Open Book Publishers logo Open Access logo
  • button
  • button
  • button
GO TO...
book cover


A Vision: The Revised 1937 Edition, edited by Margaret Mills Harper and Catherine E. Paul, The Collected Works of W. B. Yeats, Vol. XIV (New York: Scribner, 2015), pp. li + 503.
A Review Essay

Colin McDowell1

© Colin McDowell, CC BY 4.0

[R]eaders—and editors—must think for themselves
—Richard J. Finneran


Finally, with the publication of The Collected Works of W. B. Yeats, Volume XIV: A Vision. The Revised 1937 Edition, scholars have decent and readily available texts of the original 1925 edition of A Vision and Yeats’s later version of 1937, both of them edited by Margaret Mills Harper and Catherine E. Paul. When I say ‘readily available’, I do not refer to the eBook versions also published by Scribner, which are available for purchase only in the United States. Whether this is testimony to the American view that only Americans matter, or whether it is simply due to demarcation disputes amongst publishers, I do not know: I can only say that it is an irritating restriction on scholarship. Nevertheless, the edition is a monument of scholarship and well worth placing alongside other major achievements in The Collected Works of W. B. Yeats, such as the editions of Autobiographies and Later Essays. Not only has the text been carefully collated, the edition includes generous quotations from the unpublished typescripts, a lengthy Editorial Introduction, tables of ‘Proofs, Versions, Emendations, and Hyphenations’, and 170 pages of notes. With this last in particular, it is as though the prior publication by George Mills Harper and Walter Kelly Hood of A Critical Edition of Yeats’s A Vision (1925) had somewhat inhibited the textual apparatus of The Collected Works of W. B. Yeats. Volume XIII: A Vision. The Original Version, and only with this edition have the editors felt free to do as they would have wished as regards annotation. But this is mere speculation. It is also quite likely that the reviews of the earlier edition have guided the course of this one: Neil Mann’s careful appraisal in YA18 (265–96) in particular has surely had an effect for the better, as we shall see.2

In what follows, I am merely chipping away at the edges; it should not be seen as impugning the structural soundness of the achievement.

The reader new to A Vision who is looking for guidance as to what the book is about will not find it the Editorial Introduction to this edition. The reason may be that the editors did not wish to repeat material from their Editorial Introduction to the 1925 edition, which to some extent functions as an Introduction to both volumes; but a paragraph or two here may not have gone amiss.

This aside, the Editors’ Introduction is typically dense, and may reflect Harper and Paul’s over-familiarity with publication history and the ins-and-outs of manuscripts and typescripts. Reading of shorter previous studies may prepare the reader. One longs for the relative clarity of Richard Finneran’s 1977 article ‘On Editing Yeats: The Text of A Vision (1937)’ (Texas Studies in Literature and Language, 19:1 [Spring 1977], 119–34), or Connie K. Hood’s ‘The Remaking of A Vision’ (YAACTS1 [1983], 33–67); but, admittedly, these are difficult matters, things have moved on since Finneran and Hood wrote, and you can only simplify so much without distortion.

Perhaps taking to heart Neil Mann’s complaint that the editors’ A Vision. The Original Version included only a single reference to contemporary reviews (YA18 265 n. 1), the editors have over-compensated here. While some of these reviews may have contributed to how Yeats revised the work, most seem to have little relevance, and the editors don’t manage to justify their inclusion, unless they are intended to provide light relief from the textual ins-and-outs of the rest of the Introduction: they certainly don’t offer any help in coming to terms with the subject matter, but instead remind us what people say who feel they have to comment on what they do not understand. Nor do the editors mention that most of the reviews have been collected on Mann’s website in any case,, thus obviating the need for interested readers to track down musty journals. However, they do quote snippets from one review that Mann was unable to locate, and reference reviews of ‘Stories of Michael Robartes and His Friends’ that Mann did not include.

The editors also find space to indulge in some needless repetition, although perhaps it is just over-zealous signposting. For example, in two consecutive pages they inform us that ‘We have charted a conservative course, keeping our emendations minimal, noting all in Appendix 1, Table 4’, ‘Although all our emendations are noted in Appendix 1, Table 4, we explain our principles here’, and ‘All emendations are noted in Appendix 1, Table 4’ (CW14 l-li).

As for errors and misprints not related to the difficulties of the system, the volume is surprisingly free for a book of this size and complexity: I have noticed only about a dozen. The most egregious error is the identification of Henry James as one of the philosophers Yeats was reading in 1926 (CW14 xxvi), an error which necessitates an emendation to the Index as well, given that the James brothers are equally represented in this volume and the Index refers to only one of them. Nor did M. M. Rossi ‘co-edit’ Berkeley (CW14 xxviii), certainly not for the book to which Yeats wrote an Introduction: he added a philosophical commentary to what is largely a biographical study. With misprints proper, Peter Liebregts’ book is called Centaurs in the Twilights (CW14 xvi: perhaps the phrase ‘deer in the headlights’ caused contamination); the philosopher R. G. Collingwood is given as ‘R. C. Collingwood’ (CW14 394 n. 52); ‘Plunket’ is printed as ‘Plunkett’, both in the text and the notes (CW14 185, 300 n. 24, 431 n. 46, 491), as per Yeats’s misspelling (although it had been corrected, both in text and notes, by the same editors, in their edition of the 1925 version [CW13 361]); one of Yeats’s poems is referred to as ‘The Double Visions of Michael Robartes’ (CW14 394 n. 55); the Index under ‘Michael Robartes’ directs the reader to the ‘“Michael Robartes Foretells” TS, 315n28’ (CW14 492), whereas that reference is to a published passage in CW13; and the note on Gentile [CW14 350 n. 21] points the reader to ‘n34 below’, when the note is actually 35. But these are easily corrected minor matters. They do not detract from the solid achievement of the whole, and I mention them so that they may be corrected in any future printings.

The extreme shorthand of the notes gives a cramped impression and is occasionally confusing, it sometimes being unclear if the reference is to an internal page or an external source; see, for example, ‘On Phidias’s Zeus Olympios, see Pausanias, 5.11.1 (Description of Greece, Books 2–5, trans. W. H. S. Jones [Loeb, 1926], 436–37), and note 688 above’ (CW14 453), where I have no idea what ‘note 688 above’ means. The following does not make for easy reading, but is no doubt useful: ‘An Adventure and its authors appear numerous times in the AS, and WBY mentions it in AVA (136, 286–87 n126) and elsewhere. See also Plays 722; LE 115, 270, 272, 354 n35b, 452 nn36 and 36a, and 360 n10. See YVP 1:307, 319, 3:290; MYV 1:179, 224–25’ (CW14 408 n. 38; the quote from Moberly in the same note is of course from LTWBY 347–48, not L 347–48); but when the reader is faced with ‘Robartes’s death is alluded to in “The Adoration of the Magi” (Myth1 310, Myth2 202) and mentioned in notes to Michael Robartes and the Dancer (VP 821) and in Owen Aherne’s “Introduction” to AVA (lviii-lix); see also 327 n. 1 and 339 n. 60’ (CW14 341 n. 6), it is not immediately apparent to the uninitiated that the semi-colon signifies that the references which follow are to CW14 itself.

To some extent, what one chooses to annotate is a subjective judgment, as is what one chooses to include in the annotation once chosen. Many of the annotations to the 1925 edition closely followed those of CVA, adding references to Yeats’s Vision Papers where appropriate. The 1925 edition, being closer in time and subject matter to the system’s genesis, called for more references to the automatic script and their publication in Yeats’s Vision Papers, whereas this edition dictates more reference to the typescripts. However, the editors do not entirely eschew quotations from the automatic script, which is to be welcomed. Nor does interpretation loom large in the endnotes, as the editors’ brief was to provide materials for interpretation and not the interpretation itself. Nevertheless, there are several passages that stray from this self-imposed limitation, and these also are to be welcomed, e.g. the note explaining that Yeats’s ‘right to left’ in a particular passage should be ‘left to right’ (CW14 351 n. 29), or the note explaining why the astrological symbols have been reversed in the diagram of The Great Wheel from one edition to another (CW14 344–45 n. 1 #4). And given that there is a large overlap between the text of ‘The Twenty-Eight Incarnations’ and ‘Dove or Swan’ in the two editions, how do the editors handle those particular annotations? They appear to be largely the same; the changed passages are often helpful expansions or corrections pointed out by reviewers. For example, the long note on ‘fabulous, formless darkness’ (CW14 447–48 n. 60) credits Neil Mann’s review of their 1925 edition for the proper attribution, while the note on the phrase ‘the dog bays the Moon’ (CW14 356 n. 50) has added a reference to the Tarot card ‘The Moon’, also via Mann. (I do have a quibble about the way the latter note has been handled. Reference to Kathleen Raine’s Yeats, the Tarot and the Golden Dawn might have been more appropriate than reference to A. E. Waite’s popularized account. Raine includes pictures of the card from the packs of both Yeatses, and quotes the meaning it had in the Golden Dawn, where there was no reference to Waite’s misleading dog AND wolf.) In fact, most of the matters addressed in Mann’s review have been remedied, as if they had been ticked off one by one. Paul’s knowledge of Ezra Pound comes more into play in the Notes than it did in parallel passages in the earlier edition (e.g. on Dowson and Landor, not to mention Wordsworth (CW14 365 n. 102 compared with CW13 251 n. 131; CW14 354 n. 42 compared with CW13 238 n. 49; CW14 366 n. 108 compared with CW13 252 n. 137). Presumably Paul is responsible for the more obscure Ezra Pound references in the notes to ‘A Packet for Ezra Pound’ (see the reference to Massimo Bacigalupo’s article in Quaderni di Palazzo Serra, CW14 310 n. 15), as well as later ones in ‘The Completed Symbol’ (e.g. CW14 433 n. 59); it is not however mentioned that Pound is probably the source for Yeats’s initial interest in Grosseteste’s theories of light, with Duhem only coming into play later (see CW14 140, 385–86 notes 11–12). Paul’s expertise in Pound is also apparent in parts of the editorial introduction, where whole paragraphs are repeated from her 2011 article ‘Compiling A Packet for Ezra Pound’, including what I am reliably informed are faulty transcriptions of ‘questions from Mary Devenport O’Neill’ relating to A Vision (CW14 xxix).3 As an aside, Paul’s article is listed in the Abbreviations as ‘Paul’, while Russell Murphy’s essential study of the significance of Byzantium in the system, ‘“Old Rocky Face, look forth”: W. B. Yeats, the Christ Pantokrator, and the Soul’s History (The Photographic Record)’, YAACTS 14 (1996), which is referenced more often, is not so listed, while the sole Index entry for Murphy is not to the first mention of him or his article, so that someone looking for the full citation cannot find it easily. Nor is the Index complete with its references to Paul. Obviously, such problems do not occur with electronic editions of books.4

The annotations for ‘The Great Year of the Ancients’ in particular are very full, with copious quotations from Pierre Duhem’s Le Système du monde and explanations of astronomical matters aimed at an audience which may find such topics confusing.

Some of the annotations illuminate in unexpected ways. The note to ‘the Muses sometimes form in those low haunts their most lasting attachments’ (CW14 19) compares it to Yeats’s remark to Laura Riding that ‘poets were good liars who never forgot that the Muses were women who liked the embrace of gay warty lads’ (CW14 323 n. 65). However, checking Alexander Charles Sutherland’s 1978 dissertation ‘Yeats’s Revisions of A Vision: A Study of the Text, with Appendices of Textual Variants and Annotations’, Ph.D. diss., New York University, I find that he said it first. Sutherland’s dissertation is referred to in the same note, but is not given as the source of this observation.5 The fact that Sutherland was used perhaps explains why the reference is to Yeats’s letter to Dorothy Wellesley in Wade’s edition of the Letters and not to the original letter to Riding in CL InteLex 6563. I also find illuminating the note which juxtaposes a list of Indian schools of philosophy studied by Yeats with Dermott Mac Manus’ claim that the Yeatses in later life gave up spiritualism in favour of ‘the tradition of Indian thought’ (CW14 384 n. 4). This is in spite of W. J. Mc Cormack’s contemptuous dismissal of Mac Manus as ‘the pseudo-Hindoo-guru’ (‘We Irish’ in Europe: Yeats, Berkeley and Joseph Hone [University College Dublin Press, 2010], 63). (Alternatively, he is ‘the swashbuckling fascist and Higher Hindu’ in Blood Kindred: W. B. Yeats, the Life, the Death, the Politics [London: Pimlico, 2005, 13]; later, in the same book, the characterization ‘the philosophical fascist and convert-Catholic, higher-Hindu’ is used [285]. Mc Cormack is nothing if not tiresomely picturesque.) If only people and their aspirations could always be summed up so cavalierly.


Before getting on to the textual editing, I should clear a little ground. The editors dedicate their edition to Walter Kelly Hood and Connie K. Hood, who were the original designated editors of both volumes of A Vision for The Collected Works of W. B. Yeats (CW14 xi). Walter K. Hood was of course one of the editors of A Critical Edition of Yeats’s A Vision (1925), contributing the bulk of the annotations, while George Mills Harper wrote the Editorial Introduction and added to the notes where reference to the Automatic Script was required. Walter K. Hood was to have contributed historical and explanatory notes for the projected 1937 edition, while Connie K. Hood was to have edited the text. Her unpublished 1983 dissertation ‘A Search for Authority: Prolegomena to a Definitive Critical Edition of W. B. Yeats’s ‘A Vision’ (1937)’ Ph. D. diss., University of Tennessee, formed the basis for the intended textual emendations (‘Search’ vi, viii). It therefore has a direct bearing on Harper and Paul’s project.

It is perhaps worth emphasizing that Hood’s dissertation was completed before the main sources for the scholarly study of A Vision were published. The Making of Yeats’s ‘A Vision’: A Study of the Automatic Script by George Mills Harper came out in 1987 (although Hood read the first four chapters before publication), while the first three volumes of Yeats’s Vision Papers (under various editors) appeared in 1992. (The fourth and final volume was published in 2001.) It was written before O’Shea’s A Descriptive Catalog of W. B. Yeats’s Library (1985), the Compilation of the NLI Collection List No. 60: Occult Papers of W. B. Yeats, and the publication of the InteLex edition of Yeats’s Collected Letters.6 It was also completed just before the controversy erupted over the editing of Yeats’s poems, when questions of delegation arose, and the idea was mooted and energetically defended by the late Richard Finneran, that a process was begun after Yeats’s death ‘of—not to put too fine a point on it—corrupting the texts which he had worked so hard to perfect’.7 The enormity of the task undertaken by the Hoods, Connie K. Hood in particular, must be acknowledged, and Harper and Paul state that ‘[t]o observe that this edition would not have been possible without their extensive archival and contextual work is to understate drastically, and we are deeply in their debt’ (CW14 xi). This is by no means the sort of routine encomium you will find in books written by academics. Hood was nothing if not thorough; there is thus inevitably a large overlap between what she said in the historical narrative of her dissertation and what is covered by Harper and Paul in their Editorial Introduction. (They have of course updated and standardized the references.) Nevertheless, the resultant edition of the text differs from anything the Hoods would have produced, as I shall demonstrate.

Harper and Paul do not mention Finneran’s charge directly, although they gesture towards his stance by using arch quotation marks around words such as ‘corrected’, ‘authorial’, ‘intentions’ and ‘permission’ (CW14 xlviii-l). Finneran, of course, was one of the original general editors of The Collected Works of W. B. Yeats, a ‘series invested in the authorial intentions of WBY’ (CW14 xlix). A Vision, however, is a special case, as it was originally based on the automatic writing of George Yeats, who by this fact can plausibly be called ‘an author of equal standing’ (CW14 l) with her husband. Harper and Paul explain their problem as follows: ‘First is the difficulty of ascertaining [Yeats’s] intentions in the last version of the text published in his lifetime. Second is the question of what to do with emendations made to that text by trusted collaborators after his death’ (CW14 xlix). The difficulty of ascertaining Yeats’s intentions is restricted by them to manuscripts, typescripts and galley proofs: more specifically, to the fact that there is not a complete record. I would stress, rather, the fact that Yeats himself was averse to proof-reading: he did, after all, have other calls on his time, was frequently ill during the relevant time-period, his eyesight was poor, and his spelling was notoriously bad.8 Add to this the fact that A Vision is not a short piece of prose and its difficult subject matter makes for difficult proof-reading; it was also a book which Yeats could never finish to his own satisfaction.

Even if we had what Harper and Paul ask for, and lament the absence of, ‘a complete extant setting copy’ (CW14 xxxix), I do not think this would resolve all questions of authorial intention: unless, of course, one takes an extremely restrictive view of what that entails. And, obviously, this is where questions arise of delegation to ‘trusted collaborators’. Harper and Paul have decided to incorporate most of George Yeats’s corrections as marked in various copies of A Vision (1937/38). The other trusted collaborator was Thomas Mark, to whom Yeats delegated copy-editing rights during his lifetime,9 and who after Yeats’s death mailed his corrected proofs to GY for approval, apparently receiving it (CW14 xlvi, 303 n. 78). Thus, it might be argued that they had been passed by ‘an author of equal standing’, meaning George Yeats, and should therefore be incorporated into an emended text. But this would involve too many assumptions about specifics. Harper and Paul have instead chosen to relegate Mark’s corrections to the endnotes and an appendix, ‘[i]n keeping with established series policy’ (CW14 li).

However, even one of the original general editors of the Collected Works of W. B. Yeats, Richard Finneran himself, once thought that the 1962 London edition (copy-edited by Mark) was the best available (see Finneran’s ‘On Editing Yeats: The Text of A Vision [1937]’, 124). This is despite the fact that it introduced errors into the text and did not make some corrections that had been made in the 1956/61 edition, which in itself had also introduced errors. The later Finneran would not have been so lenient: he had by that stage painted himself into a corner through being forced to defend his edition of the poems, and lost no opportunity to denigrate the work of Mark, and to a lesser extent, the work of GY. The editors of A Vision in The Collected Works overstate things to their own advantage when they write that ‘We treat [in the section entitled EDITIONS AFTER 1937 (CW14 xlvi-xlviii), a brief two and three-quarter pages] the editing process for those [1956/61 and 1962] editions, which rather than producing the definitive editions they advertised introduced further errors’ (CW14 xxv), as though those editions did not also introduce some readings that were preferable to those of the original edition.10 The main point of the section appears to have been to demonstrate that the editing of the 1956/61 and 1962 editions meant that ‘we are many textual stages away from the 1937 edition of A Vision’ (CW14 xlviii), and thus may safely be ignored. I remain unconvinced by this argument. All told, while it is hard not to agree with the editors when they conclude that ‘the ideal of a perfected text [is] impossible’ (CW14 l), I think the decision to restrict Mark’s emendations to the endnotes and an appendix was not the best course of action.

Harper and Paul follow Connie K. Hood in espousing the editorial principle of least tampering with the received text. They write that their ideal is to present in its best possible light ‘the 1937 edition of A Vision, the final version of the text over which W. B. Yeats had “authorial” control’ (CW14 xlviii), adding that

We have charted a conservative course, keeping our emendations minimal, noting all in Appendix 1, Table 4. In that same Appendix, Tables 1 and 2 compile changes marked by the Yeatses in their copies of A Vision (1937, YL 2434; and 1938, YL 2435), and these tables also note changes made in the copy of A Vision (1938) in the Alspach collection. Table 3 compiles corrections proposed for or made in posthumous editions of A Vision, comparing them with the 1937 text. These apparatuses allow readers both to reconstruct different published and imagined states of the text, and to consider our own editorial practice, and thereby to ruminate over the textual authority of Thomas Mark and GY while seeing what text WBY left behind (CW14 l).

This is all admirably democratic, but perhaps it leaves too much to the reader. Hood also invokes a ‘conservative course’, writing that ‘The editorial position of this [projected] edition is extremely conservative and at the same time broadly eclectic. Changes actually made in the text must be justified from manuscript sources prior to the 1937 edition (except for the three changes Yeats himself made on his wife’s copy) except in a few cases where strong internal evidence suggests printers’ or typists’ corruptions (as “Phase 29” for “Phase 28” [cf. CW14 243, 269])’ (‘Search’ 163). It is therefore of interest to see how the Hoods’ projected edition would have compared with Harper and Paul’s actual edition.


Finneran may not have accepted what he and George Bornstein called ‘the weak argument of posthumous delegated authority’ (CW4 324), but this of course does not absolve an editor from making his or her own editorial choices. At its simplest level, you would think that this would be a matter of reading the text as printed to see that the words make sense, that the sentences are grammatical, and that syntax is not tortured: in other words, routine proofreading and copyediting. This is what Thomas Mark did, and did well. Not to use his expertise to assist your own is to operate with one hand tied behind your back; simply listing his emendations in an Appendix is not the same thing as weighing his decisions against the text and against your own. Moreover, it is not just individual sentences that have to make sense. The reader needs to put the sentences into a coherent whole: passage has to be compared with like passage and consistencies (or their converse) drawn out. If there are inconsistencies, it would be ideal to ask the writer what was intended. When the author is not available, an editor should at least make some attempt to correct those minor inconsistencies where it is apparent that lack of attention was most likely involved. This of course should not be done silently; the reader must be given the arguments. If an editor chooses not to correct inconsistencies because his or her editorial principles dictate that this should not be done, then it would still be preferable if the reader could be confident that the editor was aware of them. There is a further consideration the reader may or may not choose to take seriously. Yeats believed the system was complete but that he did not always fully understand it. His text may therefore be regarded as an imperfect embodiment of the system, which the reader is invited to complete. I myself occasionally adopt this viewpoint as a heuristic device.

Harper and Paul’s emendations, as set out in their Table 4, Emendations to the Copy Text (CW14 267–73), fall into three broad categories: spelling corrections, including titles of poems and proper names; GY’s corrections to the Yeatses’ copies of A Vision (1937) and A Vision (1938); and changes made to impose terminological consistency.

The first of these requirements, that of spelling, is I imagine one imposed by a combination of the publisher’s house style and the requirements of the fourteen-volume Collected Works of W. B. Yeats. While it is sometimes difficult to distinguish between the two, publishers don’t want to look as though they don’t know their own business: their main concern is not to alienate readers by giving them the impression that they have been content with sloppy production.

I have no problems with the changing of poem titles from italics to roman type in quotation marks; after all, this is a process begun by Yeats and carried through by GY and Thomas Mark.11 But the editors have also changed the spelling of proper names and have Anglicized Cyrillic names (CW14 li). Thus, Yeats’s ‘Tolstoi’ and ‘Dostoieffsky’, his ‘Michael Angelo’ and Leibnitz’, have been modernized. You can see the point: no publisher wants to put off readers with unfamiliar spellings. It may be noted, though, that these spellings were perfectly acceptable when Yeats wrote. If you leave them as Yeats left them, they serve as a reminder to the reader of the time when the book was written; if you change them, you lose this flavour.12 But other proper names have been corrected: Alcemon, Aeslepius, Philaus, Dionysius, Zazuki (although with the last, the editor’s confusingly leave the misspelling of the InteLex Collected Letters without a ‘sic’ when they quote that source as saying that ‘Yeats’s “Zuzuki” is written mistakenly for “Susuki”’, CW14 399 n. 72). This follows editorial procedure adopted in other volumes in The Collected Works of W. B. Yeats. However, for some unknown reason, Yeats’s ‘Grillion Club’ has not been corrected to ‘Grillion’s Club’, though the editors obviously know that the latter is correct, spelling it thus in the notes and index. Thomas Mark changed ‘Leibnitz’ to ‘Leibniz’ and ‘Grillion Club’ to ‘Grillion’s Club’, but left the others. He emended ‘Brama’ in the quotation from Keats’s ‘Endymion’ to ‘Brahma’ (cf. The Poems of John Keats, edited by Sidney Colvin [2 vols.; London: Chatto & Windus, 1920; YL 1055, 1:242), although this misspelling is left unremarked on by Harper and Paul (CW14 81, 243). Harper and Paul also change ‘Crickmaa’ to ‘Cruchmaa’ on the strength of GY’s annotation (CW14 21, 228, 236, 326–27 n. 75), whereas a quick Google would seem to support Mark’s 1962 spelling of ‘Cruachmaa’. I leave it to Irish specialists to debate that one.

Some of these spelling corrections, as opposed to modernizations, may be seen as reflections of Yeats’s intention, but it is a slippery slope. Hood, for example, suggests that ‘Aeslepius’ may be ‘a printing error based on misreading of Yeats’s poor handwriting’ (‘Search’ 230), and one can imagine that at least some of the other misspellings have a similar origin. My preference would have been for the editors to refer to drafts and typescripts before altering the text; endnotes are the place for proper identification and spelling, as they would have been the place to note an emendation that had actually been made.

There is thus the large question of whether to emend the text on the basis of what George Yeats and Thomas Mark chose to do. Their corrections to the text are available for inspection, as Warwick Gould once remarked.13 We have seen above that Harper and Paul noted in an appendix GY’s annotations in copies of A Vision (1937) and A Vision (1938), and many of them were incorporated by them into their text, presumably (although it is not stated outright) on the basis of GY’s ‘semi-authorial status’ (CW14 l). As test cases, let us examine two changes made on this basis by Harper and Paul where change might seem unnecessary.

The first is the word ‘adaptation’. Harper and Paul change WBY’s ‘adaption’ and ‘self-adaption’ to ‘adaptation’ and ‘self-adaptation’, basing their decision on ‘GY in YL 2434’. Connie Hood explains, quite correctly (‘Search’ 223), that ‘adaption’ is a valid English word, and it appears in the OED with examples from Swift and Dickens. The original 1937 edition has examples of both ‘adaption’ and ‘adaptation’, with three of the former and one of the latter (see CW14 229, 249, 270 for the places where ‘adaption’ originally appeared; the unchanged ‘adaptation’ occurs on CW14 71). One might think that the word is GY’s, since it is taken from the Automatic Script. However, WBY and GY use both spellings, if we are to trust the editors’ transcriptions in the Yeats’s Vision Papers. On October 24 1918, WBY asks, ‘What do you mean by self-adaptation?’ (YVP2 465), referring back to the session of January 3 (YVP1 192), which implies that the word first used is ‘adaptation’, and that it has been used by GY; while on February 1 1920, GY herself uses ‘adaption’ (YVP2 532). Personally, I prefer ‘adaptation’, but by crediting the amendment to GY’s annotation alone, the editors gloss over the fact that the word had been the subject of debate between GY and Thomas Mark, and had been altered in both sets of Coole proofs (CW14 249).

The second word I wish to examine is ‘cabbala’ and its derivatives. The evidence of letters and early published texts suggests that Yeats originally spelt the word with a single ‘b’, although he varied between an initial ‘k’ and ‘c’. In the 1937 edition of A Vision, Yeats used the forms ‘Cabala’, ‘Cabalists’ and ‘Cabalistic’, with the 1962 edition following this spelling consistently and the 1956/61 edition fitfully.14 Harper and Paul change all of these occurrences to a double ‘b’ on the authority of George Yeats; the source is given as ‘GY in YL 2434’ (CW14 267, 268, 273). In this case, however, we have definite proof that Yeats himself authorized Thomas Mark to make the change, at least insofar as the case of Autobiographies was concerned. Yeats had previously used ‘cabalistic’ (twice) and ‘Christian Cabala’ (once) in The Trembling of the Veil (1922) (although ‘Cabbala’ was also used once), while in the 1926/27 edition of Autobiographies the spelling varies, with the double ‘b’ having a preponderance due to addition of ‘A Biographical Fragment’ from The Criterion, with its seven usages (the inclusion of the new Section VI to ‘Hodos Chameliontos’ having added two occurrences of the single ‘b’). Understandably, Mark queried the preferred spelling when he was correcting proofs for the Edition de Luxe in 1932. Warwick Gould and Deirdre Toomey explain: ‘By underdotting “cabbalistic” in the text (Au 371) and scoring out Mark’s alternatives, WBY had indicated his preference’ (Myth 2005 xcvi and note 26). GY and Mark were thus carrying out Yeats’s decided spelling when they used the double ‘b’. This example underscores that fact that it is unwise to assume that GY and Mark were simply imposing their own preferences; it also demonstrates the seriousness with which they undertook to carry out Yeats’s expressed delegation.15 The more one looks at this matter of delegation, the more arbitrary appears the decision to include George Yeats’s annotations to copies of the 1937 and 1938 editions, but not to use the Coole proofs and the later London and New York editions. The unexpressed assumption—if there is one—must be that there is a purely one-way traffic from the annotations to the Coole proofs; in point of fact, we simply do not know precisely when individual annotations have been made. Whatever the reasoning, the editors’ decision is, at best, a half-hearted concession to those who argue that Yeats trusted others to carry out his wishes; but as the saying goes, those who try to please everyone end up pleasing no-one.

Some of Thomas Mark’s emendations to A Vision have been made to correct syntax, to clarify what is being said, or simply to assist ease of reading. For example, he inserted a comma after ‘to’ in the phrase ‘an appreciation of, or submission to some quality’ (CW14 68, 242), but he did not change the earlier ‘a sharing of or submission to divine personality’ (CW14 65). He worried about Yeats’s ‘Certain London spiritualists for some years past have decked out a Christmas tree with presents that have each the names of some dead child upon it’ (CW14 161), changing the ‘it’ to ‘them’ (CW14 256; although as Hood points out, ‘A better solution would have been to rephrase as “presents each of which has the name of some dead child upon it”’ (‘Search’ 232). He also tackled what Hood calls a ‘confusingly phrased footnote’ (‘Search’ 237) on CW14 195, which reads in part ‘“Mathematic Starlight” Babylonian astrology is, however, present in the friendships and antipathies of the Olympic gods’. (I think it is more of a headlong rush than a confusion.) Mark suggested rephrasing this to ‘“Mathematical Starlight”, Babylonian astrology, is, however, present in the friendships and antipathies of the Olympic gods’. The change from ‘mathematic’ to ‘Mathematical’ is because Mark thought, quite reasonably, that Yeats was quoting the passage that the note was attached to (the first Coole proof has ‘“Mathematical Starlight”, Babylonian astrology, is, [note: “al as in line 2?”]’, CW14 261), where Yeats writes ‘I can but see bird and woman [swan and Leda] blotting out some corner of the Babylonian mathematical starlight’. I myself would insert em dashes in the footnote, but Mark’s solution is equally plausible. I also would not capitalize ‘starlight’. Harper and Paul simply ignore the whole thing. Or take the following sentence: ‘When Passionate Body and Celestial Body give way to Mask we dwell in aesthetic process, so much skill in bronze or paint, or on some symbol that rouses emotion for emotion’s sake’ (CW14 143). Mark would have us dwelling ‘on’ aesthetic process, not ‘in’ it (CW14 254). Harper and Paul do not change ‘in’ to ‘on’, and in fact give no sign that they have thought about the matter, although I think ‘on’ makes more sense.

Towards the end of ‘The Great Year of the Ancients’ there is a fairly well-known passage which contrasts an antithetical dispensation with a primary: ‘A primary dispensation looking beyond itself towards a transcendent power is dogmatic, levelling, unifying, feminine, humane, peace its means and end; an antithetical dispensation obeys imminent power, is expressive, hierarchical, multiple, masculine, harsh, surgical’. It is the parallelism here that led Connie Hood to suggest that ‘imminent’ should be changed to ‘immanent’. ‘No manuscript source is available for this passage’, she writes, ‘but the sense of the sentence clearly requires “immanent” in contrast to “transcendent power” (263.5 [i.e. CW14 192]). Mark changed “imminent” at [the earlier passage at] 176.18 [i.e. CW14 131] but missed this instance’ (‘Search’ 237). The earlier passage at CW14 131 reads: ‘His work should neither be consciously aesthetic nor consciously speculative but imitative of a central Being—the Mask as his pursuer—consciously apprehended as something distinct, as something never imminent though eternally united to the soul’. Hood agrees with Mark here, and suggests replacing ‘imminent’ with ‘immanent’, but she bases her decision on ‘VAMS’, ‘[a] holograph manuscript of part of A Vision (1925)’ (I presume this is the document called ‘Version B’ in YVP4, an early manuscript for ‘The Great Wheel’ and ‘The Twenty-Eight Embodiments’ [NLI MS 36,263/10/1–2]. I could be wrong, but all of the references match), as well as on what the sentence seems to say (‘Search’ 97, 229). Harper and Paul’s edition of the 1925 version has a footnote relating to this sentence, but it does not take up the ‘imminent/immanent’ debate: ‘VersB adds a final sentence to the paragraph at this point: “He is a moralist” (YVP4:232)’ (CW13 89, 264 n. 226). The passage from YVP4 232 has ‘as something never imenent imanent, though eternally | united to the soul. He is a moralist’. While Yeats’s spelling is always suspect, the fact that he has changed the first ‘e’ implies that he had given some thought as to what letter should replace it. The table ‘Comparison of A Vision with Proofs’ in Harper and Paul (CW14 252) notes that both Coole corrections and the 1962 edition choose ‘immanent’ here, although the proofs as printed retained ‘imminent’. Harper and Paul print ‘imminent’ both for this instance and the later one. They certainly do not mention the possibility that the word could be otherwise.

At one stage, Yeats refers to ‘constellations [plural] of Goat’ (CW14 184), which does not seem grammatical. Certainly one of the drafts has ‘the Goat’, as Hood explains:

252.21 constellation] The word is not in the corresponding place in MBY-VB 22 [MS 36,272/6/1–2; see NLI Collection List No. 60, 41 and CW14 430 n. 41]; thus the plural form is probably a printing error, produced by an attempt to read Yeats’s handwriting on the galleys.

252.21 of the Goat] MBY-VB 22 contains ‘the’, but it was dropped in AV-B; Mark reinstated it in 62. (‘Search’ 236–7)

Harper and Paul quote the relevant passage from the typescript carbon in an endnote (CW14 430 n. 41) but do not make the correction. I suspect they quoted the passage simply because the sentence that follows is markedly different from the published version.16

There is one place where both Mark and Hood appear to have lapsed in their proofreading. On p. 151 of the 1937 edition, there is a missing letter at the end of line 3, an omission mirrored in the 1962 English reprint:

Is it an ‘s’ or an ‘n’? Harper and Paul think it is ‘n’ (CW14 112, 270), which makes a kind of sense only if you are allowed to start one sentence and end another. Hood, who is usually so thorough, does not mention the missing letter; the 1956/61 edition followed the wording of the same passage in the 1925 edition, and used the ‘s’. In their 1925 edition Harper and Paul themselves print ‘is a fragment’ (CW13 70).

Mark also missed a badly punctuated passage describing the Shiftings. Yeats writes:

‘For in a state of equilibrium there is neither emotion nor sensation’. In the limits of the good and evil of the previous life… the soul is brought to a contemplation of good and evil; ‘neither its utmost good nor its utmost evil can force sensation or emotion’. (CW14 168–69)

The ellipsis here would seem to suggest that Yeats left something out of a passage he was quoting, which means that the middle statement must also be a direct quotation. Perhaps Mark was distracted by having to correct the ‘Aeneids’ in the following sentence. Hood and Harper and Paul all quote the early version of this passage (‘Search’ 234, CW14 286), a passage which makes it clear that there are two quotations and not three, as does the equivalent passage in CW13 190. Comparison with the original passage from the Automatic Script, helpfully noted by Harper and Paul (CW13 327 n. 35, CW14 411 n. 51), shows that the ellipsis occurs because the first part occurs after the second (YVP1 491). This source also has ‘comprehension’ instead of ‘contemplation’.

Hood mentions several places where thorough proofreading alone might lead one to suspect the text needs emendation. For example, in ‘A Packet for Ezra Pound’, Yeats describes the occasion when his wife first began to talk in her sleep, ‘and from that on almost all communications came in that way’ (CW14 8). At first sight, it might be tempting to replace ‘from that on’ with ‘from that time on’ or ‘from then on’. Hood prefers ‘from that time on’, referencing ‘Yeats’s Rapallo notebook, NL 13,577’ (‘Search’ 96). However, it appears that ‘from that on’ is an Irish locution, as it was used frequently by Yeats in Autobiographies, as in the sentence ‘From time to time from that on she gave me money’ (CW3 304). Nevertheless, a note from Harper and Paul might have been welcome.
Another questionable case occurs in the description of Phase 19: CW14 112, like CW13 69, has ‘A certain actress is typical, for she surrounds herself with drawings by Burne-Jones in his latest period, and reveres them as they were holy pictures’. This would seem to require ‘as if they were’, which is the reading Hood chooses, basing herself on ‘VAMS’ (‘Search’ 97, 228; see YVP4 202 for ‘as if they were’). Hood includes numerous passages in her chapter ‘Textual Notes’ along the lines of ‘The sense of the sentence requires... but no textual authority exists for doing so’ (e.g. ‘Search’ 221, 222, 224, 226, 228, 229). A typical example is the note on ‘whirring’ [CW14 172]: ‘“whirring” refers to sound and “whirling” to motion’ (‘Search’ 235). (While we are on this topic, Yeats’s idea that the sails of a windmill whirl in opposite directions or change their distance from each other [CW14 70, 147] is incorrect, a fact which no-one seems to have remarked on. This is the sort of error textual editing can do nothing about, beyond noting it). Harper and Paul themselves followed this procedure of emending by sense in CW13 359, where they changed Yeats’s heading EXPANDING AND CONTRASTING GYRES (CVA 129; cf. CW13 104) on the grounds that ‘section elsewhere speaks of contracting, but never of contrasting’. However, as we have seen, most of the examples I have given above of errors that should have been picked up by proofreading do not rate a mention with them.

I have few problems with the editors’ standardization of the system’s technical terms. In this, they are following in the footsteps of GY and Thomas Mark, who were simply carrying out WBY’s wishes. It should be noted that there are differences in Italicization and Romanization between the editors’ 1925 edition and their 1937 edition, so the standardization between the two is not always consistent. For example, the 1925 edition has ‘When the Will predominates, and there is strong desire, the Mask or Image is sensuous’ (CW13 16), whereas the 1937 edition has ‘When the Will predominates the Mask or Image is “sensuous”’ (CW14 64). The authority for the ‘Image’ part of this change, is ‘GY in YL 2434’ (CW14 269), and of course Yeats is here defining words, as per his instructors, hence the added quotation marks around ‘sensuous’. The choice not to standardize ‘Image’ to all roman in 1925 (or to all italics in 1937) may or may not have been a conscious decision on the part of the editors; it seems as though Yeats changed his mind between the two editions (‘Search’ 111).

Harper and Paul note that there is one case where they have not imposed standardization:

Will, one of the Four Faculties (Body of Fate, Creative Mind, Mask, and Will), is not regularized with the others. Since WBY also uses the word ‘will’ in its ordinary sense, and it is not always clear whether he refers to the common concept or the specialized term, we have yielded to the authority of the copy text for this word. (CW14 li)

This is unexceptional, and they are not the first readers to have remarked on it. Another technical term, which is also in common usage, is ‘Spirit’; see CW14 171 where Yeats writes ‘The Spirits before the Marriage are spoken of as the dead. After that they are spirits, using that word as it is used in common speech’. However, one must wonder why the copy-text’s phrase ‘Spirits of the Thirteenth Cone’ (CW14 166, 167 [twice]), with its de-italicized ‘of the’, has not been taken as the template for the similar phrase ‘Spirit of the Thirteenth Cone’ (CW14 174 [three times], 175), which is entirely in italics. Although the former phrase occurs three times and the latter four, Hood admonishes: ‘The added italics in “of the” probably resulted from a printing error. “Spirits” and “Thirteenth Cone” are separate terms’ (‘Search’, 235).17 I also wonder about the phrases ‘Victimage for the Ghostly Self’ and ‘Victimage for the Dead’ (CW14 174–75). The phrase ‘Victimage for a Spirit of the Thirteenth Cone’ does not use italics for ‘for a’ (CW14 174). To emend the phrase to ‘Victimage for a Spirit of the Thirteenth Cone’ would I suppose tend to emphasize that there are three technical terms involved here, and might overwhelm the reader, whereas if we treat ‘Spirit of the Thirteenth Cone’ as a single technical term, the reader is less likely to take fright. Fright, I am afraid, is inevitable.

Of course, it is easier to impose consistency of terminology then to determine consistency of system. The former can be carried out mechanically; the latter needs engagement with what the text is trying to say. Mark is not always correct when he tries to engage with the system; he did not, after all, have access to the Automatic Script to assist him, although he did have access to George Yeats. In the ‘Table of the Four Faculties’ (CW14 71), the True Mask of Phase 4 is given as ‘Intensity through emotions’, whereas the description of Phase 18 informs us that its True Mask comes from Phase 4 and is ‘Intensity through emotion’ (CW14 108). Hood suggests ‘emotion’ for the former, giving ‘VAMS’ as her authority (‘Search’ 96; cf. YVP4 198, which has ‘by emotion’ instead of ‘through emotion’), while YVP2 465 has ‘intensity by emotion’, as does YVP3 197, thus lending support to the singular. Again, Harper and Paul leave the text as it is, saying nothing in the endnotes, although their Table 3 shows us that Mark wanted to change the later mention to ‘Intensity through emotions’, a change that was carried out in the 1962 edition (CW14 248).

When it comes to more complicated system consistency, Mark is as lost as the next person. The following examples will give the reader some idea of the difficulties involved.

The reader may well wonder whether Yeats’s phrase about Lunar South and Solar East should read ‘Lunar South in Solar East’ or ‘Lunar South is Solar East’. It is the former on CW14 138; it is the latter in Yeats’s note on CW14 146 and on CW14 183. Although Harper and Paul collate all of these references in an endnote (CW14 300 n. 31), they do not attempt to regularize the expression. That Yeats always puts the phrase in quotation marks suggests he is quoting from the Automatic Script, but I have been unable to find the phrase in Yeats’s Vision Papers. The closest match occurs not in the AS itself but in the Vision Notebooks: ‘Head, Heart, Loins & Fall do not refer to Wheel on which they are placed but mark the position of the Four Faculties at  30 on an interior wheel, when that wheel is taken to represent the Great Year. In this Wheel east on usual diagram is South… In dealing with life cone & after life cone we start at Lunar S & Solar East. Death is reached at Lunar East which is Solar N—we must therefore recognize that the Beatitude is at Solar W—CM Spirit at  & CB at ’ (YVP3 187–88). Here, the word used is ‘is’. Also CF C5 has ‘inner E is outer S’ (YVP3 250). Connie Hood quite correctly suggests emending the first usage to be consistent with the other two (‘Search’ 97). The only reason that ‘in’ was even considered is confusion caused by the phrases ‘Sun in Moon’ and ‘Moon in Sun’ (CW14 60). There, the situation is different, with one tincture consuming the other; here, the one term is the equivalent of the other, although on a lower plane. Harper and Paul’s endnote shows that they know the word should be ‘is’ (CW14 300 n. 31).

The careful reader of diagrams will also spot that there is a problem with the diagram of The Historical Cones which prefaces ‘Dove or Swan’. As Neil Mann pointed out in his review of the Harper and Paul 1925 edition (YA18 293), the word ‘WILL’ was printed there as ‘WELL’ on the line cutting the cones ‘a little below 250, 900, 1180 and 1927’ (CW13 147). (Admittedly, it is hard to see, as a diagonal line of a cone bisects the ‘E’.) Mann also noted that the colours had been transposed on the labels ‘12–13–14 (1380’ and ‘15–16–17 (1550)’. Both of these errors of CW13 have been amended by Harper and Paul in their 1937 edition, but a little above the same line they have substituted the incorrect ‘(120) 2–3–4’ for the correct ‘120 (2–3–4)’ (CW14 193). Of course, ‘120’ is the year, while ‘2–3–4’ are the phases. (Just to confuse matters further, the diagram in the automatic script on which this diagram is based used the parentheses for most of the years and left the phases naked [YVP3 61]). Jeffares in A Vision and Related Writings (London: Arena, 1989, 258) changes the correct ‘120 (2–3–4)’ that occurs further down in the original diagram to the incorrect format, but is at least consistent. The 1956/61 and 1962 editions both retain the ‘(120)’ error but have the correct ‘Will’, while the former drops ‘(19-’ from ‘20–21) 1680’ lower down on the right hand side. As explained by Hood, this error was introduced when the 1938 New York edition was produced by photolithography of the 1937 English edition: ‘Following an agreement with Macmillan of New York, the 1937 London text was copied by photolithography by the Polygraphic Company of America and published on 23 February 1938. In spite of the ostensible accuracy of the photocopying process, one printing corruption was introduced into the 1938 text: on line 9 of the diagram on page 266, the 1937 version reads “(19–20–21)” but the 1938 text has “(20–21)”. Since the 1956 text was photocopied from 1938, the corruption was perpetuated in 1956 and 1961’ (‘Search’ 141). This means that there are no editions of A Vision with this diagram printed correctly, although it may be found drawn without error on Mann’s website, at

Mann in his review of the 1925 edition wrote:

On a minor note, it is clear from the diagram of the cones that 1050 is a key date, and the editors should have had the confidence to change ‘The period from 1005 to 1180 is attributed in the diagram to the first two gyres of our millennium’ ([CW13] 164), even though it is a mistake that persisted into all versions of AVB. (YA18 294)

Despite Mann’s implied view that the editors were aware of the discrepancy here when they were preparing the 1925 edition, it is not evident to me that this was the case; however, they have emended the text in the 1937 edition, changing CW14 208.29 from ‘1005’ to ‘1050’. The reason they give is that the latter is ‘correct in The Historical Cones, p. 193’ (CW14 272). I realize you can’t put detailed explanation in a table, but this laconic entry glides over far too much: how do they know it is correct in The Historical Cones? I suspect the editors took Mann’s word for it and then searched around for a reason; certainly they took his gentle admonition as a hint to lift their game. The date of ‘1050’ is also used in CW14 on pp. 192, 199, 207, not to mention the typescript on p. 292, and as the editors inform us, it is half of 2100, which is when the ‘new Messiah’ is due (CW14 444 n. 37); but the section heading, ‘A.D. 1050 to the Present Day’ (CW14 207), alone should have given sufficient warrant, never mind the previous section heading ‘A.D. 1 to A.D. 1050’ (CW14 199). The ultimate source for the date is of course the diagram of the Historical Cones drawn by ‘Carmichael’, as printed in YVP3 61.

The tendency to fudge with the sources of the editorial decisions is also apparent with the change from ‘Domination through emotional constriction’ to ‘Domination through emotional construction’ as the description of the True Creative Mind of Phase 20 (CW14 73). The ‘Authority’ given is that it is ‘Correct on 91.24’ (CW14 268), but surely when you are faced with only two usages of a phrase in a given text, you need an external authority to be able to make a decision as to which is correct, or else you need to engage in detailed contextual analysis which may or may not be conclusive. Once again, the reader will find that Neil Mann’s review of the 1925 edition is the external source (YA18 291). Reference to the first three volumes of YVP is not entirely decisive.18 In YVP1 257, we have

20 emotional domination

not quite

constriction in emotional—in intellectual domination

whereas the table reproduced in illustration in YVP3 199 clearly has ‘construction’, with the endnote informing us that Yeats ‘also changed “construction” in CG of P20 to “constriction” in VA 32’ (YVP3 219 n. 57) (which does seem to beg the question). Hood (‘Search’ 96, 224) references ‘VAMS’ for the reading ‘construction’ (cf. YVP4 183, which has ‘True Intellect Emotional construction in intellectual dominance’).19

There are other places where the editors have left themselves some wriggle-room in their self-imposed strait-jacket. So far as I can see, both of the cases I discuss below are, at the very least, debatable. Normally, Harper and Paul relegate the corrections from the Coole proofs or Thomas Mark to Table 3, but there is one exception, and it looms large: they change ‘the thirteenth sphere’ from the 1937 edition (and the 1956/61 edition) to ‘the Thirteenth Cone’ in ‘The End of the Cycle’: ‘The particulars are the work of the Thirteenth Cone or cycle which is in every man and called by every man his freedom’ (CW14 219–20; cf. CW14 273, where the ‘Authority’ is given as ‘Coole proofs’). This change is nowhere justified or debated; the editors simply mention it in passing as having been made by Mark in order to ‘improve or correct the description of the system itself’ (CW14 xlvii). This was not a correction that Hood was going to make. It seems as though they just could not bear to forgo this change, given that it has been the subject of some debate (see, e.g., Finneran, ‘On Editing Yeats: The Text of A Vision [1937]’, 126–7).

There is an equally difficult emendation to consider. It stands out from the general uniformity of the ‘Authority’ column of Table 4, Emendations to the Copy Text, as the reason for the change is given as ‘“Version B” (YVP4:216)’ (CW14 271). In the original 1937 edition, the text ran as follows: ‘For the moment the desire for a form has ceased and an absolute realism becomes possible’; Harper and Paul change ‘desire for a form’ to ‘desire for reform’ (CW14 121), leaving the reader to find that it has been altered by consulting the table. However, if you check the source given, ‘Version B’, the text reads: ‘The desire of reform has ceased, an absolute realism becomes possible’: the word ‘of’ occurs instead of the word ‘for’ (YVP4 216). Thus, if they were going to change the text, the editors should have emended the text to ‘desire of reform’. This wording is confirmed when we check the same passage in the 1925 edition (CW13 79): the attached endnote reads, ‘The words “a form” are an error: see VersB, where the phrase reads “The desire of reform has ceased” (YVP4:216)’ (CW13 261 n. 201). Yeats does seem to use ‘desire of’ when ‘desire for’ is meant, and it is unlikely that he would have been able to articulate the difference; see, for example, ‘desire of expression’ and ‘desire of action and of command’ (CW14 92) or ‘Self-realisation attained will bring desire of power’ (CW14 196). The emendation to ‘desire for’ surely owes its existence to Neil Mann’s complaint about the same passage in the 1925 edition that ‘leaving the mistake in the text seems the wrong way round for an edited text’ (YA18 290).20 Mann makes a slip in quoting the endnote, saying that ‘it is noted that the draft in question gives “the desire for reform” ([CW13] 261)’, so it appears that the editors, in making their change, worked from Mann’s review rather than from their own sources. I can see why Mann endorsed the emendation to ‘reform’: after all, the preceding sentence of CW14 concludes that, at Phase 22, ‘there is neither change nor desire of change’. However, ‘a form’ also makes sense, given that the Mask is ‘[a] form created by passion to unite us to ourselves’ (CW14 461), and Phase 22 is not a phase that values Mask: Phase 22 is the phase split between antithetical and primary in equal measures, eventually moving towards primary.21 Here, of course, one faces the fact that detailed understanding is needed in order to establish what the correct text should be. I suspect that the Version B manuscript should be revisited with this dilemma in mind: the words ‘a form’ and ‘reform’ as written by Yeats might well be indistinguishable. Failing a definitive answer from the manuscript, and given the complications involved, my own inclination is to trust the 1937 text here, set out the pros and cons in an endnote, and make no emendation.

Some of the emendations proposed by Hood but not taken up by Harper and Paul are based on those made by WBY or GY to their copies of A Vision (1925). (They have no problem using GY’s emendations in copies of the 1937/38 edition, as detailed above.) These were first described in Finneran’s 1977 article, printed in Edward O’Shea’s A Descriptive Catalog of Yeats’s Library, and more fully captured in Harper and Paul’s edition of the 1925 version. The note in this last reads:

The Yeatses kept four copies of AVA in their library: (1) number 83 of the six hundred copies printed (O’Shea 2433); (2) number 385 (O’Shea 2433b); (3) number 366 (O’Shea 2433a); and (4) number 498 (O’Shea 2433c). As Richard J. Finneran notes in his essay ‘On Editing Yeats: The Text of A Vision (1937)’ (Texas Studies in Literature and Language 19 [1977]: 121–22), the last three of these copies have postproduction corrections made by WBY and GY. The three tables that follow present the changes made to these three copies, with indications of the changes, who made them (when it is possible to distinguish whether the marking was made by WBY or GY, or in a few instances by Macmillan editor Thomas Mark), whether the change appears in one or more of the other copies, and if the correction was carried over into the 1937 edition. Page numbers are given for the original book (in parentheses) and this edition. (CW13 339)

More should have been made of these tables in the Introduction to the 1937 version, but the references are confined to the laconic sentence ‘The Yeatses’ marks in their four [sic] copies of AVA-Laurie suggest an early eye toward a corrected version’ and its accompanying footnote (CW14 xxviii and 300, n. 19), with sporadic mentions thereafter (CW14 302 n. 59, 356 n. 52, 357 n. 55). Perhaps the editors considered that decisions to carry the marginalia over into publication were made in an orderly fashion, so that anything that was not carried over was consciously rejected.22 This of course is mere assumption. Changes that Hood would have made on the basis of marginal notes written by Yeats himself in the 1925 copies, but which were not made in any edition, include ‘return several times’ instead of the more specific ‘return up to four times’ (‘Search’ 96, 164, 222; CW13 344) and ‘The automatic script defines being as that which divides into Four Faculties’ instead of ‘By being is understood that which divides into Four Faculties’ (‘Search’ 96, 222; CW13 345). Given that these changes were not made while Yeats was alive, and that other marginal notes in these copies have the air of Yeats or GY thinking aloud, perhaps Harper and Paul’s caution is justified in these two cases; however, one does get the feeling that when they were editing the 1925 edition, they intended to make future use of ‘Corrections to the Yeatses’ Copies of A Vision (1925)’. Regrettably, any such intention got lost in the wash-up, so that an essential part of the writing of the 1937 edition is now relegated to Appendices in CW13.

One of the changes marked in one of the Yeatses’ copies of the 1925 edition relates to the table ‘Four Types of Wisdom’. This read, in 1925 (CW13 30),

At P. 4

Wisdom of Desire

At P. 12

Wisdom of Intellect

At P. 18

Wisdom of Heart

At P. 26

Wisdom of Knowledge

WBY’s note is transcribed as ‘the circled words “Intellect” and “Heart” in the grouping “Four Types of Wisdom” marked “transpose”’ (CW13 346; not in YL). Hood writes, ‘In AV-A #498 [=YL 2433c] Yeats marked the words “Intellect” and “Heart” for transposition; however, the printers misunderstood and transposed the two complete lines. See Yeats’s note at 100.17–21’ (‘Search’ 224). The passage in 1937 is emended as follows (CW14 74),

At P. 4

Wisdom of Desire

At P.18

Wisdom of Heart

At P. 12

Wisdom of Intellect

At P. 26

Wisdom of Knowledge,

and the note by WBY reads ‘I give the Four Types of Wisdom as they were given. I have more than once transposed Heart and Intellect, suspecting a mistake; but have come to the conclusion that my instructors placed them correctly, the nature of the wisdom depending upon the position of the Creative Mind (CW14 74n.). Had Hood’s emendation been adopted, the table would have read:

At P. 4

Wisdom of Desire

At P.12

Wisdom of Heart

At P. 18

Wisdom of Intellect

At P. 26

Wisdom of Knowledge

I have not yet worked out what all of this means, although I imagine it has to do with the proposed table ‘where the Four Faculties predominate’ (CW13 340, ‘Search’ 125) and the passage that was retained, CW14 141, where the placement of the Faculties differs from those proposed. The ‘wisdom of knowledge’ and its connection with Phase 26 is mentioned on CW14 109, but there is little elsewhere on this subject. However, it is worthwhile following up Harper and Paul’s endnote to ‘The Four Types of Wisdom’ in CW13 242 n. 73—there is no equivalent note in CW14—which reads ‘See YVP1:183, 2:98, 3:207, for AS and notebook entries about the four types of wisdom’. YVP1 183 uses Phases 23 and 7 instead of 18 and 26: ‘wisdom of heart at 23—wisdom of intellect at 12—wisdom of soul at 27—wisdom of desire or instinct at 4’, whereas YVP2 98 corrects the phases to those used in the 1925 edition: ‘4. Can you devide Antithetical] & P[rimary] into 4. | 4. the wisdom of heart comes at 18 | wisdom of intellect 12 | wisdom of desire 4 | wisdom of knowledge 26’, as does YVP3 207. YVP1 191 is also relevant but not specifically mentioned by Harper and Paul: ‘51. Can you explain with diagram why wisdom of heart & wisdom of intellect come so apparently out of place? | 51. no but I must go now—goodbye’. In other words, if we believe that Yeats is giving the Four Types of Wisdom as they were placed by his ‘instructors’, then Hood’s suggested emendation is incorrect and Yeats did not wish to carry through on what was written in his personal copy. However, this still means that the phases should not have been listed out of order: Phase 12 should have been listed before Phase 18.

Although, as I have said, Harper and Paul are generous with the reprinting of rejected typescripts, there are several places where Hood suggests substantial additions to the copy text itself based on typescripts. At CW14 163 a footnote by Yeats ends: ‘The Spirit is described as awakened from its sleep in the dead body’. I have always found this phrasing a trifle abrupt. Hood adds a lengthy continuation of the sentence from ‘the carbon of the typescript which was sent to the printer as partial copy for A Vision (1937)’ (NLI MS 36,272/6/1–2). The note appears to have been simply overlooked because it was typed on a separate page, and Hood’s account of how it could easily have been missed is convincing (‘Search’ 97, 209, 233). Hood would also have included a diagram illustrating ‘Lunar South is Solar East’, placing it at the end of Section VI of ‘The Great Year of the Ancients’. The source for this is a rejected typescript done after 1926 but before NLI MS 36,272/6/1–2. As Hood explains, two later unrejected typescripts, including NLI 36,272/6/1–2 itself, left a blank space for the diagram. The iteration before NLI 36,272/1–2 left a large space in which Yeats wrote ‘Diagram’ (‘Search’ 98, 236). Both of these items are of sufficient importance to have been brought to the reader’s attention, either in the text or in the endnotes.


I have mentioned numerous points of disagreement with Harper and Paul’s editing of A Vision (1937). However, weighed against what has been accomplished, these are relatively minor, and are of interest primarily to specialists such as myself. I have a different idea of what textual editing means than the one they have (more or less successfully) adhered to; others may well agree with their idea rather than with mine. The reader with a pedant’s eye for detail will still need to refer to Hood’s dissertation for further sidelights; but, all things considered, my opening statement is something I would emphasize. The richness of the annotations and the publication of extended passages from the typescripts are in themselves enough to recommend this volume. Harper and Paul’s edition of A Vision: The Revised 1937 Edition enlarges our understanding of Yeats’s achievement; there is sufficient information here to keep scholars busy for years to come.

2 The editors refer to ‘Mann, “A Vision [1925]: A Review Essay”’ several times (see CW14 381 n. 211; 382 n. 213; 448 n. 60). Since writing this paper, I have read Catherine Paul’s review of the Yeats Annual in which Mann’s critique appeared. There, she admits candidly that ‘The second part of Mann’s review has been extremely useful to us as we prepare our edition of that later version, as he enumerates errors, oversights, and misjudg[e]ments that we are grateful for the chance to consider and try to rectify’, ‘Yeats Annual Rebooted’, South Carolina Review 46:1 (2013) 211, online version at–1_paul.pdf.

3 O’Neill was not a random correspondent as the O’Neills had known Yeats for some years and Mary Devenport O’Neill had often discussed A Vision with him. See Ann Saddlemyer, Becoming George: The Life of Mrs W. B. Yeats (Oxford: Oxford University Press, 2002), 177, 351 and YGYL 34 n. 1; also MYV2 412–14. Where MYV2 has O’Neill asking if a daimon has ‘any separate existence apart from human being to whom it belongs’ (414), CW14 xxix has ‘Has a daimon any separate masters apart from human being to whom it belongs?’ Although I have not seen the original, the former reading is obviously preferable from the point of view of the system.

4 eBooks versions are essential for scholarly study, as the notes are more easily retrieved and the reader is not forced to rely on the vagaries of editorial indexing. Unfortunately, even major publishers do not seem to have thought things through with regard to scholarly eBooks, particularly when they publish older books that have had to be OCR’d: tables are printed as images because it is easier, words get run together, spaces occur where they shouldn’t, and letters are misidentified. This publisher’s sample for the eBook edition of The Collected Works of W. B. Yeats Vol. VI: Prefaces and Introductions has, in the first few paragraphs of the Editor’s Preface, the following word-joins: ‘specialinterest’, ‘Yeatsrevised’, ‘selectionof’, ‘willingnessto’, ‘revisionsmade’, ‘twosigned’, ‘never-publishedcollected’, and ‘thisto’. Needless to say, reading such a text can be wearisome. The ISBN for the eBook version of A Vision: The Revised 1937 Edition is given at the front of the hardback: ISBN 978-1-4767-9211-8. I have not seen the full version, for the reason given above. What I have seen is mercifully free of coding errors; I presume the text itself did not need to be OCR’d because of its newness.

5 Sutherland 641. I do not wish to imply anything unethical here; things get lost in the course of editing long books, particularly when space is at a premium. To take an example of what is surely an unintentional omission, ‘the Japanese interpreter of Botticelli’ (CW14 150) is identified as ‘Yukio Yashiro (1890–1975) [who] was an art historian and art critic who graduated from Tokyo University and studied in Europe from 1921 to 1925. In February 1926, just after the publication of AVA, GY gave WBY a copy of Yashiro’s three-volume study Sandro Botticelli (London and Boston: The Medici Society, 1925)’ (CW14 392 n. 42). As well as including information not in Sutherland, who tentatively identified the reference (Sutherland 872), Harper and Paul neglect to inform the reader of one of the essential sources used by them: the information about GY’s gift is not from the standard biographies but comes from the inscription in YL 2304.

6 Where I have given unsourced references to NLI numbers, I have taken them from Collection List No. 60, Occult Papers of W. B. Yeats, compiled by Peter Kenny (Leabharlann Náisiúnta na hÉireann, National Library of Ireland, [n.d.]), lists/yeatsoccult.pdf. Kenny identifies most of the sources used by Hood.

7 See Warwick Gould, ‘W. B. Yeats and the Resurrection of the Author’, The Library (1994) s6–16 (2): 101–34. The quotation is from Finneran’s Editing Yeats’s Poems: A Reconsideration (London: Macmillan, 1990), 39; it originally appeared in Editing Yeats’s Poems (New York: St Martin’s Press, 1983), 30. The epigraph at the head of this paper is from Finneran’s essay ‘Text and Interpretation in the Poems of W. B. Yeats’, in George Bornstein (ed.), Representing Modernist Texts: Editing as Interpretation (Ann Arbor: University of Michigan Press, 1991), 31. I have taken it from Gould, 122.

8 It was not only his spelling that was notoriously bad; he sometimes could not even read his own handwriting; see William H. O’Donnell’s ‘Reading Yeats’s Hand’, in YAACTS9 (1991), 87–94. Hood quotes a letter where AE admonishes Yeats: ‘Your proof-reading is abominable. You are the worst culprit I know in this respect’ (‘Search’ 165). Hood neglected to give the date of this letter, perhaps because it was from 1899; see Letters from AE. Sel. and ed. Alan Denson (London: Abelard-Schuman, 1961), 31. She suggests it got worse as he got older.

9 See Gould, ‘W. B. Yeats and the Resurrection of the Author’, 114–17, for some of Yeats’s comments on Mark. CL InteLex 5731 has Yeats’s letter of 8 September 1932 to Harold Macmillan which reads, ‘I would be very much obliged if you would give the enclosed letter to the admirable scholar who is assisting in the correction of the proofs of my new collected edition. It is partly a letter of thanks and partly an explanation of certain metrical tricks of mine which have puzzled him’. The ‘enclosed letter’ is not printed or noted here in the CL InteLex, but cross-reference should be made to CL InteLex 5733, which has ‘Mention in letter from Mark, 16 September 1932. Discussing corrections and proofs of new edition of WBY’s works’, where the reference is to LTWBY2 543–44. The date of the CL InteLex letter should be changed from ‘c. 15 September 1932’ to ‘on or before 8 September 1932’, and its explanatory note emended to include specific quotations from the letter, as per that series’ policy. Jon Stallworthy quoted sentences from it in his 1963 book Between the Lines: Yeats’s Poetry in the Making, including the unambiguous statement ‘I have never been able to punctuate properly. I do not think I have ever differed from a correction of yours in punctuation. I suggest that in the remaining volumes [of the projected Collected Edition] you do not query your own corrections’, from Gould, 115. The passage quoted by Gould from Charles Morgan’s The House of Macmillan (1843–1943) should also be added, as presumably it is from the same source: ‘For the first time there will be a satisfactory text of my work, thanks to your watchfulness and patience’. See also Note 10, below.

10 In fact, the 1956/61 edition is the only one that claimed to be definitive: its title page included the subtitle ‘A reissue with the author’s final revisions’, whereas the 1962 edition only claimed that it was ‘Reissued with corrections’. It is to be regretted that Harper and Paul’s edition does not include a textual collation such as that given in Hood’s Chapter 6, where the reader can see at a glance the differences between the 1937, 1938, 1956/61 and 1962 editions (‘Search’ 239–48). Harper and Paul’s Table 3 gives a collation with the 1962 text, but the reader is left to assemble a possible 1956/61 text from their Tables 1 and 2 (see CW14 xlviii).

11 See Yeats’s letter of 13 March 1935 to Harold Macmillan, where he writes, apropos of such things, ‘I hesitate to lay such a burdensome task [as reading proofs of A Vision] upon Mr Mark for I suppose he would be the reader. I feel that my geometrical way of expressing myself may fill him with impatience or that he may find it impossible to revise efficiently without a greater study of my philosophical ideas than he should be expected to make. If however he could read the proofs for superficial errors, for lack of uniformity in the use of italics, capitals, etc., I should be greatly obliged to him’ (CL InteLex 6199). As this letter also reveals, Yeats had the idea of trusting Frank Pearce Sturm with the task of proof-reading for system errors, but this did not eventuate. See also the afterword to this essay by Warwick Gould.

12 It may be worth mentioning that not even Finneran emended Yeats’s ‘Michael Angelo’ in his edition of the Poems.

13 See Gould, ‘W. B. Yeats and the Resurrection of the Author’, 111: ‘Whatever individual textual decisions George Yeats made, she appears to have acted in good faith. Some of them are open to challenge for the excellent reason that all can be inspected’. By ‘inspected’, Gould meant that scholars can study the roles of George Yeats (along with those of Yeats’s publishers and publishers’ readers) ‘in vast archives of proofs and letters in at least ten separate locations in Britain, Ireland, and the United States’. Gould himself has been untiring in his excavation of these archives and in publishing his findings. My own inspection is limited to what Hood and various editors have chosen to print, but that in itself is sufficient for the limited purposes of this review.

14 Of course, when referring to von Rosenroth’s Kabbala Denudata Yeats followed the spelling of the original title, as custom dictates (CW14 39). ‘Cabala’ appears to be the currently preferred spelling when discussing the Christian Cabala, ‘Qabalah’ the spelling used by occultists who wish to draw attention to themselves, and ‘Kabbalah’ that used when writing of Madonna or referring to Gershom Scholem and Moshe Idel (not to mention S. L. MacGregor Mathers and A. E. Waite) (see CW14 316 n. 30 for the editors’ awareness of the different spellings). According to the website, there are in fact 24 spelling variations, but we are concerned here with how Yeats spelt the word.

15 Gould and Toomey note that George Yeats herself expressed dissatisfaction with several of Yeats’s decisions. ‘I do not like title [Mythologies]’, she wrote, ‘but even less do I like a change he [WBY] made from “discoveries” to “Explorations” which occurs in forthcoming “Essays”’ (Myth 2005 493); however, as the reader must see, she did not impose her dislikes on the publisher. As an aside, Professor Gould advises me that the next printing of Mythologies will emend ‘Rosa Alchemica’ to ‘The Tables of the Law’ on p. xcvi, main text and n. 26. Here, I should like to thank Professor Gould for helpful comments on this review. Needless to say, any errors or omissions that remain are entirely my responsibility.

16 The hasty reader may have trouble following Table 3 (CW14 259), where the ‘1st Coole correction’ is ‘of the Goat’, ‘As printed: 2nd Coole proofs’ is ‘of the Goat’, whereas ‘2nd Coole correction’ is given as ‘of Goat’. Mark’s notebook has the correction, as does the 1962 London edition. However, Harper and Paul explain that ‘In most cases where the second-pull Coole proofs reverse a change made in the first pull, it is on the authority of the New York [1956/61] edition’ (CW14 232). This is not the first time the second-pull Coole proofs reverted to a plainly inferior reading. Hood’s suggestion (‘Search’ 231) that the earlier phrase ‘precession of the Equinox’ (CW14 149) is more usually written ‘Precession of the Equinoxes’ may be correct, but is not relevant to Yeats’s passage, which is concentrating on the Vernal Equinox alone. Harper and Paul’s endnotes on the precession are excellent (CW14 391 n. 38, 414–15 n. 1).

17 Harper and Paul italicize the word ‘spirits’ in the following passage: ‘Creative Mind clings to Body of Fate until mind deprived of its obstacle can create no more and nothing is left but “the spirits at one”, unrelated facts and aimless mind, the burning out that awaits all voluntary effort’ (CW14 138). This is, I think, a special case: Yeats is quoting from the Automatic Script, where the phrase ‘the spirits at one’ frequently occurs (see the Indexes to YVP). Assuming that it is legitimate to change quotations, which is by no means a given, the question arises as to whether the word should be capitalized. In the Section of CW13 entitled ‘The Spirits at Fifteen and One’, Yeats consistently capitalizes, although he does not italicize (CW13 198–99). The practice is inconsistent elsewhere (e.g. CW13 201 has lower case for ‘spirits’: ‘the spirits at Phase 15’).

18 I am assuming that the transcriptions in YVP are correct. Obviously, a definitive conclusion needs to be based on the manuscripts themselves, and even then inspection may not solve all difficulties. In this case, Hood admits that ‘The word in VAMS can be read as either “construction” or “constriction”; the latter choice was perpetuated into AV-B’ (‘Search’ 224).

19 Certainly there are only two uses of the phrase ‘Domination through emotional construction’ in A Vision (1937), but there is a related phrase, ‘constructive emotion’, which occurs in the description of Phase 24, where the True Creative Mind (from Phase 6) is given as: ‘Humanitarianism. through constructive emotion’ (CW14 124). Harper and Paul note that Mark changed this to read ‘Constructive emotion’ in the 1962 edition (CW14 378 n. 188; cf. CW14 251), which makes it consistent with the True Creative Mind of Phase 6 in the ‘Table of the Four Faculties’ (CW14 71). (Remember that in the table, ‘Each Faculty is placed after the number of the phase where it is formed, not after the phase which it affects’ [CW14 70]). The first Coole proofs removed the errant full-stop from the original phrase. Harper and Paul leave it untouched.

20 In this case, for their 1937 edition, Harper and Paul have agreed with Mann’s suggestion and have gone against the general series policy, which is precisely to leave the error in the text. They do not do this for the sentence ‘Though the new Husk and Mask have been born, they do not appear, they are subordinate to the Celestial Body’ (CW14 170). Instead they follow the procedure castigated by Mann for the 1925 edition. The endnote reads: ‘An error: “Mask” should be “Passionate Body”. As Neil Mann notes, “The Passionate Body gives rise to the Mask, and indeed Yeats shows how closely the two were fused in his thinking when he pairs ‘the new Husk and Mask’, a slip for Husk and Passionate Body”, “The Mask of A Vision”, YA19 [2013], 186’ (CW14 412 n. 56). One could argue, I suppose, that there was a manuscript basis to amend the text in the one case but not the other.

21 If ‘form’ is intended, one must take into account its double meaning. Cf. YVP2 468: ‘Is not the mask in subjective phases double—a form which we put on, a form which we desire, that which we become & that we would possess’, to which the answer was ‘Yes’.

22 There is one exception; this is the change from ‘Temptation versus strength’ to ‘Temptation through strength’ (CW14 73) in the Table of the Four Faculties. Harper and Paul take this change from an emendation in George Yeats’s copy of A Vision (1925), YL 2433a. See CW13 341 and CW14 268, together with CW14 243, 355 n. 44 and 356 n. 52; this is one of the annotations that does not appear in O’Shea’s YL. I confess I cannot make sense of the distinction the editors try to draw in the Introduction to the 1925 edition between ‘those changes intended as corrections—as opposed to revisions’ (CW13 xlvi). This seems to me to be specious.