"'A complicated and unpleasant investigation': The Arden Shakespeare 1899-1924" by Gabriel Egan

This paper arises from a survey of Shakespeare play editions in the twentieth century. I'm particularly interested in what those who made editions thought they were doing, how confident they felt about their work, how they thought readers would respond to the textual problems that arise in editing old plays, and how editors' assumptions about their readers were manifested in the editions that they produced. My published title in the programme covers the whole century of editions, but I'm going largely to confine my remarks to just one editorial project. For those of you who like to see the big picture first, however, I can offer a brief overview of just one of those variables I mentioned: editorial confidence [SLIDE]. I see it going like this, from a low at the start of the twentieth-century, through to a peak in the 1970s, and back to a low now. From the detailed history behind this pattern, I have room on this chart to pull just a few keys moments. [SLIDE] First, A. W. Pollard's book Shakespeare Folios and Quartos (1909) distinguished the good from the bad quartos and gave editors reasons to suppose that the good ones are textually close to Shakespeare's own papers. [SLIDE] Then in the 1940s and 50s Charlton Hinman's collation of Shakespeare Folio copies establishes the order of the setting of its formes and so laid the foundation for working out how the compositorial labour was divided amongst particular compositors.

[SLIDE] By the 1960s, the division of compositorial labour in the early Shakespeare printings was starting to offer the possibility of figuring out just how careful or sloppy each man was, and hence which parts of which plays one should emend conservatively and which parts need a heavier hand. [SLIDE] In 1969 Don McKenzie threw a spanner in the works by showing that the practice of concurrent printing--printshops working on more than one book at a time--invalidates most of the hard-won calculations about compositorial labour. This is the turning point where the rising editorial confidence starts to slow, to peak, and begin to fall. Around that peak in the late 1970s and early 1980 there were subtle influences from high French literary theory starting to be felt in literary criticism, and these reached Shakespearian textual studies in the 1980s and 1990s, accelerating the decline in editorial confidence. [SLIDE] During this decline, the 1986 Oxford Complete Works of Shakespeare was published, and was widely condemned for editorial hubris--its interventions in the received text were the most radical for centuries--and what got overlooked were the full implications of this edition's shifting of the editorial goalposts. For, this was the first edition fully to put into practice the conviction that Shakespeare's plays are better read as collaborative works produced by a theatrical team--the early-modern playing company--than as the outputs of a single mind. [SLIDE] Towards the end of the century, the Third Arden Shakespeare series was started on an entirely new, and I would argue mistaken principle, that one cannot produce an edition of 'the play', conceived as a kind of intellectual ideal, but rather one can only produce an edition of an existing, early print manifestation of that ideal.

So, that's the big picture. For this talk, I want to focus on the early part of the story [SLIDE], where we can add in some detail [SLIDE], and see how this new editorial theorizing, what is to become the New Bibiography, affects the first series [SLIDE] of the most important critical edition of the play, the Arden Shakepeare, which overlapped slightly with the grandest New Bibliographical project of the century, John Dover Wilson's New Shakespeare [SLIDE] for Cambridge University Press.

The first Arden series began  in 1899 under the general editorship of Edward Dowden, and I'll survey the 37 volumes in chronological order, trying to get a sense of editorial thinking prior to, and for the later volumes, during, the rise of New Bibliography, and looking to see how the editors conveyed to readers the complexities of the textual matters in Shakespeare. I'm interested in how the editors' working hypotheses about the textual conditions and relations of the early printings fed into their textual practices: did they want to be faithful to their copy text or did they feel free to depart from it? Did they want to speculate about their copy text's underlying printer's copy? And how far did they think discussion of such things was part of their jobs as editors?

[SLIDE] Here is the full list of Arden 1 play editions; I leave aside the two volumes that deal with Shakespeare's poetry as editorially that's a very different proposition. What can we say about all of them? As is well known, the stage-directions are in every case readerly. That is, they inherit from previous editions the belief that readers need to be told where the action takes place. [SLIDE] So, for example, Edward Dowden's Hamlet begins with the words "Act I | SCENE I.--Elsinore. A Platform before the Castle. FRANCISCO at his post. Enter to him Bernardo." This of course gives the reader more information than is unavailable to the theatregoer at this point--for example that, of the two of them, Francisco is the man on duty--but also denies the reader information. Dowden neglects to mention that these men are sentinels, a fact that all three of the early printings include in the opening stage direction and which doubtless directed the actors to be dressed and armed like sentinels. In the theatre you see two guards but don't know where they are (unless you already know the play), while in Dowden's edition you don't know that these men are guards but you do know that they are at Elsinore Castle.

But we must not leap to simply labelling this series 'readerly' instead of 'theatrical' [SLIDE] for the first 16 volumes did something that is surprising in paying little typographical attention to the division of the play into acts. To take Hamlet again [SLIDE] Dowden merely inserted a centred, all-capitals marker before the scene marker, and ran the text on so there is little visual interruption around the act break. This is in a sense the theatrical thing to do, since for most of Shakespeare's career his plays were performed without pausing between the acts. The more readerly approach, on the other hand, and the one taken in the later volumes in this series and throughout the Arden 2 series that ran from 1951 to 1982, is to start each new act on a new page, as though it were a chapter in a novel. As far as I can tell, that was an Arden innovation: the prevailing nineteenth-century norm was to run the text on without a page-break, and for the New Shakespeare Wilson went even further the opposite way and inserted no act marker at all: the scene-numbering marker simply clicked over another arabic numeral left of the decimal point, from say 2.2 to 3.1.

While we have this opening on the screen: all of the first Arden series place the collation in the middle of the three vertical-divided zones [SLIDE] of the page: dialogue, collation, and explanatory notes, an arrangement arising first, I think, in Horace Howard Furness's Variorum edition begun in 1871. In matters textual, however, the Arden editors--at least in the first couple of decades of the century--were in the shadow of the Cambridge editors of the 9 volumes of the Macmillan edition of 1863-6: William George Clark, John Glover, and William Aldis Wright. One gets the sense from their introductions that in general the Arden editors either didn't feel they had the competence, or indeed don't feel that there was the need, to do fresh textual work, and even those who do almost all credit the Cambridge with the work of collating the second, third, and fourth Folios. (Of course, that they feel that it is worth collating these derivative editions is itself indicative of the prevailing assumptions about the early printings of Shakespeare.) So, let us turn to the textual work done.

In the introduction to his Hamlet, Dowden offered the opinion that the customary act interval placings are wrong and stated that he would fix them "but for the inconvenience of disturbing an accepted arrangement, to which references are made in lexicons and concordances" (xxii). Dowden was writing as the general editor of the new series (Murphy 2003, 206-07), and is it a mark of his fairly low hopes for the project that he did not think that a fresh examination of the facts would give the Arden editions the intellectual authority to overthrow the "accepted arrangement" and establish new, more correct, locations for the act breaks. Dowden's approach to the collating of textual variants had no stated logic, but he was clearly concerned to show that he could see sense in the readings that were not in his central authority, the 1604 second quarto.

Indeed, Dowden was open to the idea that what actors do might bear upon an editor's ideas about the rightness of a variant. For example, in the opening scene [SLIDE] Dowden assigned the line "What, has this thing appear'd again to-night?" to Marcellus [SLIDE], and his collation showed that Q1 (the bad quarto) and F agree on this, while Q2 (what he calls just Q) instead gives the line to Horatio. But see how his explanatory note, as well as saying that he let the combined authority of Q1 and F overturn the authority of his copy text, remarked that [SLIDE] "'This thing' may be uttered with awe by Marcellus, or with an air of incredulity by Horatio". Those who have followed recent arguments about the future of editing Shakespeare will recognise Dowden's approach as one that has recently become fashionable again: editors should not prescribe what is right and wrong, but simply lay out for the reader the possibilities, taking into account what happens in performance.

Dowden gave up generally editing the Arden Shakespeare after the poor sales of Hamlet and W. J. Craig took over. Craig too was ahead of his time in certain ways, and in certain remarks he anticipated Pollard's foundational step of dividing the quartos into good and bad. In his edition of King Lear in 1901, Craig spotted the problem [SLIDE] that the 1607 Stationers' Register entry for the play mentions George Buc [SLIDE]: "Entred . . . vnder the handes of Sir George Buck and The wardens". The problem is that in their preliminaries to the 1623 Folio, John Heminges and Henry Condell characterized the existing quarto editions as "stolen and surreptitious" (Shakespeare 1623, A3r) and yet here is one of those quartos entered in the register under the hand of the Master of the Revels Buc, who surely is above suspicion. Craig put his finger on the problem that Pollard 'solved' in 1909 by showing the Heminges and Condell did not mean that all the existing quartos were illegitimate, only that a certain subset, the bad quartos, were stolen and surreptitious. But Craig refused to make anything of the problem, concluding "I must leave the reader to settle that question . . ." of what Buc was doing in connection with King Lear (xiii), which is asking a lot of the reader.

But since we specialist readers are in a position to pursue it, I will comment that the fact Buc is mentioned in the Stationers' Register entry doesn't mean he had anything to do with it: it might just mean that the clerk making the entry was shown a copy of the play with Buc's performance licence written in it. This was the view taken by Janet Clare (Clare 1990, 132). Richard Dutton, on the other hand, thought that Buc was licensing only books, not performances, before 1610 (Dutton 1991, 142-63), so necessarily it follows that Buc's licence was just for the printing. Taking the inevitable third view, Gary Taylor concluded that Buc "licensed both the performance and (separately) the publication of King Lear" (Taylor 1983, 111n7). We haven't an answer, then, to the question of what Buc's name is doing in this entry, but we at least we can scope the problem and know the range of possibilities. Surprisingly, however, the reader of a modern critical edition of King Lear is not so well served, for in the latest Arden 3 edition by R. A. Foakes and in the Oxford Shakespeare single-volume edited afresh from this quarto by Stanley Wells, the Stationers' Register entry is discussed in abbreviated form and the presence of Buc's name is not even mentioned (Shakespeare 1997, 90; Shakespeare 2000, 9-10).

[BLANK SLIDE] Like Dowden, Craig wrote without any sense that he might be in at the start of something big in textual studies. Of the texts of King Lear he wrote that  ". . . the ground has been too exhaustively worked by preceding editors to admit of any new discoveries of importance" (xv). Continuing the theme of editorial redundancy is Michael Macmillan's 1902 King John for the Arden: "I have not had the temerity to suggest many new readings" (v). In his 1903 Henry 5 Herbert Arthur Evans was at least interested in matters textual, and he spotted the importance of P. A. Daniel's demonstration () that the 1600 quarto cannot be Shakespeare's, or indeed anyone else's, first stab at the play that eventually became the Folio version, for the quarto has in it lines that can only be explained as garbling of the words in the Folio version (xvii-xxv). Daniel, writing in the 1870s, had anticipated the logic of Peter Alexander's seminal argument on memorial reconstruction by nearly 50 years (Alexander 1924).

Keeping up the atavistic conservativism, in his All's Well that Ends Well of 1904, W. Osborne Brigstocke assured the reader that "I have throughout endeavoured to be as conservative of the original folio text as possible" even though this is "one of the worst printed plays in the volume" (ix). H. Bellyse Baildon went even further and despite describing the quarto and Folio texts of Titus Andronicus he managed to end the introduction of his 1904 edition without having mentioned which he took as his copy text, and he provided no collation. In his 1904 Merry Wives of Windsor, H. C. Hart, like Craig in his King Lear, unearthed materials that could have overturned key aspects of the textual consensus regarding the quarto and Folio, but by steadfast adherence to unshakeable convictions he was able to resist the temptation to do so. Do we not have, he reasoned, the word of Heminges and Condell that they did right by their friend? Thus no doubt when they complained of "stolne and surreptitious copies, maim'd and deformed" that had been foisted on the reading public, they had the quarto of The Merry Wives of Windsor in mind, and so ". . . the Folio may be accepted as the text of the play in its entirety . . ." (xi). Because the text underlying the Folio was marked up for theatrical cuts and joins, there are unfortunate anomalies in it, especially around Falstaff's meeting with Mistress Ford at the end of act 3. However, wrote Hart, "These can only be set right by the alteration of words in the text. This is not, happily, the province of an editor, for it is a complicated and unpleasant investigation" (xiv). Hart had the moral fibre to resist that sort of thing: "Nothing", he wrote, "but an unswerving reverence for the Folio text enables one to withstand the temptation" to insert quarto-only passages in one's edition", but alas that reverence "did not", he complained, "belong to the early commentators" (xxiv).

Towards the end of the first decade of the century, just as the New Bibliography was emerging, the Arden introductions got increasingly concerned with matters textual. Whereas the introductions to several of the earliest volumes said nothing about the text, concerning themselves solely with their plays' themes, character, sources, and related literary topics, the later ones started to devote ever greater space to the textual problems. Thus, in his 1907 Pericles, K. Deighton admitted all the stylometric evidence that for most of twentieth century was to be ignored: "metrical tests . . . prove almost the whole of the first two Acts to be by some other author" than Shakespeare (xxv). The whole of Deighton's introduction dealt with the authorship, leaving room for nothing else, and his is the best summary of the textual situation of Pericles available until the Oxford Shakespeare of the 1980s (Wells et al. 1987, 556-60), and the best account of the debate about the evidence until Brian Vickers's recent book Shakespeare, Co-Author (Vickers 2002).

In these 100-year old Ardens there is abundant material still germane to textual scholarship. Another example [SLIDE]: the latest complete works of Shakespeare, the Macmillan/RSC edition managed by Eric Rasmussen and Jonathan Bate, is based on the Folio on the grounds that it, the Folio, has a general theatrical authority deriving from the actors Heminges and Condell working on it. [SLIDE] According to Bate ". . . Hemings and Condell authorized the publication of the Folio, and no doubt played a major role in obtaining copy, deciding on which texts to follow and which plays to include or exclude . . ." (Bate 2007, 12). [SLIDE] But as Henry Cunningham pointed out in his 1912 Arden Macbeth,  Heminges and Condell played down their role [SlIDE]: "We haue but collected them . . . . We cannot go beyond our owne powers" (Shakespeare 1623, A2v) and "[we] who onely gather his works" (Shakespeare 1623, A3r). Even allowing for conventions of modesty, Cunningham was right to stress that we have no evidence of Heminges and Condell's involvement in this publication beyond their writing of this prefatory material. In his defence of simply modernizing the Folio--of not editing the plays, in other words--Bate condemned the New Bibliography that was emerging as Cunningham wrote. [BLANK SLIDE]

Emerging, but not yet emerged. The New Bibliography is first acknowledged when J. W. Holme edited As You Like It for the Arden in 1914. Holme dated the play from its Stationers' Registry entry, which he explained using the narrative given in Pollard's 1909 book Shakespeare Folios and Quartos. The general editor R. H. Case, who took over in 1909, was quickly convinced by the new optimistic thinking, which shows in his additions to reprints of the early volumes (about which I'll say more in a moment) and in his own edition of Coriolanus in 1922. In that edition, Case indicated his agreement with the continuous-copy theory outlined by Pollard in his 1920 book, Shakespeare's Fight with the Pirates, which held that  ". . . the author's autograph copies of his plays became the prompt-copies, and that the text of many of the plays, both of those printed in quarto and those which first appeared in the folio, were set up from them" (xxv). Likewise in his 1923 Arden edition of 2 Henry 4, R. P. Crowl wrote that "On the whole, we have cause to congratulate ourselves upon the excellencies of the text of 2 Henry IV. as transmitted to us in the authorised stage version published by Wise and Aspley in 1600, and in the completer version of the Folio" (xv). Whereas the early Ardens were gloomy, seeing the text as debased but if we're lucky not too much debased, here Cowl, under the influence of Pollard, saw reason to celebrate because we have a good quarto from an authentic script, possibly even the author's script, that was improved by being used in the theatre and there acquiring the stage directions of the company.

Yet this new optimism about the textual condition of the plays did not instantly generate editorial confidence. Cowl prided himself on his textual conservativism, by which he meant not only sticking to the readings in one's copy text but also, more generally, not rocking the boat, not challenging the editorial tradition. Thus although he rightly identified the 1600 quarto of 2 Henry 4 as the most authoritative early printing, he did not think this sufficient grounds to overturn a tradition. In one place, he reported, ". . . a Folio reading has been adopted, where the restoration of the reading of the Quarto would have involved the excision of a half-line which was probably an interpolation, but which time has invested with authority" (vii). Cowl here used 'authority' virtually in the post-modern, and currently fashionable, sense that disperses authority from the singular originating writer to the many readers. And yet, although it took a little while, practices were set change because of the new theories. The last volume in the series, Grace R. Trenery's 1924 Much Ado About Nothing, began with the play's Stationers' Register entry and Pollard's 1920 interpretation of it, which she entirely accepted. The quarto, she agreed, was based on "theatrical prompt copy" (ix) and she was almost convinced by Pollard's argument that the copy was in Shakespeare's own hand. If that were proven, she wrote, we should hold the quarto "in still greater reverence and the alterations and emendations of the later editors in rather less respect" (x).

The Arden series took 25 years to do all the plays, and the early volumes were reprinted with revisions, some of them 2 or 3 times, before the final volume appeared. The commonest kind of revision was the general editor R. H. Case (who took over in 1909) adding extra footnotes that amplified, corrected, and even flatly contradicted the editor whose work was being reprinted. Thus in his 1909 edition of 1 Henry 6, H. C. Hart explained that the Henry 6 trilogy was the collective writing of Shakespeare, Christopher Marlowe, George Peele, and Robert Greene, which is why The Contention of York and Lancaster (printed in 1594) and Richard Duke of York (printed in 1595) have the distinctive words and phrases of all four men. For the second impression of Hart's edition, in 1930, Case added a footnote referring the reader to Peter Alexander's alternative explanation that the 1590s printings were cobbled together from the actors' memories and thereby bits of other plays in the repertory, recalled by the actors desperate to make something good, got mixed in with Shakespeare's words. Alexander's work, wrote Case, "seriously damages the case for attributing these plays in whole or part of other authors than Shakespeare" and thus "The whole of this Introduction should now be read in the light of" Alexander's discovery (vii n.1).

Throughout the reprinted introduction, Case inserted new footnotes disagreeing with Hart's main text. If this seems an extraordinary way for a general editor to behave, we should note that Case had at least the decency to wait for a reprint for inserting his disagreements. [SLIDE] For the same play, 1 Henry 6 in the most recent Arden series, published in 2000, the general editors Richard Proudfoot and George Walton Williams insisted on appending a note telling their editor, Edward Burns, where he'd gone wrong. [SLIDE] He ought not, they wrote, have given the character Joan of Arc the last name "Puzel" (in place of the more usual Pucelle) nor the heir to the French crown the name "Dolphin" (in place of Dauphin). Depending on how you view it, this is either a refreshing willingness to put on display the multivocal nature of a text written by W and edited by X according to the rules laid down by Y and Z, or, [SLIDE] it's a failure of general editors to live with the consequences of their avowed policy of giving each volume editor the final say.

To put this another way: should the inherent co-authorship of a modern critical edition be highlighted or effaced? It's a question we can take back to the plays themselves, for a distinctive feature of the first Arden series is its editors' acceptance of Shakespeare's collaborative writing habits. This was the widely held view until the 1920s, and two things killed it off. The first was New Bibliographer Peter Alexander's hypothesis of Memorial Reconstruction that blamed the badness of the bad quartos not on their writers--an inexperienced young Shakespeare and/or his inferior collaborators--but on the actors' poor recollections. The second was E. K. Chambers's influential British Academy lecture "The Disintegration of Shakespeare" (Chambers 1924-25), which, without mentioning it, made the Memorial Reconstruction hypothesis attractive since it explained away the evidence that had pointed to collaboration. In his lecture Chambers had somewhat to suppress his own previously-given opinions, for as Henry Cunningham pointed out in his 1912 Arden Macbeth (xvii), in the Warwick Shakespeare edition of the play Chambers himself marked off the passages he thought were Middletonian interpolations. Yet in his British Academy lecture, Chambers called the claim that Middleton had a hand in Macbeth a "heresy" of Coleridge's elaborated by Victorian editors (Chambers 1924-25, 92).

Implicitly, much that I have said concerns editorial psychology: such things as adherence to authority, predilection to innovation, reverence for traditions, and taste. Many of the editors I have quoted refused to intervene in their texts, even where they thought them faulty. An exception was Henry Cunningham, who in his 1905 edition of A Midsummer Night's Dream set out to produce "a text in advance of anything that has hitherto been published", which was possible because " . . . the text of Shakespeare is by no means 'fixed and settled.' Far from it", he wrote (ix-x). Is the textual condition of A Midsummer Night's Dream so bad that Cunningham felt he had to intervene? No, quite the reverse: as Peter Holland put it this play "does not leave the editor tearing out his/her hair in anguish" (Shakespeare 1994, 119). If anything, it seems the relative ease of his task emboldened Cunningham, and that's the problem. As Gary Taylor pointed out, ". . . editors almost invariably divide into those . . . who in practice always emend sparsely and those . . . who in in practice always emend generously" (Wells et al. 1987, 59). For all its faults, the great benefit that the New Bibliography brought to editing was an insistence that the degree of intervention ought to be set not by the inclination of the editor, but by how corrupt she thinks is the text. That is, she should emend heavily where her copy text is full of error (indeed, in the particular parts of the play that are most erroneous, if she can identify them) and lightly where her copy text seems largely correct. That the New Bibliographers themselves did not always live up to their ideal in no way invalidates it.

Works Cited

Alexander, Peter. 1924. "II Henry VI and the Copy for The Contention (1594)." Times Literary Supplement 1186 (9 October). 629-30.

Bate, Jonathan. 2007. "The Folio Restored: Shakespeare 'Published According to the True Originall Copies'." The Times Literary Supplement Number 5429 (20 April). 11-13.

Chambers, E. K. 1924-25. "'The Disintegration of Shakespeare': The British Academy Annual Shakespeare Lecture Read 12 May 1924." Proceedings of the British Academy 11. 89-108.

Clare, Janet. 1990. 'Art Made Tongue-tied By Authority': Elizabethan and Jacobean Dramatic Censorship. The Revels Plays Companion Library. Manchester. Manchester University Press.

Dutton, Richard. 1991. Mastering the Revels: The Regulation and Censorship of English Renaissance Drama. London. Macmillan.

Murphy, Andrew. 2003. Shakespeare in Print: A History and Chronology of Shakespeare Publishing. Cambridge. Cambridge University Press.

Shakespeare, William. 1623. Mr. William Shakespeares Comedies, Histories, & Tragedies. Published According to the True Originall Copies. STC 22273 (F1). London. Isaac Jaggard and and Edward Blount.

Shakespeare, William. 1994. A Midsummer Night's Dream. Ed. Peter Holland. The Oxford Shakespeare. Oxford. Oxford University Press.

Shakespeare, William. 1997. King Lear. Ed. R. A. Foakes. The Arden Shakespeare. Walton-on-Thames. Thomas Nelson.

Shakespeare, William. 2000. King Lear. Ed. Stanley Wells. The Oxford Shakespeare. Oxford. Oxford University Press.

Taylor, Gary. 1983. "Monopolies, Show Trials, Disaster, and Invasion: King Lear and Censorship." The Division of the Kingdoms: Shakespeare's Two Versions of King Lear. Edited by Gary Taylor and Michael J. Warren. Oxford. Oxford University Press. 75-119.

Vickers, Brian. 2002. Shakespeare, Co-author: A Historical Study of Five Collaborative Plays. Oxford. Oxford University Press.

Wells, Stanley, Gary Taylor, John Jowett and William Montgomery. 1987. William Shakespeare: A Textual Companion. Oxford. Oxford University Press.