 | |  |
Compositor Determination in the First Folio
King
Lear
by
I. B. Cauthen, Jr.
The methods of determining the compositors of the Shakespeare First Folio (1623) by spelling tests have become well-known to bibliographers. The first suggestion for such a determination was advanced in 1920 by Thomas Satchell, who noted that orthographically the Folio Macbeth fell in two distinct divisions, (1) from the beginning of the play to III.iii (216-col. a, 2m5), and (2) from III.iv to the end of the play (col. b, 2m5-2n4).[1] With the exception of one page (2m2), the first portion of the play contains the forms doe and goe, whereas the latter portion uses the forms do (or doo) and go. Mr. Satchell supports this division of the play by a list of thirty-three other words which are consistent in each half of the play, but differ in the two halves.[2] He was unable to determine whether these findings were best explained by supposing that two compositors worked on the play or that two copyists worked on the manuscript from which the play was printed.
It remained for E. E. Willoughby in The Printing of the First Folio (1932) p. 56ff. to demonstrate from the printing of other plays in the Folio that Mr. Satchell's conjecture that two compositors were used was correct. Using the doe/do and goe/go forms as a basis, he selected three other "especially significant words not peculiar to Macbeth": cousin/cosin (or cosine, cozen), here/heere, and traytor/traitor. Using these five words, Mr. Willoughby confirmed Mr. Satchell's division of the play: one compositor, whom he labelled "A," set from the beginning of the play through
When we apply Mr. Willoughby's test to King Lear in the Folio, it could appear at first sight that the unknown compositors were at work, as the following distribution of variants shows:
| 2q2 | 2q2v | 2q3 | 2q3v | 2q4 | 2q4v | 2q5 | 2q5v | 2q6 | 2q6v | 2r1 | 2r1v | 2r2 | ||
| doe | 1 | 1 | 1 | 1 | ||||||||||
| goe | 1 | 2 | 4 | 1 | ||||||||||
| here | 2 | 3 | 1 | 4 | 2 | 2 | 2 | |||||||
| traytor | ||||||||||||||
| do | 1 | 3 | 3 | 4 | 5 | 4 | 3 | 2 | 3 | 4 | 4 | 6 | ||
| go | 1 | 2 | 2 | 3 | 1 | 1 | 3 | |||||||
| heere | 2 | 2 | 1 | 2 | 3 | 3 | 1 | |||||||
| traitor | ||||||||||||||
| 2r2v | 2r3 | 2r3v | 2r4 | 2r4v | 2r5 | 2r5v | 2r6 | 2r6v | 2s1 | 2s1v | 2s2 | 2s2v | 2s3 | |
| doe | 1 | 1 | ||||||||||||
| goe | 3 | 1 | 1 | |||||||||||
| here | 2 | 3 | 1 | 1 | 2 | 1 | 2 | |||||||
| traytor | ||||||||||||||
| do | 4 | 2 | 1 | 3 | 4 | 3 | 9 | 6 | 2 | 10 | 1 | 6 | 5 | 2 |
| go | 2 | 3 | 5 | 3 | 1 | 1 | 3 | 1 | 3 | 2 | 3 | 1 | 1 | 1 |
| heere | 3 | 2 | 2 | 2 | 1 | 1 | 3 | 2 | 2 | 5 | 3 | 1 | 1 | |
| traitor | 2 | 5 | 1 | 1 | 2 | 3 | 1 | 1 |
Here it appears that compositor B was responsible for pages 2q2, 2q3v, 2q4, 2q5, 2r1, 2r2, 2r2v, 2r3v, 2r4, 2r4v, 2r5v, 2r6, 2r6v, 2s1, 2s1v, 2s2, 2s2v, and 2s3, where the proportion is four or more characteristic B-forms to one characteristic A-form. But in the remaining nine pages (2q2v, 2q3, 2q4v, 2q5v, 2q6, 2q6v, 2r1v, 2r3, and 2r5) we have a higher proportion of A-forms,
Another test for the determination of compositors is that evolved by Charlton Hinman, who published an account of his technique in 1940.[3] Briefly, the test gathers evidence from significant variant spellings and, by an appeal to "coincidence of position," determines the compositor who is at work. But an examination of 600 variant spellings of 142 words in the Folio text of Lear failed to bring to light any demonstrable two-compositor pattern. Indeed, from all the evidence derived from this test it appeared most likely that only one compositor was at work during the setting of the play.[4] Because characteristic forms do not appear "on distinct groups of pages" we are justified in evaluating this evidence as indicating only one compositor.
If only one compositor is responsible for Lear, the next question concerns his identity. Since the last part of the play can be shown by Mr. Willoughby's test most probably to have been set by compositor B, it follows that the whole play was set by him. However, this simple test needs confirmation. Fortunately, several additional pieces of evidence can be advanced to identify the compositor of Lear with the compositor B of the rest of the Folio. Although individually these details are not irrefutable in themselves, the cumulative weight of them all shows conclusively that this compositor is B. This evidence can be drawn from two sources, the plays surrounding Lear and from Lear itself.
First, evidence from the simple compositor-determination test shows clearly that the plays surrounding Lear were set by identifiable compositors. The following table shows the compositorial assignments in these neighboring plays:
| Timon (2g3v-2h5v) | Compositor B: 2g3v-2h5v | 21 pages |
| Julius Caesar (2k1-2l5v) | B: 2k1-2k2 | 3 |
| A: 2k2v-2k3v | 3 | |
| B: 2k4-2l5v | 16 | |
| Macbeth (2l6-2n4) | A: 2l6-2m1v | 4 |
| B: 2m2 | 1 | |
| A: 2m2v-col. a, 2m5 | 5½ | |
| B: col. b, 2m5-2n4 | 10½ | |
| Hamlet (2n4-2q1v) | B: 2n4v-2n5v | 3 |
| A: 2n6-202 | 5 | |
| B: 202v-203 | 2 | |
| A: 203v | 1 | |
| B: 204-2p4 | 13 | |
| A: 2p4v-col. a, 2p6v | 4½ | |
| B: col. b, 2p6v-2q1v | 2½ | |
| Lear (2q2-2s3) | -- | |
| Othello (2s3v-2v6) | B: 2s3v-2v6 | 30 |
| Antony and Cleopatra (2v6v-2z2v | B: 2v6v-272v | 29 |
| Cymbeline (2z3-3b6) | B: 2z3-3b6 | 31 |
Hamlet, then, ends with compositor B setting type; Othello opens with the same compositor at work; Lear intervenes. It was not a practice of the Folio compositors to change at the beginning of a play; they set by pages rather than by act, scene, or play. Their usual habit was to continue from one play to another unless they were at the end of a gathering or a section. There are only two demonstrable changes of compositors at the beginning of a play that is not also the beginning of a new gathering or section: at the beginning of 1 Henry IV, compositor A takes over from B (04), and at the beginning of Macbeth, compositor A again takes over from B (2l6); against these two instances we have at least fifteen occasions when the compositor who set the latter part of a play continued to set the beginning of the following play.[5] It is not impossible but indeed more likely that compositor B, who set the last pages of Hamlet, should continue to set the beginning pages of Lear.
Nor must one believe it would be impossible for one compositor to set a whole play. Compositor A, early in the setting of the Folio, had set Two Gentlemen of Verona in its entirety, and compositor B had set all of Twelfth Night and Timon. The situation of a compositor setting four plays consecutively is unique in the Folio but not unlikely.
From this evidence, found in plays surrounding Lear, one notes that in the four plays preceding Lear compositor B set 72 pages to A's 23 pages. Since B set more and more pages until finally he was setting three plays alone, it seems not unlikely that the pattern which is clearly established in the plays of Othello, Antony, and Cymbeline had its beginning in Lear. All of the large groups of pages from Timon through Hamlet are by compositor B: he set 24 pages (2g3v-2h5v) in his largest group. Compositor A, on the other hand, set only 5½ pages in his largest group (2m2v-col. a, 2m5). Certainly the preponderantly large groups set by any one compositor in this section are those by compositor B.
Moreover, in Lear we find only one skeleton-forme present: if we had found two skeletons in use here our conclusion that only one compositor was setting type might well be suspect. Generally speaking, if we find two skeletons in use, we may find two compositors at work.[6] If we find only one skeleton in use, we may have either one or two compositors at work, but certainly, except in very special circumstances, two skeletons more likely may be equated with the work of two compositors than with one, especially in folio printing. It would be impossible for one compositor to maintain a speed in composing, rinsing, distributing, and attending to his other duties that would allow him to keep abreast of the press in a manner necessitating the use of two skeletons; and where we find one skeleton in use we are not forced to believe that two compositors were at work. Thus if we find evidence elsewhere of only one compositor at work, we need not consider that it would be impossible for one compositor to set the whole play.
More important, however, than these technical points is the evidence that can be drawn from within the play itself. The detailed examination of the variant spellings within the play, following the method evolved by Mr. Hinman, revealed certain words where the compositor who set Lear had a very marked preference for certain spellings. To confirm these preferences, a similar spelling test was undertaken for Antony and Cleopatra, and Othello, plays definitely set by compositor B. For comparison with these spellings, an examination of compositor A's portions of Macbeth and Hamlet disclosed truly preferential spellings of both compositors. This comparison not only revealed that Lear was set by one compositor but furnished
| A Spelling | B Spelling | Lear Spelling |
| deare (17) | deare (0) | deare (2) |
| deere (0) | deere (59) | deere (26) |
| does (17) | does (31) | does (1) |
| doe's (1) | doe's (2) | doe's (1) |
| do's (1) | do's (25) | do's (15) |
| beene (13) | beene (4) | beene (8) |
| bene (3) | bene (6) | bene (1) |
| bin (3) | bin (24) | bin (19) |
| deuil (5) | deuil (2) | deuil (0) |
| diuel (0) | diuel (34) | diuel (2) |
| sirra(h) (0) | sirra(h) (5) | sirrah (8) |
| sirrha (1) | sirrha (0) | sirha (1) |
| sirra (0) | sirra (0) | sirra (1) |
| houre (9) | houre (28) | houre (9) |
| hower (3) | hower (0) | hower (1) |
| howre (0) | howre (1) | howre (1) |
| vilde (0) | vilde (9) | vilde (4) |
| vile (2) | vile (0) | vile (0) |
| vild or vil'd (2) | ||
| power (3) | power (18) | power (9) |
| powre (0) | powre (7) | powre (8) |
| breefe (0) | breefe (4) | breefe (2) |
| briefe (2) | briefe (2) | briefe (1) |
| yong (0) | yong (15) | yong (10) |
| young (2) | young (6) | young (5) |
| blood (16) | blood (31) | blood (12) |
| bloud (5) | bloud (3) | bloud (1) |
The very close correspondence between compositor B's preferential spellings and those found in Lear indicate clearly that the tentative assumption made at the first in the investigation was correct: all of Lear was set by compositor B. Where the simpler test of Mr. Willoughby failed to reveal the compositor of a third of the play, the more detailed examination of all the significant variants established the compositor of all the play.
But there remains one further point that may be suggested as accounting for certain minor divergences in compositor B's preferential spellings in Lear. In the first place, we must not expect a compositor to spell certain words in an invariable way. Mr. Willoughby has already pointed out the possibility of a compositor's varying the spelling of words, and Mr. Hinman has enlarged upon that possibility: occasional copy-spellings will come through; justifying of lines will affect preferential spellings; the diminishing of a certain sort usually employed in preferential spellings may lead him to varying his spellings; and insistence, either by editor or author, upon copy-spellings for rhyme words, or even a compositor's own idiosyncrasies, may cause variants. Finally, Mr. Hinman concludes,
Perhaps one of the factors that cause variation from preferential spellings should be emphasized here. In the plays where Mr. Willoughby's test works best, the compositor is setting from manuscript. Here in Lear, as in Midsummer Night's Dream, Merchant of Venice, and Romeo and Juliet where Mr. Willoughby conjectured unknown compositors, the play was being set from printed copy. It is probable that the printed copy has much more effect upon a compositor's spelling than has been previously recognized. It may be that if a compositor meets a word in manuscript he will set that word in his own spelling far more often than if he meets that word in print. Although this is a difficult point to prove, we must remain aware that part-muscular influences, coupled with a respect for the printed word, may be at work on the compositor when he is setting type from a printed page which is acting as copy. We may have, therefore, a situation where a compositor meeting "do" in a manuscript will impose upon it his usual spelling; but if he meets "do" in a printed text, he may accept readily the spelling that is there, either do or doe or doo. This influence of the printed word can be demonstrated in other ways, such as the use of italics, the following of unusual spellings, the printing of erroneously abbreviated speech-headings and the like, which have long been noticed by bibliographers. Thus where we have a play being reprinted from a previously
The principle that one must follow here in determining the compositor of a play set from printed copy is this: he must examine the spellings of the copy as well as those of the reprint, noting what changes have been made in the spellings. If a doe spelling is retained, it does nothing towards indicating the compositor; if, however, it is changed to do without some other factor such as line-justification intervening, we may note that as a very definite indication of the compositor. This principle proved successful in the compositor determination of Lear. [9]
But whatever the value of this principle, it is apparent, I believe, that Lear was set by one compositor, B, of the Folio. And this information is no academic exercise in a theoretical desert. With the identification of the compositor secure, we can undertake an examination of compositor B's work elsewhere in the Folio where he was setting type from printed copy; then, and only then, can we produce demonstrable evidence about his characteristic handling of copy. And that, a longer story, must be reserved for the present; but in spite of its length, it is an important one for those who wish a text of Lear that is as close as possible to Shakespeare's "fair copy." For the compositor of the Folio text of Lear did make certain characteristic changes in his text; and whatever the authority of the "Pied Bull" Quarto, it has more authority than the changes made by a Jacobean workman, compositor B of the Folio.
Notes
Seventeen of these words occur but once in each part, and therefore they may be dismissed as of little significance; seven others differ only by the doubling of a final consonant (eternal/eternall) or of a medial one (rubs/rubbes), but at times the evidence here is contradictory. Nine words, however, appear to be significant: afraid/affraid, countreyes/countries, deare/deere and dearest/deerest, maiestie/maiesty, thick/thicke, traytor / traitor, weyward / weyard, and beene/bin.
"Principles Governing the Use of Variant Spellings as Evidence of Alternate Setting by Two Compositors," The Library, 4th ser., XXI (1940), 78-94.
Another test, which confirmed these findings, is that of A. K. McIlwraith: see "Some Bibliographical Notes on Massinger," The Library, 4th series, XI (1930), 87-91. This test deals only with scene headings, stage directions, and speech prefixes which often are set in characteristic ways by compositors, but here again no demonstrable two-compositor pattern emerged.
See, for example, the continuity of compositor between The Tempest and Two Gentlemen; Two Gentlemen-Merry Wives; Measure for Measure-Errors; As You Like It-Taming of the Shrew; Shrew-All's Well; All's Well-Twelfth Night; King John-Richard II; Richard II-1 Henry IV; 1 Henry IV-2 Henry IV; Henry V-1 Henry VI; 1 Henry VI-2 Henry VI; 3 Henry VI-Richard III; Richard III-Henry VIII; see also table above for further examples. In each of these instances the compositor who was setting at the end of a play continued to set at the beginning of the following play.
I realize the exceptions that occur with such a generalization. For instance, the Q Lear (1608) was machined with three skeletons; yet only one compositor was at work: see Fredson Bowers, "An Examination of the Method of Proof Correction in Lear," The Library, 5th ser., II (1947), 20-44; and Philip Williams, "The Compositor of the 'Pied-Bull' Lear," Studies in Bibliography, II (1949), 61-68. But this is a very unusual occurrence. I hope to present in the near future a detailed examination of the use of one or two skeletons during the printing of the Folio.
Incidentally, this examination confirmed in certain particulars the conjectured preferential spellings advanced by Mr. Satchell. But because it was much more inclusive than his restricted area of one play, it furnished much more solid evidence that he could give at that time.
In order to demonstrate the validity of this hypothesis, I examined ten plays in the Folio that were set from printed copy. In six plays the compositor is not determined by the five-word test: Much Ado, Love's Labor's Lost, Midsummer Night's Dream, Merchant of Venice, Titus, and Romeo and Juliet. Two of the plays that underwent extensive revisions, Richard III and Lear, have partly determined compositors. Only two of these plays, Richard II and Troilus, have definitely determinable compositors.
This method of examination may yield some evidence concerning the compositor(s) of certain Shakespeare plays in the Folio that now are classed as by "unknown compositor(s)." If it does not identify the compositor(s), it at least will exclude the possibility that those plays have been set by the known compositors, A and B.
| | |