Memorial reconstruction is the most successful explanation in Shakespearean bibliography that was never proved. For most of the twentieth century, the first quarto of Romeo and Juliet, printed in 1597 and known as Q1, was explained by a single elegant story: actors who had performed the tragedy rebuilt it from memory for a printer, and every defect of the short, rough, paraphrasing text flowed from that act of remembering. The story explained the length, the quality, the stage directions, the errors, and the ethics of the book in one motion. It also rested on a method that its most rigorous examiner, Laurie Maguire, eventually showed to be circular: the features said to prove reporting had been collected from the very texts the theory was invented to explain, and the same features turned up in texts nobody accused of being reported. The verdict argued through this article is precise. Memorial reconstruction survives the audit as a possibility for a handful of passages and perhaps for three of the forty-one suspect texts Maguire examined, but as a diagnosis of Q1 of Romeo and Juliet it can no longer bear the weight editors once loaded onto it, and the category of the bad quarto survives only as a label for a kind of book, not as an account of how that book came to be.

The stakes are not antiquarian. Every modern edition of the tragedy rests on a judgment about what Q1 is. If the 1597 quarto is a corrupted report of the fuller text printed in the second quarto of 1599, then an editor may plunder it for stage directions and the occasional line while discarding its readings wholesale. If Q1 instead preserves an earlier authorial version, or a deliberate theatrical abridgement, or some mixture that no single story captures, then its readings carry a different authority and its differences become evidence about the play’s life rather than debris from its theft. The debate over memorial reconstruction is therefore a quarrel about proof: about what kind of evidence may legitimately convict a printed text of being a memory product, who bears the burden of demonstrating it, and how much of twentieth-century editing was built on an inference that hardened into fact through repetition. That quarrel, its invention, its evidence, its named suspects, its audit, and its rivals, is the subject of the pages that follow.

What Is Memorial Reconstruction? The Theory in Its Strongest Form

Memorial reconstruction names a specific mechanism. A play exists in performance, held in the memories of the company that performs it. One or more actors, equipped with nothing but recollection, reproduce the play for a publisher, dictating or writing out the whole text as best they can. The product is a reconstruction: strongest where the reporter’s own part ran, weakest in scenes the reporter merely witnessed, and marked throughout by the characteristic failures of memory rather than the characteristic failures of a copyist. The theory must be stated in this strong form before it can be fairly tried, because its defenders were not fools, and the pattern it predicts is genuinely distinctive.

A text transmitted by pen and paper fails in mechanical ways: eyeskip between similar line endings, misread handwriting, omitted words, transposed letters. A text transmitted by memory fails in human ways. It substitutes a synonym that preserves the sense and breaks the metre. It moves a speech to the wrong scene because the reporter associates it with a similar moment. It anticipates a famous line too early and then repeats it in its proper place. It mishears: a word close in sound replaces the word spoken. It compresses famous set pieces because the reporter remembers the effect rather than the wording, and it expands small parts because the reporter remembers personal business, entrances, and cues with disproportionate clarity. Above all, its accuracy is uneven in a patterned way, rising and falling with the reporter’s presence on stage. No rival mechanism predicts quite that signature, which is why the theory, once framed, proved so satisfying to scholars trained to think of textual variation as a crime scene.

What would a text rebuilt from memory actually look like?

A text rebuilt from memory would paraphrase where a copyist would merely err, preserving gist while losing wording, metre, and image. It would transpose speeches between similar scenes, anticipate celebrated lines before their cue, and repeat material through recollection. Its accuracy would rise where the reporter acted and fall where the reporter only watched.

The playing conditions of the 1590s gave the theory its material plausibility. Actors in Shakespeare’s company learned their roles from parts: narrow rolls containing only their own speeches and the short cues that prompted them, as Tiffany Stern’s work on rehearsal documents has established in detail. A player might perform a tragedy for years while possessing, on paper, nothing but a fraction of it. Memory was therefore not an occasional resource in the early modern playhouse; it was the medium in which whole plays lived. A company member who had acted Romeo, or stood nightly in the wings through the balcony scene, carried a version of the tragedy in his head that no document preserved. The theory asked scholars to imagine that version dictated to a printer’s copyist, and everything odd about Q1 arranged itself around the image.

Stated so, memorial reconstruction also carried a moral charge that helped it spread. The Folio editors of 1623 had already supplied the villainy in their address to readers, complaining of “diverse stolen and surreptitious copies, maimed and deformed by the frauds and stealths of injurious impostors.” A theory in which faithless actors sold their company’s property from memory gave that complaint a mechanism and a cast of culprits. The badness of the bad quarto was double, textual and ethical at once: a maimed text produced by a deformed transaction. Scholars who adopted the theory inherited both halves, and the ethical half did quiet work in persuading readers that the textual half must be true, since a text so obtained could hardly be expected to read well.

The theory also existed in softer variants, and the variants matter to its trial. A reconstruction might be the work of one actor or of several pooling their parts. It might be dictated to a scribe or written out by the reporters themselves. It might be assembled for a rival company, for a touring troupe that had lost its playbook, or for a printer who simply wanted a saleable text of a popular play. It might have been checked against a manuscript for the passages memory could not supply, the hybrid Gibbons later allowed. Each variant preserves the core claim, that memory is the principal medium between performance and print, while shedding a different set of predictions. A single reporter predicts a single gradient; a team predicts several; a manuscript-checked reconstruction predicts almost nothing at all, since any accurate passage can be credited to the manuscript and any corrupt one to the memory. The history of the debate is partly the history of this softening: the strong theory of 1910, one actor and his part, diluting by mid-century into a family of arrangements so accommodating that hostile evidence had nowhere to strike. Maguire’s audit was devastating precisely because it refused to chase the variants and instead tested the core claim in its original, falsifiable form.

One more distinction belongs to the theory’s strongest statement. Memorial reconstruction is not the claim that early play texts passed through performance, which nobody denies, nor the claim that actors’ memories influenced prompt copies and parts, which the documents support. It is the claim that an entire printed book descends from recollection rather than from any written copy of the play. The print shop received, on this account, not a manuscript of the tragedy but a transcript of remembering, and the compositor set type from the record of a mental event. That is an extraordinary provenance to assign to any book without external evidence, and the extraordinary character of the claim is what later critics kept in view while its first sponsors, exhilarated by the pattern they had found, let it pass as the natural explanation of a rough text.

Who Invented the Bad Quarto? Pollard, Greg, and the Making of a Category

The category came first; the mechanism came second; and the order matters, because the category was a judgment before it was an explanation. In Shakespeare Folios and Quartos (1909), the bibliographer Alfred W. Pollard reread the Folio’s address to readers and argued that Heminges and Condell had not meant every earlier quarto when they denounced stolen and surreptitious copies. They had meant a specific, small class of texts, corrupt in wording and illegitimate in origin. Pollard named five: the first quartos of Romeo and Juliet (1597), Henry V (1600), The Merry Wives of Windsor (1602), and Hamlet (1603), together with Pericles (1609). These he called the bad quartos, setting them against the good quartos printed from authorial or playhouse copy. The phrase stuck because it was useful: editors gained a principled reason to distrust exactly the texts that embarrassed them, and the distrust arrived dressed as bibliography rather than taste.

Pollard’s own first mechanism was not memory but shorthand. Piracy by stenography had period colour: Thomas Heywood, in the prologue to If You Know Not Me, You Know Nobody (1605), complained that the play had been snatched into print from stenographic notes, mangled in the taking. A reporter in the audience, pen moving at the speed of speech in one of the era’s shorthand systems, could in principle carry a play out of the theatre in his notebook. The hypothesis had one advantage over every rival: it described a documented period anxiety rather than an invented one. It faltered on capability. The shorthand systems available in the 1590s were slow, ambiguous, and designed for sermons and speeches delivered at a measured pace; the notion of a note-taker capturing rapid dramatic dialogue, in verse, amid stage noise, strained belief even before Maguire later examined the surviving systems and found them unequal to the task.

The mechanism that replaced shorthand came from Walter Wilson Greg, Pollard’s collaborator and the commanding figure of the New Bibliography. Editing the 1602 quarto of The Merry Wives of Windsor in 1910, Greg confronted a text he described in terms that leave nothing to inference: “there is everywhere gross corruption, constant mutilation, meaningless inversion and clumsy transposition.” Greg reasoned that the pattern of error was not stenographic but memorial, and he took the argument one step further, into detection. The corruption in the Merry Wives quarto, he observed, was uneven: the scenes in which the Host of the Garter Inn appears were comparatively well reported, while scenes without him decayed. Greg concluded that the pirate who procured the copy was the actor who had played the Host, reconstructing the comedy from his own experience of performing it. With that edition, the phrase memorial reconstruction entered the discipline, and a theory of transmission became a method for naming individual culprits inside a text.

Why did the bad quarto idea spread so quickly?

The bad quarto idea spread because it solved several problems at once and arrived with force behind it. It explained short texts, rough verse, and strange directions through one mechanism, and gave editors a reason to set aside readings that complicated their copy-texts. Pollard and Greg, its inventors, commanded the journals and editions that carried the doctrine into the universities.

The spread had a sociology as much as a logic. E. K. Chambers lent the theory the weight of his Shakespeare scholarship, and the great editions of the mid-century treated it as settled ground. When Harold Jenkins introduced his Arden Hamlet in 1982, he could praise the identification of memorial reconstruction as one of the achievements of twentieth-century textual scholarship, a compliment that registers how completely the hypothesis had been naturalized: an inference about unobservable events had become a trophy of the discipline’s scientific maturity. Leo Kirschbaum’s census of bad quartos, published in 1938, already listed twenty play texts, and the category had expanded well beyond Pollard’s five Shakespearean exhibits to plays by Marlowe, Peele, Greene, Chapman, and Heywood, each newly labelled text seeming to confirm the class it joined. Peter Alexander added the two early Henry VI texts, The First Part of the Contention (1594) and The True Tragedy of Richard Duke of York (1595), arguing that plays long treated as Shakespeare’s sources were instead his company’s reported property. Even Alexander stepped back from the full claim, retreating to partial memorial reconstruction when the features refused to align, an early sign of the elasticity that would later count against the theory: a mechanism that can be partial, collaborative, or assisted whenever the evidence thins is a mechanism that no evidence can embarrass.

The Romeo and Juliet case sat inside this expansion as one of the founding exhibits, and it is worth pausing on the physical book before the evidence is weighed. The 1597 quarto, An Excellent Conceited Tragedy of Romeo and Juliet, was printed by John Danter and Edward Allde, never entered in the Stationers’ Register, and issued without the author’s name, advertising itself as often played publicly, with great applause, by the Lord Hunsdon’s servants. The full anatomy of that book, its printers, title page, and surviving copies, belongs to the study of the first quarto of 1597 as a physical object, its printing and its title page, and is taken as read here; the concern of this article is the inference drawn from its words. Those words were strange enough. By the usual count the quarto prints some 2,350 lines, stage directions included, where the second quarto of 1599 prints more than 3,200. Roughly a third of the tragedy had gone missing, and what remained frequently read like the same play remembered rather than copied: the sense standing, the phrasing shifted, the verse limping or levelled into prose. The question the New Bibliographers posed was not whether Q1 differed but what single process could produce that profile of difference, and memorial reconstruction was their answer.

The Word Bad: A Short History of a Loaded Label

Pollard chose his adjective with intent, and the intent deserves scrutiny because labels in bibliography are never inert. “Bad” in his usage fused three judgments that have different evidences and different remedies: the text is inaccurate, the edition is unauthorized, and the producers are dishonest. A text can be inaccurate and authorized, as every carelessly printed good quarto demonstrates; an edition can be unauthorized and accurate, as any scribal transcript sold without leave would be; and dishonesty, even where it occurred, describes the transaction rather than the transmission. By welding the three into a single word, Pollard ensured that evidence for any one of them felt like evidence for all. The Folio’s phrase “stolen and surreptitious” did the same work in 1623, and it is no accident that the category was born from a bookseller’s preface: the rhetoric of theft was invented to sell the authorized volume, and the New Bibliography, for all its scientific self-image, adopted the salesman’s vocabulary as its technical term.

The label then steered the scholarship it named. Once Q1 was bad, its readings required no individual assessment; the adjective answered in advance. Editors who might have weighed a Q1 variant on its merits could dismiss it wholesale, and the dismissal was not laziness but doctrine, the doctrine doing exactly the labour it was designed for. Maguire’s replacement term, suspect texts, was more than politeness: suspicion is a state that invites investigation and admits acquittal, while badness is a verdict that forecloses both. The migration of the word across the century, from the pirates to the process to the paper, tracks the migration of the argument itself, and recovering the distinctions the word dissolved is a precondition of judging the debate. Accuracy is a question for collation. Authorization is a question for the archive. Dishonesty is a question for the courts that never heard the case. Only the first of the three has ever been answered for Q1, and the answer, that the text is derivative and theatrical, was never in dispute.

The Case for the Prosecution: What the Evidence Was Said to Show

The evidentiary heart of memorial reconstruction was a catalogue of features, assembled case by case across the suspect texts until it hardened into a diagnostic list. Maguire, auditing the tradition, counted twenty-eight features that had been used since Greg’s Merry Wives study to identify a reported text, and the catalogue deserves a fair hearing in its strongest arrangement before its trial. The features fall into families. There is paraphrase: synonym substituted for word, image flattened into explanation, a line’s meaning surviving its music. There is transposition: speeches and half-lines migrating between scenes that resemble one another. There is anticipation and recollection: a phrase arriving before its moment because the reporter knows it is coming, or echoing after its moment because it has lodged in his ear. There is aural error, the mishearing that swaps a word for its acoustic neighbour. There is metrical decay, verse set as prose and prose broken into verse, the lineation of a man reproducing rhythm from feel rather than from a page. There is borrowing, matter imported from other plays in the reporter’s repertory. And overarching all of these, there is the gradient: quality that tracks a particular actor’s presence, the fingerprint the theory most proudly claimed.

Q1 of Romeo and Juliet supplied exhibits for every family, and the prosecution’s reading of them was coherent. Consider the compression pattern. Where Q2 lingers, Q1 hurries, and it hurries most in the passages a reporter would retain as effect rather than language. Mercutio’s Queen Mab speech in Act 1, Scene 4 runs in Q2 to a sustained aria of miniature folklore, opening on the recognition “O, then, I see Queen Mab hath been with you” (1.4.53) and galloping through the catalogue of dreams she visits on sleepers. Q1 keeps the speech but thins it drastically, a précis of the set piece, as though the reporter carried away the fairy’s itinerary without her embroidery. Consider, against that, what Q1 keeps in full colour. Benvolio, deflecting the masque apologetics in the same scene, remarks in Q1 that the company will have “no without-book prologue, faintly spoke / After the prompter, for our entrance” (1.4, Q1 text). The couplet survives nowhere else, and it smells of the playhouse rather than the study: a joke about performance practice, preserved, the prosecution said, because performance was the very medium of the text’s transmission.

The balcony scene offered the gradient argument its favourite illustration. Juliet’s meditation in Q2 runs “What’s Montague? It is nor hand, nor foot, / Nor arm, nor face, nor any other part / Belonging to a man” (2.2.40-42), a line of reasoning that unfolds limb by limb. Q1 breaks the reasoning short: “What’s Montague? It is nor hand nor foot, / Nor arm, nor face,” and the argument’s completion vanishes. The prosecution read such truncations as memory’s signature, the gist retained and the elaboration lost. Elsewhere Q1 paraphrases where Q2 is exact, substitutes the expected word for the chosen one, and levels distinctive phrasing toward the commonplace, the process Greg’s school called vulgarization. The second Chorus sonnet, present in Q2, is absent from Q1 altogether, a whole formal unit dropped in a way that suited the image of reporters keeping plot and losing architecture. The detailed scene-by-scene collation of such differences, speech by speech and line by line, is a study in itself, pursued in a full comparison of what the Q1 and Q2 differences reveal; the point here is the use the prosecution made of the pattern rather than the pattern’s inventory.

Which parts of Q1 are remembered best, and why does that matter?

The best-remembered stretches of Q1 cluster where Romeo himself is on stage, and the clustering matters because it is the one prediction memorial reconstruction makes that its rivals do not. An authorial draft should differ evenly. An abridgement should cut by design. Only a reporter’s memory should wax and wane with a single actor’s presence.

H. R. Hoppe built the Romeo and Juliet prosecution into its definitive form. His monograph The Bad Quarto of Romeo and Juliet: A Bibliographical and Textual Study (1948) remains the most detailed account of how the 1597 text might have come into being, and later orthodoxy stood largely on its foundations. Hoppe worked through the quarto scene by scene, classifying variants, weighing paraphrase against mishearing against transposition, and concluded that the text had been reconstructed from memory by members of the company that first performed the tragedy, with the reporting strongest where the principal reporter’s own experience of the play was deepest. Alfred Hart’s comparative study Stolne and Surreptitious Copies (1942) had already lent the method statistical dress, measuring the bad quartos’ peculiarities against one another and against the good texts, and finding the family resemblance systematic. George Ian Duthie performed the same office for the 1603 Hamlet in The Bad Quarto of Hamlet (1941), naming the actor of Marcellus as that text’s principal reporter on the strength of the scenes Marcellus shares. By mid-century the prosecution rested: the features were catalogued, the mechanism was named, the culprits were identified play by play, and editors such as Brian Gibbons, introducing the Arden Romeo and Juliet in 1980, could endorse the account for this play while the Oxford Complete Works of Stanley Wells and Gary Taylor carried it into the reading text of a generation.

A fair summary of the prosecution must concede its real strength. The features exist. Q1 is shorter, rougher, and stranger than Q2; its variants do cluster in the families the theory describes; and nobody in the debate, then or since, has proposed that Q1 was printed from a careful transcript of the author’s manuscript. The quarrel was never over whether the 1597 text is derivative and theatrical in character. The quarrel was over whether a specific, named, unobservable process, actors remembering, had been proved by features that might have other causes, and over the confidence with which a hypothesis about the dead had been promoted into a premise of editing.

Naming the Reporters: The Detective Work and Its Limits

No part of the theory displays its character better than the naming of reporters. Greg’s identification of the Host’s actor in The Merry Wives of Windsor set the template: find the role whose scenes are reported with suspicious excellence, and the man who played that role becomes the memorial reconstructor. The method had a satisfying concreteness. Bibliography, a discipline of paper, type, and watermarks, could suddenly point at a person: this actor, in this part, betrayed his company for a printer’s fee. Duthie’s Marcellus followed the template for Hamlet, where the sentry’s scenes and the scenes Marcellus witnesses stand closer to the second quarto than the soliloquies do. For Romeo and Juliet, Hoppe’s scene-by-scene analysis traced the gradient across the tragedy and located the centre of reporting in the scenes of the leading players, the reconstruction, on his account, being the work of company members rebuilding the play around the parts they knew from the inside.

The template conceals a logical difficulty that later critics pressed hard. The gradient is identified by comparing the suspect text against the good text, and the good text is treated as the standard of what memory should have preserved. But that comparison can only show that Q1 differs from Q2 unevenly; it cannot show why. Scenes might be fuller in Q1 because a reporter remembered them, or because an abridger kept them, or because the author had not yet expanded them, or because the compositor’s copy was fuller at those points for reasons lost with the copy itself. The reporter emerges from the data only after the theory has been assumed: the scholar decides that unevenness means memory, finds the unevenness, and names the man whose part the unevenness flatters. Greg himself, in candid moments, treated identifications as the most conjectural stratum of his work, yet the conjectures were quoted for decades without their scaffolding of hesitation.

Can a reporter be identified from the text alone?

A reporter cannot be identified from the text alone, because the evidence that names him was used to invent him. The gradient of accuracy across scenes is real, but it is compatible with abridgement, drafting, and revision as well as with memory, so the step from pattern to person is narrative completion rather than demonstration.

The elasticity of the identifications tells its own story. When one reporter could not account for the whole pattern, the theory supplied a team; when the team faltered, collaboration, dictation, and assistance from a playhouse manuscript were admitted as auxiliaries. Gibbons, the theory’s careful Arden exponent for this play, conceded the softness in a sentence that deserves to be quoted whenever the doctrine is summarized: “it is convenient to speak of a ‘reporter’… though different parts of the play may have been reported by different individuals in the group, while some of the reconstruction may have been collaborative.” A convenience of speech is not a finding of fact, yet generations of students met the reporter as a person rather than as a manner of speaking. The singular culprit had become a synecdoche for a spectrum of possibilities ranging from one man’s memory to a company’s collective paperwork, and the spectrum’s width was precisely what the strong theory had promised to narrow.

There is a further embarrassment, and it is theatrical. The identifications require the reporter to be simultaneously gifted and incompetent: gifted enough to reproduce thousands of lines in near-verbatim stretches, incompetent enough to garble the remainder past recognition. Professional actors of the Chamberlain’s Men did possess extraordinary trained memories, formed by the part system and by repertories that rotated dozens of plays; but a memory that strong cuts against the paraphrase evidence, while a memory that weak cuts against the verbatim evidence. Defenders resolved the tension by distributing the text among several reporters of varying gifts, which rescued the theory at the price of making it unfalsifiable in the small: any stretch of Q1, good or bad, could be assigned post hoc to the reporter whose supposed abilities it matched. The detective work, for all its ingenuity, had built a suspect who could never be cleared because he could be redrawn to fit every new fact.

The Economics of a Suspect Book: Printers, Companies, and Property in Plays

Behind the textual argument stood a material world the theory rarely visited, and visiting it weakens the piracy narrative further. A play in the 1590s was the property of the acting company that bought it from the playwright, an asset whose value lay in performance exclusivity; print publication was a secondary exploitation that companies permitted, sold, or suffered according to circumstance. The 1597 Romeo and Juliet appeared without Stationers’ Register entry and without the author’s name, facts the prosecution read as the marks of theft. The same facts, however, describe plenty of legitimate early editions in a trade where entry was often omitted, authors rarely named on playbooks, and title pages advertised playing companies rather than writers. Q1’s title page names the Lord Hunsdon’s servants and their applause, which is the language of a selling book trading on performance fame, not of a furtive one hiding its origins; a pirate confident enough to name the victim company on the title page is a strange kind of thief.

The book’s manufacture complicates the story in a direction rarely exploited by either side. The printing of Q1 passed between two shops, beginning with John Danter and completed by Edward Allde, a division visible in the type and ornaments of the surviving copies. Danter had form in the period’s drama trade, having printed the first quarto of Titus Andronicus in 1594, and his career ended in the later 1590s amid the trade’s disciplinary weather. A hurried transfer between printers suggests a copy whose arrival and handling were irregular, and the prosecution was entitled to that suggestion; but irregularity in a print shop is evidence about scheduling and copy supply, not about the psychology of actors. The material book testifies that something unusual happened in 1597. It does not testify that the something was memory, and the habit of reading every anomaly of the object as a symptom of the favoured mechanism is the whole methodological quarrel in miniature.

The company-publication alternative, entertained quietly across the century, deserves its sentence in any fair account. Companies did sell playbooks to printers, in lean times, at changes of management, or when a play’s stage life waned, and a company selling a shortened acting text, or licensing a version it no longer guarded, would generate exactly the Q1 profile without a single dishonest actor. The absence of a register entry cuts both ways here, since a company disposing of its own property had no need to advertise the transaction. No document supports this route for Q1 either, and it is raised here not as a candidate verdict but as a standing demonstration of the debate’s true condition: the surviving evidence underdetermines the origin story, and every confident narrative, piratical or proprietary, is an interpolation into a documentary silence.

Orthodoxy at Mid-Century: Census, Statistics, and the Expansion of the List

Between Greg’s Merry Wives edition and Maguire’s audit, memorial reconstruction enjoyed the long mid-century of its acceptance, and the mechanics of that acceptance deserve attention because they explain how a conjecture became a curriculum. The first mechanic was enumeration. Kirschbaum’s “A Census of Bad Quartos” (1938) gathered twenty texts under the label, and a census performs quiet persuasion: a class with twenty members feels like a fact of the period rather than a proposal of the scholar, and each member’s membership certifies the others. Plays by Marlowe, including the A-text of Doctor Faustus and The Massacre at Paris, joined the Shakespearean exhibits, along with Arden of Faversham, The Merry Devil of Edmonton, and a lengthening shelf of apocrypha and anonymous work. The category’s expansion was read as its confirmation, although expansion is equally the behaviour of a net with a wide mesh: the looser the diagnostics, the more texts qualify, and the more texts qualify, the more established the diagnostics appear.

The second mechanic was measurement. Hart’s Stolne and Surreptitious Copies (1942) subjected the bad quartos to comparative counting, tabulating their departures from the good texts and arguing that the departures formed statistical families too regular for chance. Duthie’s Hamlet study applied the method at monograph length, and Hoppe’s Romeo and Juliet volume completed the trilogy of founding texts. Measurement gave the theory the texture of laboratory science, and the texture mattered as much as the results, because the New Bibliography’s founding ambition had been to make the study of books a science of material evidence. Yet the quantities being measured were ratios of agreement with a presumed original, and the presumption did the work the arithmetic appeared to do. A text counted as corrupt in proportion to its distance from Q2 or the Folio, and its distance from Q2 or the Folio was then exhibited as the measure of its corruption. The circle that Maguire would later expose was present from the beginning, wearing the respectable dress of tables.

The third mechanic was pedagogy. The theory entered the handbooks, the introductions to student editions, and the lectures through which each generation of readers met the early texts, and it entered them in simplified form: the bad quartos were pirated reconstructions, the good quartos were printed from manuscripts, and the Folio stood at the end of the story rebuking the pirates. Simplification stripped away the hesitations of Greg, the partial retreats of Alexander, and the conveniences confessed by Gibbons. What the classroom received was a morality tale with a bibliography attached, faithless actors, a robbed company, a rescued author. Generations of students absorbed the tale before they had read a line of Q1, and the tale then governed their reading of it: the quarto’s roughness was experienced as the roughness of theft. A theory does not need to be re-argued in every generation if it is installed in the introductions of the books every generation reads.

The fourth mechanic was the theory’s usefulness in disputes beyond itself. Memorial reconstruction policed the boundary between Shakespeare and his collaborators, predecessors, and revisers. If a suspect text was a report, its evidence about chronology, authorship, and the early history of the plays could be discounted; the Contention and the True Tragedy could be denied their claims to be source plays, Q1 Hamlet could be denied its claim to represent an early Hamlet, and Q1 Romeo and Juliet could be denied any bearing on the tragedy’s composition. The theory thus protected a clean narrative of authorship, in which the good texts descended from the author’s hand by traceable documentary paths and everything that complicated the descent was assigned to theft. Critics who later attacked the theory, Urkowitz among the first, saw the protection clearly and named its price: a shelf of documents had been exempted from serious study in order to keep a story tidy. The audit, when it came, was therefore never only about memory. It was about what the discipline had agreed not to read.

The Audit: Laurie Maguire and the Collapse of the Diagnostics

Laurie Maguire’s Shakespearean Suspect Texts: The ‘Bad’ Quartos and Their Contexts (1996) is the turning point of the debate, because it changed the question. Earlier sceptics had quarrelled with particular identifications or proposed rival origins for particular plays. Maguire audited the method itself. Renaming the exhibits suspect texts, a term that suspends the verdict the old label smuggled in, she assembled the full population of accused books, forty-one Shakespearean and non-Shakespearean play texts, and then subjected the twenty-eight diagnostic features of memorial reconstruction to the test the New Bibliographers had never performed: she asked whether the features actually discriminate. A diagnostic feature must appear in reported texts and must not appear, or must appear far more weakly, in texts transmitted by ordinary documentary means. If paraphrase, transposition, anticipation, and the rest also flourish in good quartos and Folio texts printed from manuscripts, then their presence in Q1 proves nothing about memory, however suggestively they cluster.

The audit proceeded on three fronts. First, Maguire examined bodies of material known to have been transmitted by memory: folk ballads in their variants, entries in Renaissance commonplace books transcribed from plays and sermons, and, most ingeniously, recordings of the BBC television Shakespeare productions of the late 1970s and early 1980s, which allowed actors’ genuine memory errors, caught when performers departed from the printed scripts, to be catalogued against a control text. The errors real actors made were small and local: single words added, dropped, or inverted. The large-scale phenomena the theory required, wholesale paraphrase, migrated speeches, scenes rebuilt from gist, failed to materialize in the experimental record. Second, she searched the period for evidence that the practice existed at all, reviewing Elizabethan references to the reporting of sermons, speeches, and plays, and found a culture of note-taking and shorthand complaint but no account of actors reconstructing plays from memory for printers. Third, she applied the twenty-eight features across her forty-one texts and against good-text controls, and found the diagnostics wanting: feature after feature appeared where the theory forbade it, or failed to appear where the theory required it, until the list that had convicted a shelf of books shrank to a residue too small to bear a general theory.

Why does circular reasoning haunt the diagnostics?

Circular reasoning haunts the diagnostics because the features were harvested from the accused texts. Greg and his successors studied books already presumed reported, noted what those books shared, and then cited the shared features as proof of reporting in the next book examined. No population of certainly reported plays existed to validate the list, so the criteria could never fail.

Maguire’s conclusion was measured, and its measure is what gives it force. She did not declare memorial reconstruction impossible, and she allowed that the evidence left limited circumstantial support for it in a small residue of cases, principally the first quartos of Hamlet and The Merry Wives of Windsor and the quarto of Pericles. But Romeo and Juliet did not stand in that residue. For Q1 of the tragedy, the founding exhibit of Pollard’s list and the subject of the tradition’s most detailed monograph, the audit returned no diagnostic pattern that memory explained better than its rivals, and the features invoked since Hoppe were found to be either present in good texts, absent where reporting predicted them, or consistent with several incompatible origins. The theory’s own standards, applied without favour, acquitted the text it had been built to convict.

Her subsidiary findings widened the damage. The touring rationale that Greg had built into the theory, reconstruction by actors for provincial performance, rested on assumptions about touring practice that the records did not support. The shorthand alternative collapsed on examination of what period shorthand systems could actually capture. Even the vocabulary of the debate was indicted: “bad” had described the offence alleged against the publishers, then migrated to the texts, then to the process, until a moral verdict about piracy did duty as a description of transmission. Maguire’s phrase for the whole tradition, an exercise in textual spring-cleaning, understates the structural result. After 1996, memorial reconstruction could still be argued play by play, but it could no longer be presumed; the burden of proof had moved, permanently, onto whoever wished to revive it.

Testing Memory Itself: Ballads, Commonplace Books, and the BBC Experiment

The most original stratum of Maguire’s book is its reconstruction of what memory actually does to a text, built from cases where the memorial transmission is not in dispute. Folk ballads survive in multiple versions collected from singers who learned them by ear, and their variation is the authentic record of a tradition carried in recollection across years. Renaissance commonplace books preserve speeches and sententiae copied or recalled from plays and sermons by readers whose notebooks survive to be checked against the printed sources. The BBC television Shakespeare productions supplied the controlled case: professional actors, performing from printed scripts, whose departures from those scripts in recorded performance could be classified error by error. Across these populations, a consistent profile emerged. Memory preserves structure, gist, and striking phrases. It loses connective detail. Its characteristic operations are small: the dropped article, the inverted pair of words, the single synonym, the line ending reassigned. What it does not characteristically produce, at the scale of a whole play, is the profile the bad quartos display and the theory required: sustained paraphrase of famous speeches, wholesale transposition of material between scenes, and long stretches of competent invention in the author’s manner.

The ballad evidence cuts in an instructive direction. Where a ballad tradition genuinely runs on memory, the versions diverge in narrative detail and stanza order while clinging to refrains and formulaic lines, the memorable skeleton. A memorial Romeo and Juliet should, by analogy, preserve the tragedy’s refrains, its sonnets, prologues, and famous tags, with maximal fidelity, while its connective tissue dissolved. The observed pattern in Q1 is close to the reverse in the places that matter most to the theory’s critics: the Prologue sonnet stands essentially complete, yet the second Chorus sonnet is absent entire, and celebrated arias are thinned while obscure connective passages sometimes stand fuller than the set pieces. Defenders could and did reply that a ballad singer and a professional actor are different instruments, the actor’s memory being trained, part-based, and refreshed nightly. The reply has force, but it concedes the methodological point: once the analogue populations are abandoned as inexact, the theory loses its only external calibration, and the claim that actors’ memories behave in the required way rests on nothing but the texts the claim was invented to explain.

The commonplace-book evidence adds a period flavour the theory needed and could not use. Elizabethan and Jacobean readers demonstrably extracted, transcribed, and misremembered lines from plays, and their notebooks show real memorial behaviour applied to real dramatic verse within living memory of the performances. The errors are those of the BBC actors in period dress: small, local, sense-preserving. Nothing in the notebooks resembles a play rebuilt entire. Maguire’s wider search of the period’s print for the practice itself, actors furnishing copy from memory, returned complaints about shorthand, anxieties about stolen manuscripts, and Heywood’s grievance about a text taken down by note-takers, but no description of the transaction Greg required. A practice central enough to generate a shelf of books left no trace in the records of an industry that documented its grievances energetically. Arguments from silence are weak in general; in an archive as litigious and self-documenting as the early modern book trade, this particular silence carries weight.

The Feature List on Trial: Twenty-Eight Diagnostics and Their Controls

The feature-by-feature results repay a slower walk, because the verdict on Q1 is built from them. Paraphrase, the theory’s flagship, appears throughout the Folio and the good quartos at rates that destroy its diagnostic value: wherever two authoritative texts of a play survive, as with the quarto and Folio Lear, their divergences include exactly the synonym substitutions and flattened images that were supposed to betray a reporter, yet no one proposes that the Folio Lear was reconstructed from memory. Transposition fares no better as a test, since authorial revision moves material between scenes as a matter of craft, and the Lear texts again supply the control. Anticipation and recollection, the echo of a line outside its place, occur in texts printed from foul papers, where the likeliest explanation is the author’s own second thoughts or the scribe’s eye, not an actor’s ear. Aural error, the mishearing that ought to be memory’s cleanest fingerprint, proves on inspection to be entangled with its documentary twin: a compositor misreading a secretary hand produces errors of shape that mimic errors of sound, and the two cannot be separated by inspection of the error alone.

The structural features collapse for a different reason: they describe abridgement as accurately as amnesia. Omission of whole scenes and speeches, compression of long exchanges, the disappearance of minor characters, and the simplification of doubling patterns are the deliberate acts of a theatrical adapter preparing a script for a smaller company or a shorter afternoon, and distinguishing the cutter’s knife from the reporter’s forgetting, feature by feature, is impossible without independent evidence of intention. Even the celebrated gradient, the waxing and waning of accuracy, loses its uniqueness under audit, because a gradient is what revision produces when an author expands the scenes that interest him and leaves early work standing elsewhere, and what an abridger produces when the scenes an audience loves are spared the knife. Maguire’s summary position, that very few of the twenty-eight features survive scrutiny as relevant to the primal scene of actors recalling and transcribing a script, is the sober end of a century of list-making. The diagnostics were not fraudulent; they were observations in search of a control group, and the control group, once admitted, dissolved them.

The Rival Accounts: Draft, Revision, Abridgement, and the Space Between

The collapse of the diagnostics did not leave a vacuum; it left a field of rivals, each of which had been waiting with its own reading of the same features. The oldest rival is the one the documents state in plain words. Eighteenth-century editors read Q2’s title page at face value: “Newly corrected, augmented, and amended” describes a text improved upon a predecessor, and on that reading Q1 is not a theft of the finished tragedy but an earlier state of it, the author’s first version, subsequently enlarged and polished into Q2. The first-draft theory explains at a stroke the facts that most embarrassed the prosecution. Q1 contains lines of unmistakable quality that Q2 lacks, including Mercutio’s dying expansion in Act 3, Scene 1, where the Q1 text gives him the flourish of being “fairly mounted on four men’s shoulders” while “some sexton, some base slave shall write my epitaph” (3.1, Q1 text), a sentence no paraphrasing actor needed to invent and no abridger needed to keep. On the draft account, such lines are Shakespeare’s own first thoughts, discarded in revision; on the memorial account, they must be credited to the improvising genius of the very reporters whose incompetence explains everything else.

Why does the first-draft theory refuse to die?

The first-draft theory refuses to die because it alone takes Q2’s title page at face value and alone explains Q1’s unique excellences without contortion. A draft predicts exactly what the quarto shows: a complete, actable, shorter tragedy whose best lines are sometimes the ones revision discarded. Its weakness is equally plain: nothing external dates Q1’s copy before Q2’s.

The revision theory, the draft’s sophisticated cousin, drew its charter from a neighbouring play. Stephen Urkowitz argued in Shakespeare’s Revision of “King Lear” (1980) that the quarto and Folio texts of Lear were not original and corruption but two authorial versions, and the argument’s aftershock reached every suspect text: if the very features New Bibliographers had catalogued as corruption in Lear, cuts, reassignments, transpositions, variant phrasing, could be re-read as an author revising, then the catalogue itself was neutral between crime and craft. Urkowitz pressed the institutional point as well: labelling a text a bad quarto removed it from normal discourse, excusing scholars from studying a document of the tradition as seriously as its fellows. Eric Sams carried the scepticism to its limit, contending that memorial reconstruction had no foundation at all and that Q1’s spelling habits and vocabulary were the author’s own; Hardin Craig had earlier dismissed the whole apparatus of reported texts as a modern fiction, noting the absence of any reliable historical evidence that actors ever performed the feat the theory required of them. Sams’s positive case persuaded few, but the negative point he shared with Craig has never been answered: the archive contains complaints about shorthand pirates and stolen copies, and not one record of a company actor dictating a play from memory to a printer.

The theatrical rival explains Q1 not as an origin but as a destination. Robert E. Burkhart’s Shakespeare’s Bad Quartos: Deliberate Abridgements Designed for Performance by a Reduced Cast (1975) proposed that the short quartos were cut deliberately, by professionals, for companies playing with fewer men than the London stage commanded. Deliberate abridgement predicts much of the evidence with fewer miracles: whole units dropped cleanly rather than forgotten raggedly, connective passages rewritten to bridge the cuts, doubling made possible by the disappearance of characters whose actors were needed elsewhere, and a running time near the two hours the Prologue itself advertises when it promises “the two hours’ traffic of our stage” (Prologue, Q2 text). Kathleen Irace gave the adaptation account its most powerful modern form while refusing to surrender memory altogether. In Reforming the “Bad” Quartos (1994), a computer-assisted comparison of the six principal short quartos against their fuller counterparts persuaded her that actor-reporters had indeed shaped the texts, reconstructing the plays from memories of London performances; but the same analysis showed the reporters adapting as they remembered, streamlining structure, clarifying staging, and sharpening pace. Irace thus stands as the theory’s most credible living form: memorial reconstruction demoted from sole cause to one ingredient in a theatrical process, memory working inside adaptation rather than instead of it.

A boundary must be drawn here against a neighbouring claim, because the two are often confused. Whether Q1 records the tragedy as it was actually performed, preserving genuine early staging in its directions and proportions, is a separate question from whether it was produced by memory, and the answers can vary independently: a memorially rebuilt text might record performance faithfully or falsely, and a performance text might descend from a manuscript rather than from recollection. The performance claim has its own evidence and its own advocates, examined in the study of whether Q1 is a performance text and the case for the new view, and it is not settled by anything argued here. The memorial question is narrower and colder: not what Q1 witnesses, but how the copy behind the print was made. Keeping the two apart is itself one of the debate’s hard-won clarifications, since the older literature slid between them, treating theatrical flavour as proof of memorial origin when theatrical flavour is exactly what an authorized playhouse abridgement would also possess.

The Supporting Mechanisms: Shorthand, Auditors, Dictation, and Assembled Parts

Memorial reconstruction did not hold the field alone even in its heyday; it presided over a family of transmission mechanisms, and the family’s other members clarify the main theory by their failures. The shorthand hypothesis, Pollard’s first instinct, proposed a professional note-taker in the audience, and it died of arithmetic: the systems in circulation, Timothy Bright’s Characterie of 1588 and its successors, allocated symbols by word lists incapable of following invented dialogue at stage speed, and even sermons taken by shorthand survive with the gaps and expansions of a method straining at its limits. An auditor theory in longhand fared worse, since nobody writes dialogue verbatim at speaking pace with a quill. What the period documents instead, in Heywood’s complaint and elsewhere, is anxiety about exactly these practices, which proves that contemporaries feared textual theft by ear without proving that the feat was accomplished for any surviving book.

Dictation was the family’s respectable member. A reporter, or a team, might dictate remembered text to a scribe, smoothing individual errors into a fair copy whose uniformity then concealed the memorial origin. The variant’s convenience is its danger, as with the other softenings: dictation explains the comparatively clean presentation of Q1’s verse in many passages, but at the cost of adding an unevidenced intermediary whose presence can be invoked wherever the text is too good for naked memory. The assembled-parts variant pressed a genuinely documentary object into service. Since actors possessed their own parts on rolls, a reconstruction might be built by pooling parts, the skeleton of the play supplied by authentic paper. The arithmetic of parts defeats the idea as a complete mechanism: a part contains one role’s speeches and two-word cues, so an assembly of all the company’s parts yields every speech exactly once and the cues between, a structure superficially like a play but missing the connective composition, and Q1, whatever its faults, is a continuously readable performance text, not a mosaic of cued fragments. The variants survive in the literature because each patches a hole in the parent theory, and their collective effect, a mechanism for every anomaly, is the elasticity the audit finally indicted.

Hamlet’s Shadow: The Marcellus Problem and the Most Famous Suspect Text

No account of the debate is complete without the text that overshadows it, because the first quarto of Hamlet (1603) became the battleground on which the theory’s fate was popularly decided. Q1 Hamlet is barely half the length of the second quarto, reorders major scenes, renames Polonius as Corambis, and renders the most famous speech in the language as “To be, or not to be, ay, there’s the point,” a flattening so startling that the quarto, rediscovered in 1823 after two centuries of oblivion, was received as a curiosity before it was received as a problem. Duthie’s 1941 monograph gave the memorial account its Hamlet form, identifying the actor of Marcellus as principal reporter: the scenes Marcellus attends, the platform scenes of Act 1 above all, stand anomalously close to Q2, while the soliloquies, at which Marcellus was never present, decay furthest. It remains the single most persuasive gradient argument ever constructed for any suspect text, and its persuasiveness is precisely why Maguire’s treatment of it matters: even here, in the strongest case, her audit left the evidence at the level of the circumstantial, compatible with reporting but not demonstrative of it.

The Hamlet case also displays the theory’s gravitational effect on neighbouring questions. Because Q1 Hamlet had to be a report of the Q2 Hamlet, the chronology of the play’s composition bent around the requirement, and the possibility that the short text preserved an earlier Hamlet, a version the playwright later expanded, was resisted for decades as a threat to the doctrine rather than weighed as a reading of the evidence. The essayists who reopened the first quarto of Hamlet at the start of the 1990s found a text whose differences looked less like damage the longer they were studied: a faster, differently proportioned tragedy with its own dramaturgy, whose Corambis and whose reordered soliloquies might record the play’s first form as readily as its corruption. The parallel with Romeo and Juliet is exact, and the traffic ran both ways: every concession won for the Danish quarto loosened the case against the Veronese one, and the category, which had been built as a unity, began to fail as a unity, play by play, until only the audit’s general question remained, whether the diagnostics had ever justified treating the shelf as one kind of object at all.

The Hamlet quarto even preserves the traces of a different naming world, its Polonius called Corambis, a divergence that the reporting theory filed as a reporter’s confusion and that version-based readings file as evidence of an earlier state of the play, in which the counsellor had not yet received his familiar name. No feature of the short Hamlet illustrates the debate’s structure better: the same datum, a renamed character, is corruption under one premise and composition under another, and the datum itself cannot arbitrate between the premises. Arbitration has to come from the standards of evidence a discipline accepts, which is why the quarrel over these books kept escaping bibliography into methodology, and why its resolution, when it came, took the form of an audit rather than a discovery.

The Romeo and Juliet Case Reopened: Farley-Hills and the Quatercentenary

The general audit found its specific application to this play in the same year. David Farley-Hills’s essay “The ‘Bad’ Quarto of Romeo and Juliet,” published in Shakespeare Survey 49 (1996), opened by conceding the orthodoxy its pedigree: the memorial account of Q1 had the sanction of Chambers and Greg, the endorsement of Gibbons’s Arden edition and of the Oxford Complete Works, and the detailed foundation of Hoppe’s monograph, on which, as Farley-Hills put it, modern orthodoxy largely rests. The essay then did to Hoppe what Maguire did to the diagnostic list, re-examining the Romeo and Juliet evidence piece by piece to test how secure those foundations were, and finding them considerably less secure than their reputation. The conjunction matters. In a single year, the category lost its method and its flagship case lost its monograph’s immunity, and the four-hundredth anniversary of the quarto’s printing, marked in 1997 by a wave of quatercentenary reassessment including Eric Sams’s combative reappraisal, found the 1597 text in the unprecedented position of being argued about as a document rather than dismissed as a symptom.

The timing was not accidental. The 1990s assault on the bad quarto came from several directions at once, and the directions reinforced one another. Werstine’s archival critique, published in 1990, denied the New Bibliography its documentary foundations across the whole field of textual narrative. The essays collected around the first quarto of Hamlet at the start of the decade subjected the most famous suspect text to the same re-examination, with contributors doubting the Marcellus identification and the reporting story built upon it. Merriam and Matthews applied stylometric counting to the early Henry VI texts and produced results uncomfortable for the reporting account of them. Each study was local; together they changed the presumption. Where a scholar of 1960 who doubted memorial reconstruction carried the burden of displacing an achievement, a scholar of 2000 who asserted it carried the burden of answering Maguire, Werstine, and the play-by-play re-examinations. Burdens of proof are the real currency of such debates, and the currency had changed hands.

For Romeo and Juliet the practical effect was a revaluation of the quarto’s contents. Passages that the old dispensation had filed as reporters’ invention began to be read as evidence: of the play as first staged, of the author’s first conceptions, of the company’s playing practice. The Q1-only lines at Act 1, Scene 4, the without-book prologue joke, ceased to be embarrassments requiring adoption in spite of the theory and became exhibits in an argument about the text’s descent from the playhouse by a route more respectable than theft. The dying Mercutio’s Q1 epitaph passage underwent the same promotion. None of this required adopting Sams’s counter-dogma that Q1 is simply the author’s draft, a claim the evidence cannot carry either; it required only the negative liberation that the audit supplied, freedom from the obligation to explain every excellence away. A text released from a theory can be read, and the reading of Q1 since the 1990s, in editions, in classrooms, and on stages that have begun to perform the short quarto as a viable alternative script, is the debate’s most tangible result.

The Theory’s Afterlife in the Editing Room

Theories of transmission earn their keep in editions, and the editing room is where memorial reconstruction’s decline can be watched in behaviour rather than manifesto. The mid-century editor treated Q1 as a contaminated witness: valuable for stage directions, which even a reporter transmits reliably because bodies remember business, and for the occasional reading where Q2 is manifestly corrupt, but excluded from authority over wording. Yet the practice was never as pure as the doctrine. The “without-book prologue” couplet at Act 1, Scene 4 survives only in Q1, and editors of Q2-based texts have printed it for generations, an adoption that sits awkwardly beside the claim that the quarto’s words are an actor’s paraphrase; nobody adopts the inventions of a forgetful pirate into the received text of Shakespeare and then continues to call the source corrupt without qualification. Each such adoption was a small editorial vote that Q1’s copy, whatever its route to print, carried authorial matter unavailable elsewhere.

Do modern editors still treat Q1 as corrupt?

Modern editors no longer treat Q1 as simply corrupt. René Weis, editing the Arden text in 2012 on a Q2 copy-text, consults Q1 as a witness rather than a curiosity, and Lukas Erne’s 2007 edition presents the quarto entire as a document worth reading in its own right. The editing room handles Q1 as evidence first, crime scene second.

Erne’s edition marks the institutional turn. Issued in the New Cambridge Shakespeare’s Early Quartos series, it treats the short tragedy as its first readers found it: a complete, actable, coherent play whose differences from Q2 constitute its interest rather than its disgrace. In Shakespeare as Literary Dramatist (2003), Erne had already challenged the assumption that Shakespeare wrote only for the stage and cared nothing for print, an assumption on which the piracy narrative quietly depended, since an author indifferent to publication makes theft easy to imagine and hard to resent. Once the playwright is allowed literary ambition, the short quartos stop being self-evidently stolen goods and become, at minimum, documents whose provenance must be argued rather than presumed. The practical consequence for readers choosing among texts is surveyed in a comparison of the modern editions, Arden, Oxford, and Folger, and the copy-text choices behind them; the consequence for the debate is that Q1 has acquired editors, and a text with editors is no longer merely an exhibit.

The research practice around the quarto has changed with its status. Collating twenty-eight candidate features across forty-one texts, tracking parallel passages between Q1 and Q2 scene by scene, and keeping the rival hypotheses and their exhibits in order is archival work before it is argument, and scholars build private stores for it: a research vault such as VaultBook exists precisely to keep collations, parallel-text tables, and annotated copies of the early editions in one searchable place while a textual argument is assembled. The habit matters to the verdict because the older debate ran on selective quotation, each school exhibiting the passages its theory explained best. Systematic collation, the kind a well-kept archive enforces, is what allowed Maguire to test features across a population instead of admiring them in a showcase, and it is why the debate’s centre of gravity has shifted from anecdote to audit.

Proof in Bibliography: What Would Count as Evidence?

A debate that lasts a century without resolution usually suffers from an unexamined standard of proof, and this one is no exception. The New Bibliographers worked by inference to the best explanation: assemble the features, propose the process that would generate them, and accept the process when no rival generated them as neatly. The method is legitimate in the historical sciences, but it has a condition that the tradition neglected: the field of rivals must be honestly populated, and the features must be weighed against the rivals one by one, not against a vacuum. Memorial reconstruction won its century partly because its rivals were examined one at a time and late, abridgement by Burkhart in 1975, revision by Urkowitz in 1980, the documentary foundations by Werstine in 1990, the diagnostics by Maguire in 1996. Against a vacuum, memory looked inevitable; against the populated field, it looks optional, and optionality is the audit’s whole achievement.

Documentary evidence, the kind that settles questions in bibliography when it exists, was always missing for the crucial transaction, and the surrounding documents reward a cold look. Philip Henslowe’s diary, the great surviving archive of the rival Admiral’s Men, records payments to playwrights, purchases of playbooks, dealings with printers, and the economics of the stage in granular detail across the very years in question, and it contains no transaction resembling the sale of a reconstructed play by actors. The Chamberlain’s Men’s own papers barely survive, so the silence is asymmetrical and cannot be pressed too hard; but the Folio preface, the one document the prosecution always cited, proves on inspection to be a publisher’s puff, warning buyers away from competing editions of plays the Folio’s own syndicate was selling, and general enough to cover any copy the syndicate disliked, for any reason. A case resting on a puff and a pattern was never a strong case; it was a strong story told in a discipline that had decided, for excellent reasons of its own, to trust pattern evidence in the study of paper and type, and had extended that trust, without noticing the border crossing, to the study of minds.

The contrast with genuinely documented transmission sharpens the point. When early modern texts demonstrably travelled by ear and memory, the culture left the apparatus lying around: shorthand manuals advertising their systems, practitioners advertising their speed, preachers complaining of mangled sermons, note-takers’ copy surviving to be compared with authorized print. Reported sermons can be studied as reported because the practice announced itself at every stage, in advertising, in grievance, and in artifact. The theatre’s paper trail is thinner, but it is not blank: it preserves playbook purchases, licensing, litigation over manuscripts, injunctions against printing, and the Folio syndicate’s own commercial manoeuvres. Inside that record, the complete absence of the memorial transaction, no payment, no complaint, no boast, no prosecution, is a datum and not a technicality. The theory asked the archive to be silent in exactly the place where the archive is talkative, and the request went unnoticed for decades because the internal pattern seemed eloquent enough to speak for the missing documents.

The comparative standard clarifies what proof would even look like. Where memorial transmission is genuinely documented in the period’s culture, as with shorthand sermon reporting, the record carries the apparatus of the practice: named systems, manuals, practitioners, complaints from the reported, and surviving notes whose character can be checked. Where theatrical texts are documented as stolen, the record speaks of manuscripts taken, copies bought, and rights disputed. The memorial reconstruction of entire plays by actors sits in neither archive. It is a process inferred from its supposed products, attested nowhere else, and endowed by its proponents with exactly the capacities its products required. Historians of the book have a name for such entities, and the name is not kind. The mature position of the debate, the position this article’s verdict formalizes, is that bibliography may entertain the hypothesis where a text’s features invite it, must label it as hypothesis in editions and classrooms, and may not build further conclusions, about reporters, dates, authorship, or authority, on a foundation that no document supports.

The Verdict: How Much Weight Can Memorial Reconstruction Still Bear?

The evidence now permits a judgment, and the judgment must distinguish three things the tradition bundled together: the category, the mechanism, and the local case. The category of the bad quarto, as Pollard framed it, bundled a textual observation (some early texts are short, rough, and derivative) with a legal accusation (they were pirated) and a process claim (they were reconstructed from memory). The textual observation stands; nobody disputes the profile of Q1. The legal accusation was never documented for this play: no prosecution, no register entry, no complaint names the 1597 Romeo and Juliet as stolen goods, and the Folio’s general denunciation, written a quarter-century later by men selling a replacement volume, is advertising copy before it is testimony. The process claim is the one Maguire audited, and it is the one that fails. A category that fuses a fact, an accusation, and a failed mechanism deserves to be dissolved into its parts, keeping the fact and discarding the rest.

The two cases, stripped of rhetoric, stand as follows in the InsightCrunch bad-quarto debate, the article’s findable summary of the whole quarrel.

The Case for Memorial Reconstruction	The Case Against Memorial Reconstruction
Q1 is roughly a third shorter than Q2, matching a text carried in memory rather than copied from a full manuscript.	Shortness is equally predicted by deliberate abridgement and by an earlier, slimmer authorial version.
Q1 paraphrases celebrated passages, keeping the sense while losing the wording, as in the truncated name speech of the balcony scene.	Paraphrase and synonym substitution also occur in texts printed from authorial manuscripts, so the feature cannot diagnose memory.
Q1 drops whole formal units, including the second Chorus sonnet, in the ragged pattern of forgetting.	Clean excision of whole units is the signature of a cutter with a plan, not of a memory failing at random.
The accuracy of Q1 rises and falls with particular actors’ presence on stage, the gradient Hoppe traced scene by scene.	The gradient is compatible with drafting, cutting, and revision, and the reporter is inferred from the pattern the theory itself predicts.
Greg identified the same gradient in The Merry Wives of Windsor and named the Host’s actor as its reporter.	The identification is circular: the scenes are called well reported because the theory needs a reporter, and the reporter is then cited as proof.
Actors learned from parts containing only their own lines and cues, so memory was the medium in which whole plays lived.	The part system proves actors memorized their own roles, not that any actor ever reconstructed an entire play for a printer.
Kathleen Irace’s computer-assisted comparison found actor-reporters shaping the six short quartos.	Irace’s reporters are simultaneously adapters, so her evidence supports a theatrical process in which memory is one ingredient, not the cause.
The Folio editors denounced stolen and surreptitious copies, giving the piracy narrative documentary colour.	The denunciation is unspecific, commercial, and separated from Q1 by twenty-six years; it names no play, printer, or actor.
Harold Jenkins praised the identification of memorial reconstruction as an achievement of twentieth-century textual scholarship.	Institutional prestige is not evidence, and the praise records how completely an inference had hardened into doctrine.
No documentary account of Q1’s copy survives, leaving informed inference as the only available method.	Absence of documents licenses humility, not confidence: an unobservable process cannot be promoted to a premise of editing.

The mechanism, considered generally, fails on falsifiability. A theory that explains verbatim accuracy and gross paraphrase, expansion and omission, single reporters and collaborative teams, whole reconstructions and partial ones, has arranged to be confirmed by every possible state of the evidence, and a theory confirmed by everything is confirmed by nothing. Paul Werstine’s examination of the New Bibliography’s narratives made the complementary point from the archive side: the stories told about foul papers, prompt books, and reported texts consistently outran the documentary record of the early modern book trade, functioning as what their authors needed the past to have contained. Against that background, Maguire’s restraint reads as the debate’s maturity. She left the door open where her data left it open, for Hamlet, for The Merry Wives of Windsor, for Pericles, and she closed it where the data closed, which included the tragedy at the centre of this article.

For the local case, Q1 of Romeo and Juliet, the verdict follows. Memorial reconstruction can bear the weight of a possibility and no more. It remains conceivable that memory played some part in the ancestry of Danter and Allde’s copy, as one strand in a theatrical transmission that also involved cutting and adaptation; Irace’s hybrid is the strongest form in which the idea survives, and even that form assigns the decisive shaping to deliberate theatrical work rather than to recollection. What the theory can no longer do is serve as the default explanation of the quarto, the premise from which readings are dismissed, reporters named, and editions built. The honest scholarly position is narrower than either pole of the old quarrel: Q1 is a derivative, theatrical, shortened text of uncertain manufacture, containing authorial matter of the first importance, whose precise route from stage or page to print is not known and, on present evidence, is not knowable. Uncertainty of that kind is not a failure of the discipline. It is the discipline, accurately reporting the distance between its documents and its desires.

What evidence could settle the bad quarto debate?

Only external evidence could settle the debate: a playhouse document, a printer’s record, a lawsuit describing how the copy behind a suspect quarto was obtained. Internal evidence is exhausted: every feature of the texts has proved readable by two incompatible theories. Until a document surfaces, the burden of proof rests on whoever asserts a mechanism the archive never mentions.

The myth of the article’s title, then, is not the idea that actors remembered plays; actors did nothing else. The myth is the confidence: the century-long promotion of an unexamined inference into an achievement, of a diagnostic list into a science, of convenient reporters into culprits, of a shelf label into an explanation. Dismantling that confidence does not impoverish the study of Q1. It returns the 1597 Romeo and Juliet to readers as what it demonstrably is, the first printed form of the tragedy, actable, strange, and eloquent in its differences, and it returns the debate to its proper subject, the standards by which bibliography may claim to know anything at all.

After the Verdict: What the Category Leaves Behind

The category’s remains are instructive. The phrase bad quarto has not vanished from scholarship, and its survival in a demoted sense is a useful measure of the verdict’s precision: editors still use the label as a shelf mark for the short, derivative early texts, a description of their textual character, while the explanatory cargo the label once carried, piracy by memory, is unloaded and left on the dock. Reference works hedge where their predecessors asserted. Students are increasingly introduced to Q1 as a version rather than a crime. The theory’s genuine achievements should be recorded in the same ledger as its failure: the New Bibliographers’ close collation of the suspect texts, undertaken to convict them, produced the detailed knowledge of their features on which all later work, including the audit that acquitted them, depends. Hoppe’s monograph remains indispensable as evidence even where its conclusion is set aside, and Greg’s Merry Wives edition is still the starting point for study of that text. The prosecution built the archive that freed the accused.

What replaces the old certainty is a discipline of case-by-case judgment under an admitted burden of proof. A scholar who believes memory contributed to a particular text must now show, against controls, that the features exceed what revision, abridgement, and documentary transmission produce, and must say what evidence would change the verdict. A scholar who believes a suspect text is an early version must carry the corresponding burden in the other direction. Stylometric and computational methods, of the kind Irace pioneered for these texts and Merriam and Matthews applied to the Henry VI quartos, offer the prospect of evidence that is at least independent of the reader’s prior convictions, and the debate’s future lies with such methods and with any documentary discovery the archive may still yield. The classroom and the stage have registered the change in their own currencies. Where a student edition once introduced Q1 in a paragraph of warning, the quarto increasingly appears in parallel-text selections and seminar exercises as a version to be compared rather than a corruption to be discounted, and the comparison teaches the debate’s real lesson better than any verdict could: two early texts of the same play, diverging in patterned ways, with the pattern’s cause underdetermined by the evidence. On stage, the short quarto’s proportions, its speed, its compressed arias, and its streamlined doubling have attracted practitioners interested in the tragedy at something like the two hours its Prologue promises, and a text that companies can perform entire has a claim on attention that no merely documentary exhibit can match. Kathleen Irace’s performance scholarship documented how the short quartos’ features mirror the choices working companies make when they cut and adapt, which is why the abridgement account retains its hold on theatrical imaginations: the 1597 text behaves like a playing version because, on the strongest surviving account, a playing version is what it is.

Until then, the first quarto of Romeo and Juliet stands where the evidence leaves it: printed in 1597 from copy of unrecorded origin, shorter and rougher than its successor, indispensable to editors, actable in the theatre, and innocent, for want of proof, of the picturesque crime that made it famous.

For the general reader, the debate’s outcome can be carried away in a single distinction. The 1597 quarto is certainly a theatrical, shortened, derivative text; it is not certainly a remembered one, and the distance between those two sentences is the distance the scholarship travelled in a century. Reading Q1 with that distinction in mind changes the experience of the book. Its compressions read as decisions awaiting explanation rather than as lapses awaiting culprits. Its unique lines read as survivals rather than inventions. Its stage directions read as the residue of a working theatre whose exact relation to the printed page remains, honestly, unknown. The tragedy loses nothing by the acquittal. A play that survives in two early forms, one full and one swift, is richer in evidence than a play guarded by a single authorized text, and the quarrel over the shorter form has taught the discipline more about its own methods than a dozen settled questions could have done.

Frequently Asked Questions

Q: What is memorial reconstruction in Shakespeare studies?

Memorial reconstruction is the theory that certain early printed play texts were produced when actors rebuilt a play from memory for a publisher, without access to any manuscript. The theory was framed to explain why some quartos are short, paraphrased, and uneven: a reporter remembers his own part well and the rest of the play imperfectly, so accuracy rises and falls with his presence on stage. It predicts distinctive failures, including synonym substitution, transposed speeches, anticipated or repeated lines, mishearings, and verse decayed into prose. Walter Wilson Greg named and developed the idea in his 1910 edition of the 1602 Merry Wives of Windsor quarto, and for decades it served as the standard explanation of the so-called bad quartos, including the 1597 first quarto of Romeo and Juliet. Since Laurie Maguire’s audit of 1996, the theory has been treated as an unproved hypothesis rather than an established mechanism, possible in a few cases, demonstrated in none, and no longer usable as a default explanation of a rough early text.

Q: Who coined the term bad quarto, and when?

The bibliographer Alfred W. Pollard coined the term in Shakespeare Folios and Quartos, published in 1909. Pollard argued that the First Folio’s editors, John Heminges and Henry Condell, had aimed their famous complaint about “diverse stolen and surreptitious copies” at a small class of texts rather than at all earlier quartos, and he sorted the early editions into good quartos, printed from authorial or playhouse manuscripts, and bad quartos, corrupt in text and illegitimate in origin. The phrase proved extraordinarily durable because it fused a textual judgment with a moral one: a bad quarto was badly printed because it had been badly obtained. Later scholarship, above all Laurie Maguire’s Shakespearean Suspect Texts of 1996, objected that the fusion prejudged the very question under study, and proposed the neutral term suspect texts for the same books. The label survives in a weakened sense, as a shelf mark for short derivative quartos, while its original cargo of accusation has been largely unloaded by the scholarship that inherited it.

Q: Which five quartos did Pollard place on the bad list?

Pollard’s original list named five texts: the first quartos of Romeo and Juliet (1597), Henry V (1600), The Merry Wives of Windsor (1602), and Hamlet (1603), together with the quarto of Pericles (1609). Each of the first four stood beside a fuller, tidier later text against which its badness could be measured, the second quarto in the case of Romeo and Juliet and Hamlet, the Folio in the case of Henry V and The Merry Wives of Windsor. Pericles was the anomaly, a rough text never superseded by a better early edition, admitted to the list on the strength of its internal disorder. The list did not stay at five. Peter Alexander added the early Henry VI texts, The First Part of the Contention (1594) and The True Tragedy of Richard Duke of York (1595), and Leo Kirschbaum’s census of 1938 counted twenty bad quartos, including plays by Marlowe, Peele, and Greene. Laurie Maguire’s audit of 1996 examined forty-one suspect texts in all, a population whose growth itself illustrates how elastic the category had become.

Q: Why was shorthand reporting set aside as the source of the bad quartos?

Shorthand reporting was set aside because the period’s shorthand systems could not do the work the hypothesis required. Pollard’s first instinct had been stenographic piracy, an idea with period support in Thomas Heywood’s complaint that If You Know Not Me, You Know Nobody had reached print in 1605 through notes taken in the theatre. Greg and R. C. Rhodes argued instead for actors’ memory, and later investigation settled the capability question: systems such as Timothy Bright’s Characterie of 1588 worked from limited word lists, served slow formal speech, and left even sermon reports full of gaps and reconstruction. Rapid verse dialogue, spoken at stage pace amid noise and action, lay beyond them. Laurie Maguire’s examination of the surviving systems and of documented shorthand reporting concluded that stenography was an implausible source for any play text of the period. The shorthand theory survives as evidence of period anxiety about textual theft, not as a viable account of how the bad quartos were made.

Q: Who wrote the fullest study of the first quarto of Romeo and Juliet as a reported text?

H. R. Hoppe wrote the fullest study, The Bad Quarto of Romeo and Juliet: A Bibliographical and Textual Study, published in 1948. Hoppe worked through the 1597 quarto scene by scene, classifying its variants against the second quarto under the heads the theory supplied: paraphrase, transposition, anticipation, recollection, aural error, and metrical disturbance. He concluded that the text had been reconstructed from memory by members of the company that first performed the tragedy, with the reporting strongest where the reporters’ own stage experience ran deepest, and his monograph became the foundation on which the orthodox account of Q1 rested for the rest of the century. Brian Gibbons’s Arden edition of 1980 and the Oxford Complete Works endorsed its conclusion. The study’s standing has changed with the theory’s: since David Farley-Hills re-examined its evidence in Shakespeare Survey in 1996, Hoppe’s book is read as an indispensable catalogue of the quarto’s features whose explanatory conclusion no longer commands assent, a distinction between evidence and inference that defines the modern debate.

Q: What did Laurie Maguire set out to test in Shakespearean Suspect Texts?

Maguire set out to test whether the features used to diagnose memorial reconstruction actually diagnose anything. Shakespearean Suspect Texts: The ‘Bad’ Quartos and Their Contexts (1996) renamed the accused books suspect texts and audited the method of the New Bibliography rather than any single play. Maguire assembled forty-one Shakespearean and non-Shakespearean suspect texts and listed the twenty-eight features that scholarship since Greg had treated as signs of reporting, from paraphrase and transposition to anticipation and aural error. She then asked the control question the tradition had skipped: do the features appear in texts printed from manuscripts, and do they appear in material certainly transmitted by memory? A feature found equally in good quartos and Folio texts cannot convict a quarto of being reported, however suggestive it looks in isolation. The result was a systematic failure of the diagnostics: very few of the twenty-eight features survived as relevant to actors recalling and transcribing plays, and the theory lost its evidentiary machinery while retaining, in three cases, a residue of circumstantial plausibility.

Q: How many suspect texts did Maguire examine, and how many diagnostic features?

Maguire examined forty-one suspect texts and twenty-eight diagnostic features. The texts ranged across the whole accused population, Shakespearean exhibits such as the first quartos of Hamlet, Henry V, The Merry Wives of Windsor, and Romeo and Juliet, non-Shakespearean plays including Marlowe’s Doctor Faustus and The Massacre at Paris, and apocryphal works such as Arden of Faversham and Fair Em. The features were the accumulated signs by which the New Bibliographers and their successors had identified reported texts since Greg’s 1910 Merry Wives study: paraphrase, synonym substitution, transposition, anticipation, recollection, aural error, metrical corruption, verse printed as prose, borrowing from other plays, and the patterned unevenness said to track a reporter’s presence, among others. The scale was the argument. Earlier studies had generalized from a handful of famous cases; Maguire’s census method meant a feature had to behave consistently across forty-one texts and against good-text controls before it could be called diagnostic. Measured at that scale, most features either appeared where reporting was impossible or failed to appear where reporting had been asserted, and the list collapsed accordingly.

Q: What did Maguire learn from recordings of the BBC television Shakespeare productions?

Maguire learned that real actors’ memory errors are small, local, and unlike the phenomena the bad quarto theory required. The BBC television Shakespeare series, broadcast in the late 1970s and early 1980s, gave her a controlled experiment impossible for earlier scholars: professional performers working from printed scripts, whose departures from those scripts in the recorded performances could be catalogued precisely. The errors she found were single words added, dropped, or inverted, the ordinary static of live performance. The large-scale behaviours that memorial reconstruction predicted, wholesale paraphrase of famous speeches, speeches migrated between scenes, long passages rebuilt from gist, did not materialize in the recorded evidence. She set the experimental results beside two further populations of certainly memorial material, folk ballads in their variants and Renaissance commonplace books transcribed from plays and sermons, and the same profile held: memory preserves structure and striking phrases while losing connective detail in small units. The experiment did not prove that no quarto was ever reported; it showed that the theory’s central image of how memory fails does not match memory’s observed behaviour.

Q: Which suspect texts kept any support for memorial reconstruction after Maguire’s audit?

Only a small residue kept any support: principally the first quarto of Hamlet (1603), the first quarto of The Merry Wives of Windsor (1602), and the quarto of Pericles (1609). Even for those three, Maguire’s finding was limited circumstantial evidence suggestive of possible memorial reconstruction, a long way from the confident identifications of the Greg and Duthie era, when the Host’s actor and Marcellus could be named as culprits in print. The first quarto of Romeo and Juliet, a founding member of Pollard’s original list and the subject of the tradition’s most detailed monograph, did not stand in the residue: for the tragedy, the diagnostic pattern either failed to appear where reporting predicted it or appeared equally in texts nobody suspected. The audit’s asymmetry is the sharpest measure of the theory’s decline. Memorial reconstruction had been invented for the Merry Wives, generalized across the canon, and applied with most confidence to the famous cases; after 1996 it survived, on its own chosen evidence, only as a possibility in a handful of instances, its flagship exhibits discharged.

Q: Why do critics call the memorial reconstruction theory unfalsifiable?

Critics call the theory unfalsifiable because its defenders adjusted it to absorb every contrary fact. Accurate passages proved the reporter had a good memory or stood on stage for that scene; corrupt passages proved his memory failed; a single gradient of accuracy indicated one reporter, several gradients a team; team failures were repaired by collaboration, dictation, or consultation of a manuscript. Peter Alexander, extending the theory to the early Henry VI texts, retreated to partial memorial reconstruction when the features refused to align, and Brian Gibbons conceded for Romeo and Juliet that the reporter was a convenience of speech covering possibly several individuals and collaborative work. A hypothesis consistent with every possible state of the evidence cannot be confirmed by any of it. Paul Werstine’s archival critique and Laurie Maguire’s audit pressed the point from opposite sides: the narratives outran the documentary record, and the diagnostics were circular, harvested from accused texts and then used to accuse. The charge of unfalsifiability is the debate’s deepest, because it concerns not whether the theory is false but whether it was ever framed so that it could be found false.

Q: What did Kathleen Irace’s computer analysis conclude about the short quartos?

Kathleen Irace concluded that actor-reporters shaped the short quartos, but that the reporters adapted the plays as they reconstructed them. Reforming the “Bad” Quartos (1994) subjected the six principal short quartos, the first editions of Hamlet, Henry V, Romeo and Juliet, The Merry Wives of Windsor, and the second and third parts of Henry VI, to a computer-assisted comparison with their fuller counterparts. The analysis persuaded her that the players had rebuilt the texts from memories of London performances rather than the author revising his own drafts, since the patterns of variation tracked performance experience. The same evidence, however, showed the reconstruction to be purposeful: structures streamlined, staging clarified, pace sharpened, the earliest fast-moving popular adaptations of the plays. Irace’s position is the strongest surviving form of the memorial theory, because it survives Maguire’s audit in part: memory becomes one ingredient in a theatrical process whose decisive agent is deliberate adaptation. Critics note that the hybrid inherits a burden of its own, since the analytical task of separating the remembered from the redesigned, passage by passage, remains as contestable as the original diagnosis.

Q: How does the authorial first draft theory explain the first quarto?

The first draft theory explains the first quarto as Shakespeare’s own earlier version of the tragedy, later corrected, augmented, and amended into the second quarto, exactly as the second quarto’s title page advertises. On this account, Q1’s shortness is the shortness of a first conception, its roughness the roughness of work not yet revised, and its differences from Q2 the record of the author expanding, reordering, and polishing between drafts. The theory’s strongest exhibits are Q1’s unique excellences: lines and passages found nowhere else that read as authorial rather than improvised, including Mercutio’s dying flourish about being fairly mounted on four men’s shoulders and the without-book prologue couplet at Act 1, Scene 4 that modern editions print from Q1 alone. Eighteenth-century editors held the view as a matter of course, and Eric Sams revived it in combative form. Its weakness is the absence of external evidence dating Q1’s copy before Q2’s, and the difficulty of explaining why a draft should so often read like a fuller text half-remembered, paraphrase being an odd property for an author’s first thoughts.

Q: What did Stephen Urkowitz argue about King Lear, and why does it matter for this debate?

Stephen Urkowitz argued that the quarto and Folio texts of King Lear are two authorial versions, the product of Shakespeare revising his own play, rather than an original and its corruption. Shakespeare’s Revision of “King Lear” (1980) re-read the very features the New Bibliography had catalogued as corruption, cuts, reassigned speeches, transposed material, variant phrasing, as the fingerprints of deliberate revision. The argument matters for the Romeo and Juliet debate because it destroyed the neutrality of the feature list: if the same phenomena that convicted the bad quartos could signify an author at work in Lear, then their presence in Q1 could not by itself prove actors remembering. Urkowitz also named the institutional cost of the label, arguing that calling a text a bad quarto removed it from normal scholarly discourse and excused a generation from studying documents of the tradition on their merits. The Lear case did not settle the origin of Q1, which involves different evidence; it changed the rules of inference for every suspect text, and the memorial theory never recovered its presumption of innocence for its own diagnostics.

Q: Is there any historical record of actors reconstructing a play from memory for a printer?

No historical record of the transaction exists. Laurie Maguire’s search of Elizabethan and Jacobean evidence for the reporting of plays, sermons, and speeches found a culture anxious about textual theft, equipped with shorthand systems, manuals, and complaints, including Thomas Heywood’s grievance over If You Know Not Me, You Know Nobody, but no account of actors dictating or writing out a performed play from memory for publication. Philip Henslowe’s diary, the detailed archive of a rival company’s business in the same years, records payments for playbooks and dealings with printers without any entry resembling such a sale. Hardin Craig had made the point a generation earlier, dismissing the reported-text apparatus as a modern fiction for want of reliable historical evidence, and Eric Sams repeated it in the 1990s. Arguments from silence are never conclusive, and the Chamberlain’s Men’s own records survive too poorly for the silence to convict by itself. Yet the absence matters because memorial reconstruction is an extraordinary provenance to assign a printed book, and extraordinary provenances ordinarily require at least one witness outside the texts they explain.

Q: Why do editors still print lines that survive only in the first quarto?

Editors print lines that survive only in Q1 because the lines are good, actable, and consistent with the author’s manner, and because no coherent account of the quarto can dismiss them as a forgetful actor’s improvisation. The clearest case stands at Act 1, Scene 4, where Benvolio’s remark about having “no without-book prologue, faintly spoke / After the prompter, for our entrance” exists in no other early text, yet generations of editors of Q2-based editions have admitted the couplet into the received text. Each admission is a practical judgment that Q1’s copy, whatever route it travelled to Danter and Allde’s printing house, carried authorial matter unavailable elsewhere. The practice quietly contradicts the strong form of the memorial theory under which such matter should be reporters’ invention, and it long predates the theory’s formal collapse: the editing room behaved for decades as though Q1 were a flawed witness rather than a fabricated one. Modern editors have simply aligned their stated doctrine with their established practice, treating the quarto as evidence to be weighed reading by reading.

Q: How does Lukas Erne’s edition of the first quarto treat the text?

Lukas Erne’s edition treats the first quarto as a complete play worth reading in its own right. The First Quarto of Romeo and Juliet (2007), issued in the New Cambridge Shakespeare’s Early Quartos series, prints the 1597 text entire, with annotation and an introduction that examines its provenance without presuming the piracy narrative. The editorial premise is that Q1’s differences from Q2 constitute its interest rather than its disgrace: a shorter, swifter tragedy whose proportions, staging indications, and unique passages document the play’s early life. The edition belongs to Erne’s wider argument, developed in Shakespeare as Literary Dramatist (2003), that Shakespeare cared for print and posterity as well as for performance, an argument that removes one of the piracy story’s quiet supports, the indifferent author whose property nobody needed to steal carefully because he guarded it not at all. Erne does not claim to have solved the quarto’s origin, and the edition’s restraint is its significance: the most institutional form a text can take, a scholarly edition in a major series, was extended to a book that Pollard’s vocabulary had defined as beneath editing.

Q: What is the difference between memorial reconstruction and theatrical abridgement?

The difference lies in agency and intention. Memorial reconstruction is passive and accidental: a text degrades because memory loses, substitutes, and transposes material without design, and the resulting shortness is a by-product of forgetting. Theatrical abridgement is active and deliberate: a company or adapter cuts a play for performance, for a smaller cast, a shorter afternoon, or a touring stage, and the resulting shortness is the purpose of the exercise. Robert E. Burkhart’s study of 1975 proposed deliberate abridgement for reduced casts as the origin of the bad quartos generally, and Kathleen Irace’s account of the short quartos fused the two processes, with actor-reporters adapting as they remembered. The distinction matters because the two mechanisms leave similar evidence, omissions, compressions, smoothed transitions, while implying opposite verdicts on the text’s authority: an abridgement is a professional version with its own integrity, a reconstruction a damaged copy of someone else’s. Much of the modern debate consists of attempts, none wholly successful, to find features that separate the cutter’s intention from the rememberer’s failure.

Q: Why did Harold Jenkins praise the identification of memorial reconstruction in 1982?

Harold Jenkins praised the identification because, from inside the tradition, memorial reconstruction looked like the New Bibliography’s exemplary success. Introducing his Arden edition of Hamlet in 1982, Jenkins counted the recognition of memorially constructed texts among the achievements of twentieth-century textual scholarship: a body of anomalous books had been gathered under one mechanism, their anomalies catalogued, their reporters in some cases named, and editing placed on a rational footing that distinguished authoritative copy from theatrical debris. The praise is valuable to historians of the debate precisely because of its date and its source. Jenkins was among the century’s most respected editors, writing at the end of the theory’s long ascendancy, and his compliment records how completely an inference about unobservable events had been naturalized as a finding of science. Fourteen years later Maguire’s audit subjected the achievement to its first systematic controls. The distance between the two moments, the trophy of 1982 and the spring-cleaning of 1996, measures how quickly a discipline’s settled ground can become its contested ground once somebody performs the experiment the founders omitted.

Q: What happened to the bad quarto category after the challenges of the 1990s?

The category survived as a label and collapsed as an explanation. Three converging attacks did the work: Paul Werstine’s archival critique of 1990 showed that the New Bibliography’s narratives about copy and transmission outran the documentary record of the book trade; Kathleen Irace’s study of 1994 relocated the short quartos’ origins in theatrical adaptation with memory as one ingredient; and Laurie Maguire’s audit of 1996 demonstrated that the diagnostic features of reporting failed against controls, leaving circumstantial support in at most three of forty-one suspect texts. David Farley-Hills’s re-examination of the Romeo and Juliet evidence in the same year withdrew the flagship case’s monographic foundation. The practical consequences followed in the editing room: Q1 acquired a full scholarly edition from Lukas Erne in 2007, editors such as René Weis treated it as a substantive witness, and reference works learned to hedge. The phrase bad quarto remains in use as a convenient description of a class of short derivative texts, but scholars who use it now specify that the description carries no settled account of how the texts were made.

Q: How much weight can memorial reconstruction still bear as an explanation of the first quarto?

Memorial reconstruction can bear the weight of a possibility and no more. As the default explanation of the 1597 Romeo and Juliet, it has been discharged: its diagnostic features fail against controls, its named reporters rest on circular inference, its mechanism lacks any documentary witness, and its strongest modern form, Kathleen Irace’s hybrid of memory inside deliberate adaptation, assigns the decisive shaping to theatrical intention rather than recollection. What survives is the bare conceivability that memory played some part, alongside cutting and adaptation, in the ancestry of the copy that John Danter and Edward Allde printed, a conceivability that explains nothing by itself and licenses no editorial decisions. The burden of proof has moved permanently: a scholar asserting memorial transmission for this text must now demonstrate it against the rival accounts rather than presume it from the quarto’s roughness. The honest description of Q1 is a derivative, theatrical, shortened text of unrecorded manufacture containing authorial matter of the first importance. That uncertainty is not a failure of the scholarship. It is the scholarship, reporting the exact distance between its evidence and its former confidence.