(1) The Text of 1 The Honest Whore

General

The play that later came to be known as Part 1 of The Honest Whore originally appeared in two quarto versions. The first quarto edition (Q1) has the following information on its title-page:

THE/ Honest Whore,/ With,/ The Humours of the Patient Man,/ and the Longing Wife./ Tho: Dekker./ LONDON/ Printed by V.S. for Iohn Hodgets, and are to/ be solde at his shop in Paules/church-yard 1604.

The title-page of the second quarto edition (Q2) is interestingly different:

THE/ Converted Curtezan/ With,/ The Humours of the Patient Man,/ and the Longing Wife./ Tho: Dekker./ LONDON/ Printed by V.S. and are to be sold by Iohn/ Hodgets at his shoppe in Paules/church-yard 1604.

Looking at the first three words only, one would gain the impression that these two title-pages refer to two significantly different plays, even if each contains a patient man and a longing wife. And this reaction is correct to the extent that the second title-page was indeed intended to indicate that the play it referred to was either a different play or one so substantially revised that it circumvented restrictions imposed by the body supervising printings, viz. the Stationers’ Company, which, since 1587, had limited edition sizes to between 1250 and 1500 copies. The printers of the two plays evidently from the outset planned to sell more than 1500 copies, and resorted to some deft manoeuvres to enable them to do so. The changed title-page of Q2 was part of the general camouflage. They also prepared new type-pages (i.e. ‘re-set’ ones) for slightly less than half of the new edition, while the other Q2 pages were simply impressed from type-pages composed for the first edition (i.e. left ‘standing’). This way, the printers significantly reduced the cost of production for the second edition, and at the same time, by using a significant number of re-set pages, made the second quarto look independent from the first.

Most important of all, Q2 contains a large number of important, meaningful new readings, which can be assumed to be the result of authorial revision undertaken by Dekker for the purpose of making Q2, no longer called The Honest Whore but The Converted Courtezan, look like a freshly published new book. The first person to discuss the importance of this second quarto since its appearance in 1604 was Matthew Baird, who in ‘The Early Editions of Thomas Dekker’s The Converted Courtezan or The Honest Whore, Part I’ (The Library, Fourth Series, X, 1930, 52-60) demonstrated extensively that this text is vastly superior (even in matters like punctuation) to that of Q1. He also had not the slightest doubt that it was Dekker, as author, who was responsible for the revision, and argued his case eminently. He concluded by saying that the importance of Q2 as an author’s revised text had not hitherto been recognised, adding that ‘There is therefore room for an edition of this excellent play which will embody the author’s text’ (60). Despite this plea such an edition has not appeared, and I profoundly hope that my work will, however belatedly, do justice to Baird’s wish. His important pioneering article was followed by a paper by the person still seen by many as the greatest bibliographer in the English-speaking world, W. W. Greg, ‘“The Honest Whore” or “The Converted Courtezan”’ (The Library, Fourth Series, XV, 1935, 54-60), who said about Q2 that ‘it is remarkable for a considerable number of corrections and alterations of greater or lesser importance, which there can be no reasonable doubt are to be ascribed to the author’ (55), and he deplored the fact that it was Q1, not Q2, which was reprinted by Nicholas Okes in 1615, and ‘so established to this day an inferior text [Q1] of Dekker’s most successful play’ (60). There was thus no doubt among these distinguished early scholars that (a) Q2 is, throughout, a greatly improved revision of Q1, of strikingly superior status, and (b) that the reviser was Dekker himself (with, it is implied, no-one else either likely or needed).

5The revision was, we can safely conclude, part of the creation of the supposedly new play (or at least edition) called ‘The Converted Courtezan’. Although we cannot altogether exclude that some skilled person in the printing house may be responsible for one or two good readings, those variants would not have been part of the general plan whereby Dekker was to revise his own work so as to make it appear something new. Hence, I shall in practice impute – in tune with Baird and Greg – all significant and superior new readings in Q2 to Dekker rather than an in-house editor. In general, Q2’s text is (despite misprints of its own) vastly superior to Q1’s, which included many errors. As for a number of unambiguously competent readings which two editors, Bowers and Mulholland, suggest may in theory have been the work of one or more persons other than Dekker: those scholars admit such variants, if they are improvements over Q1 readings, readily into their texts, despite their theoretical reservations. We shall see in what follows that in fact these readings are of the same quality as those which these editors do accept as authoritative; that there is no good reason for not seeing them as authoritative; and that Bowers and Mulholland offer no specific or clear suggestion as to who else than the author was responsible for them.[1]

In what follows I shall consider the nature of, the interrelationship between, and the relative authority of, the two earliest quartos offering us the text of Part 1 of The Honest Whore: (a) ‘The Honest Whore’, the first of the two quartos to appear, and (b) ‘The Converted Courtezan’.

First, however, I should explain that, several years ago, Antony Telford Moore and I published a lengthy article called ‘Breaking the Rules: Editorial Problems in Middleton and Dekker’s The Honest Whore, Part 1,’ Bibliographical Society of Australia and New Zealand Bulletin, Vol. 20, No. 4 (1996), 243-87 (henceforth referred to as ‘Daalder-Moore’). The paper contains much material that at this point in the edition cannot be reproduced or summarised but which, I am confident, still remains of value, and which is readily accessible to any interested reader.[2] Even so, the article is very long and technical, and of course more work – by several scholars – has, since 1996, been done on Q1 and Q2. I have also changed my own mind about some things. Thus I have corrected, updated and adapted the longer piece for publication in this edition. Responsibility for differences between the original article (much of which was written by Moore as my highly skilled research assistant) and the present statement is entirely mine. There are changes of some substance: I am persuaded that – contrary to what Moore and I believed at the time – most likely the printers involved are those identified by Adrian Weiss (discussed later); and, proportionately, I here present more material explicitly devoted to a comparison of the specific merits and weaknesses of the two quartos containing the text of 1 The Honest Whore. However, the longer article remains a good source for consulting detailed evidence supplied in it, to which I shall, in this essay, on several occasions refer. I would go further, and point out that the present discussion should not, by anyone seriously interested in the two quartos for specialised study, be seen as simply a substitute for the earlier paper, which, though some matters have needed updating, in certain areas remains more comprehensive than what I offer here.

The current version is more specifically concerned with textual/editorial matters as these affect readers of the plays (including of course actors, directors, etc.) than with the full detail and complexity of bibliographical issues, but those are certainly not ignored. It is the duty of an editor to try and provide some of the most important information available on these matters, which I do in this edition. Some of the main textual/editorial problems simply cannot be understood without exploration and knowledge of a number of basic bibliographical facts.

We commence with some consideration of how the text of 1 The Honest Whore, Part 1 – or rather the two different yet similar 1604 texts of that play which we must pay attention to – came into being: The Honest Whore (Q1), as it is commonly known, and The Converted Courtezan (Q2), which despite its different title is in effect a version of the same play, but which includes a significant number of authorial corrections and revisions, as well as some new printer’s errors of its own.

The First Quarto (The Honest Whore)

10At some time between 1 January and 14 March 1604, the theatre manager Philip Henslowe, acting in his capacity as manager of Prince Henry’s Men at the Fortune Theatre (where the play was to be performed), recorded a payment of five pounds to

Thomas deckers & Midelton in earneste of ther playe Called the pasyent man & the onest hore.[3]

The sum Henslowe paid was an advance, but it is possible enough that the play appeared on stage in the Fortune as early as April, when the theatres were reopened after what had been a lengthy and serious outbreak of the plague. At any rate the play referred to, now known as Part 1 of The Honest Whore, must have been completed, at the latest, by 9 November of the same year,[4] for on that date it was entered in the Register of the Stationers’ Company by Thomas Man, Jr:

Entred for his copye vnder the hand of mr Pasfeild A Booke called. The humours of the patient man. The longinge wyfe and the honest whore.[5]

Printing probably followed soon after, since the title-page of the first quarto edition (Q1) is dated 1604:

THE/ Honest Whore,/ With,/ The Humours of the Patient Man,/ and the Longing Wife./ Tho: Dekker./ LONDON/ Printed by V.S. for Iohn Hodgets, and are to/ be solde at his shop in Paules/church-yard 1604.[6]

The play was printed within a year of its composition – a reflection, perhaps, of its popularity. It would appear from the title-page of Q1 that Henslowe’s entry recorded the play’s title in abbreviated form. It is interesting that the title of the printed quarto did not match that used in the Stationers’ Register. I would surmise that the title in the Register is the one that Dekker had himself used originally, but that the play had during the time it was acted come to be known as ‘The Honest Whore’, so that it was decided to use that title for the first printed version. Moreover, the snappier and ‘sexier’ title was a much better commercial choice than the longer somewhat incoherent one that appeared in the Register.[7]

Although the title-page might suggest that John Hodgets was the publisher, in fact he appears to have been nothing other than the bookseller, and the publisher was Thomas Man, as is also strongly suggested by the title-page of The Converted Courtezan (Q2, 1604).[8] Only four copies of Q1 are known to have survived.

15In physical appearance, Q1 is a perfectly ordinary play-quarto of the period. It has 40 unnumbered leaves (80 pages), collating A-K4, the title-page (with its verso blank) being part of gathering A. Eight press-corrections have been found in Q1: three in the Outer Forme of sheet C, two in Inner G, and three in Outer K.[9]

The title-page of Q1 mentions only one printer: V.S., or Valentine Simmes. In Daalder-Moore we listed, for Q1, Simmes as the printer of the first two sheets of the play, A and B; we regarded another printer, possibly John Windet, as responsible for sheets C and D; and we judged sheets E-K to have been printed by Thomas Purfoot. Reasons for our choice of these printers can of course be found in Daalder-Moore.

We also said quite a bit about the method of printing used. Our key observation was: ‘One would perhaps expect that the text of a work that was to be shared by two or more printers would be cast off and set by formes. Such a method would not only allow the printers to determine their exact share of the manuscript and the resulting printed work; it would also enable the formes to be composed and printed in any order – an especially important consideration in a venture involving independent printing houses. There are indeed strong indications that at least five of Q1’s ten sheets were set by formes’ (246). This comment was followed by a good many more detailed remarks which describe how the text was printed, and I refer the reader who is interested in these matters to those pages (246ff.) in Daalder-Moore.

Our view on who the printers were was not eccentric or odd: on the contrary, it had in essence been widely shared by others, notably eminent bibliographers such as W.W. Greg, Fredson Bowers, and W. Craig Ferguson, and it had been based on a careful examination of actual physical evidence. But I now feel that most probably Adrian Weiss, although unnecessarily harsh in rejecting the work done on the printers by all his predecessors, is right in holding that there were four (not three) printers involved, and that they were Valentine Simmes (section A-B), Thomas Creede (C-D), Simon Stafford (E-F) and George Eld (G-K). Furthermore, Weiss seems to me persuasive in arguing that the same four printers were responsible for printing the second quarto (Q2). Hence I refer the reader to Dr Weiss’s recent critique, in the Oxford Middleton, of the Daalder-Moore view of the printers, and the articles written by him which he mentions: he provides strong evidence for his case in arguing against that proposed in Daalder-Moore. Moore and I did our best for our 1996 article, but it is in the nature of even the most careful scholarship that some of it will be corrected by later scholars, and I am happy to accept Weiss’s conclusions as to who printed Q1 and Q2. For further development of Weiss’s views and more information on printing issues, Paul Mulholland is worth consulting, although his comments are extraordinarily lengthy and at times seem to me rather speculative.[10]

The Second Quarto (The Converted Courtezan)

The first quarto of the play was followed by a second, Q2, before the year’s end, as Q2’s title-page shows: ‘THE/ Converted Curtezan/ With,/ The Humours of the Patient Man,/ and the Longing Wife./ Tho: Dekker./ LONDON/ Printed by V.S. and are to be sold by Iohn/ Hodgets at his shoppe in Paules/ church-yard 1604.’ There are two interesting changes here: (a) ‘THE/ Honest Whore’ has been altered into ‘THE/ Converted Curtezan’, and (b) Q2 makes plain that the book is to be sold by (not for) John Hodget (i.e. it is clearer, now, that he is a bookseller only). There is no doubt that Q2 was the later of the two editions: for example, it reproduces the corrected forms of all eight of Q1’s press corrections. However, of much greater importance is the considered and generally authorial character of many of Q2’s corrections of the Q1 text – a fascinating fact which gives this book a rare (though not unique) status. Q2 has the same collation as Q1. Only one of the three surviving copies is complete: that in the Bute Collection at the National Library of Scotland (in Edinburgh).

20A very striking fact is that a little over half of Q2’s pages were printed using ‘standing-type’ from Q1. In other words, after the printing of Q1, many of its type-pages were retained in their assembled form, imposed with new running-titles (except for E Inner and E Outer in the Edinburgh copy, which appear to retain their Q1 running-titles) and re-used for the printing of Q2. In some instances entire formes were retained as such for re-printing. In other instances the formes were partly broken up and re-imposed with a mixture of standing-type and reset pages. Still other formes were completely reset. The details of resetting and reimposition are presented in the table below, where ‘R’ refers to reset pages, ‘S’ to standing-type pages, and ‘S/R’ to pages of mixed character.

Standing-type and reset pages in Q2 of 1 The Honest Whore

A2r-4v Standing (Title-page S/R)
B Inner R
B Outer R
C Inner 1v R; 2r, 4r S; 3v S/R
C Outer R
D Inner S
D Outer S
E Inner S
E Outer R
F Inner S
F Outer 1r R; 2v, 3r, 4v S
G Inner R
G Outer R
H Inner S
H Outer R
I Inner S
I Outer R
K Inner S
K Outer 1r, 4v S; 2v, 3r R

As can be seen from this table, seven of the ten inner formes (including A Inner) were entirely retained in standing type, while the other three were partly retained. Most of the outer formes were completely reset. Each of the standing-type formes must have been unlocked in order to change the headlines and/or to effect alterations in the text of the play. Even the standing-type forme E Inner of the Edinburgh copy, where the Q1 headlines were retained, must have been unlocked for the purpose of substantial alterations on each of its pages – on E1v, E2r, E3v, and E4r. Thus, although eleven of Q2’s 41 standing-type pages – A2r, 2v, 3v, 4v, F1v, 2r, 2v, 3v, 4v, I2r, K1v – reveal no textual alteration, not a single standing-type page in Q2 was reimpressed without modification of some kind.

This revised copy of 1 The Honest Whore (Q2) was printed, as Weiss has shown in various ways, by the four printers who had also been responsible for the first copy (Q1). Much speculation based on the earlier belief (which had also been expressed in Daalder-Moore, 251-2) that the printers were not the same in number, or different persons, can now probably (if Weiss’s views are generally accepted) be rejected as misguided, and would probably only be embarked upon again if the identity of the printers once again became a matter of dispute.

But, although the printers of the two quartos were the same, the text of the second edition (The Converted Courtezan) differs from that of Q1 in numerous ways. There are more than 500 individual instances of variation. A great many are not recorded by Bowers[11] or anyone else, but are presented in the ‘Appendix’ of Daalder-Moore. They range from fairly inconsequential differences in spelling to significant changes in wording and punctuation, and even, in a few places, rephrasing of entire passages. We attempted, in Daalder-Moore, to list all of the Q1 and Q2 variants, and indicated whether or not they are in standing type (a distinction also made in the collation for this edition). Anyone interested in knowing all the different variants in Q1 and Q2 will find the Daalder-Moore list useful, though of course it is best is to consult the actual quartos. An advantage of the existence of the Daalder-Moore Appendix is also that it avoided the need for me to re-list in this edition huge numbers of variants beyond those called for.

There are four unusual features in Q2: its use of standing type, its division amongst four different printing shops, its many textual divergences from Q1, and its new title. It is very likely that all four features are related. Consider, first of all, Q2’s standing type. In normal circumstances, type would be distributed soon after it had been printed, and ‘would not be kept standing for any book except by a plan conceived before or very shortly after the printing of the first gathering’.[12] It is a reasonable guess, then, that Simmes and his cohorts (including, probably, the likely publisher, Thomas Man) intended a second edition of 1 The Honest Whore from the outset, perhaps because they expected larger than usual sales. Bowers, building on this possibility, conjectures that the printers decided to circumvent the Stationers’ Company restrictions which, since 1587, had limited edition sizes to between 1250 and 1500 copies.[13] With this purpose in mind, they prepared new type-pages for nearly half of the new edition, but made up the other half with standing-type pages from Q1.[14] In this way they limited the cost of fresh composition while simultaneously camouflaging the fact that more than half of Q2’s sheets were impressed from type-pages composed for the first edition. Such an unusual arrangement would presumably have required some form of private agreement with the compositors, whose employment prospects were protected by the very restrictions now being evaded. In all probability the alternative title was part of the general camouflage, in which case – and pace W. W. Greg – it should not be thought of as a ‘better’ title than The Honest Whore, and it has no real authority.[15] The contradictory imprints of Q1 and Q2, which give the impression that John Hodgets was the publisher of the first edition but not the second, may also have been part of the effort to cover tracks. At no point, either, is there any mention of the person who was probably the true holder of copyright, Thomas Man. This was a prudent move, perhaps, considering that Man's father, Thomas Man Senior, was at this time Master of the Company of Stationers.

25Significantly, the second quarto of 1 The Honest Whore is not the only second edition bearing Valentine Simmes’s imprint to contain large amounts of standing-type. Simmes’s second edition of The Malcontent (1604) is also only partly reset. So too is Edward Allde’s edition of Dekker’s own The Whole Magnificent Entertainment (Q2, 1604) and Thomas Purfoot’s edition of John Marston’s Parasitaster, or The Fawn (Q2, 1606).[16] Bowers suggests that these four editions represent a short-lived endeavour on the part of a small group of printers, publishers and, indeed, authors, to get by the Stationers’ Company restrictions. Amongst this group we must now include, if Weiss is right about the printers of Q1 and Q2, the names of three men not previously considered in this context, viz. Thomas Creede for section C-D, Simon Stafford for E-F, and George Eld for G-K. The status of Valentine Simmes (A-B) as a culprit of course remains unaffected by Weiss’s findings.[17]

The reason why the printers of Q1 and Q2 of 1 The Honest Whore collaborated was no doubt primarily that they were in a hurry to get their work finished as quickly as possible and thus to increase their profit. Also, the more expeditiously the work was done, the smaller the chance of discovery that they were breaking the law concerning edition sizes by practising deception. As well, their resources were limited, and needed to be employed effectively. For example: 43 out of 79 pages for Q2 are reimpressed. Retaining this number of pages in standing type would have severely constrained the use of type.

The key point is no doubt that the printers involved in work on 1 The Honest Whore were keen to make money from larger than usual sales. They retained slightly over half of Q1 type-pages in standing type, with the intention of producing a second edition. But they were naturally eager to create the impression that Q2 was an entirely – or at least substantially – new edition, so, besides disguising the old work with newly reset pages, they evidently asked Dekker to provide them with corrections and alterations to the text. This explains why so many of the new readings found in Q2 appear to bear an authorial stamp. The other editions featuring large proportions of standing type mentioned earlier – Simmes’s edition of The Malcontent, Allde’s edition of The Whole Magnificent Entertainment, and Purfoot’s edition of Parasitaster – also contain authorial corrections and revisions, which supports the conjecture that the authorial changes to the second text of 1 The Honest Whore were regarded (by the printers, at least) as part of the camouflage.

The basic reason, then, why Dekker undertook to revise Q1 is that this activity was part of an entrepreneurial effort to print more copies of 1 The Honest Whore than were legally allowed by making it appear that a very different publication, The Converted Courtezan, was being published. Given the opportunity to revise what had appeared in Q1, Dekker proceeded – as my analysis of detailed evidence will show – to correct a great many errors made in that printed version of his play, and in general to improve Q1’s text significantly. As much as his circumstances allowed, Dekker chose to take advantage of the situation to substantially emend, and at times even to rewrite, passages at many points in the play. Even those editors (Bowers and Mulholland) who are theoretically wary about the reset material in Q2 still admit – though often without claiming them specifically as Dekker’s – most of the significant Q2 readings into their text. As a result, it can be categorically stated that the text of the play is, in every modern edition, much closer to Q2 than to Q1.

Q1 (The Honest Whore) and Q2 (The Converted Courtezan) are the only important primary sources for an editor of 1 The Honest Whore. There is a third quarto of 1 The Honest Whore, bearing the imprint of Valentine Simmes, which appeared in 1605 (Q3). This edition, which survives in a single copy in the Dyce collection, was printed wholly from standing-type pages (some uncorrected) of the two earlier quartos. It offers nothing new to take account of and may represent a further surreptitious effort to capitalise on interest in Dekker and Middleton's play while keeping additional expenses to a minimum. A fourth quarto of the play (Q4) appeared, probably in late 1615, and a fifth edition, Q5, was published in 1635. Q4 and Q5, both printed by Nicholas Okes, have no textual authority, for Q4 is a straight reprint of Q1, and Q5 a reprint of Q4. The best these quartos can manage are a few correct emendations of Q1-2 material.

30Thus an editor trying to determine the most reliable early text of 1 The Honest Whore is obliged to concentrate on Q1 and Q2. The main formal question to be addressed is: which of these two texts, with their shared pages of standing type, their shared printing among four different printing shops, and their hundreds of minor and major variants, is most likely to provide the more accurate and authoritative version of the play? In order to answer this question, we must first try to determine the provenance of the texts preserved in Q1 and Q2.

Printer’s Copy for Q1

The nature of the copy manuscript used for printing Q1 is extensively discussed in Daalder-Moore. The section is quite self-contained and clear, and for extensive information and analysis I refer the reader to it (256-60). Here I mention only the main conclusions drawn. Bowers argues that Q1 was set up from authorial foul papers or a transcript of them (Introduction, 2-3). While this possibility cannot be ruled out, it does not give an altogether accurate impression of the text set forth in Q1. There is certainly not, for example, a large number of obvious errors such as might be present if the printing had been based on a shoddily prepared manuscript containing them. Such stage directions as are supplied in Q1 are on the whole helpful and to the point. In a modern edition, a great many stage directions do need to be added, but the quality and quantity of those which I found in Q1-2 seemed to me quite representative of what one might find in any printed play of the period based on a reasonably well-prepared manuscript. At the same time, there are no signs either that the copy manuscript was anything like a prompt-book: the manuscript used apparently provided, within its limits, an adequate text for the theatre, but still needed a good deal of additional work. Thus, all in all, it may have been an authorial or scribal fair copy which was yet to be adapted further for theatrical use. Many of the (rather large) number of errors in Q1 seem due to compositorial error rather than badly prepared copy for the press. In many cases it is not difficult to reconstruct what the author(s) in all probability originally wrote, though in the case of several names only Q2 can help us to establish the correct words.

Printer’s Copy for Q2

The question of printer’s copy for Q2 naturally focuses on (a) the alterations in the standing type pages from Q1, and (b) the reset pages of Q2, with their many divergences from the text of Q1. To begin with, it is beyond doubt that Q2’s reset pages were set up from a copy of Q1 rather than the original manuscript, because the reset pages reproduce a great many of the details of the layout, lineation and typography of the corresponding pages in Q1. The most straightforward explanation of Q2’s authoritative variants is that they were entered in this same copy of the first quarto before it was used as printer’s copy for the second edition.

However, Bowers – still widely seen as the true textual/bibliographical expert on early Dekker material – contends that authoritative alterations in the reset pages of Q2 are almost entirely limited to sheet B: ‘The fact to be faced is that authoritative alterations appear in standing-type pages [and reset sheet B] but not (except for a few very doubtful cases) in pages subsequently reset’ (Introduction, 13). If this view of what is authoritative or not were correct, obviously we would have to accept that Dekker’s changes were not entered in one particular copy of Q1 except on a curiously selective basis: the printers would, in order to incorporate his revisions, have used standing-type pages and sheet B, but nothing (or very little) else, which would then mean that – very implausibly – Dekker provided his revised material accordingly. But what Bowers proposes is based on an entirely false supposition and a bald, unsubstantiated assertion: it is not a ‘fact’ that ‘authoritative alterations appear in standing-type pages [and reset sheet B] but not ... in pages subsequently reset’. Any careful comparison of the Q1 and Q2 variants, of which I shall provide examples below and many others of which are, additionally, discussed in Daalder-Moore, thoroughly refutes that categorical claim. If it is agreed that variants occur, outside the limits which Bowers imposes, which are often likely to be authoritative, then the remarkable supposition of the printers being able to make use of the amended copy of Q1 for Q2’s standing-type pages (and reset sheet B) but not for the bulk of its reset pages can safely be rejected. Indeed, if we reject Bowers’s categorical claim it at once seems likely that Dekker used a particular copy of Q1 for his changes throughout the text. Alternatively it is conceivable that Dekker, in revising, annotated surplus sheets from Q1 and handed these to the printers, although that procedure would probably have proved cumbersome.

As Bowers himself argued that reset sheet B and (possibly) K3r of Q2 contain authoritative alterations (Introduction, 11-15), he would in theory have had to accept (though he did not do so in his pronouncements) that there was at least a possibility of other reset pages also containing authoritative changes. Contra Bowers’s adamant assertions, an editor convinced of the authority of Q2 variants in both standing-type and the reset pages mentioned by Bowers (sheet B and possibly K3r) would be willing to consider that there at least may be other such variants on other reset pages, and could thus proceed on the simple assumption that a marked-up copy of Q1 was used as the basis for alterations throughout the second quarto.

35Once one accepts that Dekker at least could have supplied, for the benefit of the printers, corrections on more reset pages than Bowers is prepared to concede, the most likely scenario would seem to be that the printers were able to keep a marked-up copy of Q1 in their shop while the standing-type pages were being corrected and the reset pages composed. In support of this scenario, I would point out that John Marston may have used the same method in revising his play, Parasitaster, or The Fawn: it is probably a copy of the first quarto of the play that Marston refers to when, in the Preface to the second quarto of Parasitaster (1606), he claims to have ‘perused this copy, to make some satisfaction for the first faulty impression’.[18]

Now we must address the question how convincing Q2’s variants are, and how many authoritative variants occur in the reset pages dismissed by Bowers.

The Authority of Q2

We shall start by briefly indicating Bowers’s reasons for regarding at least some of Q2’s text as authoritative when compared with Q1, and I do this by quoting him from pages 10-11 of his edition, as I think that passage is telling and important:

Many of the corrections are of minor matters not beyond the competence of an editor to alter; but some are certainly more scrupulous and would seem to require authority, such as the alteration of Q1 Malauella to Malauolta at II.i.91 [in Bowers’s text], and the rewriting of II.i.223 but especially of 300-302. Such a small but interesting matter as the correction of Bellafront’s Q1 What to Whaat at II.i.38 speaks for a considerable intimacy with the text (see her shaall at line 219), and there may be significance in the extremely minor change of Q1 dost at I.v.228 to the spelling doest, which is that of Dekker’s addition to Sir Thomas More though scarcely unique with him. The occasional relining of prose as verse argues for authorial care, and there would seem to be little question that the alterations stem from some authority, which was presumably Dekker. This being so, an editor is bound to accept them in all cases unless, in specific lines, there is some strong presumption that another agency was operating as well.

I find this a very impressive statement. It shows Bowers in his capacity of a precise, observant and sensitive literary critic, and I consider it a great pity that he did not consider Q2 in these terms throughout, as he then would probably have produced an excellent statement in favour of the position that Q2 is, elsewhere too, a much better text than Q1, containing many just such alterations as Bowers argues here should be accepted by an editor. However, for reasons which even after many years of study I cannot comprehend, Bowers came to claim that such alterations as he considers in this passage are limited to just some portions of the text, though I am firmly convinced that any competent committee of literary scholars would concede that equally compelling alterations are actually typical of Q2 as a whole. My purpose in much that remains in my subsequent comments on the text of Q2 in comparison with that of Q1 is to show that the superiority of Q2 is pervasive, and not just localised in such limited areas as Bowers chooses on the one hand while he – arbitrarily, I believe – on the other hand declares other areas to be systematically unauthoritative.

One of Bowers’s main procedural problems is that the passage which I have just quoted from his pages 10-11 shows that his criteria for accepting the readings which he approves of are those of a literary critic. There is no supposedly ‘scientific’ bibliographic reason for his choice in favour of these readings as against others. In essence this means that it is entirely reasonable for any literary scholar to argue that other readings elsewhere in Q2 may be just as authoritative, unless Bowers produced a scientific and binding reason as to why they cannot be. This he fails to do.

40Any objective, sober analysis of Q2 undertaken on the basis of such a literary judgement as Bowers displays in the passage I have cited would no doubt – as Moore and I as well as several readers of our 1996 article have done – come to the conclusion that there is no in-principle difference between alterations in standing pages and sheet B on the one hand, and the remaining material on the other. It is worth adding here that both Matthew Baird and W. W. Greg, whom I have mentioned before as analysts of the differences between Q1 and Q2, clearly considered Q1’s text to have been revised by Dekker throughout. Anyhow, that is the case I wish to argue, and to persuade my readers of. And, in doing so, I shall not even confine my considerations to literary judgements only.

My essential position is that Q2 is a far more authoritative text than Bowers is prepared to allow in rejecting most of the reset pages. A first question from a bibliographical point of view is whether, in Q1 or Q2, any compositors stand out immediately as having remarkably individualised features, either of a negative or positive kind, and from whose work we might thus draw some conclusions as to what Dekker wrote. There is one compositor (generally known as ‘Compositor A’) who contributed to Q1 and whose typical habits are known, thus probably enabling us to draw important comparisons between what, on the one hand, he produced in Q1 and on the other hand its counterpart in Q2. The question would be just how the revised Q2 version might differ. Would the differences be unimportant, or, on the contrary, show evidence of intelligent, meaningful, and plausibly authorial revision?

I report on a compositor study of the section of the play (A-B) printed by Valentine Simmes, based on close comparison between Q1 and Q2. There is a reasonable amount of evidence to suggest that this entire section was set for Q1 by the workman known as Compositor A, a person responsible for part or all of the quartos of Doctor Faustus (1604), Hamlet (1603) and Richard II (1597). Most of Compositor A’s known traits – identified in studies by W. Craig Ferguson and Alan E. Craven[19] – can be seen in Simmes’s section of 1 The Honest Whore. For example, this section regularly features Compositor A’s most distinctive characteristic, the use of unabbreviated speech-headings without a final stop. These occur in abundance on all pages except B4r (one instance). Simmes’s section also displays another of Compositor A’s habits: the tendency to capitalise non-exit stage directions and place them in a central position (instances of this are found on A4r, A4v, B2r, B4r, B4v). Again, Compositor A’s practice of setting parenthetical expressions within rounded brackets is seen on seven of the section’s fourteen pages: A3v, A4v, B1r, B1v, B2r, B3v, B4r. And this compositor’s slight penchant for setting normally-capitalised words without a capital letter is seen in the instances of ‘lord’ and ‘lordship’ on A3r and A3v, in ‘thurseday’ and ‘monday’ on A3r, in ‘flemmish’ on B1r, and in the numerous occurrences of ‘ile’ (rather than ‘Ile’ or ‘Ill’) on A3r, A3v, A4r, A4v, B4r, and B4v. Results from spelling tests have proven inconclusive,[20] but overall, Simmes’s section contains sufficient evidence of Compositor A’s work to make one feel reasonably certain of his presence in these pages.

One of Compositor A’s most distinctive (and worrying) practices is his habit of introducing corrupt readings which are difficult to discern.[21] Interestingly, this practice appears to have been picked up and corrected on a number of occasions in Q2. A typical example of this is Q1’s ‘Softly sweete Doctor:’ (TLN 321), a plausible phrase which is nevertheless improved on by Q2’s ‘Softly, see Doctor:’ (note the comma and the change from ‘sweete’ to ‘see’). A little later in Q1 (TLN 398), the Duke pictures Infelice hunting ‘like some gods in the Coprian groves’. Q2 makes the obvious but knowledgeable and intelligent correction of ‘Coprian’ to ‘Cyprian’; and it also changes ‘gods’ to ‘goddesse’ – a renovation which fits the image more exactly to Infelice, but which would probably not have occurred to anyone other than the author of the line. Other authoritative-looking corrections of inconspicuous, Compositor A-style errors occur in Simmes’s section at TLN 354 (‘deadst’ in Q1, which becomes ‘midst’ in Q2), TLN 355 (‘cap/cup’), TLN 373 (‘the/thy’), TLN 404 (‘it/her’), and perhaps TLN 390 (‘haunts/hurts’). Q2’s amendment of Q1’s ‘mine aunts’ to ‘my naunts’ (a dialectical variant at the time) at TLN 290 is probably a further example of such a correction. It would seem that Compositor A could deceive anyone except the writer whose words he so cunningly misconstrued.[22]

Thus, it is reasonable to conclude that it was Dekker himself who was responsible for the revised readings. I draw attention to the fact that I am not loosely generalising about the Q1 and Q2 differences in section A-B, but considering very specific examples, and giving my reasons for my firm belief that the author, in revising his text, was the person responsible for these alterations. It is difficult to envisage someone else thinking of all of these high-quality emendations: we should surely note (a) that many of them are very individualistic in nature, and (b) that Dekker was to be presented, on its title page, as the author of a supposedly new book (The Converted Courtezan) of which he was therefore the likely reviser, while no-one else was specifically required for the task.

45It may be objected that the examples here cited as corrections of mistakes typically made by Compositor A are unnecessary to my argument, as Bowers did, after all, concede the authority (indeed the authorial nature) of the changes in sheet B, where the variants just cited occur. But a good reason for accepting these alterations as authorial is that an examination of the evidence just considered reveals how their authoritative nature does not merely depend on Bowers’s declaration as a literary critic, but that there is, in this case, bibliographical evidence to support him, which, oddly, he did not mention: we can in this instance compare emendations in Q2 with the typical mistakes which a known compositor made in Q1, and, with rather more confidence than justified by literary judgement alone, explain to ourselves why the Q1 readings should be discarded, and those in Q2 be accepted as authorial.

There is another solid way of confirming that Bowers was right to have faith in Q2’s version of sheet B. We can do this by comparing the Q1 and Q2 versions and see how often not just the emendations in Q2 seem right, but how good the standard of its printing generally is. After all, it would theoretically be possible that Dekker proposed good alterations for sheet B, but that the Q2 printing nevertheless introduced a good many new compositorial mistakes of a potentially confusing nature, such as scholars like Bowers and Mulholland have feared. But if we turn to the Appendix in Daalder-Moore and collate sheet B in Q1 and Q2, we find that there would be no danger whatsoever in using Q2’s reset pages of this sheet as copy-text, for the reproduction of Q1 in Q2 is remarkably accurate, and the variants, when they occur, are often patently authoritative. Throughout the eight pages of sheet B there appear to be only fifteen Q2 variants in punctuation and capitalisation which would definitely have to be rejected as corrupt (i.e. the product of compositorial ineptitude), and which are easily recognised as such. These constitute a very small number, not least when compared with the many significant and high-quality improvements in the new text. This is not an argument Bowers himself seems to have thought of as a reason for valuing Q2 in B, yet it is another bibliographical (as distinct from purely literary) factor that he could have used to support his claims about the authorial/authoritative changes in sheet B, and the general reliability of that portion of Q2. It should, as well, have led him to ponder whether other reset pages in Q2 might not also have excellent status.

Here we come to much of the essence of Bowers’s case and its various faults (see his Introduction, especially 11-15). His key error, I think, is that he is just as dogmatic about accepting, on purely literary grounds, as ‘authoritative’ (or even ‘authorial’) Q2 variants in reset sheet B as he is about rejecting them elsewhere when they are also in reset type. This aspect of his case usually does not attract as much attention as it should. The fact remains, however, that in the case of sheet B his judgement is literary, whereas elsewhere he tries to make it bibliographical. The key difficulty which, this way, he creates for himself is that he is asking his readers to take him on trust with respect not only to the reset material in sheet B but also that which is reset elsewhere. Yet, as he accepts reset material in B on literary grounds, he ought to prove on either literary or bibliographical grounds why he rejects it elsewhere, but he does neither. Indeed, the illogicality of his position is that in practice he accepts many Q2 readings outside sheet B which in theory, going by his own supposedly ‘bibliographical’ argument, he should reject as tainted by compositorial (or at least ‘unauthoritative’) practice.

Bowers firmly, and rightly, accepts the reset B material in Q2. He also has confidence in material in standing type, though even he concedes ‘one or two minor exceptions which may be laid to the printing house’ (14). Personally I believe that this understates the inaccuracy of various readings in standing type. Admittedly, there is a good prima facie case for regarding standing-type alterations as more authoritative than those in reset pages, for the compositors had to unlock the type-pages specifically for these changes to be made; but of course that does not mean that all the changes made are necessarily correct. However, my main objection to Bowers’s thinking concerns the Q2 pages in reset type which he would in essence reject in toto as they may contain faulty, ‘unauthoritative’ material, while in principle this is just as true of sheet B itself, the authority of which he accepts.

In the event, Bowers’s actual literary discernment, as an editor, proves superior to his stated theory. An interesting example occurs in the case of sheet K. Even in his revised edition (1964, succeeding and correcting his first edition of 1955), Bowers rejected, officially, reset material ‘save for the substantive variants in sheet B and K3’ (15). As he did not have access to the Edinburgh copy of Q2 (the only complete one) while working on his first edition of Dekker’s Dramatic Works, he was at that time unaware of Q2’s major alterations in K3r, K3v and K4v. While working on his second edition, however, he was able to consult the Edinburgh copy. In each one of the variant passages contained in these pages – on reset as well as standing-type pages – he adopted the previously unseen Q2 variants for his revised text. Similarly, in the case of TLN 2193 it is interesting to note that in his first edition of Dekker’s Dramatic Works, Bowers, adhering to his theory that the variants in most of Q2’s reset pages lacked authority, adopted the corrupt Q1 reading ‘for’. In the second edition, however, the literary critic in him seems to have got the better of the bibliographer, and he substituted the Q2 reading, ‘far’, without producing any reason for doing so. Such improvements are welcome, from a literary viewpoint, but they show that in practice Bowers was not able to stick to his own theory.

50Why would that be? The simple answer is, of course, that the theory is faulty. If it had been correct, Bowers would not have felt compelled to accept numerous readings from Q2 which his own theory, if logically applied, would instruct him to reject. His attitude is one of extremism. He accepts reset material in sheet B and in K3 (Introduction, 15), but in principle categorically rejects it elsewhere. One can only guess that the rejection is based on fear that, when material was reset, printer’s errors were no doubt made which we might mistake for authorial improvements. But if that was the reason then, of course, there is no good case for making an exception for the material in sheet B and K3.

In general, especially if a reprint contains no authorial revision, it is certainly sensible, when it is based on already printed material, to be wary, notably of changes in ‘accidentals’ (some of which might be quite deceptive), but, of course, also of new ‘substantive’ errors, which could, without authorial approval, find their way into the text. However, Bowers admits, in Q2, a significant number of readings as authorial because they are part of standing-type pages, as well as others that occur in reset sheet B and in K3. It would have been logical for him to consider the possibility that an author making changes was not, and would not feel himself, restricted to these areas, so that therefore alterations could in principle occur anywhere else in the revised text (Q2). In other words, rather than close one’s mind to the possibility that readings resulting from resetting outside sheet B and K3 could be authorial, one should carefully weigh what are in fact likely to be competing claims: those of new readings which may be authorial and correct as well as those which could be errors caused by the printers.

The likelihood that at least a number of changes occurring in reset pages other than those selected by Bowers are authorial would, of course, increase if we can find, outside Bowers’s chosen territory, changes in reset material which are of high literary standing and which we would not expect a compositor, or even an in-house editor, to introduce. Let me try to give some examples of this type.

Q2 Variants from Reset Sheets other than B and K3 which Appear to be Authorial in Origin

• An interesting variant occurs in George’s response to a question from Candido’s wife, ‘What said he George when he pasde by thee?’ (G4r, Q1; TLN 1998). George’s answer reads in Q1: ‘Troth Mistris nothing’ (G4v). The same phrase on the reset page of Q2 is repunctuated as ‘Troth Mistris, nothing’ (TLN 1999). The change – which is not recorded by Bowers – is obviously likely to be correct, and typically one thought of by an author who was revising his text. So too are the following Q2 variants, all from reset pages in principle rejected by Bowers.• In the stage direction at the head of 4.2 (G3r), Q1 gives a character’s name incorrectly as ‘Poli’ (TLN 1915), while Q2 gives the correct form, ‘Poh’. In two other places where Q1 has the incorrect ‘Poli’, at TLN 2021 and TLN 2036, Q2 opts for the correct ‘Poh’ instead. Bowers points out that the correct form occurs also, in Q1, on G3v, surmising that as a result the Q2 corrections elsewhere may not be authorial. But, given the fact that in Q1 the presentation of this name had been inconsistent and we know in effect nothing about intelligent corrections made in Q2 by others than Dekker, it is much likelier that Dekker himself, when revising Q1’s text, corrected these mistakes (showing authorial consistency in the process). This supposition is the more likely as Dekker frequently had to correct mistakenly spelled, or mangled, names which appeared in Q1. There are, furthermore, several other authorial corrections of G3r readings in Q2: it is odd, therefore, to suppose that just this specific correction was someone else’s work. The situation is the same in the case of G4v.• At TLN 2193-94 (h2r), the Q1 version of the Doctor’s plea reads: ‘But be you pleas’d, backward thus for to looke, / That for your good, this evill I vndertooke’. Q2 corrects ‘for’ in TLN 2193 to ‘far’, which is surely unambiguously correct. Bowers tacitly seems to have conceded the superiority of ‘far’ when in 1964 he substituted that for his earlier ‘for’ (see my discussion above).• A little later on the same page, the Doctor speaks, in Q1, of his finger being ‘deept in blood’ (TLN 2201). Q2 changes ‘deept’ to ‘dipt’, a variant which seems preferable to the rather strained Q1 reading. (This important Q2 variant is also omitted from Bowers’s collation.)• The Doctor's reference to ‘mourning’ at TLN 2245 (h2v) is rendered incorrectly as ‘morning’ in Q1 but amended to ‘mourning’ in Q2. The correction was clearly judged to be necessary as it removed ambiguity: Q1’s ‘morning’ could (even in its context) refer to the early part of the day. It could also represent our word mourning, as OED shows that the verb mourn could c.1600 still be spelled as morn. Thus, a reader could not necessarily infer from seeing morning which meaning was intended. To remove this difficulty, the person who wished his meaning to be clear – and this must have been Dekker himself, as only he could actually know what he wanted to say – corrected morning into mourning. That did establish absolute clarity, as mourning was not, according to OED, a variant spelling for morning.

What is significant about the bulleted Q2 variants just listed above is that they all occur in reset pages other than sheet B or K3r – that is, in pages where Bowers doubted Q2 had any authority over Q1. Many more such variants can be found in the page-by-page collation of Q1 and Q2 which is provided in the Appendix to Daalder-Moore. There is in principle no reason why they are any less likely to be Dekker’s than those which we find in reset sheet B.

55It is important to realise that an inconsistent attitude to reset material in Q2 is not confined to Bowers. Mulholland, with some modification, shares it with him. He admits his adherence to Bowers, and summarises his general position as follows: ‘The analysis above [referring to his own] calls for acceptance as authoritative those alterations to standing type in SIMMES-CREEDE-STAFFORD-ELD2 (with a small number of exceptions) and, among changes in reset type, only those encountered in sheet B and the reset page of outer K. Otherwise, except in those instances in which variant readings preserved in unauthoritative reset type have been thought necessary or fitting, such variants have been rejected’ (Middleton, Comp., 511; 513).[23] There is a marked echo of Bowers here, who writes about readings in supposedly unauthoritative reset type that an editor is bound ‘to accept only those Q2 variants in reset pages which appear to him to be necessary or desirable alterations, even though unauthoritative in their origin’ (14).

What Mulholland says, in the cited passage, about his procedures for acceptance or rejections sounds very principled and well-controlled, until we examine matters more carefully. One question which immediately arises is who else than the author would have introduced ‘necessary or fitting’ readings in ‘unauthoritative reset type’, and Mulholland does not attempt to provide an answer. Neither does he explain why these ‘necessary or fitting’ readings would need to be thought of as somehow less good than those which he accepts as ‘authoritative’ or why they would have been introduced according to a process which makes them inherently more suspect than the supposedly ‘authoritative’ ones. One could, of course, suggest that another person with a mind very much like Dekker’s own (just as educated and intelligent) may have introduced these revisions. But, in that case, why would we not believe that to have happened in the case of sheet B and the reset page of outer K as well (leading to the conclusion that the author was never the reviser)? No reason is offered why the one group of emendations should be seen as authoritative and the other group should not: thus, once we actually analyse the division, it appears to be entirely arbitrary and unfounded.

Furthermore, like Bowers, Mulholland is compelled to admit that the standing type material is not invariably authoritative, which tends to diminish the very categorical difference that he sees between it and the reset material which he in principle rejects. But this is a comparatively minor matter. I have little difficulty accepting many of the standing type corrections in Q2 as Dekker’s. My problem is with Mulholland’s attitude to, and handling of, the supposedly ‘unauthoritative’ variants in reset sheets other than sheet B and the reset page of outer K. There is absolutely no reason why the many superior readings both in sheets B and K and in others should not all be seen in the same light: either all as Dekker’s, or all as someone else’s, and no matter whether just a single good revision is found on a particular page or more than one. Alternatively, if we are to imagine that there were two intelligent revisers, a reasoned attempt should be made for finding a solid way to distinguish between the two. That cannot be done by simply claiming that emendations in one reset sheet are authoritative while those in another are not. It must be remembered that the logical person to work on revision of his own text was Dekker himself, so that we create an artificial difficulty if we assume that someone else was (quite unnecessarily) called on to introduce similar revisions but without having the fiat of the author.

The Daalder-Moore essay, published in 1996, deliberately discussed a number of Q2 variants (see those bulleted above) which it considered of high quality, and which all occur in just such reset pages as Bowers in principle rejected. Going by his own pronouncement of his editorial policy, Mulholland, too, would need to reject them immediately as not authoritative. However, he does not. On the contrary, contradicting his general statement about accepting reset Q2 material only from ‘sheet B and the reset page of outer K’ (513), he says: ‘Possibly worthy of addition to Bowers’s list is the change from ELD1 ‘deept’ on h2 to ‘dipt’ (13.43 [his reference]), which may signify authorial sophistication of the order noticed in D1’ ... ‘Further, this alteration lends additional force to the possibility of authorial involvement in the change from ‘for’ to ‘far’ (13.35) on the same signature (h2) in providing evidence of more than a single instance in this forme’ (510). The fact that two changes occur on H3 does not carry any weight with me: if the emendations are authorial, they could also occur on two different signatures. However, Mulholland’s tentative preference for ‘dipt’ and ‘far’ sounds promising, though at odds with the policy he put forth on his page 513. For a moment, one wonders whether he will accept ‘dipt’ and ‘far’ as authorial/authoritative.

But this is not truly what happens: they stay in the realm of ‘perhaps’ and ultimately of what is thought ‘necessary or fitting’. On page 521, Mulholland actually turns coy. Although suggesting on page 510 that ‘dipt’ ‘possibly records the kind of sophistication associated with an author witnessed elsewhere (cf. far at 13.35)’, he now proceeds to propose that ‘dipt’ may stand as a ‘modernization’ of ‘deept’, ‘a word that, if not obsolete, probably verged on obsolescence by 1604’. And he entertains the possibility that in that case ‘dipt’ ‘may be accepted as an unauthoritative but correct emendation’ of ‘deept’. I must confess myself totally unconvinced by this kind of reasoning: ‘dipt’ gives clear, simple, and unequivocal sense, and there is absolutely no reason for thinking of it as any less authorial/authoritative than words in sheet B.

60Another point worth making is that in practice Mulholland depends a good deal more often on Q2 than he himself states. Thus, the last bulleted example quoted above (‘mourning’ vs. ‘morning’) is a case in point. The Q1 reading is ‘morning’, and this is the word that, acting on his own principles, Mulholland should accept, or offer reasons for rejecting. However, he does not even mention ‘morning’ in his apparatus, and in his text (Middleton, Works), 13.81, prints ‘mourning’, the Q2 form, as though the word ‘morning’ in Q1, which on his reasoning is to be seen as in principle preferable, does not even exist, or as though there is no such distinction between ‘morning’ and ‘mourning’ as I have discussed above.

This is something which can easily happen when editors persuade themselves that they ought to stick to a supposedly authoritative text, and must as a matter of principle reject superior readings from another text which in fact cannot be argued to be generally inferior: often, their literary inclination will ultimately steer them towards the better readings. A similarly interesting example occurs in Mulholland’s 15.61, where in his text he prints Q2’s ‘sleights’ (from I2v, which is reset) instead of Q1’s ‘slights’, which his stated policy requires him either to accept or to offer reasons for rejecting: but, again, he does not even mention Q1’s reading. More often than not he does list both variants, but in many cases he prefers (and includes in his text) a Q2 variant without explaining why, even though such explanations are certainly called for. For example, his textual note for his scene 13, line 109 mentions both Q2’s ‘bands’ and Q1’s ‘bonds’. According to his stated policy, he should prefer the Q1 form, ‘bonds’. Yet in his text he uses Q2’s ‘bands’. He gives no reason for doing so. Of course, there is, in fact, at least one good reason, and that is that ‘bands’ rhymes with ‘hands’ in line 110. The meaning of bands, too, is different from that of bonds, and fits the sense better. In other words, this is one of many instances where Q2 does provide the better reading, but Mulholland does not discuss the fact, while he simply, without more ado and against his own self-imposed rule, includes the superior reading from Q2 in his text.

Ironically, then, Mulholland regularly – and fortunately, in my view – is guided by his literary sense towards Q2 at times when according to his own theory he should reject its readings. In what follows I shall further argue the case for Q2’s general superiority.

Compositorial Inaccuracy versus Authorial Precision

I begin with the area of punctuation. It is widely believed that punctuation (belonging to Greg’s area of ‘accidentals’) was of not much concern to authors around 1600, and that the choice of punctuation marks was usually left to the printing house, with varying results. Sometimes punctuation used at the time is hard to make sense of by any modern standard, although, not infrequently, it seems to be rhetorical rather than grammatical. My contention is that Dekker cared profoundly about what he saw as accurate punctuation, and that it was no doubt he, himself, who revised much of the unsatisfactory Q1 punctuation so as to bring out syntactical sense, in a way which seems remarkably modern. The punctuation in Q2 is often much superior to that of Q1, as I shall now demonstrate. In the following examples I compare readings from the two quartos. To make matters as clear as possible, I use my own modernised spelling, but reproduce the punctuation marks from Q1-2. In each instance the Q1 version is mentioned first.

1a. ...Could her tomb
Stand whilst I lived so long, that it might rot,
That should fall down, but she be ne’er forgot.
(Q1, TLN 152-54)1b. ...Could her tomb
Stand whilst I lived, so long that it might rot,
That should fall down, but she be ne’er forgot.
(A4r, standing; Q2, TLN 152-54)2a. Duke. Which to prevent.
Doctor. ’Tis from my heart as far.
(Q1, TLN 2204-05)2b. Duke. Which to prevent –
Doctor. ’Tis from my heart as far.
(h2r, reset; Q2, TLN 2204-05)3a. Come hither, George, hie to the constable,
And in calm order wish him to attach them,
Make no great stir, because they’re gentlemen,
(Q1, TLN 676-78)3b. Come hither, George, hie to the constable,
And in all calm order wish him to attach them:
Make no great stir, because they’re gentlemen,
(C2v, C3r, reset; Q2, TLN 676-78)

Unequivocally, in 1b, 2b, and 3b the new Q2 marks provided are a significant improvement, and remarkably in tune with those which we might use now, while the clumsy punctuation in 1a, 2a, and 3a seems to be compositorial. In 1b the positioning of the comma in TLN 153 is very sophisticated, and it surely would not be a surprise if this drastic and meaningful change in punctuation originated with the author. The fact that it occurs in standing type (so that special trouble had to be taken for any revision to be made) shows how important Dekker must have found the emendation, which turns nonsense into sense. Likewise, the dash (signalling interruption) in 2b, TLN 2204, is most likely something which Dekker himself, in revising, would have thought of including, and much the same can be said about the colon in 3b, TLN 677.

65These three examples of superior Q2 punctuation are not exceptional. Daalder-Moore, 267, presents a large number but by no means all of them, and upon examination all will be found significant. I would suggest that the punctuation in the three examples above (and many others in Q2) should, to use W. W. Greg’s terms, not actually, as regards their function, be thought of as ‘accidentals’, but as ‘substantives’. Greg himself saw punctuation as belonging to the accidentals of a text. As all editors of Renaissance texts know, Greg distinguished between (a) the significant or as he called them ‘substantive’ readings of the text, those namely that affect the author’s meaning or the essence of his expression, and (b) others, such in general as spelling, punctuation, word-division, and the like, affecting mainly its formal presentation, which he regarded as the ‘accidents’, or as he proceeded to call them ‘accidentals’, of the text (see his ‘The Rationale of Copy-Text’, Studies in Bibliography 3, 1950, 21). The difficulty in this way of thinking is that it suggests that Renaissance punctuation affects mainly the formal presentation of a text, and not the meaning or essence of the author’s expression. Yet the punctuation of 1b, 2b, and 3b very clearly does carry meaning. Hence an editor who chooses Q1 as the copy-text would ignore important pointers to the meaning, in all three instances above, which are provided by Q2, but not by Q1. And the fact that the punctuation of the revised text, Q2, is generally speaking superior to that of Q1, and more often makes sense, establishes one powerful reason for viewing Q2, and not Q1, as the more authoritative copy-text (the text which ‘may be supposed to represent most nearly what the author wrote’, to use Greg’s words). At this point, I feel I need to point out that the superior punctuation of Q2 is almost never recorded, or (at least explicitly) paid attention to, by Bowers and Mulholland. My own view is that the so-called accidentals of a reprint like Q2 should be just as closely studied as its substantives, as they are not necessarily less important in conveying meaning.

Spelling errors are rife in Q1, and Q2 corrects a very large number of literal errors, showing, in relation to this matter, how slovenly or ignorant the Q1 compositors often were, and how precise Dekker himself was in revising Q2. Daalder-Moore offers an extensive list of such errors (266), including examples like ‘pocker/poker’, ‘tempred/tempted’, ‘iusttuctiion/instruction’, etc. Indeed, whole words are often changed, such as ‘carter/courtier’, and ‘praise/phrase’: in such cases we are not just faced with misspelled words in Q1 as a matter of ‘accidentals’, but as what, in the printing of that quarto, had become ‘substantives’ other than the author apparently intended, and which he for that very reason subsequently corrected. Again, Dekker himself rather than anyone else is the person most likely to have both spotted and corrected such mistakes.

Those printing Q1 very often had quite serious problems with names, which are very often incorrect. Several of the Q2 corrections of Q1’s erratic names occur in sheet B or in pages that are in standing-type, but they are widely distributed throughout the text, so it would be odd to think that Dekker would not have wanted to correct them in all reset pages, and he did in fact do so in e.g. TLN 1976 (‘Myllan’ Q1 / ‘Millan’ Q2), and TLN 2472 (‘Chastrucio’ Q1 / ‘Castruchio’ Q2). The frequent changes in names, throughout the new text, whether in standing or reset type, are punctilious and consistent. Even the numerous Q2 emendations in standing type confirm that the printers made a mess of the names in Q1, while it was no doubt Dekker himself who corrected them in Q2. Indeed, how, if faced with misspelled names, would others than the author himself necessarily have known which names he intended to use? The very quantity and quality of his corrections of names in standing-type pages should lead us to take very seriously similar spelling changes that we find in reset pages. Obviously, if Dekker found it important to correct names and other words extensively in standing pages, it is reasonable to assume that he wished to correct them in reset pages as well. For examples of emended names, see Daalder-Moore, 267.

Further categories of errors which show the mistaken practice of Q1 printers on the one hand and an intelligent reviser on the other are omissions of words and mistaken interpolations in Q1 which are corrected in Q2. Q2 often provides more careful pointing for ellipses. Q2 furthermore corrects a number of Q1’s mishandlings of lineation and prose/verse-setting. Quite often Q2 adjusts capitalisation according to the requirements of grammar or lineation. Many specific examples of Q2’s superiority over Q1, in these areas, are listed in Daalder-Moore (266-67).

More Sweeping Changes in Q2

In yet other cases changes go well beyond mere corrections of errors or comparatively modest changes in meaning. For example, in TLN 289, Q1’s ‘olde dames’ is replaced by Q2’s ‘mad-caps’. This is clearly a case of an author thinking of a completely new alternative: ‘mad-caps’ could hardly be due to a printer’s error or audacity. Indeed, at times whole passages are rewritten in a way which almost certainly someone in the printing house would have no reason for considering. To verify that this is not a hollow assertion I would draw attention in particular to the paragraph on p. 264 of Daalder-Moore which starts with ‘The reset page K3r contains some of the most extensive changes in Q2. Here, two entire passages (V.ii.395-7 and 401-10) are rewritten ...’. The paragraph gives a good indication of the nature and extent of such more substantial and extensive changes. I think that anyone studying both the passages mentioned and what is said about them would concur that these significant examples of drastic re-writing were very probably undertaken by the author himself. Furthermore, and perhaps especially, the evidence of much careful and extensive attention being given, by the reviser, to the lines of the eponymous heroine in her big scene ( 5.2 ) reinforces the impression of authorial contribution to Q2’s text.

Final Methodological Observations

70Greg’s idea of a copy-text was that, ‘in the case of printed books, and in the absence of revision in a later edition [my italics], it is normally the first edition alone that can claim authority, and this authority naturally extends to substantive readings and accidentals alike’ (‘The Rationale of Copy-Text’, Studies in Bibliography 3, 1950, 22). It is commonly held to be the text which ‘may be supposed to represent most nearly what the author wrote’ (Greg, 21), or, as others might or do put it, ‘the text which is seen as in principle the most authoritative’. Commonly it is this text which editors most closely follow, with the least possible alteration, and accepting deviations from the text only when they believe there are compelling reasons for doing so, as in the case of e.g. an emendation which turns sense into nonsense (the common belief being that authors wrote something which they intended to make sense).

Where a printed text is the first edition, it will, if not followed by reprints, of course usually be seen as the copy-text. Naturally this does not necessarily mean that it will be faultless, but it will be held to be generally reliable unless there is evidence to the contrary. In the case of texts printed around 1600, it is widely held, and on good grounds, that every reprinting will bring with it the risk of one or more newly introduced errors if the text is re-set (although occasionally corrections also get included).

Thus the common reason for, usually, not accepting the authority of a reprint is the conventional bibliographical rule that in many cases the copy-text chosen by an editor will/should be the earliest printed text in the series, because this text is, as Bowers put it, ‘closest to the ultimate authority of the lost manuscript’.[24] Certainly, even in the case of Q2, not only have mistakes from Q1 been preserved in the reprinting, but there are also some new errors. Anyone who thinks that Q2 will invariably and in all respects be a better text than Q1 is wrong and foolishly optimistic. I do use it in principle as my copy-text in this edition, but necessarily with caution.

While Greg said that ‘it is normally the first edition alone that can claim authority’, it is nevertheless important to note that in the same sentence he added ‘in the absence of revision in a later edition’. It is the presence of revision in a second edition which, in principle, might always to some extent weaken the claim of a first edition’s right to be considered the ideal copy-text, while it may make people wonder, at least, whether revision in the second edition is so important and pervasive that it, in fact, becomes the text which ‘may be supposed to represent most nearly what the author wrote’, particularly if we believe that the author improved the first text greatly (notably if that contained many mistakes), and, as well, that in principle we ought to respect an author’s second thoughts if the changes are not just corrections but carefully considered alterations. I believe that if a modern audience is offered a choice between the kind of faulty first edition and a superior second version to which I am referring, it will (and should) prefer the revised version.

The evidence which we have considered unequivocally establishes Q2 as a carefully, extensively corrected and revised text which is greatly and comprehensively superior to Q1. Interestingly, the consistent superiority can be seen as establishing a new authority which, to use Greg’s words, ‘extends to substantive readings and accidentals alike’. Even Bowers, though in theory not convinced of the authority of a number of Q2 readings which he nevertheless accepts as ‘necessary or desirable’, took the view that, in the revised text, ‘Certain of the alterations made in standing type appear to be of such a revisory nature as to surpass the competence of a printing-house editor, and in this case there is no need to conjecture the presence of a publisher’s editor other than the author’ (14). Needlessly and unconvincingly, Bowers (followed by Mulholland) contended that revisions in standing type and in reset sheet B and K3 are authoritative, but that other Q2 variants which he accepts in his edition are, although ‘necessary’ or ‘desirable’, nevertheless ‘unauthoritative’ (18). In fact, however, the distinction which Bowers makes between ‘authoritative’ superior readings and supposedly ‘unauthoritative’ ones is, as we have discovered, wholly arbitrary and unfounded. In practice, Bowers accepts so much material from Q2 that, substantively speaking, his edition is actually based on that text rather than Q1. Hence, rather than undermining the authority of Q2, Bowers in practice corroborates it.

75What we see, then, is that Bowers’s and Mulholland’s choice of 1 The Honest Whore as their copy-text rests methodologically on a mistaken basis. These editors in essence ignore Greg’s caveat that ‘in the absence of revision in a later edition [my italics], it is normally the first edition alone that can claim authority’. Q2 is most certainly a significantly revised text, and there is no reason for doubting that it was revised by Dekker, as both Baird and Greg already believed in the 1930s. Part of the difficulty in the Bowers-Mulholland view of matters is that they realise that some of the revisions, at least, could only have been made by Dekker, but that they decide to see them as occurring only in a very limited way on certain specific pages without realising that revisions of the same kind as those which they accept can be found throughout the text of Q2. If they had been more open-minded about Q2 as a whole, they might have come to consider that text as an authoritatively revised quarto, which they could safely have made their copy-text, although, of course, as in the case of any early seventeenth century text, any editor must look out for some new unintended misprints (which in Q2 are actually easy to spot).

In the event, we see that Bowers and Mulholland decide to choose Q1 as their copy-text, ignoring – in theory – many superior variants which abound in Q2. They regard those variants as ‘unauthoritative’, except in a small number of instances, which for no good reason they see as confined specifically to certain pages. To quote Mulholland, he accepts ‘as authoritative those alterations to standing type in SIMMES-CREEDE-STAFFORD-ELD2 (with a small number of exceptions) and, among changes in reset type, only those encountered in sheet B and the reset page of outer K. Otherwise, except in those instances in which variant readings preserved in unauthoritative reset type have been thought necessary or fitting, such variants have been rejected’ (Middleton, Comp., 511; 513).

A number of difficulties arise from this outlook. The most important one is that the basis for judging sheet B and the reset page of outer K is a literary one, which leads to the problem that similar literary judgements might be made concerning material outside sheet B and the reset page of outer K. By deciding to view the revisions on sheet B and the reset page of outer K as legitimate, Bowers and Mulholland claim a specific connection between certain pages and what they regard, on literary grounds, as authoritative revisions while no bibliographical evidence is offered to prove such a connection. In theoretically rejecting good revisions outside the pages mentioned, Bowers and Mulholland do not really quite know what to make of those. In theory, these editors should reject all revisions outside their chosen pages as ‘unauthoritative’, because they occur outside the boundaries they have selected. However, as literary scholars, they can in practise see that many readings outside those boundaries are very good and will improve their texts if incorporated. They hence come to the view that such readings should be included, but on the vague grounds that they are ‘desirable’ (Bowers), ‘fitting’ (Mulholland), or ‘necessary’ (both scholars). It needs no arguing that inevitably these words point in the direction of very subjective individual choices from a text which is not their copy-text, whereas an editor who, as I do, uses Q2 as the copy-text is methodologically consistent, systematic and clear about the choice of text made. (I do not select readings from Q1 on the basis of taste, though occasionally I am compelled to incorporate a correct Q1 variant when Q2 contains a misprint.) Bowers and Mulholland in practice – because they choose their Q2 readings well – produce a text which substantively is in effect identical to mine, and they would have made a more logical and defensible choice by using Q2 as their copy-text. In Greg’s terms, Q2, as a drastically re-written second text, contains new authorial substantives and accidentals to which authority naturally extends alike, and thus it meets the requirements of a perfectly satisfactory copy-text.

So far what has been missing entirely is an edition directly based on Q2, which includes a great deal of high-quality, consistent and pervasive correction and revision by Dekker himself, as Matthew Baird and W. W. Greg already realised in the 1930s. This is what I have aimed to provide, in the full conviction that it is both responsible and necessary to use Q2 as a copy-text. My reason for this conviction is that Dekker himself revised his Honest Whore text very extensively, and did so with great competence, as it was meant to offer would-be buyers the opportunity to purchase a publication written by him which needed to look quite different, had a new title (The Converted Curtezan), and was, in fact, a significantly corrected and revised work: there was a strong incentive for Dekker to produce a volume which could lay claim to being an improvement over 1 The Honest Whore. Indeed, he could, if questioned by an owner of 1 The Honest Whore about the similarity of the two texts, legitimately claim that the second impression was a much-amended version. As well, he probably derived personal satisfaction from the process of revision, and he took great pains to improve numerous details very punctiliously.

At this point I should like to mention an additional reason why I believe it is preferable to use Q2 as the copy-text. While adherence to Q2 must not be slavish, as new printer’s errors do occur (though far fewer of them than had occurred in Q1), a preference for Q2 avoids the risk that superior new ‘accidentals’ (of which there are many), or for that matter substantives, are overlooked. As T. H. Howard-Hill notes in his paper ‘Modern Textual Theories and the Editing of Plays’ (The Library, Sixth Series, Vol. XI, No. 2, 1989, 89-115), a tendency to adhere passively to a text like Q1 as one’s copy-text in the matter of accidentals while one only accepts the odd substantive reading from Q2 is likely to be not only illogical but harmful, as one may too readily reject or overlook, in a thoughtless way, Q2 material (especially accidentals) that may be authorial. This may be particularly important if accidentals are overlooked which help to establish the author’s meaning, as punctuation often does. In the case of Q1 and Q2 of The Honest Whore that seems to me a very real risk. There also are, as Howard-Hill argues, many cases of what may superficially seem to be ‘indifferent variation’ between the original and the reprint where in fact the latter should be preferred as even a slight difference may be authoritative: ‘For a revision the qualitative possibility of recapturing the author’s second, perfecting touches [in Q2] is arguably more important than the minimal protective goal of preserving a greater quantity of his rejected readings in the copy-text [Q1]’ (100). All in all, then, in this case it is safer to use Q2 as one’s copy-text than Q1.

80Methodologically, we need to examine a little more closely what type of text Q2 is. It belongs to a small group of texts, in the English Renaissance, which were deliberately, extensively, thoroughly and authoritatively revised before they were reprinted. A number of such works have already been mentioned earlier in this discussion. An extreme example would be, as Greg points out, Ben Jonson’s Every Man in his Humour as that appears in Jonson’s 1616 folio. To quote Greg:

[It] was not set up from an independent manuscript but from a much corrected copy of the quarto of 1601. That Jonson revised the proofs of the folio has indeed been disputed, but Simpson is most likely correct in supposing that he did so, and he was almost certainly responsible for the numerous corrections made while the sheets were in the process of printing. (‘Rationale’, 34)

Dekker, too, may have revised proofs of The Converted Curtezan, although it would be hard to prove that he did. What we can nevertheless safely assume is that he worked very closely with the printers in what was a conspiratorial project designed to produce, in secret, two closely connected quartos, the second of which was in part based on revised standing type for the first, and otherwise on equally carefully revised reset type: comparing the two quartos we find, in Q2, consistent correction and revision of substantives and accidentals. We may particularly note very precise correction in Q2, especially of such things as names and punctuation, about which the author perhaps consulted very directly with the printers to ensure absolute accuracy. (Names, in particular, had often been unrecognisable in Q1.)

At any rate, authorial involvement in the revision – and perhaps even in the process of printing – must have been very close. Theorising about close and extensive revisions of printed texts, Greg’s predecessor Ronald B. McKerrow describes something close to what is applicable in a consideration of Q1 and Q2:

We are not to regard the ‘goodness’ of a reading in and by itself, or to consider whether it appeals to our aesthetic sensibilities or not; we are to consider whether a particular edition as a whole contains variants from the edition from which it was otherwise printed which could not reasonably be attributed to an ordinary press-corrector, but by reason of their style, point, and what we may call inner harmony with the spirit of the play as a whole, seem likely to be the work of the author: and once having decided this to our satisfaction we must accept all the alterations of that edition, saving any which seem obvious blunders or misprints.[25]

Greg adds to this that McKerrow failed ‘to add the equally important proviso that the alterations must also be of a piece (and not, as in The Unfortunate Traveller, of apparently disparate origin) before we can be called upon to accept them all’ (‘Rationale’, 25). In the case of Dekker’s revisions in The Converted Courtesan, it seems that Greg did accept that the revisions were, in fact, ‘of a piece’, for at the end of his article ‘“The Honest Whore” or “The Converted Courtezan”’ he laments the fact that Nicholas Okes reprinted, in 1615, Q1 ‘and so established to this day an inferior text of Dekker’s most successful play’, with the result that – purely through ‘ill luck’ – the text which Greg unambiguously saw as superior, that of Q2, fell into neglect.[26] I profoundly hope that my edition will now, to use Greg’s word, ‘establish’ the authority of Q2.