 | |  |
Notes
The date of registration was scarcely more than a month after Marlowe's murder. Naturally, the query has been made why Jones waited not less than five months, and possibly more, after acquiring title before he published the play if 1594 were the actual first-edition date. It is interesting also that the Register entry repeats the wording of the printed title though without the statement of acting, an indication, perhaps, of some preparation for the press.
For convenience the book is here referred to as a quarto (the actual method by which its cut halves of double sheets were machined on the press) although technically it is an octavo, distinguished as a 4°-form octavo.
To this statement (p. 61) Dr. Welsh adds a footnote which is pertinent to the argument, "We should also have to suppose, of course, that between the printing of A inner and the imposition of B inner headline V was moved to the opposite quarter of the skeleton."
For the basic discussion of this matter see R. B. McKerrow, An Introduction to Bibliography (1927), pp. 184 ff. This early examination, however, puts excessive weight on the question for a first edition whether a particular book may begin with the text in sheet A or in a later sheet, leaving A or arbitrary signs or other expedients for the preliminaries. For a useful corrective, see R. C. Bald, "Evidence and Inference in Bibliography," English Institute Annual: 1941 (1942), pp. 162-163. In fact, Marlowe's and Nashe's Dido (1594), and Doctor Faustus B-text (1616) were also printed with sheet A last, like Edward II.
This printer, on the evidence of the ornaments, and the initial on A2, was Robert Robinson, not Richard Bradock as conjectured by Greg (no. 129). See Robert Ford Welsh, "The Printer of the 1594 Octavo of Marlowe's Edward II," SB, 17 (1964), 197-198.
One may speculate, however, that he had not perhaps cast off the whole text since he left himself with blank leaf M4. Because he was not setting by formes, or by simultaneous setting and printing in two halves, an exact casting-off was not required.
The question of expertness may just possibly be involved in the reason why Compositor X was brought back at the end to set cast-off sheet A, in addition to the partial equalizing of the stints that resulted. Professor Hinman has found that in the Shakespeare First Folio the apprentice Compositor E was not entrusted with the setting of head-titles and the text around initial letters. Since especial expertness would be required for the more difficult task of composing a balanced titlepage, it may be that one of the causes for the return of Compositor X was the need to set up the title on sig. A1. Of course, he may have started composition of sheet A while Y was finishing sheet M and thus speeded up the conclusion of the typesetting.
For the difference with which two compositors could treat the lining of difficult copy, see the Textual Introduction to Cupid's Revenge in The Dramatic Works in the Beaumont and Fletcher Canon, ed. Bowers, 2 (1970), 324 ff.
It would then follow that if a lost edition of 1593 was not the copy for the Dyce manuscript leaves, the readings permit only the 1594 quarto to have been the source. In the seventy-two lines of stage-direction and text transcribed in the manuscript leaves Greg lists sixteen substantive differences between manuscript and the 1594 text, all of them manuscript errors so far as can be told. Many of these are egregious and make a hash of the sense, such as Its for 1594 As (22), bakt for Rakt (23), eate for dart (44), and gasing for grazing (62). It is not properly evidence, but one cannot help observing that a mistake in the date is no more careless than some of the errors in the text. Greg remarks, however, that whereas the date 1598 in a printed edition might be mistaken for 1593, the evidence is clear that the transcript was not made from a copy of the 1598 quarto. See, for instance, manuscript and 1594 horses but 1598 horse in line 31 and an for 1598 the in line 63. Also, significantly, 1594 Porpintine and manuscript Porpentine but 1598 Porcupine in line 43. In the Introduction to the Methuen edition of Edward II (1933) the editors Charlton and Waller (pp. 3-5) agree with Tucker Brooke that the manuscript variants thine, dinner, and Sylvan for 1594 thy, dinner time, and Sylvian (11,34, 61) are preferable, but this is much a matter of opinion and leans heavily on two manuscript normalizations and one mending of the metre. If these were indeed the readings of a 1593 edition, the rather extraordinary departures from them of a reprint in 1594 (away from normality) are difficult to account for since in the same breath the editors argue for the serious manuscript errors of Its, bakt, eate, and by (22,23,44,69), to which should be added gasing (62) and tantum (24), though corrected in 1594, as 1593 "mistakes which might easily be made by a printer working hastily from manuscript to catch a public still excited by Marlowe's death," and conclude, "It will be noticed that the six cases adduced by Tucker Brooke as inferior readings in the manuscript are really stronger evidence for the existence of a 1593 edition than the three in which the manuscript seems preferable; for the latter might be improvements made by the scribe himself, whereas Its, bakt, and eate require explanation by some antecedent corruption." There is difficulty in accepting this view. Superficially, of course, it is easier to account for the Dyce manuscript corruptions not as scribal errors in transcribing printed copy but as compositorial errors setting from manuscript. But in the seventy lines of manuscript text plus two lines of stage-direction there are sixteen substantive variants from 1594, and if only three are to be taken as certain 1594 corruptions, then we are left with six clear errors in the Dyce transcript purporting to come from the 1593 printed copy but corrected in 1594. To these, however, must be added seven more indifferent substantive variants that in the nature of the case cannot all be 1594 reprint corruptions but must in some part at least be taken as present in the hypothetical 1593 edition. To extrapolate that high proportion of one substantive error for at least every ten lines of 1593 throughout the rest of the text would be a fantastic proposition scarcely to be supported by the evidence of 1594, no matter how much haste might have occurred in the 1593 setting. The logic of this argument will not bear scrutiny.
| | |