University of Virginia Library

The first edition of H. W. Garrod's The Poetical Works of John Keats (Oxford, 1939) was published with some 135 errors and omissions[1] in the apparatus criticus to "The Eve of St. Agnes." About twenty substantive variants were left unrecorded, and there were some forty misprints and mistakes in transcription, thirty-five instances of wrong or incomplete sigla, and forty other errors in description of the various manuscripts underlying the printed text. While it might be supposed that some of these would have been corrected in the recent second edition (1958), the fact is that not a single remedial change was introduced. But there is a more worrisome matter concerning the soundness of the text itself—not only Garrod's but all printed texts of Keats's poem—a matter that depends not so much on a rechecking of the manuscripts (for the facts surrounding the text have for some time been available) as on an editor's judgment. At least since the publication of Amy Lowell's biography (1925) it has been known that Keats's publishers enforced changes in the language of the poem; but so far no editor has attempted to repair Keats's text or in any way depart from the version first printed, in the Lamia volume of 1820.

Keats first drafted the poem during the last two weeks of January (and perhaps also the first week or so of February) 1819. His original manuscript (Garrod's H, now in the Harvard Keats Collection) was twice copied by Richard Woodhouse, legal and literary adviser to Keats's publishers, in the

transcripts designated by Garrod as W 1 and W 2 (both are at Harvard). The evidence for the order of these independent copies is ambiguous, but W 1, showing a few more errors and more blank spaces where Keats's manuscript could not be read, would seem to be the earlier; W 2 is dated by Woodhouse 20 April 1819. Early in September Keats revised the poem and had it "copied fair."[2] His fair copy, presumably the one sent to the publishers, is now lost, but in the following January, perhaps on the 15th (see Letters, II, 243), George Keats copied it in the transcript known as E (British Museum), and sometime before or afterward—probably both[3] — Woodhouse read it and entered corrections and variant readings (designated w by Garrod) between the lines and opposite the text of W 2. "The agreement . . . of E and w," writes Garrod (2nd edn., p. xli), disregarding the question of variants between them, and actually intending to include the agreement of E with the other manuscripts as well, "represents Keats' fair copy. . . . Any divergence of 1820 from Ew must be interpreted as a change made in proof either by Keats himself or by his publishers." Not counting copyist's errors, there are more than forty such substantive "divergences"; it is a nice question, in each instance, whether Keats or his publishers were responsible.

Concerning a few of them, however, I think it is time we reached a decision. On 12 September 1819 Keats read the revised form of the poem to Woodhouse, who gave the following report to the publisher John Taylor in a well-known letter of September 19 (Letters, II, 162-163):

Of the three revisions here specified (apart from the change of the hero's name), only one should cause us difficulty, the alteration of the last three lines. In all extant versions, from the first draft on, Angela is brought in (in a sense) "dead stiff & ugly," and always accompanied by the Beadsman, whose death is a similarly grotesque affair. In HW 1 W 2 and the 1820 text, "meagre face deform" describes Angela; through a change in punctuation the corresponding phrase in Ew ("with face deform") is made to apply to the Beadsman. But in no extant version is Angela "only" introduced. One could suppose a lost ending, but I rather think that Woodhouse, who heard the revised version but obviously had not yet read it, and who, furthermore, had been given plenty to think about by one of the earlier alterations that he describes, simply misunderstood the revised ending. In any event, the alteration (of whatever nature) has no place in the final text of the poem: beneath the variant ending (w) in W 2 Woodhouse wrote "Altered 1820." and two words in shorthand that are best read as "before March."[4] Presumably, before March 1820, when he was preparing his poems for publication, Keats restored the original conclusion of the poem.

The other two alterations are a simpler matter. The "additional stanza" inserted "early in the poem" is of course that given in the Ew transcripts between the present stanzas VI and VII:[5]

The alteration that Woodhouse "abused for 'a full hour by the Temple clock'" was a revision of the present lines 314-322 to read:[7] Because Woodhouse calls it an "alteration . . . of about 3 stanzas," Lowell, Garrod, and others have felt that the revised version to which he objected has been lost. But the lines just quoted fit all the other details of Woodhouse's description ("winds by degrees his arm round her, presses breast to breast, and acts all the acts of a bonâ fide husband, while she fancies she is only playing the part of a Wife in a dream. . . . there are no improper expressions but all is left to inference"). Woodhouse wrote "about 3 stanzas"; the revised lines affect two stanzas. If we again recall that Woodhouse had heard, not read, the poem when he wrote to Taylor, and also consider that the revised text, by setting off a train of uncomfortable thoughts in his mind, quite possibly would have seemed longer than it really was, it should become clear that the Ew lines we now possess are those that Woodhouse condemned as making the poem "unfit for ladies."

In his reply to Woodhouse of September 25 (Letters, II, 183), Taylor confessed that the account of this last revision excited in him "the Strongest Sentiments of Disapprobation," and he added: "Therefore my dear Richd if he [Keats] will not so far concede to my Wishes as to leave the passage as it originally stood, I must be content to admire his Poems with some other Imprint." Clearly the publishers forced the restoration of the original lines 314-322, and it is almost as certain that they forced the rejection of the additional stanza inserted between VI and VII. Once the possibility of sexual references had been opened, the lines describing "More pleasures . . . in a dizzy stream," "virgin morn," and "weeping Magdalen" (very

likely an allusion to the deserted unwed mother of Book VI of The Excursion, who is called "a weeping Magdalene" and "a rueful Magdalene" in lines 814, 987) would similarly have rendered the poem, by the publishers' standard, "unfit for ladies."

Just as clearly, the revised lines and the additional stanza should be restored to the text of the poem. In a critical article[8] I have suggested ways in which these passages heighten the irony of Madeline's self-deception and clarify Keats's condemnation of "dreaming" (Madeline's engrossment in superstitious ritual to the point of losing touch with reality). More to the point here is Keats's recoverable intention in the matter of text. On the one hand, we know that he vigorously opposed Woodhouse's objections: "He says," Woodhouse noted in the same report to Taylor (Letters, II, 163), "he does not want ladies to read his poetry: that he writes for men—& that if in the former poem [i.e., the original version of the consummation] there was an opening for doubt what took place, it was his fault for not writing clearly & comprehensibly—that he shd despise a man who would be such an eunuch in sentiment as to leave a maid, with that Character about her, in such a situation: & shod despise himself to write about it &c &c &c— and all this sort of Keats-like rhodomontade." On the other hand, to balance this, we have only Woodhouse's cryptic note in W 2 that "K. left it to his Publishers to adopt which [alterations] they pleased, & to revise the Whole." One can imagine with what willingness (and in what tone of voice) Keats surrendered that privilege.

Since an editor must always act according to principles, the problem facing an editor of Keats's poem is what, if he includes the Ew revisions so far discussed, he should do about the rest of the late manuscript readings that were rejected in 1820. A single example will serve to illustrate. In all extant transcripts the poem is called "Saint Agnes' [or Agnes] Eve" (Garrod's first textual note is wrong), which was always Keats's form of the title in his letters (see Letters, II, 58, 62, 139, 157, 174, 234, 294; so also Charles Brown, II, 276). Woodhouse and Taylor use forms of the title "The Eve of St. Agnes" (Letters, II, 162, 182), which appeared at the beginning of the poem, in the running heads, and on the title page and a half-title in the 1820 volume. Can we assume, then, that Keats's publishers altered the title? There is a strong possibility that they did alter it. Can we assume, if they did, that Keats disapproved of the change? No—because we know that he read proofs of the poem, and that in at least two instances (see Letters, II, 294-295) he insisted that his manuscript readings be restored. We can, I think, assume that he had a free hand wherever the publishers did not object specifically on moral or religious grounds.[9]

To only three other readings could the publishers have objected on such grounds. (1) At line 98, for "Mercy, Porphyro!" the manuscripts (some of them without the comma) read "Mercy, Jesu!"—which, aside from being metrically preferable, lends force to Angela's suspicions (detailed in stanza XIV) that the Porphyro confronting her may be an evil spirit. (2) At line 143, again in Angela's speech, for "Go, go!" the manuscripts (with minor variations) read "O Christ!"—a natural enough reaction just after the "cruel . . . impious . . . wicked" Porphyro has proposed his "stratagem." (3) At lines 145-147, for

the manuscripts (with minor variations) read —which, owing to Paul's association with chastity, embodies an ironic oath especially appropriate to Porphyro's plot against Madeline, but would have aroused the same disapproval from the publishers as the sexual overtones of the additional stanza between VI and VII.

The nearly forty other late manuscript readings that were ignored or rejected in 1820 have nothing in them offensive from a moral or religious point of view. The alterations are primarily stylistic, and in every instance we cannot be sure that Keats himself did not make the change in proof, or at least concur in the change if it was made by someone else. It is this circumstance that prevents us from always accepting the Ew agreement as Keats's final text: he could have been responsible for every single one of the stylistic alterations, from the title on.

Employing the principle that a proper text of the poem will embody the latest readings intended by the poet, including those that there is good reason to think were rejected by the publishers against the poet's wishes, future editors of Keats's poem, whether making a scholarly text or putting together selections for an anthology, should restore the Ew version of lines 314-322 and the additional stanza between VI and VII—this last even though it will result in the subsequent renumbering of stanzas and lines through most of the poem. Editors may wish to restore the manuscript readings at lines 98, 143, and 145-147; at present, while I myself favor them, these seem a matter of individual option. Otherwise (saving for the need to tidy up the apparatus criticus in the next Oxford edition) the text we now have will serve.

Keats at one time thought enough of "The Eve of St. Agnes" to request that it appear first among the poems in the 1820 volume (Letters, II, 276). It seems less than fair not to accord it the best text possible; certainly the only slightly more innocent version we have always had, whether or not fit for ladies, has been often enough misunderstood.