Erdman (786-87) and Bentley (xliii-xliv) both discuss the problems. Erdman admits that he is "inclined . . . to read commas or periods according to the contextual expectations." Their versions of The Book of Los (which exists in only one copy) differ over punctuation largely because Blake's marks cannot be directly transcribed into standard typography. Differences of interpretation or representation between the two editions can be readily multiplied: see Murray's review of Bentley, which concludes, "In the long run, the problems and the contradictory solutions available for them probably exceed even a theoretic comprehension, much more any set of workable editorial principles" (160).
 In Experience, ll. 1 and 5 have Blake's common non-syntactic full stop: "When the voices of children. are heard on the green / Then come home my children. the sun is gone down." The corresponding lines in Innocence, verbally exactly the same, both omit the stop. A similar apparent indifference about punctuation is suggested by different punctuations for parallel clauses in adjacent stanzas of "A Cradle Song":
Sweet smiles Mothers smiles
All the livelong night beguiles.
// [. . .]/[. . .]/
Sweet moans, sweeter smiles,
All the dovelike moans beguiles.
(The differences between the two stanzas are present irrespective of how the punctuation of the original is rendered.)
 Peter Middleton presents a confused argument for the importance of what he takes to be Blake's punctuation, but on the assumption that the forms and significances of the punctuation in the originals are those of letterpress. In the five-line passage mainly discussed he contrives to misquote the punctuation, the lineation, the paragraphing and the words.
 Johnson and Grant's selection, based on their own studies of original sources, also draws on both Erdman's edition and the William Blake Archive transcriptions: see its "Textual Technicalities," 599-602.
 For example, Blake's use of the question mark: contrast its complete absence from the questions of "The Lamb" with its copious presence in the questions of "The Tiger"; or, in "Earth's Answer," its absence from questions in stanza 3 with its presence in the middle of questions as well as at their end in stanza 4.
 See, for example, The Marriage of Heaven and Hell, "A Song of Liberty," Chorus: "Let the Priests of the Raven of dawn, no longer in deadly black, with hoarse note curse the sons of joy." Blake's meaning is evidently "[. . .] no longer, in deadly black [. . .]": priests, indicatively black-gowned, should no longer curse (not, as the punctuation on a modern interpretation implies, priests, wearing something other than their usual black gowns, should continue to curse).
 Fuller, Poetry and Prose. This contains the major works, mostly complete, each with an individual introduction and detailed annotation, including accounts of designs, and similarly annotated selections from The Four Zoas, Milton, and Jerusalem.
 "A Cradle Song." This hypothetical lay-out is based on Bentley's reading of the original punctuation:
Sweet moans. dovelike sighs.
Chase not slumber from thy eyes.
Sweet moans. sweeter smiles.
All the dovelike moans beguiles.
The stanza exemplifies the difficulties of rendering the punctuation of the original in letterpress. Erdman (12) and Lincoln (pl. 16) both represent the punctuation of this stanza differently from Bentley (37) and from each other.
 Cf. "The Garden of Love," where Blake marks internal rhyme by an a-syntactic comma: "And Priests in black gowns, were walking their rounds, / And binding with briars, my joys & desires." (On the rhyme "gowns" / "rounds," see my Poetry and Prose 96.)
 Cf. "Holy Thursday," Songs of Experience, stanza 1, in which each line ends with a syntactically redundant full stop. Also Songs of Experience, "Introduction," "That might controll. / The starry pole" (ll. 8-9), and "Earth's Answer," "Break this heavy chain. / That does freeze my bones around" (ll. 21-22)—though both of these are also examples of the ambiguous forms of Blake's punctuation. In the first case Bentley (174) gives no punctuation (perhaps interpreting the mark undoubtedly present in some copies as a spatter); in both cases Erdman gives a comma (18, 19), Lincoln a full stop (pls. 30, 31).
 Contracted forms are not retained by Stevenson, but are retained by Mason, and also in the more partial and conservative modernization of Keynes.
 One exception to Blake's ed indicating èd in his metrically regular verse comes in the "Introduction" to Songs of Innocence: "So I piped with merry chear, / ... / So I piped, he wept to hear" (ll. 6, 8). Here Blake may have thought of the "e" as necessary to modify the "i" and to distinguish "pipe" from "pip": cf. Songs of Experience, "The Little Vagabond," l. 3, "use'd" (Notebook, p. 105, "usd"), and Milton, pl. 11.47, "tone'd" (where the "e" is retained to modify the "o"), both cases of other verbs identical in form with nouns. In these cases the apostrophe is apparently used to indicate that the "e" should not be pronounced. See also my note to "Ah, Sunflower," Songs of Experience, l. 5: Poetry and Prose 95.
 In the fair copy Blake actually marks with an accent the "e" of "asked" in l. 1 (to distinguish it from monosyllabic "asked" in l. 3), and he maintains the distinction from the Notebook version between ed (l. 2) and 'd (ll. 7, 8, 12). "Turned" (l. 2) must be disyllabic: l. 2 is not otherwise the trimeter required by the ballad meter stanza structure to match l. 4.
 For example, "Nurses Song," l. 17, "The little ones leaped and shouted and laugh'd," where the same ed / 'd distinction is made in both versions.
 Here the distinction is reproduced with almost complete consistency. The single exception is "The Angel" where in the Notebook l. 6 has "wiped" (which, since it is not metrically impossible, may indicate only that in engraving the poem Blake changed his mind). In the draft of "My Pretty Rose Tree," "But my rose was turned from me" was altered to "But my rose turnd away with Jealousy" (l. 7); in the draft of "London" "The german forged links I hear" was altered to "The mind forgd manacles I hear" (l. 8)—both changes that clearly indicate Blake observing the 'd / ed distinction.
 See, for example, The Book of Urizen, pls. 10-13; The Four Zoas, pp. 54-55; Milton, pl. b: despite the changes of form between The Book of Urizen and The Four Zoas (short to long lines), and the overall abbreviation of the passage in Milton, there is only one single change of a verb form.
 See, for example, The Four Zoas pp. 39.17-19, 40.2-20, 41.1-18, and 42.1-19, which became Jerusalem, pl. 29.33-82: in a context of several minor changes the d (or 'd) and ed distinction is reproduced with complete consistency. Exceptions can be found: America pl. 8.6-12 is repeated verbatim in The Four Zoas p. 134.18-24, except that one ed becomes d. Whether this indicates Blake's occasional inconsistency or a change of mind about the rhythm is, of course, impossible to tell.
 The distinction is ignored by Stevenson and Mason. The only edition actively to bring it out in the text is that of Sloss and Wallis, which retains the non-syllabic forms and renders the syllabic form èd. Johnson and Grant indicate syllabic ed in their annotation.
 New Literary History 30 (1999), 25-56.
 On my basic sympathy for McGann's long-running arguments about the subjectivity of criticism, and an attempt to draw some different conclusions, see my "Keats and Anti-Romantic Ideology."
 On the Aesthetic Education of Man, in a Series of Letters, ed. and trans. by Elizabeth M. Wilkinson and L. A. Willoughby, Oxford: Clarendon, 1967; throughout, but see especially Letter 15.
 For an attempt to give greater currency to critical practices broadly of this kind, synthesizing analytical, critical and creative work, and offering the reader a range of interactive strategies for structured play in re-writing texts, see Rob Pope, Textual Intervention: Critical and Creative Strategies for Literary Studies.
 Tamburlaine the Great, Parts 1 and 2. In The Complete Works of Christopher Marlowe, vol. 5 (Oxford: Clarendon P, 1998).