Volume 18 · Issue 4

begin page 216 | ↑ back to top

The Four Zoas: Intention and Production

by ROBERT N. ESSICK

Paul Mann’s challenging theory in “The Final State of The Four Zoas” offers an opportunity for others to test their sense of Blake’s plans for his long and various manuscript poem. Mann very kindly allowed me to read his essay before publication and accepted my suggestion that I send to this journal a companion piece in the hope that the two articles together will produce more light than would either in isolation. I wish merely to pursue a few lines of speculation that support his basic theory but evolve into a hypothesis about Blake’s first plans for his text.

I trust it has become something of a truism that Blake was much concerned with the processes and materials of publication. A man trained to the book and print business of his time who also invented a number of unconventional ways of producing and distributing images could hardly avoid such interests. A corollary proposition is that Blake’s works are media specific—that is, his images (whether pictorial or calligraphic) differ in some observable ways because of the different technologies used to produce them. Given the practical exigencies of image production, as well as Blake’s insistence on the radical unity of conception and execution,¹↤ ¹ See, for example, Blake’s statement that “Invention depends Altogether upon Execution or Organization” in his annotations (c. 1808) to Reynolds’ Discourses and similar statements in the Public Address (c. 1811). The best and most recent study of this central principle of Blake’s aesthetics is Morris Eaves, William Blake’s Theory of Art (Princeton: Princeton Univ. Press, 1982), Chap. III. it is reasonable to assume that the preliminary stages in the development of an image may reveal the medium in which Blake intended to produce the final form of that image. I take this to be the perspective, as well as the context in which we are to understand the meaning of “intention,” implicit in Mann’s essay. I share that perspective and usage, and thus I will also view The Four Zoas manuscript not only as the “text” of a “poem” (i.e., the physical embodiment of an aesthetic experience) but also as a preliminary—perhaps even a mockup—for a subsequent work never executed.

Is The Four Zoas manuscript the signifier of an unrealized intention to produce the poem in some other form? The physical condition of the work as we know it, with its deletions, erasures, marginal insertions, and designs rendered as pencil sketches, immediately suggests as much. Mann has argued cogently that Blake did not intend to create a single illuminated manuscript—either the one we have or another never executed—as an end in itself. Such works are generally produced on commission only, much as Blake was hired by Flaxman, Butts and others to make watercolored illustrations to the works of other poets, but it is difficult even to imagine a late eighteenth-century sponsor for an epic like Vala or The Four Zoas. This does not, of course, allow us to rule out completely the theory that Blake originally planned to create a beautiful illuminated manuscript as his final product and only later allowed it to degenerate into a working draft. Yet, as Mann suggests, Blake’s entire career as a publisher indicates a consistent desire to find a wider audience for his poetry than the very limited number of people who could see, read, and be moved by a unique manuscript. From the prospectus of October 1793 to the first chapter of Jerusalem (c. 1804-1820), Blake addressed himself “To the Public”²↤ ² The Complete Poetry and Prose of William Blake, Newly Revised Edition, ed. David V. Erdman (Berkeley and Los Angeles: Univ. of Calif. Press, 1982), pp. 145 (Jerusalem, pl. 3), 692 (1793 prospectus). and tried an extraordinary range of media and publication methods to reach them.

The development of Blake’s publishing efforts in the mid- and late-1790s provides some sense of the range of media he had available for the publication of his illustrated poetry. In 1795, Blake published The Book of Los and The Book of Ahania (illus. 4). Both have texts etched in intaglio, whereas the other illuminated books were printed from letters etched in relief.³↤ ³ The one exception discovered to date is the title page of The Song of Los (1795), which appears to have been printed planographically. See Essick, William Blake, Printmaker (Princeton: Princeton Univ. Press, 1980), pp. 128-29. This title page, like the evolution of the 1795 color printed drawings, offers further evidence of Blake’s movement away from relief etching in the mid-1790s. The visual differences between intaglio and relief letters are less significant than the considerable changes in production procedures. It is not at all clear why Blake abandoned his own invention, relief etching, which he trumpeted as both a practical and beautiful method of self-publication in his 1793 prospectus, and returned to a far more conventional medium.⁴↤ ⁴ There are a number of English eighteenth-century books with etched and/or engraved texts, including several illustrated editions of Philip Ayres, Emblems of Love. Several Continental publications of considerable length were completely engraved, including a Virgil in five volumes (The Hague: Henry Justice, 1753-1767) and an edition of Ovid’s Metamorphoses in 141 plates (Paris: Basan et Lemire,^[e] plates dated 1767-1770). The Books of Los and Ahania are also notable for their paucity of interlinear decoration, as in letterpress, even though the intaglio medium would not in itself prohibit the root and vine motifs punctuating The Book of Urizen or even larger and more elaborate designs set anywhere on the page. Clearly, Richard Edwards’ 1797 edition of Young’s Night Thoughts, with a typographic text surrounded by intaglio designs on slightly less than half the printed pages, bears a special relationship to The Four Zoas. But we should also keep in mind The Book of Los and The Book of Ahania, for they indicate the medium Blake chose for the production of his own poetry published closest in time to the beginning of his labors on Vala or The Four Zoas.⁵↤ ⁵ G.E. Bentley, Jr., Blake Books (Oxford: Clarendon Press, 1977), pp. 454-55, and Erdman, ed., Poetry and Prose of Blake, p. 817, agree that Blake worked on the manuscript c. 1796-1807. Blake Books summarizes the information Bentley first presented in his facsimile edition of Vala or The Four Zoas (Oxford: Clarendon Press, 1963) and repeated in his edition of William Blake’s Writings (Oxford: Clarendon Press, 1978), II, 1721-32.

Turning to the manuscript itself, we can immediately see that some pages are written in what Bentley calls a “Copperplate Hand” (illus. 1), others in a related begin page 210 | ↑ back to top

begin page 211 | ↑ back to top begin page 212 | ↑ back to top begin page 213 | ↑ back to top begin page 214 | ↑ back to top begin page 215 | ↑ back to top begin page 217 | ↑ back to top but “Modified Copperplate Hand,” and still others in Blake’s usual plain manuscript hand (illus. 2-3, 5-6).⁶↤ ⁶ Bentley, Blake Books, p. 455, where pp. 1-14, 17-18, 23-30 are listed as written in copperplate hand and pp. 15-16, 31-42 in modified copperplate. Erdman, ed., Poetry and Prose of Blake, p. 816, records pp. 1, 3-18, 23-42 as containing all the “fine copperplate script.” Thus the only discrepancy is p. 2, the verso of the titlepage, which bears only a pencil sketch and a rough pencil inscription. In what follows, however, I will include that page among the “fine script” sections since it clearly accompanies pages written in that hand. Our sense of an important structural distinction between the copperplate and plain hands is reinforced by stab holes indicating that Blake “had once sewed together into one group all the sheets with elegant script [i.e., both copperplate and modified copperplate] and into another group those sheets . . . in his usual [plain] hand.”⁷↤ ⁷ Erdman, ed., Poetry and Prose of Blake, p. 817. If we now consider the manuscript as a mockup, it is reasonable to begin with the assumption that the three hands may indicate at least two different intentions to publish. In what follows I will also assume that the deletions and marginal additions to pages bearing fine script, as well as designs that intrude into fine text areas (on page 25, for example), were made after the abandonment of whatever intentions for publication the copperplate hand(s) may embody.

Through a process of elimination and the weighing of practical probabilities, we can reach some conclusions about the intentions that may lurk behind those sections of The Four Zoas written in fine script. There would have been no reason to write in such a hand if the method of publication was to be either letterpress or relief etching. A typographic printer must of course select one of the available typefaces—or go to the great expense of having a new font case—regardless of the author’s manuscript hand. All that is wanted is a clear script, as in the Tiriel manuscript and large portions of The Four Zoas itself. As for relief etching, that is a direct process in which the forms of the letters are composed directly on the copper. Here again, the manuscript needs to be no more elaborate than a fair copy. Blake may have even avoided that stage of production by moving directly from a finished working draft, of the sort found in his Notebook, to the copperplate because corrections can be made easily on the plate at any time before acid is applied. The one extant mockup for pages in a published illuminated book is a rough pencil sketch with only horizontal lines to indicate the approximate position of the text.⁸↤ ⁸ A drawing (sheet 22.4 × 31.7 cm.) for The Book of Thel, pl. 6, and an alternative design for pl. 7, recently acquired by the British Museum. See Martin Butlin, The Paintings and Drawings of William Blake (New Haven and London: Yale Univ. Press, 1981), I, no. 218, and II, pl. 249. There is no extant manuscript for any of the illuminated books with lettering even vaguely like the copperplate hand(s) in The Four Zoas. This negative evidence cannot prove that such manuscripts never existed; but, given the production characteristics of relief etching, there are practical reasons why Blake need not have written such manuscripts. Even if Blake used a transfer method for applying his texts to the surface of the copper (which I very much doubt), the letters would have to be written on specially treated paper in a glutinous, acid-resistant material.⁹↤ ⁹ Such a method was invented by Ruthven Todd and W.S. Hayter and first described by Todd in 1948. See Todd, “The Techniques of William Blake’s Illuminated Printing,” reprinted in The Visionary Hand: Essays for the Study of William Blake’s Art and Aesthetics, ed. Essick (Los Angeles: Hennessey & Ingalls, 1973), pp. 35-36. No part of The Four Zoas manuscript meets these requirements and it is difficult to conceive how they could be added after writing had begun.

We are left then with intaglio publication of text and illustration. As practiced in Blake’s day, intaglio was an indirect process in that the image was generally developed to a fairly high level of completeness and finish in materials—usually pencil, ink, or wash on paper—preliminary to the graphic medium itself. This work was then transferred, often by mechanical or semimechanical means, to the copper. Blake describes a completely mechanical transfer process in his Notebook, but like most it requires an image executed in “black lead pencil”¹⁰↤ ¹⁰ “To Engrave on Pewter,” Poetry and Prose of Blake, ed. Erdman, p. 694. Joseph Viscomi has demonstrated that Blake used this type of procedure for transferring the drawing for Jerusalem pl. 26 to the copper for etching in white line; see Blake: An Illustrated Quarterly, 16 (1982), 46. or chalk. In a letter to George Cumberland of 6 December 1795, Blake outlines a soft-wax technique that works very well for transferring an intaglio print to copper for reengraving.¹¹↤ ¹¹ Poetry and Prose of Blake, ed. Erdman, p. 700. Blake’s method is very similar to the one described in Sculptura Historico-Technica: or, the History and Art of Engraving, Fourth Edition (London: J. Marks, 1770), pp. 99-100. Lines printed in intaglio are raised slightly above the surface of the paper, and thus it is possible to make an impression of them in a soft material. The copperplate hand in The Four Zoas is of course in ink on paper, and I suspect that even a soft-wax process would not permit its transfer.

If we must rule out mechanical transfer, the sections of The Four Zoas in copperplate hand could still serve a purpose for an intaglio etcher/engraver. After composing his script with some care on paper, Blake could look at the manuscript in a mirror and copy it backwards onto an etching ground or blank plate, following the manuscript letter-forms as closely as he wished within the limitations set by a nonmechanical method. This two-step process avoids the problems of composing (as distinct from copying) a handsome script backwards on a varnished plate or copper surface. Throughout the manuscript, Blake has consistently limited illustration to the margins and avoided interlinear designs on major pictures set between sections of the text on a single page. This format follows what we find in the intaglio plates of The Book of Los and The Book of Ahania. If the copperplate hand sections of The Four Zoas, before deletions and insertions, are in some sense a mockup for a publication, the most probable technology would seem to be a text etched and/or engraved in intaglio with surrounding intaglio designs. This method of production would allow Blake to produce the entire book using techniques in which he was thoroughly skilled and not require him, as would letterpress, to acquire the services of journeymen with tools and materials he did not himself possess.

Where would Blake acquire enough large copperplates to etch or engrave a poem anywhere near the length of Vala or The Four Zoas? Mann’s suggestion that Blake may have acquired the Night Thoughts coppers from Richard Edwards is compelling. It is certainly possible that Blake was paid partly in copper for his labors on Night Thoughts, regardless of who retained the rights to publish the images first etched on the plates. By writing much of his poem on unused leaves of Whatman paper also left over from the Night Thoughts project, Blake insured that the size of his text blocks and accompanying designs would fit with minimal adjustments begin page 218 | ↑ back to top onto the plates illustrating Young. We need not assume that Blake intended to use the engraved Night Thoughts designs themselves in The Four Zoas. The proofs used in the later parts of the manuscript may be there because of similarities in format—a text panel surrounded by designs—rather than symbolic or thematic conjunctions between Young’s poem and Blake’s. The presence of proofs from the plate printed on page 73 of Night Thoughts, showing Christ crowned with thorns, on pages 59 (illus. 3), 111, and 115 (illus. 5) of The Four Zoas creates an unusually repetitious sequence of illustrations. It is unlikely that the present arrangement of Night Thoughts designs in the manuscript represents the format for an intended publication of any sort. If Blake did have the Night Thoughts coppers, he could simply scrape and burnish away the designs to Young’s poem and etch or engrave new images. The backs of the plates could also be used, as in many of the illuminated books, for a total of eighty-six etchings and/or engravings—a good many more than the thirty-eight pages of The Four Zoas bearing fine script. Or, if one side only was to be used, perhaps it is significant that the number of fine script pages is just five short of the number of Night Thoughts coppers.

Pursuing this line of speculation a bit further, we should even consider the possibility that Blake abandoned the idea of using all the Night Thoughts plates for The Four Zoas between c. 1804 and 1807 and began using them for Milton and Jerusalem. By then the manuscript poem may have grown too long for the forty-three known Night Thoughts coppers,¹²↤ ¹² There appears to have been at least one further plate, for a “leaf from Night 5” (presumably illustrated) was offered for sale by the London bookdealer Francis Edwards in a catalogue of c. 1927-1928, item 44 (£26). There may of course have been other plates acquired by Blake or Richard Edwards in contemplation of publishing further Nights of Young’s poem. forcing on Blake some alternative publication plans if he had not already made them before extending the poem. Exact comparisons of the sizes of the Night Thoughts, Milton, and Jerusalem plates are not possible because of paper shrinkage and stretching and the unavoidable loss of some copper when a plate is cut and its edges beveled. It is certain, however, that the Night Thoughts plates could have supplied more than enough copper for both illuminated books. If pieces of copper were made from seventeen Night Thoughts plates by cutting along a horizontal line at a point 13.5 to 17 cm. from the top or bottom edge (ie., just short of the middle of the plate), each of the pieces could be cut into thirds to make three Milton coppers. If all seventeen pieces were cut this way they would provide a total of fifty-one plates for etching on one side.¹³↤ ¹³ We can assume that Jerusalem was etched on both sides of its copperplates because of paired platemark dimensions and the presence of platemakers’ marks. (Many Milton plates can also be paired, but I know of no evidence of platemakers’ marks). Each of the remaining, slightly larger pieces from these seventeen Night Thoughts plates could be cut in half to yield two Jerusalem coppers each, for a total of sixty-eight pages in the illuminated book if etched recto and verso. This procedure would require no further cutting, except possibly for minor trimming or filing of edges, and would waste very little copper.¹⁴↤ ¹⁴ Bentley, Blake Books, p. 225 n. 2, suggests that the Jerusalem plates could have been made from the text panels of the Night Thoughts coppers. This procedure would have wasted a good deal of copper, leaving large but awkwardly shaped and probably useless fragments. Other Jerusalem plates could have been cut from the remaining Night Thoughts coppers, although we know that at least plate 96 of the illuminated book was made from a quarter-piece of the “Moore & Co’s Advertisement” Blake designed and engraved c. 1797-1798.¹⁵↤ ¹⁵ The telltale fragments of white lines from the Moore engraving were first pointed out and identified by Erdman in “The Suppressed and Altered Passages in Blake’s Jerusalem,” Studies in Bibliography, 17 (1964), 36-37. Other plates in Jerusalem may have been made from other pieces of the Moore copper—or for that matter from any used intaglio plate. The presence of the Moore plate in Jerusalem indicates that Blake may have retained possession of similar commercial engravings, clearly executed on commission, and reused them for his own productions. The possibility that Blake could have used Night Thoughts coppers for Milton and Jerusalem lends some (admittedly circumstantial and speculative) support to Mann’s theory that Blake might have retained those very plates and once contemplated using them for Vala or The Four Zoas.

The plain-hand portions of The Four Zoas offer even less evidence than pages in copperplate hand concerning methods of publication. The intended medium might have been relief etching (although that still seems unlikely), intaglio etching/engraving, or letterpress. It is most tempting to conclude with Mann that the change from fine scripts to plain might record a change in publishing plans, roughly parallel in chronology to the shift from an unrecoverable poem entitled Vala to an unfinished poem entitled The Four Zoas.¹⁶↤ ¹⁶ Bentley accepts the usual assumption that the copperplate-hand portions were composed before the plain-hand pages and states that a reversal of this sequence “cannot . . . satisfactorily account for such diverse factors as symbolism, stab holes, corrections, and handwriting” (Blake Books, p. 455 and n. 5). Erdman notes that “on certain copperplate pages a distinctly late style appears in the script, marked by the g which Blake adopted after Nov. 1802” and states that “it is safer to conclude that the copperplate pages range in inception from 1797 to 1803” (Poetry and Prose of Blake, p. 817). Yet even if some of the fine script was written as late as 1803, the sequence of copperplate hand to plain is not thereby upset. The contextual and historical evidence Mann assembles in support of his letterpress hypothesis comes from what is generally taken to be a middle or later period in the development of the manuscript—the years with Hayley, 1800-1803. Blake’s work with the provincial printer Joseph Seagrave on a book (the 1802 Ballads) with a typographic text and intaglio illustrations may have suggested a return to that combination of techniques, also used in Edwards’ Night Thoughts edition, for the publication of The Four Zoas. Such a change could account for the appearance of Night Thoughts proofs in the plain script pages of the manuscript and the use of their text panels as a mockup for a projected publication with the same format. At this point, Blake could have cut up some of the Night Thoughts coppers for other purposes if the new plans for The Four Zoas required fewer than one design per page of text, again as in Night Thoughts. The first Night Thoughts proof appears on page 43 of the manuscript (illus. 2), immediately following the last page in fine script. If the plain-hand revisions to pages first composed in copperplate script are later revisions, as seems probable, the differences between pages 42 and 43 may record a significant shift in Blake’s intentions for publication.¹⁷↤ ¹⁷ Bentley places page 42 as the last in a group he dates 1797 and page 43 (illus. 2) the first in a group he dates “?1802” (but see Erdman’s statements on the date of the fine script in note 16, above). These are also the last and first pages in groups once separately sewn, as the two distinct patterns of stab-holes indicate (see Bentley, Blake Books, pp. 454-56).

As attractive as Mann’s theory—or the combination Mann/Essick theory—may at first seem, the manuscript pages on Night Thoughts proofs exhibit a few recalcitrant features. The number of lines written in the text panels ranges between twenty-eight and fifty-seven (not including pages bearing part-titles or those on which Nights end short of the lower margin), the text begins above the panel on fourteen pages (illus. 6), and extends below it on two (illus. 5). Accommodating marginal additions would cause further dislocations. If these portions of the begin page 219 | ↑ back to top manuscript are a letterpress mockup, they form a rather loose one. Blake may have planned to create somewhat larger text panels in a variety of sizes. If the panels were not altered, a sympathetic printer might be able to manage difficulties by shifting the text about as required, although this would probably disrupt some text-design relationships. A minor form of this problem occurred in the production of the 1797 Night Thoughts, with the result that the lines illustrated by the design on page 54 begin on page 53.

The treatment of the Night Thoughts proof on page 137 of The Four Zoas (illus. 6) throws into doubt at least two of our theories. Blake has drawn-in the back of one of the wrestling figures right through the text panel. The text carefully follows the new lower left margin created by the outline of the back. If this drawing was composed as part of the development of The Four Zoas (as distinct from a drawing executed earlier for some other purpose), it strongly suggests that at least this one Night Thoughts design was to be used in Blake’s poem. At the same time, the drawing argues against a letterpress text. It would not have been impossible to adjust the left margin of a typographic text according to the pencil line of the back, but such a format would have been highly unusual for an early nineteenth-century letterpress book. An etched or engraved text, however, would have accommodated the drawing without technical difficulties or the violation of printing conventions. On page 72 of the Edwards’ Night Thoughts, a figure’s head intrudes slightly into the lower right corner of the text panel, but this in no way disrupts normal letterpress margins. In spite of his introduction of printed designs with rigid text panels, Blake’s extensions and revisions of The Four Zoas forced the manuscript ever further from a recognizable mockup.

At some point Blake must have given up any plans for eventual publication, and perhaps that occurred before he ceased fussing with the manuscript. If this is indeed the case, then the suggestions presented here can be summarized as a three-step process of changing plans for a manuscript reflecting those changes as follows:

1. (1) a manuscript and production mockup for a poem (Vala?) intended for publication of text and designs as intaglio etchings/engravings (i.e., the Blake of Los and Book of Ahania format), tentatively dated c. 1796-1800;

2. (2) a manuscript and production mockup for a poem (The Four Zoas?) intended for publication as a letterpress text accompanied by intaglio etched and/or engraved designs surrounding selected pages of text (i.e., the Night Thoughts format), tentatively dated c. 1800-1804;

3. (3) a working manuscript unrelated to any specific publication intentions (i.e., The Four Zoas manuscript as we know it today), tentatively dated c. 1804-1807.

The foregoing has clearly proven only one proposition: it is impossible to prove any theory about Blake’s production plans for The Four Zoas. Yet I believe that these companion essays may contribute to a better understanding of the manuscript and the intentions it never completely reveals. Our speculative arguments may have the additional benefit of providing a more comprehensive view of the evolution of Blake’s activities as a poet, engraver, and printer from 1796 to 1807.