.” Most editors, however, including the Oxford, follow Pope in emending to “Epidamnum.” But the place in question is Epidamnus, the setting of Plautus' Menaechmi. Why the classicist Pope should have chosen the accusative form of the noun (which occurs in various declined forms in Menaechmi) is mildly puzzling. Feeling slightly giddy at venturing where no fool editor has ever rushed before, I would propose “Epidamnus,” the correct Latin name for the city Shakespeare apparently had in mind. But does it matter? In performance, not a whit. A fictitious place, a mere name on two romance characters' Mediterranean itineraries, that is all it is. But for the editor, it must matter, however insignificant it is: something must be printed and justified against the contending alternatives. In performance, almost anything can be said and no one will blink.

The Courtesan's house is called the Porpentine. This is an archaic spelling of “porcupine,” so a modernizing editor should prefer the modern form. Shakespeare used only “porpentine” though “porcupine” was already current; it occurs eight times in his works, five times in Errors. But it is a proper name here. Is that then an argument for retaining its archaic form? No. That is to opt for quaintness, a practice that mars the otherwise splendid Riverside edition. A quaintness quotient has no place in a scholarly modernizing editor's set of guidelines. We modernize other spellings, so why not this? Curiously, it has been relatively modern editors, starting with Aldis Wright in the famous nineteenth-century Cambridge Shakespeare, who have reverted to “Porpentine,” while Rowe modernized the spelling and was followed by editors until Wright. The Oxford, like eighteenth-century editions, has “Porcupine” (and thus obviates the need for a gloss). No editor, I believe, has retained the Folio spelling “Tyger,” another house-name, at III.i.96, nor “Centaure” for the inn where Antipholus of Syracuse lodges at I.ii.9. And in the context of those other recognizable beastnames which abound in the play, “Porpentine” sounds odd in performance as well as looking odd on the page.

Another problem facing editors, but which causes little or no difficulty in the theatre, is the unmetrical line, whether short or long. The New Cambridge editor of Errors rhetorically asks à propos of one such, “Need we be worried by a line which is metrically short?”³¹ Editors, pace Dorsch, usually are, assuming that Shakespeare always wrote regular iambic pentameter and that verse lines which contain fewer or more than the standard number of syllables (excluding feminine endings) must be faulty. But perhaps Dorsch is right and the assumption needs reexamining; Shakespeare, like Homer, must have nodded now and then. Metrifying Shakespeare is harder when a syllable or a word is missing than when there are too many; adding something requires that we invent or reconstruct Shakespeare. The line “A meane woman was deliuered” (I.i.54) has been regularized by most editors, usually by adding F2's “poor” before “mean,” though some recent editors, from Peter Alexander (1951) to Bevington, leave the line as it stands in F. In performance it is easy and natural for an actor to pause a beat before saying “mean” or to emphasize it to mark the difference between this woman and Egeon's own wife, about whom he has just been speaking, and thus fill out the line. The same holds for many other such lines, short by a mere syllable. A good actor will not chant verse or mark the metre obtrusively anyway (nor will he reduce it to prose), and a missing syllable here and there, provided the sense is clear and the surrounding flow of the verse is maintained, is hardly going to be disastrous. But again, an editor, producing a printed text, may feel the lack more keenly and will probably at least consider whether to supply it, even if he finally decides not to, or writes a note asking whether we need be worried by it.

Gary Taylor's eloquent advocacy of invention by editors must be endorsed with caution and caveat (note 10 above). It will seem to many chillingly like an unrestricted licence. The naturally conservative editor and the naturally, or supernaturally, inventive poet seldom cohabit in one mind, and Taylor is right when he observes: “It is because those who have the facility seldom possess the judgement to restrain their inclination that those with a gift for emendation … invariably indulge in it too often” (“Inventing,” p. 43). Though no harm may be done by adding a word to fill out a line of verse, will enough be gained to warrant the in(ter)vention? Taylor's own “mean-born” in the line under discussion seems to me to be a scant improvement over F2's redundant “poor mean.” The condition of her birth is not relevant, and to refer to it here points away from the birth that is relevant, that of her “burden, male twins” (I.i.55). The point is that she is now “mean,” that is poor, of low estate, and so is willing to sell her twin sons to Egeon to be servants to his. If the compulsion to metrify proves too strong, I would favour “A mean young woman was deliverèd,” but, like many editors, most actors, and any audience, I can live with the Folio's mere nine syllables. In general, editors seem less troubled by F's long “Vnwilling I agreed, alas, too soone / wee came aboard” (I.i.60), which some retain, or the very short one resulting from breaking it up into two: “Unwilling I agreed. Alas, too soon / We came aboard” (following Pope).³² Sometimes when the latter choice is made, a note solemnly remarks on the rhetorical effect of such a short line at this decisive moment in Egeon's narrative. Such “effect” has been imported by the editor, of course, in breaking up the long line, and an actor can impart rhetorical effect in his delivery by pausing, sighing, whatever, if such an effect seems appropriate at this point, whether or not the line is printed as one or as two in his script.

The inventing editor will find somewhat more fertile ground in Dromio of Syracuse's frantic outburst at II.ii.192-3: “This is the Fairie land, oh spight of spights, / We talke with Goblins, Owles and Sprights.” Here is another short line, lacking two syllables this time. The Second Folio, that anonymous first edited text of the First Folio, recognizes the problem, but does not get it right somehow, reading “and Elves Sprights.” Pope changed “Elves” to the unmistakably bisyllabic “elvish.” Theobald transformed “owls” to “ouphs.” Most modern editors, however, have stuck with F's three unmetrical monsters. Even the Oxford retains an octosyllabic line, but modernizes Theobald's “ouphs” to “oafs,” a lexically legitimate move all right, but one that creates a misleading and therefore undesirable secondary meaning for modern readers and audiences (as any modernization may run the risk of doing). The short line invites expansion. Is it not plausible that, by haplography, Compositor D conflated “oules and elues and” in his copy to “Owles and”? “We talk with goblins, owls, and elves, and sprites” seems appropriate to Dromio's terrified state, his fevered brain coining monsters pellmell.³³ But, of course, the actor has even more reason here to pause, engage in business, break the line up, and hence stretch it that extra foot, than was the case with Egeon's line in the first scene discussed earlier. If in performance, where and only where his text can become a play, it does not matter if a word is missing, is the editor justified in indulging the inventing itch? Of course, we edit for readers, who can interrupt their reading of the text, who indeed are invited to do so, to jump to the fine print at the bottom of the page where we discuss the options and defend our decisions, as theatre-goers at a performance cannot. For the play editor's peace of mind, the ideal reader of his edition will be a literary reader and not a theatrically-minded one, and will read it as he or she reads a novel, a poem, Johnson's Dictionary, or Lawrence's letters. When Shakespeare was but the prince of poets, happy drudges lost less sleep.

Lacunae of a whole line or more, even in a rhyming verse passage where it seems clear that something is missing, are dealt with in performance, while the editor sweats and strains and, maybe, invents iambic pentameter. A case in point occurs in Errors III.i., at the height of the furious row between those inside the Phoenix and the rightful occupants and guests outside. In a long rhymed passage, immediately following the line in which Luce's name appears twice, the Folio reads as follows:

ANTI[PHOLUS of Ephesus].
Doe you heare you minion, you'll let vs in I hope?
LUCE.
I thought to haue askt you.
S. Dro.
And you said no.
E.DRO.
So come helpe, well strooke, there was blow for blow.

Theobald emended “hope” to “trow” in the first line, producing a triple rhyme, of which there are four others in the passage: lines 19-21, 64-6, 67-9, 76-8. But he produced no more sense. Some modern editors have followed him, including Cuningham, Foakes (new Arden), Wells (New Penguin), Levin (Signet), and Dorsch; the last is peremptory in dismissing Malone's conjecture that a line rhyming with “hope,” perhaps ending with “rope,” had dropped out (p. 68 n.). Just as many, however, have preferred to retain F's “hope,” usually citing Malone's conjecture: Wilson (1922 and 1961), Alexander, Jorgensen (Pelican), Riverside, Bevington, Tetzeli. Only the Oxford, though, both retains “hope” and leaves a space in brackets to indicate that a line is missing before “Do you hear … I hope?” Gary Taylor admits (“Inventing,” p. 43) to not having the temerity to insert his own line into Shakespeare's text in another play, but says further that he would feel no compunction about marking a lacuna and mentioning the conjecture in a note. This he did in the present case, recording “E. Dro. Thou wouldst answer well to hanging, if I had a rope.” This supplies the rhyme, but leaves Luce's (in the Oxford text, Nell's) “I thought to have asked you” still unattached. What did she think to have asked whom? Bevington thinks the missing line should follow rather than precede the “hope” line, but does not conjecture.³⁴

As textual adviser to the Bristol Old Vic production of Errors in 1989, I discussed the lacuna with the director, who, while using the Arden edition as her script, studied the text very carefully, consulting several other editions. Prior to one rehearsal, I composed several alternative lines, one of which she might choose to insert in the gap. Alas for my inventions, I arrived to find that she had decided to ignore the lacuna, keep the Arden's “trow,” and try to make sense of what was there. The actors had invented business to that end. Nell's and Dromio of Syracuse's half-lines were a resumption of a hypothetical previous conversation, a further sally in the former's attempt to seduce the latter. A little personal drama was simmering away indoors even as the larger, more public drama boiled over outdoors. The large Nell spread-eagled the small Dromio against the door. In the fever of the moment, it worked. Dromio's “And you said—no?” became a plea for mercy. No one gave a further thought to the dread lacuna, which in any case was filled from outside when the other Dromio, in mime, thrust a privy member through the letter slot, and the outraged, frustrated Nell inside applied a vacuum cleaner to it. The audience roared its amusement at the mayhem which ensued, and the rejected inventing editor had, willy-nilly, to join in. An emboldened editor may invent, but if he cannot insert his invention in his text, and performers who know about it do not want or need it, it can only survive as a conjecture, buried in a note. What then is its status or point? Inventing Shakespeare only to lose him in the apparatus seems an unprofitable expense of spirit. Yet Taylor's plea is a powerful one, and it enhances that task of helpful drudgery to which editors earnestly commit themselves. It urges the editor to recreate as well as to recover, to become Shakespeare in some sort, momentarily. That his invention, which may not become text, will be only if it is spoken in performance and if it is, will be only for an instant, are the absurd odds against which he plays. Taylor's description of such inventive emendation as game is reminiscent of the late Philip Brockbank's advocacy of “festive scholarship.” Ludic editing serves Shakespeare, the play, and the reader, not just the black signs on the white pages of F. Ludus—the medieval Latin word meant game and play, as students of medieval drama well know. But games have rules and boundaries, as Taylor reminds us. Because he cannot cheat and write his invention into Shakespeare, the gaming editor offers it and hides it wistfully, playfully, at the same time. It may be in print all right, but it is out of bounds, off stage, below the line. A director like John Barton may write hundreds of lines of “Shakespeare” in his adaptation and they get spoken at every performance. The inspired editor invents, and directs furiously in his head.

May an editor adopt a reading recorded only in acting editions? If we mean what we say about the primacy of performance, why not? A crux or a confusion may be clarified by an actor or director who has to get or make sense out of it, and that reading may be passed down in playhouse tradition, unknown to scholarly editors who collate those dozens of other scholarly editions. In Errors I.ii., Antipholus of Syracuse speaks his first soliloquy, quoted earlier (p. 238). In the Hull adaptation of the late eighteenth century, revised by John Philip Kemble in 1811, the Folio's “falling” in line 37 is emended to “failing.” Thus the parallel drawn by Antipholus between the water drop and himself is exact: it seeks its fellow in the vast ocean, and failing to find it, loses (confounds) itself; he, seeking his family in the wide world, unhappy (unsuccessful) in his quest, loses himself. When this emendation was suggested to Owen Teale who played Antipholus in the Bristol production, he grasped it immediately, perceiving the logic and clarity it achieved. An eighteenth-century theatrical emendation lived again in performance 200 years later. It is a tiny change, to be sure, an i for an l, and the improvement in sense is slight if real. But should we continue to resist or ignore it in modern editions because an actor not an editor first invented it, and when a simple explanation, compositorial misreading of i for l, is available anyway? Which reading would Shakespeare have opted for, the actor's, the compositor's, or the editor's? His own—but that begs a few questions.

We have, it would seem, come almost full circle, from making a text by remaking another text which never was, and never will be, what it is meant to be, unless it is performed, and then will be something else quite, to realizing that performance not only makes the play, but can, and often does, make the text itself. Of course, the example just given is a very small one, one word, one letter, in an entire play. We do not edit performance, or base our scholarly editions solely or mainly on texts derived from performance—at least, we do not say that we are doing so. But what of those Shakespeare revisions preserved in the Folio? A modern edition of Hamlet based on the Folio text may, it would seem, be said to be derived from a performance text. But we as editors unmake and remake that particular early text upon which we base our editions, wherever it emanated from, by modernizing its spelling and punctuation, correcting its obvious misprints, spelling out speech prefixes, adding and expanding stage directions, even reconstructing, as the Oxford editors reconstructed the Pericles Quarto, and by emending cruxes, mending lacunae, inventing Shakespeare. In each repeated effort to fix our text, pin it down, get it right, we make the text more, not less, unstable. For each edition is another text, different from all others, be it by no more than a few commas (it is of course always more than that), as each production is a new text, in that larger sense of the word, as well as enunciating a new text in the narrower sense. Every edition is, and is not, definitive. And each performance of a single production has its peculiarities of rhythm, mood, tempo, “feel,” as performers are always telling us.

Actors and directors talk readily about remaking Shakespeare when they do a new production. Indeed, most believe that is what the theatre is about.³⁵ Editors are more reluctant to acknowledge that they too remake Shakespeare, even when they are engaged in producing diplomatic texts or facsimiles. This implies—and this is no earth-shaking discovery of mine—that that TEXT of Shakespeare which we believed, avowedly or tacitly, it was a duty to attempt to recover in toto and exactly, not only is not there to be recovered, that that concept itself is faulty, but that if it were it could never be “restored” in our editions, old- or modern-spelling, diplomatic or inventive. Editors of multi-text plays such as Hamlet and King Lear have come to realize that, if they have the courage of their convictions, they cannot have all the Hamlet or Lear that Shakespeare may have written in their edited texts. They cannot print the complete words of Hamlet and call it Shakespeare's play. To print all the Hamlet that Shakespeare wrote at different times, on first, second and subsequent thoughts, even if one is convinced that he did write all of it, is to produce something under Shakespeare's name that he never invented.³⁶ We can only edit texts, finally, not plays, not authors. Once we accept that our control-text, be it Folio Hamlet or Errors, Quarto Lear or Pericles, is itself only a version, and perhaps a partial one, of Shakespeare's own total Hamlet, Errors, Lear, or Pericles, and that the very existence of that entity is doubtful and unprovable, we may find ourselves freed from the old idolatry. And if we further accept that all our texts are remakes of versions of uncertain provenance, well, we shall be in no worse company than that of the poets banished by Plato from the commonwealth for making counterfeit copies of imitations of the Forms.

This freedom should in turn help us to overcome the “inhibition of seriousness,” as Taylor calls it, the po'-facedness of scientistic textual scholarship which has prevented editors from realizing the playful truth that the object of the quest is given much of its substance and shape by the quester himself. The editor does not return with the captured TEXT in hand, but emerges at the other end of the labyrinthine way with the text that he has found, trouvé.³⁷ Such a concept of the editor's art, partway between setting out to find and retrieve a determinate object that is known to exist a priori out there, and the free, frivolous invention, the parlour-game that Taylor, as he predicts, will be accused of encouraging, seems to me a fruitful one. What I find will, of course, be partly determined by what I look for, and it is no game of blind man's buff that I play, or random hunt I set out upon. Not just any treasure trove will do; ghastly roadside warnings like The Other Shakespeare and the deformed corpses of A. L. Rowse's brood litter the way. But what I cannot come back with, however I may “struggle for the vision fair,” is the one and only Comedy of Errors by Shakespeare.

I might have multiplied the examples of textual puzzles and problems in The Comedy of Errors, and proposed or speculated about them at length. I have avoided the most notorious crux in the play, Adriana's speech beginning “I see the jewel best enamelléd” (II.i.108-12). Whatever an editor decides to print in these lines is bound not to be right. The actress Rosie Rowell managed to make the passage sound quite meaningful as it stands in most modern editions when she played Adriana in Bristol in 1989. Editing inevitably complicates such cruxes, while performance, also inevitably, simplifies them: it must. As Stanley Wells suggests, “For some reason—perhaps because an edition can be annotated—one is more willing to confront a reader than a playgoer with nonsense” (Re-editing, p. 49). But often, what is nonsense on the page is given sense on the stage, or at least an audience is easily persuaded that it makes sense. The moment comes and goes in seconds, and an audience which is not going to worry whether it is “Epidamium” or “Epidamnum” or “Epidamnus,” or fret over Luce or Nell, will be swayed by the gist of what Adriana says as she grieves at her husband's supposed desertion. They will not see the collations and commentaries, or the whole articles devoted to emending and explaining the five lines spoken in fifteen seconds in performance.³⁸ To return to one of the issues raised at the beginning, it is because he must print this or that but not both that the editor has to resolve cruxes, but he sets about it knowing that he can canvass, collate, comment, and explain also, while the actor has to say it, play it, and be done. Shakespeare, of all people, knew that reading a speech and playing it were worlds apart, that obscurity can evaporate in the action of the stage. May it be that some of the famous cruxes are our own inventions, as readers, critics, and editors? Shakespeare did not count on us getting in the way.

I am not yet sure whether to stick to the Folio's double nativity” in the Abbess's final speech (V.i.403-9, quoted above, pp. 240-1). Double nativity at the end of a play about two pairs of twins lost at sea, separated, then reborn, in the words of their new-found mother, is fitting. Besides the play was performed, probably for the first time, at Christmas 1594. Nativity was seasonal. Johnson proposed “festivity” in the final line, Dyce adopted it and was followed by the Oxford edition, but Hanmer's “felicity” fits the line and sums up the tragicomic action best: “After so long grief, such felicity.” Whichever word is spoken in performance, the joy, the felicity, the festivity of comic dénouement and romance rebirth are ambient. Editors will make various decisions; performance will variously fix that particular word at that moment, but will unfix and remake the very text it speaks, playing and showing nativity, festivity, felicity, and more, where the edition can read one of them, collate others, comment on all, but convey none. Text and performance merge, and the printed word, apparently always the same but, as a famous son of ancient Ephesus, Heraclitus, would have known, always in flux, is confounded by the act that gives it being but is itself evanescent, rushing headlong to its own closure, the final curtain.

Romance and farce merge at the end of The Comedy of Errors, I have argued. The local absurdities, cruxes, and confusions of the latter are confounded in the eternal improbabilities and incredibilities of the former. Both genres flaunt unlikelihood, the one calling us to witness with our own eyes that it is true, the other telling us even as it shows us that it cannot possibly be true. To attempt to create a coherent whole out of such unlikely components seems doubly unlikely, ludicrous. But Shakespeare brazenly pulled it off. Setting out yet again to edit a Shakespeare play-text may appear an unlikely, quixotic venture, but we do it all the time, put our names to it, package it and sell it—Wilson's, Foakes's, Wells's, Bevington's, Dorsch's, Whitworth's Errors—claiming, or at least acquiescing in publishers' claims, that we've got it, the play, right here. Performance too is risky, volatile, ephemeral, over as soon as it is done, and just as brazen, each new show—the Lord Chamberlain's Men's, Kemble's, Komisarjevsky's, Nunn's, Noble's, Lloyd's Errors—implying each time that for that time, this one is it, the play, Shakespeare's Comedy of Errors.³⁹ Rectify Shakespeare's Errors? Perform Shakespeare's, the whole of Shakespeare's, and nothing but Shakespeare's Errors? Not very likely. Yet on and off we and they go, doing it because it must be done, repeatedly. None of the products of these efforts, the myriad performances, the endless editions, though they stretch out to the crack of doom, can be Shakespeare's Errors. All of them, hypothetical and actual, future, present, and past, not one before another, may be so called.

Notes

1. A prior question, in the poststructuralist era, might be to do with how “text” in such a context should be construed. I shall use the term in a conventional sense, to denote the written or printed pages which collectively make up a single work of dramatic literature, a play as it is read in book form. The differences between a play text as edited and read, and a play as performed and seen/heard are crucial to this discussion.

2. Among the fuller arguments for this general case is Stanley Wells's Reediting Shakespeare for the Modern Reader (Oxford, 1984); one of the most recent and vigorous is T. H. Howard-Hill's “Modern Textual Theories and the Editing of Plays,” Library, Sixth Series, II (1989), 89-115. Especially pertinent to the concerns of this essay is Howard-Hill's “Playwrights' Intentions and the Editing of Plays,” TEXT, 4 (1988), 269-78, in which testimony is adduced from an unexpected quarter: C. S. Lewis, half a century ago, anticipated central issues in the current text-performance debate.

3. I do not mean that an editor should interpret in the text itself. An example of editorial overfixing is the Oxford Shakespeare's “understand” in the speech of Dromio of Ephesus at Comedy of Errors, II.i.51-3 (all references will be to the Oxford Complete Works, in modern spelling, ed. Stanley Wells and Gary Taylor (Oxford, 1986), unless otherwise indicated). In modern as in Elizabethan English, understand is not hyphenated. To hyphenate it is to interpret reductively, to accord priority, with a nudge and a wink, to the secondary meaning, “stand up under his blows.” An editor as annotator may of course do that in his commentary, if he judges it necessary; it is an actor's job to do so in delivering the line, if he judges it necessary.

4. The New Penguin Shakespeare is the RSC's “house” edition and is frequently mentioned in programme credits. Directors there often use others as well or instead; the Arden is a favourite because of its copious annotation, which, however, actors often admit, can be either an encumbrance or irrelevant to their work. Antony Sher writes that the company used both New Penguin and Arden editions when rehearsing Richard III in 1984: “and when there are discrepancies we'll choose whichever is more useful for our purposes” (Year of the King [London, 1985], p. 156).

5. In his 1988/9 production of Hamlet for the RSC, director Ron Daniels transposed a scene in accordance with the “bad” Quarto of 1603. He also cut 900 lines from the New Penguin text, itself based on the “good” Quarto of 1604/5, with liberal helpings from the Folio.

6. On hearing versus seeing a play in Shakespeare's time, see Andrew Gurr, Playgoing in Shakespeare's London (Cambridge, 1987), pp. 85-97.

7. A good example of this occurred in Deborah Warner's 1988 production of King John, when Salisbury (Edward Harbour), as he spoke the lines “My arm shall give thee help to bear thee hence, / For I do see the cruel pangs of death / Right in thine eye” (V.iv.58-60), to the mortally wounded Count of Melun, reached behind his back to receive a misericord from his companion Pembroke, with which he put Melun out of his agony. The stage business not only conferred a novel meaning upon “My arm shall give thee help to bear thee hence,” but was appropriate in the context, a soldierly act of mercy to a dying, noble enemy who had done the English lords a good turn by warning them of treachery.

8. At least three noteworthy editions of The Comedy of Errors have appeared since I undertook my own in 1985: that in the Oxford Complete Works, T. S. Dorsch's New Cambridge (1988), and David Bevington's Bantam (1988). There was a bilingual German-English edition by Kurt Tetzeli von Rosador in 1982, and the script used in the BBC television production was published in 1984. Meanwhile, a variorum edition is in progress in the United States, and the Arden is being revised, etc.

9. The classic essay “The Rationale of Copy-Text” is reprinted in Greg's Collected Papers, ed. J. C. Maxwell (Oxford, 1966), pp. 374-91. Howard-Hill challenges the rationale and the editorial tradition since Greg in the first article cited in n. 2 above.

10. Besides works already referred to, I would include in this body of recent work in the immediate light of which my (and others') current editing of Shakespeare is being conducted: Stanley Wells and Gary Taylor, Modernizing Shakespeare's Spelling, with Three Studies in the Text of “Henry V” (Oxford, 1979); the same authors' (with John Jowett and William Montgomery) William Shakespeare: A Textual Companion (Oxford, 1987); Wells, Shakespeare and Revision, Hilda Hulme Memorial Lecture, University of London, 3 December 1987 (London, 1988); Gary Taylor and Michael Warren (eds.), The Division of the Kingdoms: Shakespeare's Two Versions of “King Lear” (Oxford, 1983), and several articles written in response to this volume; Gary Taylor, “Inventing Shakespeare,” Jahrbuch 1986 (Bochum, 1986), 26-44; Taylor, “Revising Shakespeare,” TEXT, 3 (1987), 285-304; Taylor, other articles by Taylor and the other members of the Oxford team, and reviews of their books and edition. To these could be added a host of articles, both theoretical and on specific matters relating to the Shakespeare text, such as Folio and Quarto printers and compositors, published in the last decade by such scholars as Fredson Bowers, G. Thomas Tanselle, D. F. McKenzie, Paul Werstine, Howard-Hill, and many others. Among all of these, the Oxford Textual Companion is the one for a desert island, or for the editor-errant travelling (moderately) lightly. It is a monumental work of scholarship, and while editions continue to bloom and fade, it will fertilize them and fodder their editors and critics for generations to come.

11. Preliminary consultations on the text were held in November and December 1988, when Miss Lloyd was Sir Barry Jackson Fellow in the School of Performing Arts, University of Birmingham, with practical work on text and performance with drama students. Several such sessions also took place with the Bristol Old Vic company during rehearsals of the production, which ran from 16 February to 11 March 1989 at the Theatre Royal, Bristol.

12. Aristotle, typically, complicated this straightforward distinction: for him, in drama, the poet's agents, the actors, imitate for him by taking on the characters of the persons; thus drama is in one sense a mediated form (Poetics, 1450a-b). But since for Aristotle action is more important than character, the immediacy of the representation or imitation gives the dramatic genre, tragedy, its superiority over the narrative one, epic (1462a-b). Conversely, Homer is a good narrative poet to the extent that he speaks little in his own person and a great deal in the (assumed) persons of his characters. Another Aristotelian tangent worth pursuing would be the consequences for the subsequent study of drama as poetry of Aristotle's relegation of the elements of performance—spectacle, music, etc.—to positions of minor importance.

13. This dimension of Peele's play is discussed more fully in the introduction to my New Mermaids edition (London and New York, 1996), pp. xxvi-xxvii.

14. I realize that I have oversimplified both romance and farce for the sake of my argument. There can be more to farce than buffets and pratfalls. But critics who label The Comedy of Errors “farce” also oversimplify the genre in focussing on those elements, and thus fail to do justice to the play.

15. Reprinted in Gāmini Salgādo, Eyewitnesses of Shakespeare (New York, 1975), pp. 68-9.

16. But this is as nothing compared with the hash made of the play by one reviewer of the 1989 Bristol Old Vic production. Among other things, he names the wrong actor in the part of Angelo, refers to Adriana's entrance in a swimming pool when it was Luciana who appeared thus, says that the Abbess comes out leading the Ephesian Antipholus, and that Shakespeare unforgivably marries off the Abbess and Egeus (a double howler); that Egeus (again) is “bailed” at the end when, of course, he is pardoned and released unconditionally. Finally, he says, we shall never know which Dromio ends up with Nell (as that production called Luce), when it is perfectly clear a few lines from the end that the Syracusan resigns her with relief to his Ephesian twin (Financial Times, 21 February 1989).

17. Preface to Johnson's edition of The Plays of William Shakespeare (1765), reprinted in W. K. Wimsatt (ed.), Dr. Johnson on Shakespeare (Harmondsworth, 1969), p. 62.

18. The fullest, often irksomely exhaustive treatment of sources for the play is T. W. Baldwin's On the Compositional Genetics of “The Comedy of Errors” (Urbana, Illinois, 1965).

19. Johnson on Shakespeare, p. 64.

20. Phyllida Lloyd, in her 1989 Bristol Old Vic production, had the courage, as most modern directors have not, to put Egeon alone on stage with the Duke for the first scene. Egeon was lit by a single white spot. Some of the Duke's questions were given to recorded voices off, as of a crowd or press corps, but the theatre audience were not distracted, as so often, by a stage audience busily listening and reacting.

21. Johnson on Shakespeare, p. 67.

22. The word “money” occurs twenty-six times in Errors, more than in any other play in the canon. Marks and mart also occur more times than in any other play. Gold/golden are found more times only in Timon of Athens; ducats and merchant(s) more times only in The Merchant of Venice.

23. Plautus, incidentally, set one of his plays, Miles Gloriosus, in Ephesus, and another, Curculio, in Epidaurus, mentioned in Egeon's narrative; most are set in Athens. Epidamnus was in Illyria.

24. On the reading “failing” for F's “falling,” see above, pp. 250-1. Antipholus' lines about seeking and failing to find his mother and brother are an abbreviated romance narrative on the same theme as Egeon's: another link between the plots. The inner plot, the farcical comedy itself, sounds a romance chord. In Menaechmi, the romance story is told briefly in the Prologue, outside the play.

25. Pericles, xxi. 183-5; Cymbeline V.vi.369-71.

26. Wells, Taylor, et al., William Shakespeare: A Textual Companion, p. 39.

27. “‘Papers’ and ‘Prompt Books’: Printer's Copy for Shakespeare's Comedy of Errors,” Studies in Bibliography, 41 (1988), 232-46.

28. Textual Companion, p. 266.

29. But see the note in its defence (Textual Companion, p. 267).

30. Compositor C also set “Epidarus” at another place on the same page where “Epidamium” appears twice. F2 corrected the former to “Epidaurus.” But Shakespeare might have gone back and changed it if he had revised his foul papers, because Emilia later says that she and the children with her had been picked up by men of “Epidamium,” not Epidaurus (V.i.357).

31. T.S. Dorsch (ed.), The Comedy of Errors (Cambridge, 1988), p. 64 (n. to II.ii.181).

32. Henry Cuningham (old Arden) thought “We came aboard …” an incomplete line and proposed to complete it thus: “… and put to sea, but scarce,” an intelligent conjecture, if one accepts the short-line hypothesis.

33. Gareth Roberts has strengthened the case for “owls” and against replacing it with “elves,” arguing that it is quite plausible for Dromio to fear being sucked black and blue by a strix, a screechowl's body housing a witch or other malign spirit (“The Comedy of Errors II.ii.190: ‘Owls’ or ‘Elves’?,” Notes & Queries, N.S. 34 (1987), 202-4).

34. The Comedy of Errors (New York, 1988), p. 29n.

35. Though the kind of informed, sensitive inventiveness that Gary Taylor urges editors to exercise is sometimes in the theatre, unfortunately, usurped by directorial arrogance, duncicality, or perversity.

36. Ergo, we cannot edit Hamlet, but either the bad Quarto or the good Quarto or the Folio of Hamlet. If Hamlet is an eclectic edition, Shakespeare did not write it. If Shakespeare wrote different versions at different times, which survive as Quartos and Folio, eclectic editions are amalgamations and adaptations, misleadingly labelled. A comparable case-history from Elizabethan nondramatic literature is that of Sidney's Arcadia(s). He may have written all of the “old” Arcadia and all of the incomplete “new” Arcadia which breaks off in mid-sentence in Book III, but he did not write the composite Arcadia published in 1593. That was constructed after Sidney's death by his literary executors, who also rewrote some of the “old” Arcadia used to piece out the “new,” and, with a link passage by yet another author, Alexander, was the version read for more than 300 years: the Countess of Pembroke's Arcadia more than her brother's, in a significant sense. Yet even after the discovery and publication of the complete “old” Arcadia early this century, some editors (e.g. Maurice Evans) and critics (e.g. C. S. Lewis, Walter R. Davis) have edited and written about the composite version, insisting, in some cases, on its primacy (see Lewis, English Literature in the Sixteenth Century (Oxford, 1954), pp. 331-3).

37. As in Provençal trobar—find, make, invent, compose, as the troubadours did.

38. The latest and most detailed discussion, of many, is Gary Taylor's “Textual and Sexual Criticism: A Crux in The Comedy of Errors,” Renaissance Drama, N.S. 19 (1988), 195-225. My use of masculine pronouns throughout this essay when referring to Shakespeare editors reflects the fact observed by Taylor at the beginning of his, namely, the virtual absence of women from the field.

39. I have deliberately excluded film and television from this discussion. They are quite different media from the stage, and Shakespeare's plays were not written for them. Film is a cool medium, its message frozen.