"The first 5 years of the Arden Shakespeare: 8 plays, 1899-1903" by Gabriel Egan
This paper arises from a project to examine the theory and practice of editing Shakespeare in the twentieth century. The New Bibliography that emerged in the first years of the century had by the 1930s clearly affected the way that editions of Shakespeare characterized and responded to the textual situation. My original plan was here to compare the first Arden series (1899-1924) to Cambridge University Press's The New Shakespeare (1921-66) in order to see how quickly and in what ways the new ideas began to shape editorial practice. This turned out to be more work than can be squeezed into an SAA paper, so I hope that seminarians will indulge my giving only the beginnings of the historical survey. I will consider just the 8 play editions that appeared almost at once (within a span of 5 years) at the start of the first Arden Shakespeare series.
The Arden Shakespeare was to become the most successful mass readership Shakespeare edition in the twentieth century, articulating recondite textual matters to scholarly readers while also providing commentary and notes that the general reader, the student, and the theatre practitioner would find helpful in trying to understand drama from an historically distant culture. To discover how the Arden Shakespeare later came to this profitable satisfaction of diverse markets, I will attempt to infer from the introductions a sense of the editorial thinking at work in these first 8 volumes by 6 editors, published just before the emergence of the New Bibliography that can for convenience be dated from the publication of A. W. Pollard's Shakespeare Folio and Quartos (Pollard 1909).
I will try to characterize these editors' sense of self-confidence in their work: how much closer to what Shakespeare wrote did they suppose their labour could take them, and how much better than the early printed texts did they think their editions could be? I will record what I think are their underlying textual assumptions and where detectable the unconscious biases. Specifically, did they attempt to be faithful to their copy texts or did they feel warranted in departing from them? How far, if at all, did they speculate about the manuscript printer's copy for their copy text? Finally, and most easily stated of all: to what extent (counted in pages) did they discuss matters textual in preference to the many other concerns that might appear in the introduction to a single-volume-single-play edition of Shakespeare?
This work has been made easier by the publication of our seminar leader Andrew Murphy's magisterial Shakespeare in Print by Cambridge University Press in 2003. Murphy has 'scoped' the field and laid out a map of the terrain that allows those of us who follow to supply detail without losing sight of the wider perspective. I shall proceed with the editions in the order he lays them out in his Chronological Appendix (Murphy 2003, 367) with two exceptions where I dispute his dates, as explained below.
*
The first of the Arden Shakepeares was Hamlet edited by Edward Dowden (Shakespeare 1899), which had an introduction of just 20 pages. Dowden explained that Q2/F was the basis of his edition, and that every important variation from each of them would be recorded (p. ix). He thought Q1 was surreptitious, made by a stenographic record of the dialogue that was perhaps fleshed out by referring to the theatre-library copy of the true text (p. xvi). Of the differences between Q1 and Q2/F Dowden gave a wonderfully succinct but detailed account, focussing on such crucial matters of the plot as Gertrude's being made aware in Q1 of Hamlet's suspicions about his father's death (p. xvii). Dowden rejected the idea that there is behind Q1 an old Hamlet play, perhaps by Thomas Kyd (p. xvii), and observed that "The general style of the Hamlet of 1603 is much more like that of an ill-reported play of that date than like the style of a play of Kyd's and Marlowe's time . . ." (p. xix).
There are parts of Dowden's introduction that must have been (and still are) difficult for a non-specialist reader to follow, for they assume considerable theatre-historical knowledge. For example, Dowden explained the omission in Q2 of "the passage about the boy actors in II.II" as being "probably made, as Professor Hall Griffin suggests, because it would be unbecoming in the King's servants to show hostility to the children, who were servants of the Queen" (p. xix). The Folio text of Hamlet Dowden thought "cut for the purpose of stage representation", so "more theatrical, but less literary" than Q2, and he noted that the Folio editors tried to "modernize certain words which were regarded as old-fashioned" (p. xx). Dowden considered the customary act interval placings to be wrong and would have corrected them in his edition "but for the inconvenience of disturbing an accepted arrangement, to which references are made in lexicons and concordances" (p. xxii). The inherited body of scholarship necessarily embodies standards and the conveniences these afford prohibit interference in the received Shakespearian text: the likely improvement would not be worth the trouble.
There was no stated logic to Dowden's selection of variants from F and Q2, which latter he confusingly called Q, the 1603 edition being not good enough even to receive its usual label. Or rather, in his introduction he did this: in the collation the 1603 text is called Q1. A glance at the collation for 1.1.14-21 gives a flavour of Dowden's eclecticism:
14. ho! Who is ] who's F. 16. soldier ] F, souldiers Q. 17. has ] F, hath Q.
21. Mar.] Q1, F; Hora. Q.
Of these choices, only the last is discussed in an explanatory note. Regarding his reassignment of "What, has this thing appeared again tonight?" from Horatio (to whom Q2 gives it) to Marcellus (to whom Q1 and F give it), Dowden thought the agreement of Q1 with F1 (and indeed the later Folios) was strong evidence. He nonetheless suggested in his note how the line would bear different dramatic weight in the mouth of the scholar Horatio and the mouth of the soldier Marcellus. It is worth noting that in not closing down the theatrical possibilities (while necessarily choosing one of them for his text) Dowden displayed an openness about performance and potentialities that is again popular, on this the far side of the New Bibliographical revolution.
Dowden resigned as general editor of the series after publication of his Hamlet edition, to be replaced by W. J. Craig (Murphy 2003, 207), but he continued as editor of particular volumes. The second in the series was Dowden's Romeo and Juliet, with an introduction of 31 pages (Shakespeare 1900). As before, Dowden labelled as simply Q the quarto of greatest authority, which we would call Q2 (p. x). This would be his copy text, but "the corrections of the later Quartos and of the Folio are valuable aids towards ascertaining the text, while in not a few passages Q1 lends assistance which cannot elsewhere be found" (p. xi). So, his approach can here too fairly be characterized as eclecticism. Q1 Dowden thought useful for giving us sight of original readings that Shakespeare later revised out, and was itself "made up partly from copies of portions of the original play, partly from recollection and from notes taken during performance" (p. xii). Dowden was explicit in his belief that Shakespeare substantially revised a play after completion of a first version, and he gave an example of an entire speech in Q1 rewritten to make the improved version in Q2 (p. xiii). Thus Q1 was "an imperfect representation, piratically issued" of substantially the same play as shown in Q2, but before Q2 appeared Shakespeare rewrote particular passages (p. xiii). The rest of Dowden's introduction was taken up with matters of themes and characterization and, rather jarringly, with explanatory notes that he confessed to writing too late to put into their correct places in the body of the book. The collation shows essentially the same eclectic attitude towards the early texts as shown in his Hamlet.
The following year appeared the new series editor W. J. Craig's edition of King Lear with a 58 page introduction (Shakespeare 1901). Craig could not know--because W. W. Greg did not publish the discovery for another 7 years (Greg 1908a; Greg 1908b)--that Nathaniel Butter's edition dated 1608 was in fact published in 1619 (pp. vii-xi). Craig did however know (because the editors of the 1863-66 Cambridge edition had established by collation) that the Butter edition was merely a reprint of the Pide Bull quarto of 1608. Yet, despite admitting this Craig thought that the Butter edition "contains some valuable corrections of the Pide Bull" (p. xii). Craig spotted that Nathaniel Butter and John Busby's Stationers' Register entry for King Lear in 1607 has George Buc's authority behind it--"Entred for their copie vnder th[e h]andes of Sir GEORGE BUCK knight and Th[e ]wardens A booke called . . ." ()--and yet in the Folio preliminaries all the preceding quartos appear to be characterized as stolen and surreptitious (Shakespeare 1623, A3r). As Master of the Revels, Buc ought to be beyond suspicion of such things, but rather than offer a solution to the mystery Craig wrote "I must leave the reader to settle that question . . ." (p. xiii).1
Craig thought that Q's differences from F were the result of authorial or non-authorial revision but it is "impossible to say with any certainty" which is the original and which the revised text even though the "superiority of the Folio text" is beyond doubt (p. xiii). This suggests that superiority is not to be determined by authoriality (since even if it is the site of non-authorial revision, F is better), but Craig did not make clear what the alleged superiority consisted in. He was aware that in part F depended on Q since it repeated more of its errors than can be attributed to mere coincidence. At this point Craig shifted his ground and observed that in a case of authorial revision an editor ought to follow the revised text even if she thinks the revisions worsened the work (p. xiv). Craig decided that the Q/F differences were not authorial but he was overtly uncertain, and so rather than choosing one of them as his authority--necessarily F he claimed, without saying why--one "must needs be eclectic". In selecting a Q or F reading where they differ "fitness and positive superiority, or what in my judgment I deem to be such, are the only guides . . ." (p. xv).
Craig's collation showed that he was indeed eclectic in his selection of variants. This, he noted, was how previous editors had worked on the play and it was not surprising that his text was like theirs since ". . . the ground has been too exhaustively worked by preceding editors to admit of any new discoveries of importance" (p. xv). Clearly, Craig felt himself to be working near the end of a long and fruitful tradition (with only the tidying up left to do) rather than setting out to develop a new one that would break fresh ground. That the Arden Shakespeare should start with such a strong sense that there was little textual work left to be done is surprising, for it would seem to confine an edition to neglect. More surprisingly still, neglect ensued: to read Craig's introduction in the 100-year old copy in the specialist Shakespeare Institute library I had to cut open the pages.
According to Murphy's dating the next Arden Shakespeare, Morton Luce's edition of The Tempest, appeared in 1901 (Murphy 2003, 367) but according to the history of the reprinting and revision given in later printings it was published in 1902 (Shakespeare 1902b). This edition need not detain us as its 62 pages of introduction contain not a word on matters textual. The same year appeared the Arden Shakespeare Julius Caesar (Shakespeare 1902a), in which Michael Macmillan reported that he had "not had the temerity to suggest many new readings" (p. v). Like Luce's, Macmillan's 81 pages of introduction made no mention of the textual situation. And as with Luce's The Tempest the availability of only the Folio text did not tempt the editor to speculate about the provenance of the manuscript copy for F, which speculation was to become characteristic of New Bibliographical thinking.
Dowden's final contribution to the series was his edition of Cymbeline with an introduction of 37 pages (Shakespeare 1903a). Dowden began with comments on the difficulties of F and regrets at not having a Q "to test the readings of the Folio and correct some of its errors" (p. vii). Noticeably, he did not entertain the possibility of basing his edition on such an imaginary quarto if he found it was only slavishly reprinted in the Folio. Dowden was apparently convinced, without saying why, that the Folio would in all events be preferable. (This cannot be a simple preference for the Folio in every case, for as we saw above Dowden based his Romeo and Juliet on Q2.) Dowden somewhat boastingly reported that when collating F1 with its reprints he used his own copies of F2 and F4. Dowden's reassurance that would edit conservatively (p. vii) is borne out. He came up with Imogen's (as he named her) "Think that you are upon a lock, and now | Throw me again" (his 5.5.262-3) to improve on F's awkward "upon a rock", supporting this from the newly-published New English Dictionary (NED) and adding a note about it at the end of his text. Yet Dowden did not adopt this reading in his text: being quite certain and having strong evidence in support of an emendation was not enough, it seems, to warrant interference in the received text.2
That Dowden was nonetheless interested in what we would consider matters textual is indicated by his reference to metrical tests that help identify the chronology of Shakespeare's works (p. x). The remainder of Dowden's introduction is concerned with Shakespeare's sources--or rather, more widely, his inspirations--and this interest in the mind rather than the texts is evident in all the introductions in the series so far. An isolated exception concerns the bad verses spoken by the ghosts of Posthumus's family between their entrance while he sleeps and their being silenced when Jupiter enters on an eagle. Dowden supposed that the ghosts were originally silent throughout but that audiences wanted more from them so one of the actors cobbled together this doggerel and "In the theatrical copy of the play from which the Cymbeline of the Folio was printed these lines naturally were found, and before 1623 they had become an accepted portion of the whole" (p. xxxix). Here, theatrical interpolation is brought in to explain badness, and in order to support this there had to be editorial comment on the nature of F's copy, which otherwise would not be mentioned.
In a similar vein, Dowden had explained in his Hamlet edition that the prince's dying "O, o, o, o" in the Folio was an example of "actors' addition" (Shakespeare 1899, xx). The low importance he gave his "Idle conjectures" about such things is shown by his not acting on them--he does not remove the alleged interpolations--and in his calling his conjectures "harmless" because "not insisted on" (Shakespeare 1903a, xl). In the case of Hamlet's "O, o, o, o" Dowden left out the actors' interpolation, but that simply meant adopting the Q2 reading, whereas in Cymbeline he was clearly conscious of having no alternative authority so he did not interfere. The same belief held equally strongly (that both were actors' interpolations) led to different editing practices, depending on whether there was an alternative early printing to support the supposed correct reading.
By contrast to the preceding editions in the series, Herbert Arthur Evans's 40-page introduction to his Henry 5 quickly and deeply immersed its reader in consideration of the textual situation (Shakespeare 1903c). Evans called Thomas Millington and John Busby "two piratical booksellers [who] had succeeded in getting hold of a garbled version" to make the quarto of 1600 that is "a very imperfect and clumsy representation of the text of the play as curtailed for some particular performance" (p. xii). Evans thought that the Stationers' Register order of 4 August 1600 that printing of Henry 5 was "to be staied" () was at the players' request--but who would care about them in Stationers' Hall?--and he skated over the problem that the register 10 days later recorded that the play had already been, or was being, printed by Thomas Pavier (). This 10-day gap convinced Peter W. M. Blayney, in an unpublished argument summarized by Cyndia Susan Clegg, that the staying order referring to Henry 5 was actually directed at 2 Henry 4 (Clegg 1999, 478-79). To Evans it was simply that Busby and Middleton made Pavier a front for their dishonesty (p. xiv).
Evans devoted an entire section of his introduction (pp. xvii-xxv) to the relationship between the quarto of 1600 and the Folio text of Henry 5. The theory of Alexander Pope and others that Q is Shakespeare's first stab at the play had been exploded, Evans reported, by P. A. Daniel in his introduction to the New Shakspere Society parallel text edition. Daniel showed that Q has phrases in 1.2--"Hugh Capet also" and "the foresaid Duke of Loraine"--with no grammatical antecedents, so unless a dramatist unaccountably wrote this nonsense and the necessary antecedents were added later to make the text underlying F (which is hard to believe), Q represents an imperfect shortening of what underlies F (Shakespeare 1877, xi-xii). Likewise, having only outlined Capet's case, Q has the bishop sum up with "So . . . King Pippins title and Hugh Capets claime, | King Charles his satisfaction all appeare", despite having said nothing about Pépin and Charles. This cannot be explained other than by omission and confusion, and Daniel's use of this evidence to show that Q represents a mangling of the play that underlies F anticipated Peter Alexander's demonstration by the same logic of the same relation between Q and F 2 Henry 6 nearly 50 years later (Alexander 1924).
Although he cited Daniel, Evans was not as knowledgeable a bibliographer, else he probably would not have claimed that Q prints prose as verse in order to waste paper and so make itself look large (p.xviii). Evans thought that Q is short mostly because based on a short performance at court or on the public stage (p. xviii), and that it was made by Middleton and Busby sending "a shorthand writer to the theatre" (p. xx). (To reconcile those last two statements, we must of course exclude a court performance since no-one would risk surreptitious stenography in the presence of the queen.) Evans held Q in contempt but thought its readings useful for emending where F has blundered or where it could support conjectural emendations already generally accepted into the play (p. xxi). Unusually for this series so far, Evans ended his introduction with a note making clear the textual basis of the edition, which was the Folio text modernised in spelling and punctuation and "not . . . departed from without reason" (p. xlvii). "For any apparatus criticus of the text", he wrote, "the Cambridge Shakespeare remains the fountainhead . . ." (xlvii-xlviii).
Reminding us of the limited resources available to an editor 100 years ago, Evans mentioned how valuable it was to have the first volumes of the ongoing NED, lamenting that he "only had the advantage of its company through the first half of the alphabet" (p. xlviii). Looking through his collation, it is clear that Evans often (on average once in 10 lines) adopted a preceding editor's emendation, but adopted a Q reading only three times in the first act and always where the editorial authority of Pope or Nicholas Rowe had preceded him to the adoption. Thus, he was sticking to F as much as he could--he was conservative not eclectic--and where he departed from it he wanted an illustrious predecessor's support.
Finally for this snapshot of one group of Shakespeare editors at the start of the last century, there is H. C. Hart's edition of Othello, the introduction to which ran to 36 pages (Shakespeare 1903b). Hart declared F better than Q on the opening page (p. ix), and then laid out the textual situation most carefully. Showing remarkable assiduity, Hart alerted the reader that press variants between examples of Q1 would be noted wherever Q1 was referred to in the critical notes (p. ix n. 2). Hart thought Q1 was printed from "an independent MS., which has been an early acting copy" whose non-deletion of oaths indicates its predating the act of 1606 (p. x). Because we have to trust the Folio preliminaries' claim about its contents' provenance ("true original copies"), F is our authority "unless the Quarto can establish a prior claim" (p. xi). Hart saw no such claim established.
F has 160 lines not in Q and it corrects many errors in Q, but Hart admitted that the reverse is also true in places (p. xii). Having done more collation than was strictly needed for his edition, Hart characterized this labour as "all to the good for an intimate knowledge of the text" (p. xiii). So, how come Q lacks those 160 lines? Since Shakespeare did not revise, Hart reported, they cannot be additions (over and above the play underlying Q) that created the play that underlies F. Hart enjoyed using phrases from the play to characterize the textual situation: Thomas Walkley published Q1 because he knew F was coming and he saw a chance of "putting a little money in his purse" (p. xii). The superb F lines that are not in Q cannot be intentional cuts: "There never could have been an Indian base enough to throw such pearls away intentionally . . ." so perhaps here F is showing "later insertions" (p. xiv). This appears to contradict Hart's assertion that Shakespeare did not revise, although if we confine the scope to just a few 'pearls' that were added (that is, not a wholesale insertion accounting for the entire set of 160 lines) then the contradiction is not great. Another seeming contradiction is Hart's assertion that in the Folio text of Othello "There is no reason to assert in any place that the reading is either certainly corrupt, or undoubtedly spurious, as there is in many of Shakespeare's other plays" (p. xiv). Yet he had already conceded that on occasion Q is right and F is wrong (that is, in his terms, Q 'corrects' F), and one senses here a reluctance to declare F indubitably faulty. Hart was avowedly conservative: "In this edition the text is practically that of the Globe, keeping the Folio reading rigidly paramount where there was left an open question" (p. xvi).
*
Extending this survey by one year would have brought in another four plays published in 1904, and extending it a further year to 1905 would have brought in another four plays again, making a total of 16 plays or nearly half the canon and thus providing useful contrasts to the editorial work sketched above. But to keep to length I will end here with a few tentatively conclusive remarks. I have excluded H. C. Hart's edition of Measure for Measure that Murphy dates to 1903 (Murphy 2003, 367) because the copy I examined claims first publication in 1905, beyond my scope here. Even from these 8 editions, it is clear that the Arden Shakespeare did not anticipate the New Bibliographical principles that were about to emerge: there is no sense on this evidence that these were principles whose time had come.
Two of the six editors surveyed, Morton Luce (The Tempest 1902) and Michael Macmillan (Julius Caesar 1902) wrote nothing about the textual situation they encountered, and extending the survey would have brought in others who likewise confined themselves to themes, characterizations, the plays' time-schemes (a now-neglected concern that all the editions surveyed took up enthusiastically), and the sources of Shakespeare's inspiration for his writing. The other four editors, Edward Dowden (Hamlet 1899, Romeo and Juliet 1900, Cymbeline 1903), W. J. Craig (King Lear 1901), Herbert Arthur Evans (Henry 5 1903), and H. C. Hart (Othello 1903) were about equally conservative in relation to the editorial tradition, eclectic and unsystematic in their selection of variants from multiple early printings where available, and unlikely to make an emendation that could not be supported from at least one of the early printings (irrespective of their view of its relative authority).
Repeatedly these editors asserted the superiority of one of the early printings over the others without indicating the bases for their views. It is hard to tell if this is because they did not think the reader could bear the weight of minute detail needed to establish the facts, or because they felt entitled to be trusted in such specialist matters, or because they did not themselves know why they felt that particular texts were superior. I suspect the last of these is the correct explanation, but even a wider survey is unlikely to reveal their motives. Looking at David Nichol Smith's personal papers about the Arden Shakespeare 1905-10 in the Beinecke Library (as I intend) may throw light on this. Only series editor W. J. Craig displayed the kind of sharply discriminatory thinking--such as placing weight on coincidence-in-error when determining the King Lear texts' genealogies--that came to characterize the New Bibliography, and he took up a matter of importance that later editors have neglected, the appearance of Buc's name in the Stationers' Register. Revisiting these editions from around 100 years ago it is clear that certain principles that are now widely discredited were held as unremarkable--that piracy was rife amongst publishers, that actors' interpolations could easily get into the script--and yet they also considered unremarkable other ideas that now meet hostility from some quarters, such as Shakespeare routinely revising his plays and (although I have not had space to go into it here) Shakespeare routinely writing collaboratively.
Notes