University of Virginia Library

Variation, Accident, and Intention in William Blake's The Book of Urizen
by
Robert N. Essick

The relationship between the interpretation of a printed image and its means of production takes on unique twists and turns in William Blake's illuminated books. Because they were etched in relief and white line on copperplates, the printing bodies of these works have the stability and resistance to revision of stereotype plates.[1] In contrast, the exigencies of relief inking and printing and Blake's tendency to add hand work to individual impressions introduce a great many variations into the process. He was also in the habit of arranging his plates in different sequences and producing apparently complete copies of his books with fewer plates than others. A good case in point is The [First] Book of Urizen, etched in 1794 and known in seven copies with between 24 and 28 plates printed on rectos only.[2] Six of these copies (A-F) are color printed. This technique, as well as the 1794 watermarks in copies B and D, strongly suggest that all six were printed ca. 1794-1796. The last copy printed (G), on paper watermarked 1815, is the only one colored entirely by hand rather than a combination of color printing and subsequent work on the impressions with pen and brush. Only two copies (A-B) have all 28 plates; no two copies have the plates arranged in the same order. Copies D and F have the same number of plates, but both lack plates 4 and 16, the latter a full-page design the absence of which does not affect the narrative sequence. Copy C also lacks plate 16, contains a second impression of plate 2, and is without two text plates, 7 and 8. Both copies B and C contain plates printed in two colors of ink, suggesting that they are composites of two press runs. The inclusion of plates 8 and 10 in all but copy C adds a further anomaly, for the text on both these plates begins a chapter headed "IV."[3] Copy E lacks plate 4, three full-page designs, and the

final plate, the absence of which might be explained by some accident, or action taken by a binder, that altered the original composition of the volume. The last copy (G) is unique in having 27 plates, with only plate 4 absent. As in copy D in 26 plates, Blake foliated copy G in one continuous numbered sequence, thereby indicating that he considered it a complete copy of the book.

Among these many major variations, the absence of plate 4 from four copies, two of which are continuously foliated in Blake's hand, is the most difficult to explain. That plate includes the end of verse 3 of Chap. II, begun on plate 2, five additional verses from the second chapter, the heading for Chap. III, and one and a half following verses (see Illustration). Without plate 4, the incomplete sentence on plate 3 can be completed grammatically (and perhaps even narratively[4]) by the first two lines of text on plate 5, but the verse numbering becomes redundant with a verse "3" indicated at the last line on plate 3 and again at the third line of plate 5. Further, we are left without a heading for Chap. III until the running head of "Urizen C: III" above verses numbered 9 through 11 on plate 6. Unless these inconsistencies are rectified in some way, the reader knows immediately that something is missing.

One might surmise that plate 4 contained, in its text or design, something at such variance from Blake's final intentions for Urizen that he felt compelled to leave it out of later copies (perhaps D-F, certainly G) of the book—in spite of the resulting disruptions in his chapter and verse numbers. Much of the text on plate 4 is devoted to Urizen's only speech in the poem. It has generally been taken to be a central passage because it introduces important images of Urizen's self-alienation from the Eternals, his "books formd of metals," and his desire to impose his "one command" on all existence. The design at the bottom of the plate, showing a man seated behind a curtain of dark rain or flames and grasping his hair with both hands, is completely consistent with other tormented figures portrayed in the book, including the full-page design (plate 12) of Urizen submerged in water which may illustrate lines 21-23 on plate 4. In short, much in the text can be considered essential, not peripheral, and nothing in the design leaps to one's attention as a reason for cancellation. Further, in the list of publications Blake sent Dawson Turner on 9 June 1818, he described "Urizen" as having "28 Prints."[5] If Blake's statement in this letter, written long after the initial etching and printing of Urizen, represents his final intentions for the book, then it seems

reasonable to assume that copies with fewer than the full complement of 28 plates may have been printed earliest, before all the plates had been etched. This line of reasoning apparently stands behind Bentley's belief that the absence of plates 4 and 16 from copies D-F indicates that they were printed before copies A-B with all 28 plates. But if the presence of 28 plates indicates Blake's later, perhaps even final, intentions for the completion of the book, why does copy G—the most highly finished and the last printed—lack plate 4?

A previously recorded, but generally ignored, loose impression of Urizen plate 4 provides some answers to this question (see Illustration). More importantly, its unique characteristics offer insights into Blake's working methods and the ways they influenced the contents of his books and our interpretations of them.

In the "Addenda" to Blake Books, p. 953, Bentley lists an impression of Urizen plate 4 in "An Anonymous Collection" and suggests that it was "perhaps . . . intended for copy G," now in the Rosenwald Collection of the Library of Congress.[6] Several features of this print, which I acquired privately in February 1984, indicate that it was almost certainly printed by Blake as part of copy G but rejected by him just before the final stages of production. The paper is the same type of off-white wove found in copy G. Like twenty leaves in that volume, this impression of plate 4 shows no watermark. The pumpkin-orange ink is of the same color and grainy texture as in copy G.[7] The only other extant Urizen plate printed in a similar color is a single impression of plate 1 among a group of three unfinished proofs (Beinecke Library, Yale University). That plate, however, is color printed, a medium which immediately distinguishes it from the separate impression of plate 4 and from copy G. In both, the text areas are very carefully printed but the designs are rather thinly and unevenly inked, as is typical of Blake's nineteenth-century relief prints he intended to color extensively by hand. The hand coloring in blue, orange, and reddish-orange on plate 4 matches those colors on several plates in copy G. Most importantly, the separate print and all plates in copy G are surrounded by orange water color over pencil framing lines between 5 and 9 mm. from each side of the plate image. No other extant Urizen prints have framing lines of any sort.

One feature of the separate impression of plate 4 distinguishes it from copy G and indicates why it was not included with its companions. The plate image is not properly aligned relative to the sides of the sheet. Although this problem is difficult to see in a monochrome reproduction, the etched margins in the original clearly show that the printed image is tilted about 5° to the right. The hand coloring Blake added around the periphery

of the plate constitutes a rather awkward attempt to correct the improper alignment by painting in new margins. The framing line parallels these hand-painted (as distinct from printed) margins. At some point prior to completing the coloring of the design and numbering the plate, Blake apparently decided that his salvage operation was unsuccessful. He then numbered consecutively the other impressions pulled for what thereby became copy G, complete in 27 plates, and omitted this or any other impression of plate 4. In its stead, Blake placed the full-page design of plate 9 and numbered it "4." To rectify the resulting inconsistency in verse numbers, he scratched out, and drew a tendril over, the "3" preceding the last line of text on plate 3 so that the same verse number on plate 5 (the next text plate) is not redundant. This, however, did not correct the problem with chapter numbers, and copy G has no indication of a Chap. III until the running head above verses 9-11 on plate 6, placed and numbered as leaf "7" in copy G following another full-page design.

Misalignment is a common problem when printing with an engraver's rolling press.[8] The solution is generally simple enough: either the plate is immediately reprinted, or the edges of the sheet are trimmed parallel to the margins of the plate image. It is evident that Blake did not take the first of these options in the production of copy G. Perhaps he did not notice the error until after he had disposed of his ink, used all available paper, or altered his press—and thus believed it would be quickest and most economical to attempt to rectify the problem on the misaligned impression itself. But why did Blake not exercise the second option and simply trim the edges of the sheet? The only reason for not doing so is that he must have determined the leaf size for his book, and cut at least plate 4 to that size, either prior to printing that plate or, at the very least, prior to discovering its misalignment. Any cutting of the plate would have created a leaf smaller than all others planned for the book. Oddly enough, that physical inconsistency appears to have been less acceptable to Blake than the cancellation of plate 4, at least after he had disrupted its clear printing with attempted repairs.

As Bentley notes, the leaves of copy G of Urizen have been "trimmed" to 28.9 x 23 cm. and "gilt."[9] The separate impression of plate 4 is on a sheet 29 x 22.8 cm. and shows no evidence of having ever been gilt or bound with copy G. None of the edges of the sheet are deckled, and it is very likely that Blake printed plate 4, like all plates now in copy G, on a larger sheet. It may be mere happenstance that the dimensions are so close as to suggest that

the plate 4 sheet was at some point cut to the same size as copy G and only subsequent shrinking and stretching, or minor trimming during the gilding process, led to the slight differences. But we need not invent such scenarios, or make any assumptions about original sheet sizes, to account for Blake's failure to realign plate 4 by trimming the edges of the sheet. The reduction of the leaf from a predetermined size to which plate 4 had already been cut is an unavoidable consequence of alignment by trimming no matter what the original size of any of the sheets on which copy G was printed. Nor would the basic circumstances affecting Blake's treatment of plate 4 have been any different if he had printed more than one Urizen plate on each sheet, as distinct from printing each on a separate leaf cut to size.[10] In either case, only a predetermined leaf size would have prevented Blake from trimming plate 4 and instead prompted him to attempt alternative measures with pen and brush.

The separate impression of plate 4 indicates that Blake originally intended to produce copy G of Urizen in 28 plates, but that a simple error in alignment and Blake's inability to correct it successfully led to the cancellation of a plate bearing 10% of his text. The graphic processes Blake used in his illuminated books were inherently prone to countless minor, unintended variations in inking and printing, but the final form of copy G demonstrates the extent to which accident shaped his texts. It is difficult to imagine that any conventional printer or publisher would allow mechanical accident to play such a large role in selecting the verbal content of a book. Similarly, a conventional understanding of the relationships between meaning and the means of its (re)production would lead us to entertain only conceptual and fully intentional reasons for the major textual variations among copies of The Book of Urizen. Yet Blake's own aesthetic doctrine of the unity of conception and execution, when understood in its most radically literal sense, offers a rationale for the impact of the material and mechanical on his art. In his Public Address of ca. 1810, Blake insists that "Execution is only the result of Invention."[11] Urizen copy G demonstrates that the reverse is also true, and that the means of execution can shape the inventions of the poet. Because Blake the author was his own illustrator, etcher, printer, and bookseller, his daily labors offered the means for a free intermingling of mental creation and material production, each generating the other as a practical corollary to his aesthetic theories.

In light of Urizen copy G and its cancelled plate 4, we should be cautious in our assumptions about the reasons for other major variations in the formats of Blake's books. Those reasons may have as much to do with the contingencies of Blake's presswork as his transcendental ideas. Indeed, to his way of thinking there is no absolute distinction between those realms of act and

thought. The format of Urizen copy G and the way it came about dramatize a truth about the potential impact of material production on all texts, a truth of central concern to the bibliographer.

Notes