article
begin page 192 | ↑ back to topThe Changing Order of Plates in Jerusalem, Chapter II
As everyone knows who has ever gone from a text of Jerusalem edited by Keynes to one edited by Erdman, the second chapter of the work exists in two arrangements of plates.*↤ I wish to express my gratitude to Morton D. Paley, David V. Erdman and G.E. Bentley, Jr., for stimulating and enlightening exchanges on the subject treated in this essay and for other generosities during its preparation. Certain inconveniences have attended this fact—a clutter of bracketed numbers to designate the variant plates, confusions in the classroom when students show up with the differing texts, difficulties for readers of the Doubleday edition when they use the Concordance and Damon’s Dictionary. More troubling, the existence of these widely differing arrangements tends to feed the still common suspicion that the order of Jerusalem is arbitrary, capricious, and perhaps incoherent. Blake criticism, however, has generally shied away from this textual problem, troubling or not. It remains, in the words of one important critic, “a vexing question that deserves a complete analysis never accorded it.”1↤ 1 Stuart Curran, “The Structures of Jerusalem,” in Blake’s Sublime Allegory: Essays on The Four Zoas, Milton, and Jerusalem, ed. Stuart Curran and Joseph Anthony Wittreich, Jr. (Madison: Univ. of Wisconsin Press, 1973), p. 334.
A complete analysis is perhaps not feasible; the external record is barren of pertinent information, and we still know too little about the principles that inform Blake’s architectonics to settle the matter conclusively by an appeal to those principles. The best we can do is to interpret the available data with hypotheses that are plausible, consistent, and widely applicable to the art of Jerusalem. Mindful of the speculative nature of the enterprise, I will venture in this essay to provide such hypotheses along with the evidence that supports them. If the following arguments are detailed, it is because I believe that the question of Chapter II is important enough to warrant close pursuit, not only for its intrinsic interest but also for the suggestive light it sheds on the creative forces that shape Jerusalem as a whole.
The questions raised by the reordering of plates in Chapter II are questions of form and structure, not questions of thematic meaning. What kind of work is it, we are forced to ask, that can suffer its internal sequences to become transposed without any radical disturbance to the equilibrium of the whole? Among the possible answers to this question are two that fit the particular characteristics of Jerusalem: it is a work composed of sub-units that are discrete and autonomous,2↤ 2 Karl Kroeber has cited Blake’s penchant for repetition to account for Jerusalem’s structure: “Reiteration permits a fluidity of structure; to a degree plates can be rearranged because reordering does not upset patterns of allusion” (“Delivering Jerusalem” in Blake’s Sublime Allegory, p. 352n). I would put the matter the other way around: fluidity of structure permits repetition. Because the units are discrete, not enchained in a consecutive argument, they tend toward internal comprehensiveness, and comprehensiveness within the part virtually guarantees repetition in the whole. and it is a work where coherence of visual design often rules arrangements at the expense of narrative and thematic continuity. In Chapter II these characteristics are somewhat easier to observe than they are elsewhere in the poem. In transposing pls. 29-41 with 43-46 (as numbered in the First Order3↤ 3 The terms “First Order” and “Second Order,” used in this essay to refer to the plate orders of copies A-C-F and copies D-E, respectively, are borrowed from G.E. Bentley, Jr., Blake Books (Oxford: Clarendon, 1977), p. 234. Plate numbers belonging to the First Order are printed in roman type, Second Order numbers in italics, but only when they vary from the First Order numbers. All editors agree that the arrangement of Chapter II found in copies A-C-F is the earlier. For more on the subject, see the seminal essay of David V. Erdman, “The Suppressed and Altered Passages in Blake’s Jerusalem,” Studies in Bibliography, 17 (1964), p. 40 ff. of copies A and C) when he assembled copies D and E, Blake revealed the seams and faultlines of the chapter, the boundaries of its tectonic components, units that remain undivided themselves in either arrangement. Three extended sets of plates emerge to view, 29-41, 43-46, and 47-50 (I refer to them hereafter, respectively, as Sequences 1, 2, and 3). Once we become aware of the seams between these internal units, we are quick to notice other seams even where the “stitching” has remained intact. The thirteen plates of Sequence 1 soon resolve themselves into two more or less autonomous sets, 29-32 and 33-41, the latter containing the centerpiece of the chapter, the parley of the cathedral cities gathered to save Albion. Each of these separable units in the chapter maintains its own autonomous coherence both in verbal content and in framing visual design. As poetry, these units take the form of detachable dramatic setpieces, as in pls. 33-41, or more typically, of radically condensed epics, self-sufficient synoptic visions of Blake’s total myth;4↤ 4 I use the term “synoptic” here in its sense of “constituting a synopsis,” not in its New Testament sense of a story harmonizable with other accounts, the sense employed, for example, by Joanne Witke in “Jerusalem: A Synoptic Poem.” Comparative Literature, 22 (Summer 1970), 265-78. To document extensively the conception of the individual sequences within Jerusalem as autonomous synoptic visions would take more space than can be spared here. A brief glance at Sequence 2 (pls. 43-46 in the First Order) may suffice for the time being as an illustration. The sequence starts with a memory of the prelapsarian Eden (“I elected Albion for my glory,” the voice of the Divine Vision begins [43.6]), but Albion turns into “darkning rocks” (43.28) and there ensues a long account of the initial strife of Albion and Luvah (43.33-82), quoted directly from The Four Zoas, Night III, followed by the escape of Enitharmon and the Spectre of Urthona from the wreckage of the fall (44.1-4) and their assimilation in the bosom of Los (44.16-17)—here we see a quick synopsis of much of the action in the five middle nights of The Four Zoas. The sequence then quickly moves us into the present: fortified by his spiritual reunion, Los (Blake) explores London in search of the Minute Particulars (45.2-38) while European war rages (“the strife of Albion & Luvah is great in the east” [45.55-56]), and Albion’s sons aggregate ominously into Druidic powers, bringing the cycle back to where it started: “All things begin & end, in Albions Ancient Druid Rocky Shore” (46.15). Quotations of Blake are from The Poetry and Prose of William Blake, ed. David V. Erdman (Garden City, N.Y.: Doubleday, 1970), henceforth cited within the text in parentheses. the work as a whole comes to look like a compilation of brief epic lays devoted to the same cycle of myths, with the inevitable variations overlaps, and repetitions that such compilations display.5↤ 5 Jerusalem, considered as a collection of synoptic tales, would have ample precedents in an age filled with compilations and redactions of mythic cycles and bardic remains. The antiquary Edward Davies, whose influence is evident in Jerusalem, thought that Genesis itself might be regarded as a “collection of documents” (with Moses as inspired redactor), especially the first few chapters with their overlapping considerations of the same events and introductory tags that “point out the beginning of detached compilations” (Celtic Researches on the Origin, Traditions & Languages of the Ancient Britons [London: J. Booth, 1804], p. 40). Curran, in “The Structures of Jerusalem,” p. 331 ff., is illuminating on the repeating events and overlapping sequences in the work. See also Witke, “Jerusalem: A Synoptic Poem,” passim. At the same time the visual designs that penetrate and frame these lays array themselves in demonstrably coherent patterns, and each pattern is specific to the unit it frames, the key signature, as it were, of that unit’s autonomy. It appears that the small narrative units acquired their visual shapes and their etched state long before they fell together into the order we find in the earliest extant copies of Jerusalem.
With so much potential autonomy granted to the work’s smaller units it is a wonder that the larger divisions of Jerusalem, its four chapters, maintain as much stability as they do from copy to copy. Critics often speak of “thematic juxtaposition” as the principle that binds sequence to sequence within the poem,6↤ 6 See Harold Bloom, Blake’s Apocalypse: A Study in Poetic Argument (Garden City, N.Y.: Doubleday, 1963), p. 432. and that principle has fueled enough commentary to make further elaboration of it superfluous here. Less well recognized and more in need of attention is the massive role that patterns of visual design play in forging links between the different sections of each chapter. To cite instances from Chapter II only, sequences begin page 193 | ↑ back to top rich in large designs7↤ 7 By “large” designs, I refer in this essay to illuminations individually occupying a quarter of a plate or more, set off from the text in black (copy E excepted), heavily inked backgrounds, as opposed to small, unframed interlinear or marginal designs. The latter, though often attractive and iconographically significant, do not intrude on the text or strongly demarcate its boundaries as a visual block and are thus left out of my arguments here. There are thirteen large designs, as defined here, in Chapter II of Jerusalem: two each on pls. 31, 33, 35, and single blocks on pls. 28, 37, 41, 44, 46, 47, 50. The interlinear design of the plowman on pl. 29 and the headpiece of the blacksmith on 32 are too narrow by far to counteract the visual effect of overwhelming textual presence. The somewhat larger headpieces on pls. 39 and 40 are borderline cases, but I do not consider them to have a significant effect on the balance of distribution in Chapter II’s illuminations. are juxtaposed for the sake of balance with sequences that are relatively poor (e.g., pls. 33-41 and 29-32, respectively); iconographically related designs belonging to independent sequences gravitate to one another and bind their respective sequences together (as in the ironic juxtaposition of illuminations on pls. 46 and 47, portraits of Jerusalem and Vala in composure and disarray8↤ 8 Leopold Damrosch, Jr., in Symbol and Truth in Blake’s Myth (Princeton: Princeton Univ. Press, 1980), p. 174, notes that Jerusalem is shown as “Michelangelesque” on pl. 46 and as “contorted in mannerist extremity” on 47. ); visual echoes and correspondences appear widely dispersed in the chapter (as in the four half-page illuminations in the chapter, on pls. 28, 37, 41, and 50, the first and the last posted at the extremes of the chapter, the middle pair standing like sentinels on either side of its center—pl. 37 nine positions removed from pl. 28, pl. 41 nine positions removed from pl. 50). Such evidence recalls us to the necessity, splendidly demonstrated by W.J.T. Mitchell in Blake’s Composite Art, of considering the Blakean illumination “as a picture in a world of pictures,”9↤ 9 Blake’s Composite Art: A Study of the Illuminated Poetry (Princeton: Princeton Univ. Press, 1978), p. 5 (Mitchell’s italics). a world that is essentially independent of the world of text. No attempt to account for the ordering or reordering of plates in a poem as richly illuminated as Jerusalem can afford to proceed without that insight as a guide.
Nor may we content ourselves merely with tracing iconographic relationships. The true autonomy of Blake’s designs inheres not in the likelihood that they “mean” something that the text also means, but that they break the text, provide it with spacing, that is, open spaces within it to an antitextual universe. Indeed a study of his work from a purely visual standpoint must see the text itself as a visual element, just as the walls between windows are part of the visual facade of a building. Thus we must attend to the formal layout of Blake’s plates 10↤ 10 A start in this direction is provided by Kay Parkhurst Easson, “Blake and the Art of the Book,” in Blake in His Time, ed. Robert N. Essick and Donald Pearce (Bloomington & London: Indiana Univ. Press, 1978), pp. 35-52. and consider the location of the design on the plate, the proportions of text space and picture space, and patterns in the distribution of these spaces within the plates taken collectively. Independent of the sense of the texts they accompany, these purely formal patterns often seem to follow laws of their own. As Blake reshapes his work and new patterns arise in his mind, they are quite capable of encroaching on the domain of verbal sense and interfering with its continuity. At the same time local patterns of design become subject to effacement by new influxes of verbal invention. To quote Mitchell again, the “relationship [of text and design] is more like an energetic rivalry, a dialogue or dialectic between vigorously independent modes of expression.”11↤ 11 Blake’s Composite Art, p. 4. Out of this rivalry (with design often getting the upper hand) come the shape of Chapter II as initially found in the first printed copies and the subsequent changes in order as well. To pursue this argument further, however, and to enforce the general observations made so far, one needs to take a closer look at particulars.
The Early Core of Chapter II: A Reconstruction
One useful way of approaching[e] the changes that Blake introduced into the order of Chapter II when he printed copies D and E is to view these changes as the last visible manifestation of an evolutionary history lost to record except in its very late stages. A certain amount of that history, at least in its latter stages when Chapter II was beginning to coalesce, may be reconstructed, however, from what is actually available to analysis in copies A-C.
The long thirteen-plate Sequence 1 is a good place to start. It contains the center of the chapter’s dramatic action, the parley of the cathedral cities. It also contains the most striking example of extended visual patterning in the chapter. Plates 31-37 in particular display as a group a conspicuously formal regularity in the distribution of their designs. Throughout this series plates laden with illuminations alternate consistently with plates fully covered by text (peripheral designs aside); of the illuminated plates, three (31, 33, 35) display a common layout, a miniature pattern of alternation, designs above and below divided by a central band of text; the fourth (37) completes the series on a balanced note, all text in the upper half, all illumination in the lower. Intricacies of correspondence in the size of various blocks of text and illumination in the series augment its visual formality.12↤ 12 The proportionate measurements of the illuminations in Jerusalem is a matter that has not been adequately studied. The proportions of text to illuminations on some plates show clear deliberation, e.g., pl. 35, which is vertically symmetrical, the upper and lower designs 2½” in height each, the middle block of text 3¾”. There are apparent relations in proportion between different plates as well. Another 3¾” block of text appears on pl. 37 and the single 5” illumination on that plate equals the combined total of the two illuminations on pl. 35. The middle block of text on pl. 35 (3¾”) is roughly twice the height of the middle block on pl. 33 (1¾”). The headpiece design on pl. 31 exactly fits the text space on 37, and the text and tailpiece on 31 fit the space of 37’s large design. This sort of correspondence is evident elsewhere in Chapter II: in Sequence 2, for example, pls. 44 and 46 are like positive and negative images; where text takes up the bulk of pl. 44, a design of nearly identical height assumes its place on pl. 46, and where a design appears at the top of 44, on 46 we find a residue of text. Like any good designer, Blake apparently strives for regularity without rigidity: pleasing variation and harmonies sometimes unobtrusive, sometimes bold. There is no need to attempt a reading of the various illuminations in the series 31 to 37 in order to discover patterns in their arrangement; one has only to imagine these plates laid side by side to observe a display of visual order that Blake must have devised deliberately (see illus. 1). Once we notice the pattern, however, we soon realize that it raises as many problems about the nature of its devising as it appears to solve, that it conceals anomalies in the continuity of text and perhaps effacements of earlier patterns. Though unified by its pattern of layout, the series of plates from 31 to 37 proves not to be coextensive with any continuous narrative unit but, rather, overlaps two adjacent units (29-32 and 33-41) that differ in subject matter and in script (an indication that these units were designed independently). Linked to pls. 33 and 35 in its visual design and style of lettering, pl. 31 is lodged in the middle of the adventures of Reuben on pls. 30 and 32, even though the text of 31 has no more self-evident connection with that story than the Reuben episode itself has, graphically or thematically, with the series beginning on pl. 33. Elsewhere in Sequence 1 other oddities appear: the continuous narrative that begins at pl. 33 and includes the parley of the cathedral cities fails to end where the visual pattern of formal alternation does, at pl. 37, but flows on to pl. 41, passing through a distinctly different pattern of layout. In the instance of pl. 31, a formal pattern extends itself at the expense of continuity in subject matter; in the second instance, subject matter spills beyond the confinements of formal pattern apparently designed to contain it.
To account for this apparent misalignment of subject matter and containing formal design, we must attempt to recover what we can of earlier stages in the evolution of the middle plates of the chapter towards their final form. Only if we understand the principles that shaped and modified Blake’s early material into the form we now possess can we begin to hypothesize the reasons for such later revisions of begin page 194 | ↑ back to top
The aesthetic integrity of the early five-plate version of the “Divine Family” scene is manifest on several levels (see illus. 2). The tone is sustained in its mildness. The narrative, typically synoptic, tells a tale of Albion “turning from Universal Love” (J 34.7) and of Los summoning the twentyfour cities of the Divine Family in emergency session, and it culminates in a catalogue of the cities and a momentous point of crisis. The designs tell a tale of humanity seemingly in the Spectre’s power though actually reposing in the arms of divine love (pl. 33), of the apparent triumph of the Spectre as he rides on high (pl. 35), and of the timely arrival of an apocalyptic chariot (pl. 41), immensely complex in its iconography, but certainly carrying the Bride in the arms of a God that is “Not an Avenger” (see 41.28).15↤ 15 A critical tradition is beginning to form that considers this design parody, indeed as self-parody. Thus in The Illuminated Blake (Garden City, N.Y.: Doubleday, 1974), David Erdman is equivocal; he notes the presence of apocalyptic emblems—the beasts are “ox-hooved, lion-maned, man-headed” (p. 320), but at the same time “Blake knows that the whole contraption would look ridiculous beside a fiery Pegasus” (p. 321). Mitchell, in Blake’s Composite Art, goes further: “The chariot is an impossible rattletrap of old mythologies. . . . This monstrosity is a self-parody, a satiric vision of Blake’s epic machinery” (p. 216). See also John E. Grant, review of Damon, A Blake Dictionary in Blake (Winter 1980-81), p. 132. Grant considers the design “pessimistic and parodistic.” These views are invigorating and perhaps correct. But in reading Blake’s designs we may be wise to prefer a cautious approach, separating that which is indisputably there, such as the presence of a combination of apocalyptic emblems, from what is critical inference, based on notions of the grotesque that may not be Blake’s. The iconography of the chariot is not without its precedents in sources available to Blake, as Morton Paley points out in “ ‘Wonderful Originals’—Blake and Ancient Sculpture,” in Blake in His Time, p. 175. But the most striking pattern of these plates is found in their visual layout. Here are five plates, five blocks of text in strict alternation with five blocks of design, a study in doublings and synoptic concentration: a doubling of doubly-illuminated plates, a doubling of full-text plates, and a final plate that visually emphasizes the duality of the two modes.16↤ 16 Plate 33 is in itself a symphony of doublings and oppositions: there is of course the vertical symmetry of designs above and below a medial text, and within the designs there are the upper pair of wings and the lower pair of wings, a scene of spiritual support above and spiritual disease below, two figures in the upper design, two in the lower, a tree on the left and a tree on the right in the upper, sun on the left and moon on the right in the lower.
It is pleasant to speculate that this miniature illuminated epic, so internally perfected, was the grain about which the rest of Chapter II crystallized.17↤ 17 There is some evidence to suggest that this series cohered as a unit from a relatively early period in the coalescence of Jerusalem’s plates. On plate 34 of copy A, in the upper right corner, there is a residual trace of a still earlier stage in the assembling of the poem, a small engraved “6” that Blake has only partially managed to cover over. Versions of pl. 34 that lack this “6” are likely to date from an earlier stage still, and in the Rosenwald Collection of the Library of Congress there is in fact a proofsheet of the plate, printed in black, lacking the “6” (a second proofsheet of pl. 34 printed in green, also in the Rosenwald Collection, shows the “6” clearly). Yet even this proofsheet with the missing “6” shows the catchword “By” that links it to pl. 35. These two plates probably cohered from a distinctly early period then, nor can we doubt that pl. 33, which is needed to make the start of 34 intelligible, also cohered to it at that time; context guarantees the connection of 35 and 36. For discussion of the “6” on pl. 34, see Erdman, “Suppressed and Altered Passages,” p. 49; see also n. 20 below. But if Blake once intended the five-plate core to cohere as an aesthetic unit, clearly this intent did not survive into the printed copies. Blake’s second thoughts habitually impinge on and dislocate older stable forms (witness the textual crisis that overtakes The Four Zoas between the Sixth and Ninth Nights). Four plates, 37-40, in Copies A and C, now separate the once adjacent 36 and 41. This situation is apparently the product of two distinct interpolations, pl. 40, first, and then, possibly much later, pls. 37-39 (these latter plates go together as a unit; their script is uniform, syntactic parallels connect pls. 37 and 38, and a catchword, as well as continuous sense, ties 38 to 39). As far as one can reconstruct these revisions, sometime after etching the core unit Blake decided to add pl. 40, possibly to augment the role of Bath (perhaps under the sway of Richard Warner’s War Inconsistent with Christianity).18↤ 18 See David V. Erdman, Blake: Prophet Against Empire, rev. ed. (Garden City, N.Y.: Doubleday, 1969), pp. 476-77. Blake linked the new plate to pl. 36 via the catchword there but left it detached from 41. Possibly he considered using 40 as a replacement for 41. To do so would avoid the non sequitur of the list of cities on 41, now disjoined begin page 195 | ↑ back to top
By bypassing pl. 41, Blake might also obtain a smoother link with 42. The first line of 42 (“Thus Albion sat, studious of others in his pale disease”) might appear to proceed from 40.36 (“[Albion] frown’d on All his Friends, counting them enemies in his sorrow”). At the same time, however, the sacrifice of pl. 41 would mean the loss of Oxford’s speech there, and, more seriously, the disappearance of the conclusive half-page design of the apocalyptic chariot. Blake apparently decided finally to include both plates, engraving the present final line of 40 to give the opening lines of 41 a new rationale. There was still a cost, however: a disturbance in the symmetrical layout of the five-plate core unit, now grown to six. Here then is an instance, one of many, where Blake’s matter and verbal expression managed to override the claims of his form.
The insertion of pl. 40 into the core sequence has relatively few aesthetic consequences other than its effect on the visual layout, for it is relatively homogenous with its surroundings. By contrast, pls. 37-39 massively intrude on the continuity of the core unit as originally ordered, drastically altering its visual form and noticeably modifying its tone. Although Blake is careful to provide a narrative link to pl. 40 at the bottom of pl. 39, the sequence 36-37 is markedly disjunctive. The script changes abruptly on pl. 37 from the smaller, more graceful, more cursive style of pls. 33-36 and 41; Bath, “benevolent” in the last line of 36, becomes “the poisoner: the best and worst in Heaven and Hell” in the second line of 37; and pls. 37-39 introduce a harsh, congested mass of Blake’s late mythic paraphernalia—the Spectre, the Four Zoas, an infusion of biblical names into English settings—matter virtually absent from the more sweetly rendered account of the Divine Family on adjacent plates and only loosely related to it. Plates 37-39 differ more strongly in these respects from pl. 40 than 40 does from 33-36, 41. It is likely that pls. 37-39 allude to a rather different set of historical circumstances from those which inform the original account, circumstances by no means clear but probably reflected in such things as the darker treatment of Bath and the curious account of an attempt to bear Albion back to Eden “against his will” (39.3). These plates may be very late indeed; if the evidence of early suppressed numbers scratched on the copperplates provides any clue, Blake may have already adopted a tentative order for Chapter II before he even etched 37-39.20↤ 20 Some impressions and proofsheets of Jerusalem show faintly engraved numbers that differ from and were superseded by the numbers added in pen in copy A. In Chapter II, Erdman reports besides the “6” on pl. 34 (see n. 17 above) a “1°” or “18” or “12” on pl. 28, a “31” (pl. 31), “36” (pl. 33), “9” (pl. 40), and “19” (pl. 50) (“Suppressed and Altered Passages,” pp. 48-49). Recently the Morgan Library has acquired an early proofsheet of pl. 28 (a reworked version of the well known first-state pull of that plate already in the Morgan’s possession) which confirms the engraved number there as “12.” I cannot read any number on pl. 33 in the copies I have consulted but have been able to corroborate the “6” in various copies, the “9” in copy A, and the “19” in copies D, E, and (faintly) the posthumously printed copy I. These numbers may shed light on the early order of the chapter, for the latter three appear in the same order and at roughly the same intervals as the plates themselves do in the extant First Order (the engraved number on pl. 31 is irrelevant to the early order, since it is likely that this plate is a quite late interpolation in the sequence, and the “36” on 33, if it exists, must belong to a different bout of numbering from the one that produced “6” on pl. 34, for it is highly improbable that 33 was ever physically separated from 34). Indeed, if one numbers Chapter II, starting with pl. 28 as the first and leaving out the probably late arrival pl. 31, one discovers that pl. 34 is indeed the sixth plate in the sequence; plate 40 is indeed the ninth plate if pls. 37-39 are omitted as a late interpolation, and if 40 is “9” it is easy to see why pl. 50 would be “19.” The “12” which we now know that Blake inscribed on a plate already designated as the start of Chapter II supports Erdman’s suggestion that “at one time the chapters were under ten pages in length” (“Suppressed and Altered Passages,” p. 49)—or, more precisely, it suggests that the first chapter once consisted of ten plates exactly (“To the Jews” would be pl. 11). This “12” does make the early “6” on 34 and the “9” on 40 problematic, though not insolubly so. At some early point Blake may have provisionally situated the six plates of the extended core sequence, 33-36, 40, 41, in a ten-plate Chapter I, following the extant pls. 1-4 (a workable combination; pl. 41, moreover, has both the visual appearance and the rhetorical effect of a Jerusalem chapter end-plate); in this case, pl. 34 would still be the sixth and pl. 40 the ninth in the series. Alternatively, the numbers “6,” “9,” and “19” may simply have provided Blake with a private guide to the internal ordering of the plates that make up Chapter II (before late interpolations) without regard to their ultimate numerical place in the poem as a whole. Since this alternative makes better sense of the “19” on 50, ten plates after the “9” on 40, it seems to me more likely that the early engraved numbers represent, for the most part, the internal ordering of Chapter II and that at the time of the numbering the plates had already assumed much the same arrangement as we find in Copy A. By this reckoning the interpolation of pl. 37-39 would not be merely late but very late.
So large an interpolation as that of pls. 37-39 now quite effectively obliterated the early verbal-visual design once intended for the episode of the cathedral cities, but it also provided room for new elements of visual design to integrate themselves with patterns established in plates already etched. Thus the large half-page illumination on pl. 37 of Albion seated in despair occupies the same niche, with respect to preceding plates, that was once filled by the marginally smaller design of the apocalyptic chariot, now displaced; in size, the design on 37 is precisely equal to the sum of the designs on pl. 35; its subject continues the theme of humanity in the Spectre’s power that was displayed in the designs of 33 and 35 and is explicitly enunciated in the begin page 196 | ↑ back to top mirror-script on 37. Next, by etching pl. 38 as a monolithic block of text, Blake continues a pattern that has prevailed in the even-numbered plates since pl. 30. Further, the designs at the top of 39 and 40 are identical in size and contrapuntally related in subject; both are water-visions, one conveying redemptive possibilities, the other fallen states.21↤ 21 See Erdman, The Illuminated Blake, pp. 318-19. Thus a new five-plate pattern of layout, 37-41, is added to the old, and the whole episode grows to nine plates. In this expanded sequence, plate 37 becomes a new pivotal fifth; in the thematic sequence, it provides a nadir vision of despair, equidistant from images of the savior at the beginning of the sequence (pl. 33) and the apocalyptic chariot at the end (pl. 41); as an element of visual structure, it anchors 41, the ninth plate, as well as 33 and 35, (the first and third). Thus new patterns supersede old without wholly effacing them, like the overlapping rings of lunar impact craters. Blake’s mode of introducing new matter in his work is often nearly as violent, yet at the same time his instinct for form persists.
How the rest of Chapter II fell into the arrangement extant in Copies A-C is a matter for conjecture only. At some point Sequence 2 (43-46) locked into place with Sequence 3 (47-50) through the nexus of the great designs of Vala and Jerusalem on 46-47. How this series of eight, with its powerful visual center, attached itself to the core sequence is not clear. The somewhat rough connections between pl. 42 and its immediate neighbors suggest a more complex history of joining than we can now recover (42 may be the residue of a longer series of plates pared down at both ends). Meanwhile the five-plate central core had at a relatively early stage incorporated pl. 40. Somewhat later, probably, the augmented core acquired the three-plate sequence 29-30, 32, as a proem.22↤ 22 Plates 29-30, 32 do not appear to have been etched at the same stage as pls. 33 ff., for the style of script on these three plates is uncharacteristically large and elaborate compared with the Jerusalem norm. This script appears in J 56 and J 61, palpable interpolations both, leading Bentley to describe them as “apparently late” (Blake Books, p. 225). A more telling argument for lateness is the close resemblance of this script to that of The Ghost of Abel (1822). Although Erdman describes the script on J 29-30 as “an early format” (The Illuminated Blake, p. 309), the evidence leans slightly toward a relatively late date for these plates in the etching of Jerusalem. But even a very early date of etching is no obstacle to relatively late attachment to the core sequence of the cathedral cities, with which pls. 29-30, 32 share little in subject matter. The chapter now had the form of a double set of nines, anchored together at what became their medial pivot, the appropriately encompassing design of the apocalyptic chariot on 41, equidistant from the beginning and end of the chapter. The latest developments presumably were the interpolation of pls. 37-39 and of pl. 31. The case of pl. 31, long recognized as an intruder,23↤ 23 See Erdman, “Suppressed and Altered Passages,” p. 21. bears further comment. A self-contained doctrinal passage, its text may never have formed part of any extended sequence in the poem, and for the reader it continues to remain aloof from its surroundings. For the viewer, however, pl. 31 coheres very well with what follows it: it displays precisely the same visual layout as pls. 33 and 35 and certain iconographic connections with these as well (particularly in the repetition of flying or hovering figures). By placing 31 where he does, Blake extends forward the pattern of alternating textual and illuminated spaces so that every plate from 30 to 36 (and, with the eventual addition of 37-39, every plate to 38) is made part of the pattern. Here then is evidence that it is often not the sense of the text but the look of the page that determines the location of some plates in Jerusalem. In form, like apparently gravitates to like, creating a pattern that cuts into narrative continuity and extends itself across the boundaries of discrete narrative units.
The Rearrangement of Sequences in Copies D-E
This hypothetical reconstruction of some stages in the prehistory of Chapter II brings us finally to the order of the first printed copies. But Blake’s concern for visual form, for a handsome distribution of illuminations, could not have ceased once these printed copies came into being. Certainly Blake’s previous treatment of successive copies of longer works—his rearrangement of full-page illuminations in The Book of Urizen and Milton, for example—suggests the opposite. Indeed, Jerusalem is actually conservative in the general stability of its order from copy to copy. Yet Chapter II, already the product of much careful designing and many disruptive second thoughts, gets substantially rearranged in copies D and E. In these copies Blake moves pls. 43-46 to the front of the chapter, just after the chapter opening on pl. 28, and he inserts plate 42 between what were in the First Order begin page 197 | ↑ back to top pls. 37 and 38. Any explanation of these changes should base its inferences on the same principles that seem to have motivated Blake’s earlier planning in the chapter.
A quest after possible thematic motives for the rearrangement of the chapter seems to me the least promising way of explaining the problem. Commentators on this point have generally pronounced both orders thematically satisfying, and indeed they are. In one order (A-C-F), after the proem of plate 28, the narrative begins with Albion repudiating the Divine Vision; in the other (D-E), it begins with the Divine Vision proclaiming its early intimacy with Albion. The narrative in the earlier order shows Los undertaking a solitary quest to save Albion after the group effort of the cathedral cities fails; in the later order the cathedral cities appear to convene (though not immediately—the Reuben episode and other matter come first) after Los’s individual failure. In terms of thematic significance, these alternatives offer little to choose from, since they both yield the same gloomy outcome, a stubbornly unredeemed Albion. But when we turn our attention from theme to visual form, very different consequences attend the two orders. The rearrangement of narrative sequences demands rearrangement of illuminations as well and an altered set of relations between blocks of illumination and blocks of text. Given Blake’s evident concern for such matters, we may plausibly suppose that he found the distribution of illuminations in copies A and C unsatisfying, even though the arrangement of sequences formed a reasonable, even a desirable, narrative order. If this motive did in fact figure significantly in the shifting of sequences, it provides us with a magnified instance
The concern for form in this instance may, however, signal only a desire to rectify imbalances in layout that earlier intrusions of content may have caused in the first place. In planning the layout of the smaller units of the chapter, Blake seems to have pursued a kind of medial symmetry as his goal, sequences with strong pivots at the center and balanced elements flanking it on the imaginary left and right. The same goal appears evident in the plan of Chapter II as a whole. If we visualize its order before the interpolations of pls. 31 and 37-39, we find a series of nineteen plates (omitting the address to the Jews) with large illuminations so distributed that their center of gravity falls in the middle (an appropriate distribution for a middle chapter: large illuminations in Chapter I cluster towards the beginning of the chapter; in Chapter IV they cluster towards the end24↤ 24 Six of the nine large block-illuminations in Chapter I occur in the first half of the chapter, indeed within the first eight plates of actual verse; eight of Chapter IV’s thirteen large illuminations appear, in crescendo fashion, on its last seven plates of verse, less than a third of the whole chapter. Large illuminations are infrequent in Chapter III. (The five full-page designs in Jerusalem are excluded from this count.) For a table showing the distribution of large designs in Jerusalem, see Bentley, Blake Books, p. 231. ). The half-page illumination of the apocalyptic chariot occupies the tenth position (pl. 41), the exact center of the series, flanked at opposite ends by the half-page illuminations on the first (pl. 28) and the nineteenth plates (pl. 50). The other large illuminations are distributed on either side of the center at roughly corresponding intervals from it, five before and four after (the greater massiveness of the designs on the later plates balances out the slightly greater frequency of designs on the earlier). A well-balanced visual order would thus have prevailed at one time in the chapter as a whole as well as in its individual illuminated units.
When Blake expanded the parley of cathedral cities by three plates (none massively illuminated except pl. 37) and
For all the presumed defects of Chapter II as a visual sequence, Blake nevertheless produced two printed copies of the chapter in this order before he began to have second thoughts. The second thoughts seem to have coincided with his plans for a full-color version of Jerusalem (copy E), a version in which the claims of visual design would have preeminence.26↤ 26 Copy D, uncolored and sometimes poorly inked, of course also displays the Second Order. But there is little reason to suppose that D is earlier than E. Both are watermarked 1820, and the length of time that the production of E would need for completion argues for an early start on its planning. If any perceived deficiencies in the sequential design were to be remedied, a project such as the production of copy E would certainly have provided the appropriate occasion, and Chapter II, which is particularly rich in large designs, would be likely to have received special attention. An inspection of the new order of plates in the D-E version of this chapter indeed assures us that the imbalances present in the A-C order have been rectified. If we divide the twenty-four plates that compose the enlarged chapter at its median, we find seven large designs in the first half and six in the second, a considerably more even distribution than the eight and five split of copy A. If we divide the chapter in thirds, the effect of medial symmetry is even more pronounced (see illus. 4). In the first eight plates there are three large designs, in the next eight, seven, and in the final eight, three (a similar calculation applied to copy A yields five, four, and four, even-sounding enough, until we observe that the distribution sequence within the middle third is itself considerably lopsided [see illus. 3]). The medial symmetry that may have been present in an earlier hypothetical nineteen-plate chapter is now restored.
In the transformation of pls. 43-46 into pl. 29-32 there is one other consequential gain in pictorial symmetry. With the migration of the former plate 46 towards the beginning of the chapter, Vala and Jerusalem no longer change at the turn of a page from elegant aloof emblems into fleshy women, falling with and falling for a man as contorted as they. An ironic narrative sequence loses the link between its two parts, but at the same time, by undoing this heavy conjunction of massive designs, Blake restores a balance. In D and E the two versions of this female pair, one version depicting eternal states and one depicting fallen vegetative selves, are as symmetrically disposed in their locations in the chapter as the emblems of St. Paul’s and Westminster Abbey are on the first of the panels themselves. Plate 32, formerly 46, now concludes the first movement in the chapter, an epyllion of four plates (not including “To the Jews” and pl. 28); it is more than ever the mirror image of pl. 47, which introduces a final movement of four plates. Both plates are nearly equidistant, on opposite sides, from the midpoint in the chapter (if one leaves “To the Jews” and 28 out of the reckoning, 32 and 47 are precisely equidistant). The mirroring effect is particularly apparent in E, where Blake employs an identical color scheme for the bodies, clouds, and background sky in both pictures.27↤ 27 This repetition of the color scheme is more exact in the Mellon copy itself than in the Trianon Press facsimile. The parallels and the radical contrast of these pictures are enforced by the symmetrical relations of their location. They are fit panels to flank the chapter’s visual center, itself filled with symmetrical doublings and oppositions, visual displays of hard choices, of succor and disease, begin page 200 | ↑ back to top of soaring and sinking, where the “best and the worst, Heaven and Hell” often inhabit identical pictorial frames.
The Status of Plate 42
One curiosity in Blake’s rearrangement of Chapter II remains to be discussed. Why is pl. 42 the forty-second plate of Jerusalem in both Orders, even though the plates adjacent to it shift radically? This plate once directly followed the nine-plate sequence on the cathedral cities that ended with the picture of the apocalyptic chariot. In the Second Order it appears deep within the nine-plate sequence between plates numbered 37 and 38 in A-C. Since pl. 37 moves four places back with the transpositions of Sequences 1 and 2 and 42 moves four places forward, the latter needs no renumbering. It is hard to imagine why this one plate should remain fixed in its place like a rock in a stream while a total of seventeen plates before and after shift around it. Fond as Blake is of symbolic numbers, the number 42 does not seem to be among them. The plate has no large block-design. Although it presents a powerful speech by Los and some fairly typical rumblings from Albion, there is nothing in the content that marks it as demanding singular treatment. Finally, most curious of all, to remain pl. 42 in the Second Order it must interrupt the continuity between pls. 41 and 43 without supplying any new plausibility; indeed it interrupts a series of parallel clauses in apposition:
They saw Albion endeavoring to destroy their Emanations (41:31)In the Second Order, two non sequiturs in the sense appear where there were none before.
They saw their wheels rising up poisonous against Albion (43:1)
In its original location in the First Order, this plate had no inevitable connection to the plates immediately adjacent to it. It cannot have been conceived as part of the scene of the twenty-four cathedral cities that preceded it, for on 42 the parley abruptly disappears, and Albion states that Hand and Hyle “have seized the Twenty-four rebellious ingratitudes” (42:48), though nothing of the sort occurs in the previous nine plates.28↤ 28 The plate must have been etched relatively early, however. The “rebellious ingratitudes” are twenty-four, not twenty-eight or eighteen and ten (see n. 19 above). Plate 42 may owe its oddly disconnected character to the loss of a plate or plates once adjacent to it. Nor does it connect obviously with the succeeding plate, since 42 concludes strongly on a line of summary finality (“This is the Net & Veil of Vala, among the Souls of the Dead” [81]), and 43 presents a powerful fresh scene with its first line (“Then the Divine Vision like a silent sun appeared”). Even in the First Order, then, pl. 42 is inessential to narrative continuity. The most that might be said for its earlier location is that its text provides at least an adequate transition from the collective effort of the cities to the individual effort of Los to redeem Albion. What it does not provide, however, is a fit introduction to that last, immensely powerful synopsis of Blake’s historical myth that bursts upon us on pl. 47, as “Luvah [tears] forth from Albions Loins . . . in rivers / of blood over Europe” (47:3-4). Yet this is precisely what pl. 42 must introduce when Sequence 2 is moved to the beginning of the Chapter in the Second Order. Far better is the juncture that actually appears[e] in D-E, “O God thou art Not an Avenger” (41 [46]) implicitly judging “that Holy Fiend / The Wicker Man of Scandinavia” on the next plate (47:5-6), just as the design of the apocalyptic chariot on 46 implicitly judges the disarray of Albion, Vala, and Jerusalem on 47’s adjacent design.
Once again Blake seems to cater to the claims of coherent sense. In that case pl. 42 must be removed, but where? It cannot be pushed further back, since 47-50 form a single, sweeping, unbroken movement from the first line of 47 right up to the etched words “End of Chap. 2” on 50. Thus it must move forward. Oddly, it comes to rest at a point where it severs a scene in the middle and interrupts a passage of verse previously bound together by syntactic parallels; no catering to sense here. But from a pictorial standpoint, the resting place is not so odd. What appears to have happened is that the large design of the seated Albion on the lower half of 37 [41] has arrested the forward drift of 42, or to put it another way, the first lines of 42, “Thus Albion sat, studious of others in his pale disease: / Brooding on evil,” gravitate toward their precise illustration, tugging the whole plate with them. That the dislodged plate finally locates itself in the forty-second position of the reordered Jerusalem turns out to be fortuitous. What is important, rather, is that the power of a design appears, once again, sufficiently compelling to Blake to determine the position of a plate. Moreover, just as the claims of narrative continuity failed to deter the interpolation of pl. 31 between 30 and 32 at an early stage of the chapter’s arrangement, these claims once again fail to deter the interpolation of 42 between the former 37 and 38. In the latter case, more than purely visual elements enter into Blake’s choice, but even here, the continuity or bond that prevails is one between word and picture, not word and word. In the tug of war between the visual and the verbal that goes on throughout the evolution of Chapter II, the visual element appears to prevail.
Epilogue: 1826
In the long run, however, no one component of Blake’s composite art can claim a permanent victory. If Blake altered the arrangement of the second chapter in copies D and E, arguably on grounds of felicity of design, not long before his death he changed his mind again, and in the Pierpont Morgan copy of Jerusalem (F) he restored the First Order of page numbers. His deliberations on the matter evidently occupied him to the very last, for many of the pages in Chapter II of F show signs of erased or altered numbers, inscribed after the printing (one of them on paper watermarked 182629↤ 29 Plate 30[34]. ) but now superseded by the extant numbers. As far as we can decipher these erasures they invite us to believe that they are the remains of a Second Order foliation like that of copies D and E (for more on this subject, see the Appendix, “Emendations in the Foliation of the Morgan Copy, Chapter II,” below). The erasures and alterations prove conclusively that Blake could not have begin page 201 | ↑ back to top stumbled back into a First Order foliation accidentally but must have done so deliberately.30↤ 30 The evidence of Blake’s painstaking emendations to produce First Order numbers invalidates Bentley’s hypothesis that Blake intended the Second Order scheme for the plates of Chapter II in copy F “but carelessly copied the former order when he numbered them with a pen” (Blake Books, p. 234). A decisive endorsement of the First Order seems to have been Blake’s last word on the form of Jerusalem.
We cannot know for certain why he took this last step. Perhaps the restoration of the original order reflects Blake’s tendency to commemorate the firstlings of his conception and belongs with such manifestations of the tendency as the restored “12” on the title page of Milton and the date 1804 on the title page of Jerusalem. But it is also possible that verbal and narrative continuity asserted new claims for Blake’s favor. Certainly copy F, though elegantly touched up with China white, is a drab thing compared to the extravagance of E, but as a reading text it is far clearer.31↤ 31 “The retouching of rude mendings . . . is carried through more conscientiously in the Morgan than in any other copy and rather less conscientiously in the Mellon than any other. . . . Blake took more pains to make the text legible in the Morgan copy than in the Mellon” (Erdman, “Suppressed and Altered Passages,” p. 41). Many plates are faintly printed, some (e.g., pl. 15) scarcely legible, in copy E. The orangish ink also contributes more to the visual splendor of the book than it does to ease of reading. We may surmise that Blake considered E the viewer’s Jerusalem and F the reader’s Jerusalem. In that case, a sequence agreeable to the viewer might cause discomfort to the reader. The location of plate 42 in the Second Order, for example, is wholly illogical from the standpoint of narrative continuity, a point that Blake might have perceived more forcefully as shapeliness of visual design receded somewhat in importance to him.
Although we can only speculate about Blake’s motives for restoring the First Order in F, we can refer to the evolutionary history of Jerusalem, where a competition of two autonomies, the visual and the verbal, goes on continually. Perhaps modern editors of Jerusalem, faced with the choice of the two variant orders, should first consider what sort of choice Blake himself was making as he moved from one order to the other, what alternative goals he had before him, and they should make their own editorial decisions accordingly. To establish a “preferred” order of plates on extrinsic, mechanical grounds—the order represented in the majority of copies, the one with the most colors, the one produced last—is to neglect the principles of ordering that would have mattered to Blake himself and to evade their problematic nature. If there are such things as a reader’s Jerusalem and a viewer’s Jerusalem, then perhaps the wisest thing a future editor might do is to follow the order of plates that best suits the alternative adopted. If, for example, the edition is to be furnished with reproductions of the illuminations (as in the recent Bentley edition)32↤ 32 William Blake’s Writings, ed. G.E. Bentley, Jr., 2 vols. (Oxford: Clarendon, 1978). , then the order of copy E, with its superior disposition of designs, would be the one to follow; but if the edition is an unilluminated letterpress text, then the order with the greatest verbal coherence, such as we find in copy F, would be more appropriate. There is no abstract “correct” order to the second chapter of Jerusalem, but neither should our choice be arbitrary; the appropriate order of the poem is profoundly related to the medium of its transmission.
It may be disquieting to some to propose, in place of “the” poem or “the preferred” poem, a plurality of Jerusalems offering faces that accord with the differing desires of viewers and readers. But differing desires, pluralities, unfinished competitions, these are the elements in which Blake breathes. The ordering and reordering of sequences would be impossible in the first place if the poem were not composed of dozens, perhaps scores of smaller poems, internally autonomous but often covering portions of the same ground as the others do: in short, competing with one another to tell the same story, to offer the reader alternate versions of the same truth. And as scene competes with scene, designs come to compete with text, new conceptions with old conceptions, late copies with early copies, and the surface of the poem is dislocated but then reorganized each time. Organization becoming firm in one part often means disorganization somewhere else. This process is perfectly imaged in the figure of “Golgonooza the spiritual Four-fold London eternal . . . ever building, ever falling” (Milton 6.1-2), evading a petrified order. In most of Jerusalem the building and rebuilding has arrested itself, leaving only archeological traces of its earlier turbulent development. Chapter II is a gate into the place where the building and rebuilding goes on, allowing us to clarify our conjectures about what the process as a whole must have been like. It is the place where Blake still allows a glimpse of his unappeasable appetite for an unending conversation between strong text and strong design. “Go on, builders in hope” (J 12.43), it seems to proclaim.
Appendix: Emendations in the MS Foliation of the Morgan Copy, Chapter II
The matter of erased pen numbers in F needs a somewhat detailed exposition, for the erasures and other emendations have not been completely reported hitherto. Not all the plates subject to shuffling in earlier arrangements of Chapter II show evidence of emendation beneath the extant page numbers, but the mendings are more extensive than previously realized. Erdman finds an erased number, “possibly ‘33’ ” and mendings in the digits of “30,” “31,” and “32” (“Suppressed and Altered Passages,” p. 51). Close attention to scouring marks on the paper, aided by an ultraviolet lamp, has enabled me to collect more complete information (I am also indebted to Thomas Lange and the staff of the Morgan Library for help in deciphering the evidence presented here).
The listing below follows the Second Order arrangement of copy F’s present binding (of Victorian date):
| pls. 28, 43[29], 44[30] | no evidence of earlier numbers |
| 45[31], 46[32] | beneath extant First Order numbers |
| 29[33] | extant “29” written over 2 erased digits, almost certainly “33” |
| 30[34] | “0” written over an erasure and in lighter ink than the “3” |
| 31[35] | “1” written over an erasure |
| 32[36] | “2” written over an erasure and in lighter ink than the “3” |
| 33[37] | no apparent emendation beneath extant “33” |
| 34[38] | “4” written over a heavy erasure |
| 35[39] | no apparent emendation |
| 36[40] | both digits written over severe erasing; to the left of the extant “3” the imprint of a “4” left by the pressure of the pen is distinctly visible in ultraviolet light; a second earlier digit to the right of the “4” illegible beneath the extant number |
| 37[41] | no apparent emendation |
| 42 | written in heavy black ink (unlike emended numbers, written mostly in a grayer ink) |
| 38[43] | “3” written over an erasure; the “8” is mended in gray ink from an earlier “3” written in black ink |
| 39[44] | no apparent emendation |
| 40[45] | faint evidence of erasure under “0” |
| 41[46], 47, 48, 49, 50 | no apparent emendation; numbers inscribed in gray rather than black ink |
As the table above makes clear, not all the plates show evidence of early fossil numbers beneath the extant First Order numbers. But although the early patterning of numbering was intermittent it was by no means random. Plates 43[29]-46[32] and 47-50 show no signs of earlier numbers; pl. 29 [33]-32[36], on the other hand, display such signs on every plate; pl. 33[37]-41[46] show a pattern of early numbers on alternate plates in the series (pl. 42 is included here as the fifth in the series), with the odds (assuming a likely Second Order arrangement) skipped in the first half and the evens in the second. One plausible way of accounting for these variations is to assume that Blake numbered these plates in different batches and at different times. It is interesting to note that these different batches of plates, 29-32, 33-41, 43-46, and 47-50 are virtual replicas of our old acquaintances, those pre-fabricated units that came to form Chapter II of Jerusalem in the first place: respectively, the four-plate introductory unit of Sequence 1, the nine-plate central panel of the cathedral cities, the highly mobile Sequence 2 that ends with the tableau of Vala and Jerusalem, and the immobile concluding Sequence 3. Did a sense of the autonomous integrity of these units persist so distinctly in Blake’s mind, even to the last, as to influence the way he initially went about compiling copy F?