The Shakespeare Conference: SHK 27.039 Thursday, 4 February 2016
Date: February 3, 2016 at 6:22:31 PM EST
Subject: R3 and J. K. Walton
I wrote most of this in 2015, when I began to appreciate the webs R3 assumptions weave. Only a long article can clear the air, which I ain’t up to. Some of the mysteries were man-made long ago; some are modern in origin. I hadn’t expected such length—but then, I never do. If you haven’t thought about any of this stuff, lucky you; unless some of it forms the basis of your opinion.
J. K. Walton thought Q1 King Lear was a memorial reconstruction, though he bucked the trend to allow a chance it was a shorthand report (as I believe). He denied any Q2 Lear involvement in F’s printing, opting instead for Q1 as F printer’s copy. The consensus now is that Q2 was heavily used; I agree with the few who think F copy was primarily a manuscript redacted from Q1 Lear. It’s surprising that the Lear and R3 cases are ultimately similar; yet while Walton assigns Q1 as F Lear copy, he allows no Q1 R3 influence on F R3, except through Q3. His rationale in leaving Q1 out of the Q3/Q6 debate seems to influence others (unless I missed their reasons for ignoring Q1 (and QMS) as possible direct influences on F production). All agree that printer’s copy for most of F was Q3, Q6, or both. It’s sort of a Matthew, Mark, or Luke New Testament derivation thing: comforting to the faithful to drop it at God or Shakespeare, but unhistorical.
Walton attempts to use the methods of classical textual scholarship (study of manuscript transmission) to determine which quarto in the series of reprints served as F copy. His first error is to reject the notion that more than one R3 edition could have been used (Q3, Q6, or both? Most scholars nowadays accept evidence for both). Further, Walton restricted the evidence to “definite errors” introduced in any quarto of the series that are repeated in F; too much useful evidence is eliminated by this method. For example, Smidt saw a correlation in the use of parentheses in Q1 and F: they may not be errors, they may not be ‘definitive,’ yet they got Smidt’s attention, and rightly so; coincidental readings of any kind are matters of probability and evidence.
But another error seems fundamental. Walton remarks a “special feature” of playtexts like R3: for the purpose of determining which quarto served as F printer’s copy, the “archetype [‘the exemplar in the copying of which the first split in transmission originated’] is the same as the original” (65). That is, we can assume Shakespeare’s manuscript is immediately behind a Q1 because “the manuscripts used for printing his plays were either originals or transcripts not far removed from them.” It is wrong to think early theatrical transcripts are of no account; but his point is that if two copy-texts were practically the same, “we may be misled by the number of errors common to Q and F texts [to conclude] that F was printed from corrected quarto copy.” In other words, when Q1 and F are alike, F does not reprint Q1 if a later quarto is shown to influence F’s printing. That mistake stipulates a “common original source” for a container-ship load of Q1/F evidence if R3 is in this “special Shakespeare” category.
On one page, Walton observes that Q R3 “is seriously corrupt on account of having gone through a memorial stage” (80); on the next, he states that “reprints of bad quartos” are not relevant to his study (81). Why isn’t Q1 R3 a bad quarto? According to Walton, “goodness or badness of the text considered in terms of literary value is, after all, the basis for the whole concept of ‘good’ or ‘bad’ quartos . . . ” (16). No one doubts that textual inferiority led to defining bad quartos as necessarily resulting from memorial transmission. The cause of the inferiority is inferred: memorial “stages” create bad quartos. Walton fails to see the connection but his consequent rejection of Q1 evidence has been accepted, possibly because his discussion is out of joint; one must page back and forth (takes one to know one). However, editors assuming F R3’s authority are more apt to welcome than to question supportive argument; they dearly want a nearby original (or two?) in Shakespeare’s hand.
In each case, printer’s copy for a bad quarto (memorial report) is far removed from its originating texts: spelling, misreading, punctuation, actor-error (in many forms), lineation, made-up or missing s.d.’s and s.p.’s, additions, revisions, and copying and printing errors all contribute to differentiate the genealogy of the texts; if F is much like a bad quarto, the “original” is long gone. Walton recognizes that a classical “archetype will be free of errors arising after the [first split in the transmission of the surviving texts], but, with classical and medieval texts, it will rarely be the same as the original, and will therefore contain errors of its own arising from the process of transcription” (44). R3, if memory serves, represents a ‘new’ archetype very, very different from the original text. If one insists that F reproduces the authorized text and that it reprints a quarto, then Q1 must be allowed as an F-copy candidate.
The R3 Ms. that existed with all certainty, and which we should know a lot about, is Q1 printer’s copy; the memorial report (I say). If QMS survived to be recopied or revised, or if it served to annotate one or more quartos, it is an FMS to be reckoned with. Some textual problems can be addressed by corrective use of QMS as the simplest solution. However, its use would not preclude the existence of another FMS; the alternatives should be weighed.
It seems possible that FMS was a separately reported text. Since shorthand reporting is what shorthand reporters did, that’s not so unlikely an explanation for added F errors, its inability to correct or overrule Q, and printers’ preference for printed copy (why struggle with another problematic text?) Alternatives should be addressed in order; I’ve not got that far.
If Q1 is reported, the printers knew it and were free to correct, or not, as each saw fit; their object was a readable, popular, and economical book. Success was six editions by 1623, with little editorial attention. That changed when F (the dreaded one-legged A, in a different font) came along.
According to Arden3, “Folio [R3] is remarkable for changing a play that had been more or less constantly in print for twenty-six years” (442). It would be more remarkable if the exceedingly corrupt playtext had not been changed. Heminges and Condell presented F as ‘cured’ of stolen text maladies printed before; Q R3 was not acceptable “as is.” However, as with Lear and so many other texts, quartos are reprinted in new dress and often by different methods. In R3, F arbitrarily “fixes” meter and lineation; it tries hard to correct some text; but other changes seem more like whimsy than correction. An extreme number of substitutions have no rationale other than to be different (‘that’ becomes ‘which’). It’s wrong to take F readings as Shakespearian on no other ground than their appearance, or their better appearance, in F.
I’ve argued that F R3 derives from Q1 (from my own perspective, which included some casual reading); but I was unaware of Walton’s exchanges with one of my favorites, Andrew S. Cairncross (who gets little respect nowadays). In 1957, J. K. (Library) disagreed with A. S. (RES), who argued that F was printed from corrected copies of Q1, Q3, and Q6, “for the convenience of the compositors and the editor-corrector.”
Cairncross reasoned from Q1/F ‘agreements in error’ (against Q3 & Q6) that “they would have either to be set down to pure coincidence, or to be explained by some common manuscript origin, that is, some bibliographical link between a manuscript assembled memorially by the players (Q), and the authentic playhouse manuscript (F)—bibliographical links which would further have to be maintained contrary to the quarto copy used. Neither supposition will suffice. . . . Q1, Q3, and Q6 must all be considered as possible sources of copy for F throughout.”
Walton disagreed, somewhat contemptuously, often irrelevantly, but not without effect. Intermingled use by compositors of three different quartos as F copy-texts doesn’t make full sense by the evidence. However, Cairncross nearly describes my working hypothesis (were I working). He factors in like cases: “Both quartos (Lear and Richard III) contain a memorial element, and there seems no sufficient reason not to relegate them to the ‘bad’ quarto class, except perhaps that they are, or rather seem, better than the other bad quartos so far recognized. It is a matter of degree, and the degree may be more apparent than real.” As Cairncross implies, if F reprints Q1 it can’t indicate the extent of Q1’s ‘badness.’
To eliminate objections to F’s use of three interchangeable quartos, I would alter his proposition: “the apparent alternation of careless correction with meticulousness, suggests that we are dealing, not with the vagaries of an individual corrector, but with alternate types of copy. The meticulousness may be nothing more than use of Q1 [as F printer’s copy], and Q1F agreements [against Q6] not corrections but adoption of uncorrected or partly corrected Q1 [as F copy]. The agreements of F with Q6 would thus be attributable to the intermittent use of Q6 copy; the absence of Q6F agreements being due, naturally, to the use of Q1 copy.”
The actual “Q1 copy” for F may not have been the printed Q1 but a transcription either of Q1 or of Q1’s own printer’s copy (QMS); the transcription will have incorporated corrections, large and small, found now only in F. Alternatively, QMS identity (or partial identity) with FMS can’t yet be ruled out. On these considerations, Q1/F agreements occupy their places alongside text from the printed F copy-texts, Q3 and Q6 (just as the redaction of Q1 King Lear was augmented by Q2 in the printing of F Lear; R3 would not be the only F text printed from manuscript revision of a bad quarto and other steps in the process I envision are also not unprecedented).
Cairncross elsewhere theorized a method of printing from exemplars of early editions. His description was poorly received and poorly understood, yet it has the merit of granting printers an often-denied intelligence: multiple individual copies of early editions were combined with corrected manuscript text such that each printed page had its corresponding written, corrected text pasted on the opposite page (necessitating preparation of another exemplar to correct the pasted-over pages). Thus the compositor (and corrector) had the means quickly to set the type for extensive redaction assisted by printed copy (according to their professional lights, rather than to modern editorial principles). For example, one-sixth of F R3 is printed, with little alteration, from two Q3 sections. There’s no way to know if reliable, separate printer’s copy was on hand but Walton may rightly suggest that “the absence of variants in the two passages is due to the failure of the collator to make the corrections an intact manuscript required him to make.” Yet required is too strong a word; the very existence of quarto printer’s copy for F R3 (or any reprint) in addition to manuscript copy implies some freedom in its use.
Arden3 notes that Alice Walker found 300 substantive F/Q1 agreements against Q2-Q6. Collation shows that F/Q1- and F/Q1Q2-only readings are much more numerous than exclusive F agreements with the Q3 and Q6 reprints. If a redacted Q1 was the primary F printer’s copy, use of Q3 and Q6 was (in effect) on top of the manuscript descendant (and a largely irrelevant convenience): the “raging” argument over the later editions will have missed the significant evidence.
The relevant texts are Q1 and F; the evidence must be related to them in its proper order. However, it’s a good idea to keep some obstacles in mind. For example, A. E. Craven (SB 26) lists substantive variants compositor ‘A’ made in printing Q2 Richard 2 from Q1, where “only 5 changes of the total 155 produced readings that are patently unsatisfactory.” Although ‘thundering smoke’ (Q2) is no ‘thundring shocke’ (Q1), it works; but it also warns us not to suppose that variants are equally significant. Apparently, the same compositor set the Valentine Simmes portion of Q1 R3. In the same vein, a multitude of arbitrary alterations in F R3 can’t be trusted as meaningful without good reason. Yet there is plenty of evidence.
To disallow a probability that Q1 or QMS was directly involved in F’s printing, Walton assumed that FMS and QMS were practically identical. In his roundabout way, Jowett agrees: Q1 “forces us to recognize that there is something suspect about the assumption that memory is a world apart from transcription” (Perplexed, 233). The question is again one of degree. Jowett relegates extensive evidence of memorial transmission to scribal or compositorial error, when a fair Q1 assessment gets beyond those causes. Memorial reporting is not transcription; shorthand reporting is a world apart—too much of an original text is lost.
Jowett argues that a minor speech prefix crux indicates QMS derivation from FMS. Of the vast evidence of memorial transmission, prefix error is among the strongest. Shorthand reporting inherently mistakes speech ascriptions that subsequent agency can’t effectually repair. There is no plausible explanation of such error in authorial or transcribed texts when correct prefixes are essential to all stages of production in the playhouse. To assume that Q1 inherits its mix-ups from F is merely to repeat Walton’s mistaken view that Q1 reproduces F so well that it can serve as F printer’s copy. Strange to say, their next trick is to assert that Q1 could not have been F copy because it matches F too closely.
In reply to Walton, Cairncross cites classical scholars to observe that it “is taken for granted that the establishment of the genealogy of documents ‘depends on the principle that identity of reading implies identity of origin.’ The only alternative is where coincidence operates. ‘Accidental coincidences do occur, and have to be reckoned for: but except where an alteration is very plausible and tempting, the chances that two transcribers have made the same alteration independently is relatively small’” (RES). We must apply a form of this principle in respect of R3: A memorial report—especially a shorthand report—made so many changes to a text, many of which would not be recognized as errors without authentic text to compare, that “identity of reading” throughout a second text could not possibly occur unless the second text derives from the report. The idea that a theatrical report might accurately reproduce an authorized text is fantastic. It could record a ‘good’ performance, but comparison to previously written text would reveal great differences; spelling, punctuation, errors, speech assignments (right or wrong in substance, s.p.’s would differ), set directions, and many other features (whether substantive or ‘accidental’) would attest that one text copied another.
All agree that F R3 reprints Q1, through Q3 and Q6. But with 300 Q1-only agreements with F, and many Q1/Q2-only agreements, given the nature of theatrical reporting, F must reprint Q1 itself. Except for its additions and some substitutions, F is Q1—a shorthand report. The whole of the project implies the lack of authorized text. It may be that a mistaken assumption of good manuscript copy of the play available to the publishers is repeated by assuming authorized text for the additions and corrections.
I got on another tangent by reading Alice Walker’s “300” commentary. It is interesting (to me) how close early investigators get to textual solutions, only to veer off when assumption dragons get too hot. If not before, I apologize now for the numerous long postings. I look behind assumptions mostly to satisfy my own curiosity. With moderate results I suppose others might be interested. That mistake is compounded by one thing leading to another.
I’m reminded of an old movie’s textual problem: dive-bomber Fred MacMurray said good-bye to his dog Piggy as he headed toward a smokestack; the media thought he said “Peggy” and the plot thickened.
Gerald E. Downs