Volume 36 · Issue 2

Michael Phillips. William Blake: The Creation of the Songs from Manuscript to Illuminated Printing. London: The British Library; Princeton: Princeton UP, 2000. [xi] + 180 pp. 37 monochrome and 72 color illus. £16.95/$29.95 paper, £30.00/$55.00 cloth.

Reviewed by Alexander S. Gourlay

Michael Phillips’ engaging account of the origins, early history, and production of the Songs of Innocence and of Experience bristles with startling insights, fresh perspectives, and new discoveries, but important aspects of its most prominent arguments are simply wrong. The idea behind it is outstanding: to review the publishing, political, and aesthetic contexts in which the Songs arose, to reconstruct as far as possible the stages in which they were composed, and to detail the physical processes by which Blake actually produced the books, all accompanied by compelling images of relevant documents, tools, and other visual materials, many in color. In many ways the promise of this concept is admirably fulfilled. The writing is clear, bracing and energetic, the typography and layout are attractive and readable, and the illustrations are fresh, plentiful, well chosen and clear. The book includes more than two dozen large full-color images of pages from particular copies of Songs, many of which have not been reproduced elsewhere. Given all these positive characteristics, it is unfortunate that it is marred throughout by major and minor errors in interpreting the complex evidence about how Blake created the Songs of Innocence and of Experience.

begin page 67 | ↑ back to top

Phillips begins to establish the formative context and origins of these poems by adducing two interesting books of prints from his personal collection that Blake may have owned and doodled on as an adolescent. These aren’t shown to be especially relevant to the Songs, either in themselves or because of Blake’s putative additions to them, but they do effectively suggest the potential range of his early interests, and it is tantalizing to see a few pages from them. Phillips then focuses more narrowly on Innocence, reviewing early Blake poems copied (by someone other than Blake) into a copy of Poetical Sketches and also lyrics in the dramatic satire An Island in the Moon, of which Phillips published a sumptuous and scholarly facsimile edition in 1986. Certain of these are versions of poems that later appeared in Songs of Innocence, and others explore related themes in ways that illuminate Blake’s developing sense of the topography of both Innocence and Experience. Although some poems probably had versions even earlier than the fairly polished drafts we see recorded in these mss., and any or all could have been adapted to the contexts in which they appear rather than being steps in a straightforward evolutionary process culminating in the Songs, Phillips’ account of the poetic process is generally persuasive and consistently illuminating. Knowing what a poem used to be or where it came from should not determine our understanding of what it became, but this chapter makes an excellent case for reading poems in their developmental context.

Some troublesome issues arise two chapters later as Phillips conducts the reader through the much more complicated documents that underlie the etched texts of the Songs of Experience. These are more clearly drafts in development even if some may have been copied into Blake’s Notebook from still earlier unrecorded versions, and most contain discernable (but not easily distinguishable) layers of revision in addition to plentiful false starts, dead ends and other expected signs of poetry caught early in the creative process. When the late David V. Erdman documented these changes in his editions of Blake and of the Notebook, he only described the sequence of marks that overlap, and didn’t specify a chronology of creation and revision or attempt to sort out broad layers of changes made about the same time. Phillips undertakes to do both, but the resulting account of Blake’s process of composition and revision is rife with dubious assertions—the broad outlines of the process Phillips describes are plausible and the narrative as a whole provides a very useful perspective, but much of what he specifies about the mechanics and chronology of composition and revision is unlikely and most is impossible to confirm.

Part of the problem is that although Phillips’ account is not overtly polemical, it is more tendentious than it first appears to be: throughout the book, the descriptions of Blake’s creative and printing processes are skewed by misconceived reaction to the portrait of Blake as a maker of illuminated books promulgated by Robert N. Essick in William Blake, Printmaker (1980) and William Blake and the Language of Adam (1989) and even more emphatically by Joseph Viscomi in Blake and the Idea of the Book (1993). Essick and Viscomi represent Blake as an heroic and ingenious inventor of means by which the poet/artist is freed from the drudgery and mechanization that dominate creativity in conventional book production. In their view, Blake’s methods maximize artistic control, spontaneity, and the ratio of productivity to uncreative labor. When Viscomi described the process of making illuminated plates in 1993, he was especially concerned with debunking critical corollaries to the persistent myth that in making his plates Blake mechanically transferred finished calligraphic texts to the copper after they were actually written on something else. In that context Viscomi noted that when writing on the copper Blake could have recorded spontaneous effusions of verse or created designs directly on the plate, without any drafting or premeditation. But when Viscomi wrote of Blake regularly “composing” text and pictures on the copper he referred not to literary or graphic invention (as Phillips seems to think) but to the arrangement (and adaptation) of pictorial and textual material on the page.¹↤ 1. See esp. chapter 3, “Illuminated Printing as Composing Process,” Viscomi 26-31. Some of the shaggiest passages of such later prophecies as Milton and Jerusalem could have been invented wholesale as Blake recorded them on copper, but nobody has ever argued that Blake wrote the Songs without drafts. The critical point that both Essick and Viscomi were trying to make was that every stage of making a Blake book was a fresh performance by the artist, whereas in conventional bookmaking there is almost always at least one stage that is at best a simulacrum of a performance that has been fixed at an earlier stage and in another medium, as when a commercial engraver copies a design by an artist or a compositor makes a copy in type from fair copy in the poet’s hand.²↤ 2. The engraver or compositor in conventional book production may be an artist in his or her own right, and may have considerable control, but nothing like what Essick and Viscomi call Blake’s “autographic” command of processes and content.

In a move that is characteristic of this book, Phillips undertakes to set Essick and Viscomi straight about the means of production by showing how hard Blake appears to have worked on drafting the proto-Songs in the Notebook, as if they were claiming that Blake’s poems, including the Songs, were all communicated to him by spirits while he worked on the copper. Just as he seems to misunderstand “composition” of texts and designs, Phillips treats the word “autographic,” used by both Essick and Viscomi to describe the way Blake produced the printed texts and designs for his books, as a synonym for “automatic,” as in “automatic writing”: Phillips writes, “The manuscript drafts, both in An Island in the Moon and in the Manuscript Notebook, make clear that Blake did not write autographically, unpremeditated or from any form of dictation divine or otherwise” begin page 68 | ↑ back to top (111).³↤ 3. Ominously, Phillips has communicated his confusion to the reviewer Vincent Carretta, who concludes that “the evidence for the evolution of the Songs . . . renders this notion of ‘Blake’s autographic mode of graphic composition’ no longer tenable” (443). The word “autographic” actually means “written in the author’s own hand,” and to the extent that it is related to the issue of volition it means that the author is more, not less, in control of what appears in the published book.

The misdirection of the argument about composing on copper is complemented by systematic errors in interpreting the manuscript drafts of the Songs that make it appear that Blake, far from being in control, was both inattentive and irrational. Ironically, because Phillips is so diligent about specifying what he supposes to be the minutiae of the creative process (and so careful to provide the materials that we need to corroborate his narrative), it quickly becomes apparent that he is not entirely reliable as a guide through the editorial thickets of the Notebook drafts. In most of these erroneous reconstructions, Phillips treats tentative moves by Blake that were rejected immediately or overwritten as if they were still part of the context as Blake moved on with an entirely different thought in mind. To take one example, in reconstructing the “first draft” of “I laid me down upon a bank” (36), Phillips has Blake writing “Triven” for “Driven” in the last line. Are we to suppose that Blake’s first inspiration was to invent a nonsense word, then thought better of it after he finished the poem? Isn’t it more likely that he abandoned a word beginning with “T” and wrote “Driven” on top of it, then went on to the rest? This isn’t an important error by itself, but as the instances multiply one realizes that Blake is being treated more like an oblivious telegraph clerk than a poet.⁴↤ 4. To take another example, in reconstructing the “first draft” of the poem “I asked a thief” on page N114 of the Notebook, Phillips offers a neat draft of the whole poem ending with a nonsense phrase, “And twixt earnest & game / He enjoyd the dady.” Later, Phillips reports, as part of a general scheme of revision, Blake revised it to read, “And twixt earnest & joke / Enjoyd the Lady,” correcting “dady” to read “Lady” (38-39). I think Phillips has grievously misinterpreted a harmless over-reading by Erdman, thereby producing editorial hash. In transcribing the poem for his 1965 edition of Blake, Erdman supposed he saw vestiges of a “d” at the beginning of the word “Lady” and guessed, in his impetuous way, that Blake must have first written “dame” to rhyme with “game” (itself a very speculative reading, perhaps suggested by an initial “g” that seems to be underlying the “j” of “joke”) and then later changed it to “Lady.” In his edition of the Notebook and in his 1981 revised edition of Blake, Erdman corrected his account of the revision, recognizing that the letters “me” were never written, but he still reported “da[me],” that is, an abortive “dame” rejected even before the poet finished writing it, with “Lady” written on top of it. I think Erdman’s “da[me]” was a figment of his fertile editorial imagination: more likely, Blake initially wrote “lady” and then, either immediately or sometime later, capitalized it with his usual majuscule “L,” which has a leftward loop in the lower corner, thus producing something that looks like a “d” made up of that loop and the superimposed stems of the “l” and “L.” The hypothetical Blake who inhabited Erdman’s head when he examined manuscripts was creative, logical, and purposive, his mind bursting with synonyms and potential rhymes as he writes. The hypothetical Blake who writes “enjoyd the dady” is as unaware of what he is producing as a numb-fingered recruit from the typing pool. Similar improbabilities are specified as fact in transcribing “Thou hast a lap” on N111 (Phillips 52), where Phillips leaves “can” in line 3 of the second stanza, though it is much more likely that the word would have been deleted before he wrote it again in the next line. In line 3 of “The Earth’s Answer,” also on N111 (52), Phillips has Blake writing “Her eyes fled dead” in the position of a tri-syllabic line; even if we assume that Blake had not yet determined the metrical structure when he drafted the first stanza, he would at least have settled on “eyes fled” or “orbs dead” (another rejected reading) before writing the rest of the poem, and it makes no sense to offer a “reconstructed” draft of the poem containing meaningless and mis-metrical scraps of both readings at once. Once one starts looking closely at these editorial reconstructions one can quickly find as many more that are down-right impossible, as well as numerous instances in which more plausible alternative reconstructions are neglected. It does not help that there are also several annoying typos in Phillips’ transcriptions that suggest the intervention of an officious electronic automaton rather than flawed editorial methods: e.g. “new moan hay” on pages 39 and 41, “jealously” on both 52 and 53. (In his “Corrigenda” of 2001 Phillips lists other errors in the book—see “Works Cited.”)

In spite of the problems in interpreting manuscript particulars, Phillips’ reconstruction of the broad stages of composition of Songs of Experience is usually sharp and suggestive. Examining the reconstructed drafts helped me to understand the poems better even when I was most skeptical about his specific reasoning, and if taken with a grain of salt this part of the book would be a very useful resource for anyone studying the Songs at any level. The elaborate foregrounded evidence of pens, ink, and writing sessions is a minor distraction from the more important story Phillips has to tell about the these poems.⁵↤ 5. I agree with Phillips that the Notebook drafts reflect several stages of revision, but his stroke-by-stroke account of multiple pens and inks is more positive in its assertions than my experience with sharply pointed quill pens and period inks would allow me to be. Such pens and such inks can produce writing with highly variable characteristics, even at a sitting, if the pen is set down, held aloft, trimmed, recut, or even shifted in the hand. Furthermore, Phillips has a gift for finding and presenting compelling illustrative evidence, and many of the photographs, flyers, and other historical documents that he offers in conjunction with the primary texts are important and illuminating as well as eye-catching. (A different problem arises with one of these—see Appendix.)

Following each of the chapters detailing the development of the Songs in draft are chapters describing Phillips’ version of the processes by which Blake produced the printed versions of Songs of Innocence and then Experience. The material here has already generated considerable controversy, so I will not attempt more than a perfunctory review of the problems as I see them. Phillips’ account of the publication process appears at first to be an objective exposition building upon previous scholarship and experimentation, and much of it is exactly that, but once again a misguided, partially covert, polemic against Blake and the Idea of the Book bubbles up at critical points in the discussion. Viscomi emphasized not only the “autographic” aspect of Blake’s begin page 69 | ↑ back to top methods of creating printable text on plates, but also the efficiency of the techniques used to print from those plates, techniques which similarly freed the artist from mechanistic drudgery. Viscomi pointed out evidence of small-scale mass production everywhere in Blake’s illuminated canon, and showed that even though copies of illuminated books are often quite different from each other, Blake’s method permitted him to create individuated copies efficiently without having to become a slave laborer in his own mills of reproduction. Viscomi may have exaggerated the efficiencies achieved by systematic production while downplaying some of the ways in which Blake worked unsystematically and inefficiently, but once again Phillips’ account of the printing process goes far overboard in the other direction, and again he sometimes has Blake acting in ways that don’t make much sense.

Phillips’ description of how Blake wrote, drew, and etched text and design on the plate basically follows Viscomi and Essick, but whereas they were especially interested in the “composition” stage of the process, Phillips glides by it in a few sentences. All three agree that Blake wrote and drew backwards directly on the plate with a quill or brush loaded with an acid-resistant substance, and then etched or engraved further on some plates after initial etching. Phillips does not dwell on the fact that as Blake worked with the resist he could stop and erase errors or make improvements fairly easily (we don’t yet know exactly how easily, since the recipe for the resist itself is still the subject of speculation), so improvising the layout on the copper and freely creating or modifying text or design as Viscomi described would be practical if not inevitable. As part of the campaign against Blake and the Idea of the Book, Phillips emphasizes the existence of some preliminary drawings of elements used in the etched images, but even these are very sketchy designs copied freely to the plates, and hardly prove that Blake never created major images on the copper. The fact is, we have very few sketches of design elements that appear in any illuminated books, and even fewer combining text and design, and those circumstances suggest strongly that Blake usually composed (in Essick’s and Viscomi’s sense) his pages directly on the copper. In spite of Phillips’ arguments, it is also likely that some primary designs for the Songs never existed in any material form before Blake drew them in resist upon the plate.

Although Phillips does not say as much as he could about the process of preparing plates, photographs of beautiful facsimiles of Blake’s relief plates “by Michael Phillips” are sprinkled throughout the book (frontispiece, figs. 33, 34, pls. 30, 31, 32, 35, 65) and were featured prominently in the Tate Britain/Metropolitan Museum Blake shows of 2000-2001 (co-curated by Phillips). These are probably accurate replicas, and they are impressive to see, important to show, and inadvertently misleading, because they make it appear that the author has personally mastered Blake’s plate-making techniques. When asked, Phillips freely acknowledges that they were created photographically, using a modern light-sensitive resist to reproduce Blake’s writing and drawing on the plate, then etched in stages, as Blake did. At least two reviewers of this book have supposed that these plates were created using Blakean techniques rather than photographically generated.⁶↤ 6. Carretta was misled (443), as was the astute John Windle in his online review.

When Phillips turns to inking and printing from relief plates, he again pointedly emphasizes how hard and how long Blake must have worked, as if Essick and Viscomi had represented Blake churning out prints effortlessly or carelessly. Phillips makes much of an 1839 account by John Jackson (not the color printer John Baptist Jackson), who says that Blake spent longer wiping the ink off his plates after inking than he did inking them (20). But even if we accept this belated report at face value, the amount of time for the whole inking process would have been brief: if Blake was adept at manipulating the printer’s ball⁷↤ 7. Or a dabber, used in intaglio printing, as Viscomi proposes. that he probably used to ink his plates, basic inking (exclusive of loading the ink on the ball) might have taken at most five to ten seconds per plate, so if he spent twice or three times that much time wiping the border of the plate and removing an occasional spot of ink from the etched areas, the total time spent on inking would still be measured in seconds rather than minutes, and probably didn’t require removing the plate from the bed of the press. Unlike Phillips, I believe that Blake usually didn’t foul the etched “whites” when inking, or if a plate did get fouled occasionally it would usually happen in the middle of a large open area where wiping would be almost as easy as cleaning the border. We need not assume, as Phillips seems to, that every inking was followed by several minutes of tedious wiping (and perhaps reinking) before Blake could take a clean impression. I am not a skilled printmaker, but I can’t imagine any self-respecting artisan who would work that way—usually, one strives to do things right the first time, operating briskly with care and grace, rather than blundering through and then laboriously cleaning up the mess. When we see an illuminated page printed pristinely, with no residual ink smudges in the white areas, we are probably seeing an impression from a plate that was inked carefully and efficiently in a few seconds, without any fouling at all, and wiped only on the border edge (if the edge was wiped, as it often was not in Blake’s later practice). The entire printing process for ordinary relief printing could have averaged as little as one or at most two minutes per impression. That is more time than it would take a commercial printer to print a simple relief woodcut of the same size, but such a rate hardly “belies” (Phillips’ word) Viscomi’s estimate that Blake and his wife could in theory have done the printing for all the first copies of Innocence (700 pages) in a single week, especially if they printed two separate leaves at a time.⁸↤ 8. See Phillips 21, where Viscomi’s week shrinks to “a few days.” In fact, even printing a single page at a time at the two-minute rate, it would have been possible to complete the whole job in three eight-hour days of continuous work—“a few days” by any standard. Phillips also challenges the idea that Innocence was printed in “conjunct pairs,” which as far as I know nobody has proposed, rather than pairs of separate leaves, which Viscomi does mention as a possibility in discussing how efficient the process could have been; there was probably room in Blake’s press to print two separate leaves side by side, and conjunct pairs were certainly used, for instance, in printing MHH; see Viscomi 107, 250.

begin page 70 | ↑ back to top

For the first copies of Songs of Experience, which were printed in color as well as washed with watercolors, Phillips insists upon a printing process that is both cumbersome and extremely improbable. In their earlier writings (and in their recent refutation of the two-pull theory promoted in this book⁹↤ 9. Essick and Viscomi, “An Inquiry”; I have drastically shortened this review, originally written before I saw their piece, so as not to repeat their arguments. ) Essick and Viscomi argue that for most color printing, Blake applied a variety of thick color pigments to some areas of the relief plates in addition to his usual printing ink and printed from the etched and unetched surfaces of the shallow plate at once. He could then replenish the ink and/or colors as needed and print again, often without cleaning or even removing the plate from the press. The process would have taken only a little longer than relief printing in monochrome—if that took a minute or two per page, the most careful color printing might take twice or at most three times as long, depending on the amount of color involved. If two or more impressions were taken from each inking and pages were printed in pairs (both probable if not inevitable), the per-page rate would have been much faster than that. I think it is likely that runs from a given plate between cleanings would typically be shorter than in monochrome printing, since the plate would get messy fairly quickly, but he could still take several impressions in succession before stopping to clean accumulated ink and colors from the plate, a tedious process. In contrast, Phillips supposes that Blake must always have color-printed the Songs plates in two stages, once in monochrome and once in colors, working one leaf at a time, and twice cleaning the plate completely for every single impression. With the exception of the obviously double-printed “Nurses Song” of Experience Copy E, the isolated, slender and ambiguous evidence that Phillips produces as evidence of universal two-stage color printing—occasional suggestions of overprinting and isolated “registration” Pinholes¹⁰↤ 10. These “pinholes” do not exist (as Phillips now acknowledges), and wouldn’t mean very much if they did. See Essick and Viscomi “An Inquiry” 88-90 and “Blake’s Method” and Phillips’ response (and Martin Butlin’s) in this issue. I do not believe that Phillips or Butlin have effectively answered any of Essick’s and Viscomi’s objections to the two-pull theory. —are much better explained as accidents in one-stage printing or the consequence of subsequent hand-coloring. We would need much more and much clearer evidence before we could conclude that Blake always or often or ever worked in this extremely inefficient way.

Nevertheless, before we can resolve these disputes decisively, several more rounds of careful experiments with Blakean media are due—someone will have to show that it is possible to make Blake-like plates from scratch (with something like his lettering) and print truly Blake-like impressions in color from them efficiently. Though Viscomi’s experimental plates and the prints from them are very good¹¹↤ 11. See Blake and the Idea of the Book and the illustrations to the Essick/Viscomi article (“Blake’s Method”) in this issue. and strongly indicate that he is on the right track, I hope we can eventually get even closer to Blake’s methods. The stakes in the debate about printing are significant, for these technical questions affect virtually every aspect of Blake studies. We can’t, for instance, edit illuminated books authoritatively without understanding how he made them, and we can’t read them well without good texts. We will better understand the books when we know more about their projected audience, and, as Phillips cogently argues, understanding the ergonomics and economics of publishing illuminated books is essential to understanding his concept of audience (and to Blake’s biography, dating and establishing the illuminated canon, restoring and preserving Blake’s art, and much more).

Some of the new information about Blake’s methods and media that Phillips has uncovered is both sound and important, especially the dramatic evidence of lead white and its deterioration in some copies of Experience and his reports of other new physical and chemical analyses conducted by several other researchers. The description of Blake’s watercolor palette, with nicely reproduced samples, is likely to be a useful resource for anyone studying Blake works that include watercolor. Also compelling are Phillips’ explanations of the ways in which the technical aspects of Blake’s procedures intricately relate to contemporary historical and political circumstances and the metaphors and images in his poems and designs. All in all, this book represents a significant, albeit significantly flawed, contribution to our understanding of Blake’s Songs and the processes that went into them.

Appendix: Phillips’ Annotated Edition of Paradise Lost

In 1995 G. E. Bentley, Jr., accepted Phillips’ copy of the 1732 edition of Paradise Lost edited by Richard Bentley as a probable “Blake Book,” adding it to the carefully tended “Books Owned” section of his Blake Books Supplement; he judged that its two annotations are “persuasively signed ‘W.B.,’ probably by the poet” (322). Aside from its appearances in the Tate Britain/Metropolitan Museum Blake shows, the reproduction of the bottom of two pages in The Creation of the Songs is the first public appearance of this annotated Milton. Only one of the two annotated openings is actually reproduced, and the same pages were displayed in the London and New York exhibitions. The encircled, begin page 71 | ↑ back to top slightly botched, monogram “WB” that appears there (Phillips 56-57, fig. 25) does indeed closely resemble the very distinctive one with which Blake sometimes signed his art. But this monogram is easily imitated, and I don’t think Blake ever used it to sign a text—only visual art. Further, the rest of the lengthy annotation is definitely not in Blake’s hand and its sentiments are at odds with his usual opinions, so the monogram must be either a forgery or an improbable coincidence, the work of some other “WB.”¹²↤ 12. Probably not the lawyer William Blake, whose hand is quite different. See Bentley, “William Blake and His Circle” 177, and Phillips, “William Blake and the Sophocles Manuscript Notebook” 64.

The annotation’s tone and content are not Blakean. It begins sarcastically: “I cannot enough admire the hardiness of Bentley, who would expunge these two last Lines, as proper and surely as beautiful as any in the whole Poem, and substitute cold expressions foreign to the Author’s [Judgement del.] probable and natural meaning . . .” (56). Blake undoubtedly would have rejected Richard Bentley’s editorial intervention if he were ever aware of it. But I don’t expect to find Blake making an argument in which an appeal to what is “probable and natural” trumps all, and the icy contempt in the annotator’s use of “hardiness” (i.e., audacity) is hardly consistent with Blake’s thoughts about inspiration (or his vocabulary elsewhere).

But the handwriting in the body of the annotation is the most powerful evidence that the annotator was not Blake. It shares some features with some of Blake’s many hands, as does almost any eighteenth century roundhand. Blake occasionally used a very formal “copperplate” roundhand (as in the original title page of Vala), modulating into a variety of more or less informal hands that incorporate some of its characteristics—if it were Blake’s, this handwriting would be one of the more formal ones. But this hand is much too tight (the bead-like minuscules, particularly), too metronomically rhythmic, and too consistent in its extreme rightward tilt to be like any of Blake’s that I can remember seeing anywhere. Further, a damning cluster of improbabilities quickly becomes evident when one examines particular letterforms and common ligatures or quasi-ligatures. For example, in every one of Blake’s hands from every period, in all but one instance¹³↤ 13. The solitary exception I have ever found is in the word “should” five lines up from the bottom of page f9r of An Island; see Phillips’ edition, facing page 60; the same brief passage, however, contains four ligatures in the usual form as well as many examples of the usual up-right initial “d.” of the letters “ld” at the end of a word that I have found featuring a retrograde (uncial-style) “d,” it crosses or is too short to reach the ascender of the “l,” but the annotator’s similar retrograde “d,” occurring twice in the words “would” and “cold,” loops carefully up, over and around the “l” that precedes it. Further, the annotator always used the same retrograde “d” in all positions, whereas Blake almost always used an upright “d” in initial positions, often used it medially (and if medial and retrograde it usually swoops forward again), and he sometimes used it in final position—Blake’s most formal hand in The Four Zoas features an upright “d” in all three positions, though there are a few retrograde examples in final position.

Similarly, Blake rarely used the formal roundhand version of a Roman majuscule “A,” almost always preferring a capital that looks like a large version of the minuscule “a,” and when he did use the Roman-style “A” it invariably has a strong horizontal crossbar; the annotator’s “A” has an elaborate curved crossbar, so obviously practiced that it must have been part of his usual “A.” And so on.¹⁴↤ 14. Bentley reviews some other features of Blake’s hands in the course of refuting another mistaken Phillips attribution. See “William Blake and the Sophocles Enigma.” The “Sophocles” manuscript features several un-Blakean hands, none of them anything like the one used in the Milton annotation. Given all these uncharacteristic features of the annotator’s hand, I see no reason to believe that the poet William Blake had anything to do with this book.

Works Cited

Bentley, G. E., Jr. Blake Books Supplement. Oxford: Clarendon, 1995.

—. “William Blake and His Circle: A Checklist of Publications and Discoveries in 1994.” Blake/An Illustrated Quarterly 28:4 (spring 1995): 142-86.

—. “William Blake and the Sophocles Enigma.” Blake/An Illustrated Quarterly 31:2 (fall 1997): 65-71.

Carretta, Vincent. “Exhibition Review.” Eighteenth-Century Studies 34:3 (spring 2001): 440-45.

Erdman, David V., ed. The Complete Poetry and Prose of William Blake. New York: Doubleday, 1965; Newly rev. ed., Berkeley, CA: U of California P, 1981.

Erdman, David V., ed., with Donald K. Moore. The Notebook of William Blake. 1973. Rev. ed. New York: Readex, 1977.

Essick, Robert N. William Blake and the Language of Adam. Oxford: Clarendon, 1989.

—. William Blake, Printmaker. Princeton: Princeton UP, 1980.

Essick, Robert N. and Joseph Viscomi. “An Inquiry into William Blake’s Method of Color Printing.” Blake/An Illustrated Quarterly 35:3 (winter 2002): 74-103 and online at http://www.blakequarterly.org

Phillips, Michael, ed. An Island in the Moon. Cambridge: Cambridge UP, 1986.

Phillips, Michael. “William Blake and the Sophocles Manuscript Notebook.” Blake/An Illustrated Quarterly 31:2 (fall 1997): 44-64.

—. “William Blake: The Creation of the Songs from Manuscript to Illuminated Printing: Corrigenda and a Note on the Publication of Gilbert Imlay’s A Topographical Description of the Western Territory of North America.” Blake/An Illustrated Quarterly 35:1 (summer 2001): 30-31.

Viscomi, Joseph. Blake and the Idea of the Book. Princeton: Princeton UP, 1993.

Windle, John. [Online review of Phillips]. http://www.worldbookdealers.com/articles/br/br0000000283.asp