INTRODUCTION

   Not too long ago, the 1597 first quarto (Q1) of Romeo and Juliet was simply considered a ‘bad quarto’, and its alleged badness largely disqualified it from critical and editorial circulation. It was referred to and discussed in introductions to editions of Romeo and Juliet, and editors ransacked it for what it could tell us about the play. Those who closely engaged with it realised that it was in some ways a text with remarkable qualities, but as long as labels such as ‘bad quarto’, ‘memorial reconstruction’ or ‘piratical publication’ clung to it, it remained in the margins of Shakespeare scholarship and was rarely mentioned in the criticism.

   In more recent times, however, the New Bibliographical paradigm which alleged the text’s badness has come under attack from various quarters, and, as a result, our view of the ‘bad quartos’ is about to change. It is becoming more generally recognised that certain of these texts have important things to tell us about early modern drama. Stephen Orgel has pointed out that there is ‘very little evidence that will reveal to us the nature of a performing text in Shakespeare’s theater; but there is a little. There are the “bad” quartos, whose evidence, in this respect, is not bad, but excellent’.¹ Despite the critical and scholarly turn in Shakespeare studies towards performance since the last decades of the last century, editorial tradition may have prevented us from recognising the true importance of the ‘bad’ quartos insofar as they reflect, however imperfectly, the plays as they were performed. As Orgel adds, ‘If we were less exclusively concerned with establishing texts and more concerned with the nature of plays, these [the ‘bad’ quartos] would be the good quartos.’² Similarly, Gary Taylor has asserted that a ‘bad’ quarto ‘may represent a more finished, dramatic, socialized phase of the text than that preserved in an edition printed from Shakespeare’s foul papers’.³

   Among the ‘bad’ quartos, Q1 Romeo and Juliet is a particularly important publication, with a carefully printed text, exciting stage directions which may well shed light on the play’s early modern staging,⁴ a shorter, tightened text with considerable pace, and an important number of intelligent alternative readings. Indeed, Q1 Romeo and Juliet is in some respects the best witness we have for the dramatic and theatrical practices of Shakespeare and of his company.

   Q1 Romeo and Juliet is a crucially important witness precisely because of its difference from the longer, better-known version. A writing process is rarely stable, and the textual life of a play in the theatre is even less so: early modern play-texts were corrected, revised, abridged, changed.⁵ The textual material still extant today is naturally a very imperfect record of the literary and theatrical genesis and transformations of Shakespeare’s plays. A printed text preserves only a part, and often a small part, of the history of a play, and its relationship to the words spoken onstage by Shakespeare and his fellow actors is one we cannot hope to guess, on the basis of the extant printed texts. All these problems and limitations granted, Romeo and Juliet, by preserving not only the quarto on which modern editions are usually based but also the first quarto, may allow us to glimpse, in the gap that separates one quarto from the other, some of the transformations a Shakespeare play underwent in the first years of its existence.

   Accordingly, this edition, even though it is an edition of The First Quarto of Romeo and Juliet, does not advocate study of the first quarto for its own sake. Q1 and Q2 throw interesting light on each other, and it is at their peril that students study Q1 without awareness of Q2 – or Q2 without awareness of Q1. Therefore, throughout introduction, collation, and commentary, I propose to study the first quarto by constantly keeping an eye, as it were, on the second quarto. By providing separate critical editions of the first two early quartos, the New Cambridge Shakespeare series invites comparative study of the different versions of Romeo and Juliet.

   What, broadly, is the difference between the first and the second quartos of Romeo and Juliet? The difference in length is conspicuous: with some 700 lines more than Q1, Q2 is almost one third longer than Q1. The two versions have the same characters and dramatise the same events in the same order. Contrast the case of Hamlet and King Henry Ⅴ: Q1 Hamlet contains a scene unique to it which summarises material present in other scenes in Q2 and the Folio; Q1 King Henry Ⅴ omits a number of notable characters present in the Folio; Romeo and Juliet presents a different case. The Chorus at the end of Act 1 (1.5.144–57) is not in Q1, and a short sequence with the Capulet Serving-men (1.5.1–14) is similarly absent from the shorter text, but, apart from these two, all other dramatic movements in Q2 have their equivalents in Q1. Some of these movements are considerably shorter, though, and a number of Q2 speeches disappear entirely from Q1. Q1 also does not include at all or condenses a great many short or very short passages present in Q2. Each Q1 scene is shorter than its equivalent in the longer text, Scene 11 being less than half and Scene 2 being more than 90 per cent of Q2, with the relative length of the other scenes being somewhere in between. In fact, the only part of Q1 that is longer than Q2 is the stage directions, which tend to be more detailed and more numerous in the shorter text.

   Length constitutes an important but by no means the only difference between the two versions. Q1’s language often departs from Q2’s, the difference being at times a matter of isolated words but at others substantial. In general, the relationship between the two texts is relatively close in the first seven scenes (1.1 to 2.4) but becomes less so in the remainder of the play. Here is a representative example of parallel passages from early on in the play, excerpted from Romeo’s conversation with Benvolio in the first scene:

   Love is a smoke made with the fume of sighs,
Being purged, a fire sparkling in lovers’ eyes,
Being vexed, a sea nourished with loving tears.
What is it else? a madness most discreet,
A choking gall, and a preserving sweet.

   (Q2, 1.1.181–5)

   Love is a smoke raised with the fume of sighs,
Being purged, a fire sparkling in lovers’ eyes,
Being vexed, a sea raging with a lover’s tears.
What is it else? A madness most discreet,
A choking gall, and a preserving sweet.

   (Q1, 1.118–22)

   Here, by contrast, are parallel passages, also from Romeo’s part, more typical of the latter part of the play:

   I do remember an apothecary,
And hereabouts ’a dwells, which late I noted
In tattered weeds, with overwhelming brows,
Culling of simples; meagre were his looks,
Sharp misery had worn him to the bones;
And in his needy shop a tortoise hung,
An alligator stuffed, and other skins
Of ill-shaped fishes, and about his shelves
A beggarly account of empty boxes,

   (Q2, 5.1.37–45)

               As I do remember,
Here dwells a ’pothecary whom oft I noted
As I passed by, whose needy shop is stuffed
With beggarly accounts of empty boxes;
And in the same an alligator hangs,

   (Q1, 18.28–32)

   Despised, distressèd, hated, martyred, killed!
Uncomfortable time, why cam’st thou now
To murder, murder our solemnity?
O child, O child! my soul, and not my child!
Dead art thou. Alack, my child is dead,
And with my child my joys are burièd.

   (Q2, 4.5.59–64)

   Cruel, unjust, impartial destinies,
Why to this day have you preserved my life?
To see my hope, my stay, my joy, my life,
Deprived of sense, of life, of all by death?
Cruel, unjust, impartial destinies!

   (Q1, 17.85–9)

   All of the above examples are in verse, but parts of Q1 and Q2 are in prose and, at times, one text prints a passage in prose which the other text has as verse. Juliet’s mother has several prose speeches in Scene 3 of Q1 which are printed as verse in Q2. The same applies to the Nurse in Scene 4 (1.5). Conversely, Q1 prints the Queen Mab speech (Sc. 4; 1.4) as verse, whereas Q2 has prose. When it comes to the texts’ dramatic verse, Q1 and Q2, like Shakespearean drama in general,⁶ are predomi- nantly in iambic pentameters but with a share of short and long lines. It is noticeable, however, that Q1’s lines, on the whole, are metrically more irregular than Q2’s. The parts of Capulet and the Nurse in scene 14 (3.5) might serve as good examples to show the occasional irregularity of the verse in Q1 in comparison with Q2.

   This preliminary survey of differences between Q1 and Q2 of Romeo and Juliet has so far concentrated on length and language, but other elements could be added to this. The two versions usually assign the same speeches to the same characters, but on a few occasions they do not. For instance, in the first scene, two Q1 lines given to Montague’s Wife are spoken by her husband in Q2, while in Scene 5 (2.1), Mercutio is given a short passage attributed to Benvolio in the longer text. Characters usually enter and exit in the same order and at roughly the same point in the two texts, but here, too, there are exceptions. In the final scene, for example, both Balthasar and Friar Laurence are arrested and brought onstage by the watch, but while Q1’s Balthasar precedes the Friar, in Q2 Balthasar follows after him.

   In comparing Q1 to Q2, I have consciously chosen to present a value-neutral comparison, a comparison that refrains from judging the relative merit and explaining the likely reasons for the differences between the two texts. I have done so because responses to the first quarto have too often been marred or at least inflected by a parti pris regarding their relative value. If we are predisposed to think of Q2 as a ‘good’ and of Q1 as a ‘bad’ text, i.e. a text that has – say – been cobbled together by dishonest pirates, then we will be tempted, when the two versions differ, to prefer the presumed ‘good’ to the ‘bad’. The opposite of this prejudiced assumption about Q1 has also happened in the reception history of the texts of Romeo and Juliet: in the early twentieth century, Theodor Eichhoff edited the first quarto and went on to write a comparative study of Q1 and Q2 in which he argued for the superiority of the first quarto over the second.10⁷ Once Eichhoff had convinced himself of the texts’ relative merit, he was able to find reasons in virtually every parallel passage why the first quarto is much to be preferred to the second. Eichhoff is an extreme case, but it seems clear that an analysis of the first quarto and its relationship to the second can gain from establishing first what there is in the two texts and what the differences are between them. Naturally enough, this assessment leads to the question of why the first quarto came to take on the form it has, and why it differs from the second quarto in the way it does, and it is these complex questions that I now address.

Textual provenance

A CENTURY OF BAD QUARTOS'⁸

Scholarly thinking about the provenance of Q1 Romeo and Juliet is intimately related to the ‘bad quartos’ more generally, so I first wish to survey past thinking about this group of texts.⁹ It will be well to start by considering two publications that appeared at the end of the first decade of the twentieth century. In Shakespeare Folios and Quartos, Alfred W. Pollard invented the textual category which he labelled ‘bad quartos’.¹⁰ This invention followed from his revisionary reading of Heminge and Condell’s address in the First Folio ‘To the Great Variety of Readers’. While earlier commentators had believed that their reference to ‘diuerse stolne, and surreptitious copies’, with which the readers had previously been ‘abus’d’, had been to the quarto editions in general, Pollard insisted that only the ‘bad quartos’ were meant. Since several ‘good’ quartos had been used as copy when the First Folio was printed, Heminge and Condell, so Pollard argued, could hardly have claimed to have ‘cur’d’ all the ‘surreptitious’ and ‘maimed’ quartos. The texts he identified as ‘bad quartos’ were Q1 Romeo and Juliet (1597), Q1 King Henry Ⅴ (1600), Q1 The Merry Wives of Windsor (1602), Q1 Hamlet (1603), and Q1 Pericles (1609). In the following year, W. W. Greg refined that part of the narrative to which Pollard had paid little attention: the agency behind the manuscript copy of the ‘bad quarto’ that allows us to account for the text’s relationship to the ‘good’ Folio text. Noting ‘the very unusual accuracy with which the part of mine Host is reported’ in the ‘bad’ quarto of The Merry Wives of Windsor and ‘the comparative excellence of the reporting of those scenes in which the Host is on the stage’, Greg concluded that ‘the pirate who procured the copy . . . was none other than the actor of the Host’s part’.¹¹ Even though Greg cannot be credited with having invented the concept of ‘memorial reconstruction’ (see below), he seems to have provided both the label and the first detailed investigation of it.

   The twin theories of ‘bad quartos’ and ‘memorial reconstruction’ thus having been put forward in the space of two years by two of the leading scholars of their time, much of the scholarship in the following decades went into consolidating Pollard and Greg’s publications of 1909 and 1910 and applying them to other plays. In 1915, H. D. Gray suggested that Q1 Hamlet was also a memorial reconstruction, undertaken by the actor who played the role of Marcellus.¹² Once ‘memorial reconstruction’ had been applied not only to Q1 Merry Wives but to all of Pollard’s Shakespearean ‘bad quartos’, the theory spread beyond the bounds of the Shakespeare canon. Comparing the quarto edition of Orlando Furioso with Edward Alleyn’s extant part of the title character, Greg concluded in 1923 that the quarto is ‘a version severely abridged . . . for performance by a reduced cast’ and that the text ‘is based almost throughout on reconstruction from memory’.¹³ By 1930, E. K. Chambers fixed the canon of Shakespeare’s ‘bad’ quartos by endorsing the cases advanced for 1 Contention, Richard Duke of York, Romeo and Juliet, King Henry Ⅴ, The Merry Wives of Windsor, and Hamlet, while rejecting The Taming of a Shrew.¹⁴

   After Chambers’s influential pronouncement, the ‘bad quarto’ and ‘memorial reconstruction’ theories remained largely unchallenged in their broad outlines for roughly half a century.¹⁵ Summing up in a few sentences a critical territory on which much ink was spilt inevitably results in simplifications. It is certainly true that there was considerable argument as to what exactly was memorially reconstructed, by whom, and to what purpose. A popular reference work such as F. E. Halliday’s Shakespeare Companion can give a good impression, however, of how orthodox the assumption had become. The six Shakespearean ‘bad quartos’, Halliday wrote, ‘were reconstructed from memory by one or more actors who had played minor parts in a London production, and tried to reproduce the play for a provincial performance. This theory is generally accepted.’¹⁶ With the orthodoxy thus in place, ‘memorial reconstruction’ became an explanation with which scholars tried to account for textual features of a growing number of early modern play-texts, a number that, by the 1990s, reached a total of more than forty.¹⁷

PAST THINKING ABOUT Q1 ROMEO AND JULIET

Theories about the first quarto of Romeo and Juliet started long before the ‘bad quartos’ came into being as a textual category. Eighteenth-century editors usually considered Q1 as a first draft and Q2, in keeping with its title page (‘Newly corrected, augmented, and amended’), an improved and completed text on which they based their editions.¹⁸ Yet there were exceptions: Pope drew eclectically from Q1 and Q2, choosing whichever version he preferred, at times combining the two in the same passage or even line. Warburton, even though he did not base his text on Q1, preferred it to the later version, because the ‘trifling and bombast passages are in [the longer text] far more numerous’. He considered the additions in Q2/F to Q1 to be a result of the practice lamented by Hamlet, with actors speaking more than has been set down for them: ‘as a proof’, Warburton writes, that Shakespeare could not escape this practice, ‘in the old editions of Romeo and Juliet there is no hint of a great number of the mean conceits and ribaldries now to be found there’.¹⁹

   Prior to the 1840s, editors agreed, however, that Q1 was an authorial version and that it was the earlier of the two. The first editor who opposed this view was John Payne Collier in his edition of 1842–4. He wrote that Q1 had hitherto

been treated as an authorised impression from an authentic manuscript. Such, after the most careful examination, is not our opinion. We think that the manuscript used by the printer or printers . . . was made up, partly from portions of the play as it was acted, but unduly obtained, and partly from notes taken at the theatre during representation. Our principal ground for this notion is, that there is such great inequality in different scenes and speeches, and in some places precisely that degree and kind of imperfectness, which would belong to a manuscript prepared from defective short-hand notes.²⁰

   For the remainder of the nineteenth century, the traditional ‘early draft’ theory and Collier’s new theory co-existed, some scholars preferring the former (e.g. Knight, Dyce, Staunton, Ulrici, and Hudson) and others the latter (White, Mommsen, Furness, Daniel, and the Cambridge editors). Minor variations of Collier’s theory included first and still tentative considerations that the genesis of Q1 might be partly memorial, thus anticipating the ‘memorial reconstruction’ explanation prevalent in the following century. Writing about Q1 Hamlet and Romeo and Juliet, Tycho Mommsen, in 1857, held that both texts ‘abound with every kind of shallow repetition – now of set phrases, oaths, expletives, then (which is strongly indicative of interpolation) of certain lines and passages of peculiar energy, such as would impress themselves more literally upon the memory of the hearer’ (my emphasis).²¹ Four years later, White similarly considered ‘the unmitigated failure in the memory’ as a possible explanation for parts of Q1.²² Finally, a few years later still, the Cambridge editors argued for stenographic origins of Q1 but recognised that ‘the text of (Q1) is more accurate on the whole than might have been expected from such an origin’ and added that ‘possibly some of the players may have helped [the short-hand writer] either from memory, or by lending their parts in MS’ (my emphasis).²³

   The idea that memorial agency had something to do with the genesis of Q1 thus existed prior to the watershed publications by Pollard and Greg in 1909 and 1910, but it was thereafter that the memorial reconstruction theory fully established itself. In 1919, Pollard and J. Dover Wilson co-authored an article about the provenance of Q1 Romeo and Juliet which built on Pollard and Greg’s publications but was considerably complicated by their belief in the existence of an earlier, non-Shakespearean play. Q1, they thought, ‘represents an abridged version of Shakespeare’s first revision of an older [non-Shakespearean] play eked out by what a pirate could remember of the later version’.²⁴ When Chambers reviewed their convoluted narrative in 1930, he soberly concluded: ‘I do not find this theory satisfactory.’²⁵ Nevertheless, Chambers had no doubt that Q1 was ‘one of the bad Quartos’, held that it is ‘certainly a “reported” text’, and repeatedly referred to ‘the reporter’.²⁶ The fullest investigation of the provenance of Q1 Romeo and Juliet, which was to establish the dominant view for decades to come, was H. R. Hoppe’s book-length The Bad Quarto of ‘Romeo and Juliet’ in which he proposed ‘to demonstrate that Q1 of Romeo and Juliet is a memorial reconstruction of a version that Q2 represents in substantially correct form’.²⁷ This theory was endorsed by Greg in The Shakespeare First Folio and, among recent editors, Brian Gibbons, G. Blakemore Evans, and the editors of the Oxford Shakespeare.²⁸ The merits of and problems with the ‘memorial reconstruction’ narrative are explored below, but it seems fair to say that it had a crippling effect on scholarly engagement with Q1. This is exemplified by the following quotations from a 1955 survey of ‘Recent Work on the Text of Romeo and Juliet’ by Wilson: ‘it is a pirated edition and . . . exhibits all the stigmata we have learnt to associate with such texts . . . And that is practically all an editor needs to know about this text.’²⁹

THE EARLY DRAFT/REVISION THEORY

One argument which has recently been revived that conflicts with the memorial reconstruction theory is that Shakespeare revised his own plays, including Romeo and Juliet. As pointed out above, the first sketch/authorial revision view had many adherents in the eighteenth and nineteenth centuries. Its modern revival followed in the wake of Ernst Honigmann’s The Stability of Shakespeare’s Text which investigated ‘the possibilities of authorial “second thoughts” before its delivery to the actors’, resulting in ‘two copies of a play, each in the author’s hand, disagreeing in both substantive and indifferent readings’.³⁰ In the 1970s and 1980s, the authorial revision debate centred around King Lear, which many scholars argued was extant in two substantially different authorial versions. As Greg’s argument for memorial reconstruction had created a search for ever more memorially reconstructed plays, so the theory of an authorially revised King Lear triggered a reconsideration of Shakespeare’s possible revision of several other plays. As part of this reconsideration, the possibility that the ‘bad quartos’ in fact represented first versions which authorial revision turned into the longer and better-known plays was again examined. Steven Urkowitz supported this theory, arguing that ‘each of the multiple texts may represent a different stage in Shakespeare’s and his acting company’s composing and revising process’.³¹ That Q1 Romeo and Juliet is an authorial first draft has also been argued by Y. S. Bains: ‘Shakespeare composed the First and Second Quartos as two versions for the stage, and the second is superior to the first because he revised it.’³² An obstacle to belief in this theory resides in the fact that the derivative nature of most of the ‘bad quartos’ is strongly suggested by passages whose sense only becomes entirely clear after a comparison with the ‘good’ text. A couple of examples will suffice as illustration.³³ Q1 reads:

It is not entirely clear what ‘All this’ refers to. The second quarto adds a passage between lines 44 and 45 with which Juliet’s concluding words make better sense:

   Back, foolish tears, back to your native spring,
Your tributary drops belong to woe,
Which you mistaking offer up to joy.
My husband lives that Tybalt would have slain,
And Tybalt’s dead that would have slain my husband:
All this is comfort . . .

   (3.2.102–7)