University of Virginia Library

Compositorial Practices in Tourneur's The Atheist's Tragedy
by
MacD. P. Jackson

Although Cyril Tourneur's The Atheist's Tragedy has been edited in the scholarly Revels Plays series,[1] no bibliographical analysis of the original Quarto (1611), upon which any modern edition must be based, has hitherto appeared. Yet the Quarto is the sole source of our knowledge about Tourneur's practices as a dramatist and is thus of crucial importance in the

prolonged controversy over the authorship of The Revenger's Tragedy, one of the most fascinating of all Jacobean plays and a key work in any account of English Renaissance drama. In its linguistic minutiae The Atheist's Tragedy has been shown to match The Revenger's Tragedy much less closely than does the typical play by Thomas Middleton.[2] As critics of this sort of evidence have remarked on the unfortunate aptitude of compositors and scribes to alter an author's colloquial contractions and preferred linguistic forms, the Quarto of The Atheist's Tragedy seems an especially appropriate text for bibliographical investigation. Anything we can discover about the setting and printing of the play is of potential relevance to inferences concerning Tourneur's orthographical style and comparisons between that style and what we encounter in The Revenger's Tragedy.

The Atheist's Tragedy was printed by Thomas Snodham for John Stepneth and Richard Redmer. It collates A1 B-L4 (—L4). The title page, with a list of characters on the verso, is in fact the last leaf of sheet L, transferred to the beginning.[3] Careful inspection and measurement of the headlines reveals that two skeletons were used: skeleton I imposed B(o), C(o), D(o), E(i), F(i), G(o), H(i), I(i), K(i), and L(i); skeleton II imposed B(i), C(i), D(i), E(o), F(o), G(i), H(o), I(o), K(o), and L(o). This implies normal two-skeleton printing, with one skeleton imposing inner formes, another imposing outer formes, the association within B-E of skeleton I with outer formes and skeleton II with inner formes being reversed in E-F, and re-established in G, to be reversed again in H-L.

The second of these reversals—that in sheet G—appears, even on a preliminary inspection of the text, to coincide with some bibliographical division in the text, though at first it is difficult to say precisely where the change occurs. The contraction I'll (as it is in modern spelling) appears 25 times up to the end of sheet F, 24 times as Ile (C2v-F4v), only once as I'le (on C2). It next appears once on G2 as I'le, but then 3 times on G2v and once on G3 as Ile. Then from 2 appearances on H4v to 2 appearances on L2v the invariable spelling (15 times) is I'le. The change from Ile to I'le thus seems to occur within G or H. Within H we find that for the first time many exit directions are preceded by a dash: beginning with a single instance on H2v and ending with one on L3v there are 16 altogether within sheets H-L; 3 entry directions are also preceded by a dash (on H3v, K1, and K2v). Another change occurs a little earlier in the Quarto. Up to and including G1 there are 41 abbreviated speech prefixes for Castabella: 39 of these, including 5 on G1, are printed as Casta, only 2 as Casta (on C2v and G1). On G1v there are 5 instances of Casta, none of Casta, and the figures for G1v-L3v are Casta 16, Casta 13. Both forms appear on I1, I1v, and I2. Finally, there is a shift from the indiscriminate use

of -e and -ee endings, or a preference for -ee, in he, me, she, and we, to a strong preference for the modern spelling.[4] The following table makes this clear:            

	SHEETS B-F	SHEETS H-L
me/mee	42/42	69/16
he/hee	24/42	20/5
she/shee	1/12	5/1
we/wee	6/10	9/4
TOTAL:	73/106	103/26

Sheet G has been omitted from the above table, for though the overall figures for F (22/36) contrast strikingly with those for I (42/7), the difference between G (19/13) and H (19/8) is barely perceptible.

Fortunately one further item of evidence not only enables us to locate the exact point of change in the Quarto, but also allows us to be quite certain that we are dealing with a change of compositors rather than with some variation in the nature of the manuscript copy. The Quarto contains an exceptionally large number of semi-colons. These are regularly followed by a space. Within sheets B-F the semi-colons are not normally preceded by a space: of 221 semi-colons in B-F only 7 are preceded by a space, and 4 of the exceptional 7 appear in full prose lines. Within sheets H-L, in contrast, semicolons are normally preceded, as well as followed, by a space: of 240 semicolons in H-L all but 18 are preceded by a space, and 17 of the exceptional 18 occur in full prose lines.[5] The following table shows the situation within sheet G; for each page the number of semi-colons not preceded by a space is followed by the number of semi-colons which are preceded by a space:

   

1	1v	2	2v	3	3v	4	4v
4/0	0/3	0/3	6/0	4/1	0/10	0/14	2/0

The pattern is obvious: for pages of the outer forme the total is 16/1; for pages of the inner forme the total is 0/30. Clearly, the Quarto was set by two compositors: Compositor A, who preferred not to precede semi-colons by a space, set sheets B-F and the outer forme of sheet G; Compositor B, who preferred to precede semi-colons by a space, set sheets H-L and the inner forme of sheet G.

The evidence previously cited is consistent with this division. Casta on G1 is anomalous (though outweighed by 5 instances of Casta), but the other 7 instances of Casta in sheet G belong to Compositor B's pages. Compositor A always uses Ile, except for one deviation into I'le on C2, whereas Compositor B always uses I'le. Returning to the -e/-ee spellings in G, we find that for the outer forme the figures are 5/11, in accord with Compositor A's frequent use of -ee spellings within the earlier sheets, while for the inner forme the

figures are 14/2, in accord with Compositor B's tendency to avoid -ee spellings within the later sheets.

The division of sheet G between compositors by formes proves that this sheet at least was set by formes, not seriatim, and suggests that the whole Quarto may have been set by formes. Miscalculation in the casting-off of copy, and a consequent need to conserve space in setting the text, might thus account for some of the "extraordinarily large number of verse passages which have been set as prose," and render unnecessary Ribner's postulating of a scribal fair copy intervening between Tourneur's holograph and the printed text.[6] The neatness with which copy was made to fit exactly nine sheets suggests that a fairly careful advance estimate was made of the amount of space the printed text would occupy. However, there is no convincing evidence from type shortage to support the supposition that the whole Quarto was set by formes, the mislined passages are not confined to, or even especially prevalent on, pages where a compositor'r manœuverings to compensate for miscalculations in casting-off might be expected, and no other undue cramping of the text or wastage of space is apparent, though the occasional use of a blank line above or below a stage direction may represent an attempt at minor adjustment. Actually, Tourneur's frequent shifts from verse to prose and back again within a single scene might well have proved bewildering to compositors, especially in view of the bold enjambments of his verse, which often make it resemble chopped up prose, and there is much to be said for Nicoll's conjecture that revision, especially to D'Amville's role, was a source of confusion.[7]

Price and Murray both believe that the Quarto of The Atheist's Tragedy was set from Tourneur's holograph and point to orthographical links between The Atheist's Tragedy and Tourneur's non-dramatic works which suggest that Snodham's compositors were conservative in their handling of the forms in their manuscript copy. Murray further shows that while The Atheist's Tragedy shares certain significant spellings and forms of colloquial contractions with Tourneur's poems printed in other shops, Ben Jonson's The Alchemist (1612), the one play printed by Snodham within a year of The Atheist's Tragedy, exhibits spellings and colloquial forms differing from Tourneur's but characteristic of Jonson's late comedies in general. He concluded that Snodham's compositors were fairly faithful to the minutiae of their copy. We are now in a position to refine his argument somewhat, for of the two compositors who set The Atheist's Tragedy one, Compositor B, appears to have set the whole of The Alchemist. Semi-colons are again common, though less so than in The Atheist's Tragedy, and throughout the text they are preceded by a space. The single -e spelling is normal in pronouns (except in such forms as hee'll, wee'll, etc.). Spacing practices with respect to punctuation marks other than semi-colons agree with those found throughout The Atheist's Tragedy.[8] The orthographical differences between The Alchemist

and The Atheist's Tragedy (at least its second half) cannot, therefore be purely compositorial in origin.

More significant is the uniformity of the two parts of The Atheist's Tragedy with regard to orthographical features which distinguish Tourneur's play from The Revenger's Tragedy and Middleton's plays. Examining The Revenger's Tragedy, The Atheist's Tragedy, five plays by Middleton, five plays by other authors which were printed, like The Revenger's Tragedy, by George Eld in 1607-8, and also Snodham's Quarto of The Alchemist, Murray "determined both the frequency of occurrence and the spelling for all colloquial contractions of prepositions, articles, personal pronouns and verbs, and their combinations with one another" and also determined the frequency with which has, hath, does, and doth occur.[9] He undertook separate analyses for preferred linguistic forms and preferred spellings of them, presenting his data in five tables, the first of which records the spelling of the forms studied, w'are and we're, for example, being regarded as different forms of the contraction for we are. The upshot is that "in every one of the eighteen spelling conflicts between Tourneur and Middleton the pattern of spellings occurring in the RT is closer to Middleton's than to Tourneur's" (pp. 165-166).

The present compositorial analysis of The Atheist's Tragedy gives these findings added weight. With the exception of Ile/I'le, none of the eighteen forms is treated in a significantly different fashion by Snodham's two different compositors. Murray shows, for example, that the contracted form of it is appears in The Atheist's Tragedy as t'is 50 times, tis once, 'tis not at all. In The Revenger's Tragedy, and four representative plays by Middleton, the strong preference is for tis, with 'tis as the second most common form. Of the 50 examples of t'is in The Atheist's Tragedy exactly half were set by Compositor A, half by Compositor B. In The Alchemist Compositor B regularly uses 'tis (38 times, tis once)—the form which characterizes The Revenger's Tragedy and Middleton's plays; so presumably he set t'is in The Atheist's Tragedy because that form appeared in his copy. Murray shows that whereas the verb does always occurs in its modern spelling or as doe's in The Atheist's Tragedy (16 times), the normal spelling in The Revenger's Tragedy and in four Middleton plays is dos or do's. Of the 16 instances of does or doe's in The Atheist's Tragedy 9 are in Compositor A's stint, 7 in Compositor B's. Yet in The Alchemist Compositor B shows himself perfectly willing to set the form which prevails in The Revenger's Tragedy and the Middleton plays: do's occurs 19 times, dos 4 times, does and doe's once each. In The Atheist's Tragedy 'em is the only contracted spelling of them—12 times set by Compositor A, 6 times by Compositor B. In The Revenger's Tragedy and the Middleton plays e'm and em are also common. In The Alchemist Compositor

B invariably sets the Jonsonian 'hem—51 times. Even the Ile/I'le spellings, over which the two compositors differ in The Atheist's Tragedy, contribute to Murray's case, because the only alternative to Ile in The Revenger's Tragedy and Middleton, used about half as frequently, is ile. In The Alchemist I'll is normal (77 times), though Ile (16 times) and I'le (twice) also occur.

I have been referring to The Alchemist as the work of Compositor B more confidently, perhaps, than the evidence warrants. If the play was in fact set by Compositor A, the argument in support of Murray's case remains unaffected, since both men treated the forms cited by Murray in almost exactly the same fashion. Even if The Alchemist was set, wholly or in part, by a third compositor, the Jonson play at least affords evidence that the uniformity with which two compositors spelled colloquial contractions in The Atheist's Tragedy was not due to the imposition of some strict printing-house style. The natural inference is that Snodham's Compositor A and Compositor B both reproduced the linguistic minutiae of their manuscript copy with some care.

Notes