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

11. A Tale of Two Texts: Or, How One Might Edit Virginia Woolf’s To the Lighthouse

© Hans Walter Gabler, CC BY 4.0

The Hogarth Press edition is our text, say the British readers, critics, editors. The Harcourt, Brace edition is ours, say the Americans. But the two editions together are Virginia Woolf’s public text of To the Lighthouse.1 The doubling holds theoretical, critical and editorial challenges that the present essay will explore.

The source materials for the novel, though not consistently preserved, are well defined. There is a complete holograph draft, and there are the first proofs from the Edinburgh printers, Clark & Clark, whom the Hogarth Press regularly employed. One set of these, marked up by Virginia Woolf herself, and a fragment of a second set, from page 273 to the end of the book, also marked up by her, have survived.2 They were sent as printer’s copy to Harcourt, Brace in New York. The set that carried the mark-up for the Hogarth Press edition, on the other hand, is no longer extant. That both main sets were identical, however—the preserved printer’s copy for New York and the lost proof set used for a first round of corrections and revisions for the Hogarth Press edition—may be inferred from the significant textual identity in substantives as well as accidentals between the New York and London editions. More decisively, it is demonstrable from the close typographical congruence between the extant Clark & Clark first proofs and the Hogarth Press first edition. Unless text has been changed, or spacings have been adjusted by meticulous compositors, their line and page breaks fully coincide.

Since the novel’s holograph draft survives,3 the proofs make it possible, by retrospective exploration, to compare draft and revision states in terms of the structuring of the novel and the composition of its text. Although this would be a fascinating field to explore, I do not intend to do so here. Prospectively, the proofs document the textual point of departure towards the novel’s public appearance. In the case of To the Lighthouse, this was a double appearance, manifested in the London and New York editions. These editions were published simultaneously—quite literally so, as they appeared on the same day, 5 May 1927, on both sides of the Atlantic. But, as is well known, they are not identical, since in a significant number of instances Virginia Woolf marked up the first proofs differently for each of them. She furthermore continued to revise the text for the Hogarth Press edition alone, making additional changes to the Clark & Clark revised proofs; this final stage of revision has not before been clearly distinguished, but I shall argue for it below.

To the Lighthouse was thus given to the public in two distinct texts. In terms of their difference, the two first editions constitute two versions of the novel. In terms of their simultaneous appearance, these must be termed simultaneous versions. This is a new, or certainly an unaccustomed, category for the textual critic and editor, to whom versions are commonly consecutive. Versions as Siamese twins, simultaneous versions, have not been much reflected upon in textual scholarship. By contrast, criticism has in specific cases, such as those created by Virginia Woolf for a surprising number of her publications, sometimes at least shown itself aware of them—if only as an irritant. Yet such simultaneity holds a critical challenge. So it is under an angle of its critical implications, in the first place, that one should approach the text-critical as well as editorial problem of simultaneous versions.

From the outset, though, it is important to look at the problem from the perspective of textual materiality. In the case of To the Lighthouse, before the anomaly of the simultaneous versions takes effect, there is plentiful evidence of the normal identity, or else the simple difference, between the proofs and the published text. The Lighthouse proof text and the London and New York published texts are to a large extent identical. Where they differ, the proofs record a textual state that did not reach publication, while the British and American editions together manifest precisely the (one and only) public-text alternative to the pre-publication state. Simple acts of revision have created simple textual alternatives—or, looked at the other way round, the acts of revision have left textual alternatives altogether behind in the work’s pre-publication state. To the extent that the two first editions of To the Lighthouse conform to such normality, they do so because the mark-up for revision (that is, the author’s mark-up) on the two first-proof exemplars (basically identical in themselves) was identical. Nonetheless, however, the two editions present two distinct text versions of the novel. They do so because the mark-up of the proofs was not only identical; it also differed. Where it differed, we may distinguish three mark-up patterns. Sometimes, the proof text was revised for the New York edition but left unchanged for the London edition; sometimes, the proof text was revised for the London edition but left unchanged for the New York edition; and sometimes, although the proof text was altered for both editions, it was revised differently for each.

From these three mark-up patterns, two reasons for the versional difference follow. The London and New York texts differ either because the proof text was doubly, and differently, overwritten; or else, because it was only half overwritten; the difference thereby created meant that one line of the transmission from the proof-text state to publication retained unaltered a reading changed in the other. The double alterations with a difference have the same effect as has the identical mark-up of the printer’s copies for the New York and London editions: they leave the proof text altogether behind in the novel’s pre-publication realm of existence. The double and divergent overwriting of the proofs might be termed the active cause of the versional distinction. On the other hand, the one-edition-only revisions also contribute to establishing the public texts’ versional difference; but here, the difference arises because the proof text is not tracelessly overwritten. Textual elements of the proof text thereby become public text. They do so, however, not because the work’s text at the instances in question remains invariant, but because, paradoxically, it fails (as it were) to be touched and changed for either the one or the other of the two simultaneously published editions.

Taken in all, the situation is not easy to deal with. It is in fact impossible to contain it in terms of text-critical and editorial orthodoxies focused upon authorial intention. In terms of these orthodoxies, editors seek to establish an unambiguously perfected text—something that, as its circumstances of composition and transmission will suggest, is hard to determine for To the Lighthouse. The editorial aim so defined is posited, moreover, on the assumption of an authorial intention that is itself conceived of as directed towards an unambiguously perfected text. Implicit in this alliance between an intentionalist orientation and a teleology of the text is a notion of the closed text. This is a concept, however, that recent theories of literature and text would hesitate to uphold; as a matter of fact, it was the practice of modernist writing in the twentieth century in particular that induced literary theory to question its viability. But if this is the case, how can textual criticism and editing take legitimate guidance from it? And how, specifically, could text-critical and editorial justice be done to a modernist text in the light of a concept that literary theory has relinquished? The transmissional situation of the two simultaneous versions of To the Lighthouse should therefore not only suggest pragmatic non-intentionalist solutions to what is clearly a challenging editorial problem. To face its critical and theoretical implications may also lead, beyond pragmatics, to an adjustment of critical thinking and of the methodologies of criticism to both a modernist sense of the literary text and to some fresh conceptualisations in terms of theories of literature. The acts of revision as they result in the simultaneous versions of To the Lighthouse may help to suggest fresh ways of thinking about the very notion of text, of textual processes, and of both the construction of, and the construction of meaning in, texts.

Examples should advance the argument. Let us begin with the (deceptively) easy cases of identical revision of the proof text for the London and New York editions.4 Changes such as:

(A1, 16; E1, 19)


(A1, 164; E1, 169)

or—a reenvisioning revision this—:

(A1, 112; E1, 117)

suggest an uncomplicated process whose results almost automatically prompt an evaluation. A wish to perfect the text must have been at work: one might, if so inclined, pronounce the revised phrasings the better ones. With Virginia Woolf, as with any author, it would indeed be as difficult to overlook the desire for improvement—however precarious in any given instance the determining of the appropriate criteria for evaluation might be—just as it would be foolish to close one’s critical eye to writing and rewriting evidently intent on bettering the text. Yet we need to recognise that the critical attitude so taken is an author-centered one. It focuses on the author in control, and thus views the text and the given revision as the outcome of acts of writing realising an authorial creative impulse.

Yet, if we consider especially the third example above: with only a slight shift in the interest of our enquiry, we may identify in the change a reading response to the text under revision. The proofs, by way of two loosely strung questions, textualise an imagined first view of Constantinople. Of the two questions, the second one is vague and undefined. At the same time, it holds a wealth of possibilities of what might be seen looking out over the city. It is therefore textually adequate in terms both of narrative and of character. But still: just what is it that the eye fastens on? If this is a question that the text’s question ‘What’s that?’ elicits, it is a reading-response question. It appears that it was the author herself, rereading for revision, who first read it as such, and that it acted on her for the moment as an extra-textual stimulus. It enticed her readerly imagination to envision what else, beside Santa Sofia, might meet the eye on first looking out over Istambul. Hence, it was through the specification of a particular meaning—and therefore by way of a reader-response construction of the text read—that ‘Is that the Golden Horn?’ was written in as a replacement question, narratively as appropriate as the question it superseded. What it loses in terms of a multiple potential to mean, it gains in terms of both the specificity of the vision, and the alacrity of the character narrated.

Our attention is thus drawn to revision as the outcome of acts of reading, and thus as the response of a creative imagination to potentials of meaning inherent in the text. Such a view of the acts and processes of revision is text-centered. It does not leave the author out of the account but focuses on the revising author as reader. Author, author-as-reader, and reader are thereby placed at a common point of reference, which is language itself. For, writerly creativity notwithstanding, texts and their meanings are ultimately constituted in language, and it is in the nature of language and texts to have potential for multiple meanings. Therefore, any response to the reading of a text by overwriting it in the interests of revision—rather like interpreting it in the interests of criticism—performs, in the very act of overwriting, only one of a potential multiplicity of retextualisations.

Revision is thus doubly controlled. In the secondary, because selective, respect, it is controlled by an author’s intuition, intelligence, judgement and taste. But in the primary respect, it is controlled, since it is engendered, by the text’s—the written, but as yet unrevised, text’s—potentials to mean, and (indeed) to signify. To progress from a dominantly author-centered to a dominantly text-centered understanding of the nature of revision thus helps us not only to appreciate the author as reader even in the very act of composition (and, incidentally, to remain undisturbed by instances of mis-revision: the critical evaluation as to bettering or worsening loses its relevance where the author can on occasion be just as good—meaning also: just as bad—a reader of the text as you and I). Above all, the shift to a dominantly text-centered understanding of revision also proves capable of avoiding a fixation on authorial intention when considering its acts and processes. In terms of a text-centered understanding, revision is recognised as being less the result of exclusively willed writerly decisions than of the playing of a text’s potentials of meaning against one another. Revision releases, deepens, shifts, or suppresses these potentials in a tendentially limitless, and thus theoretically indeterminate series. Pragmatically, it is true, the theoretically indeterminate play of language and meaning is always determinately embodied in the revisions actually carried out. In practice, revision will thus be radically determinate, since as experienced, and documented, it tends to be unique, or at most successively singular. In the case of Virginia Woolf’s reading response to her own writing in To the Lighthouse and other works, however, the documented revisions are simultaneously, and thus coexistently, double, or even multiple. Dealing even-handedly with that doubleness or multiplicity provides text-inherent openings to the text’s potentials for meaning which a forcing of the simultaneous variation into hierarchies governed by imputedly overriding intentions would precisely foreclose. Thus text-inherently conceptualised, moreover, the authorial acts of revisional reading may, in further consequence, be seen as initial steps towards their analogous continuation in the reading public’s reading of the texts as well as in the (theoretically speaking, again limitless) analytical and reinterpretative performances of criticism.

To return, then, to readings from To the Lighthouse regarded in such a light: in many cases, the double revision, or the half-revision, for that matter, is relatively unspectacular. Yet, freed from fretting over the question of which among alternative readings the text of To the Lighthouse should foreclose upon, we may find ourselves entertained, even thrilled by the play of equally possible alternatives. The American edition’s revision against the proofs and the British edition

(A1, 159; E1, 164)

may, it is true, be no more than a gesture towards an American audience unfamiliar with the significance (if any) of memorising the numbers on British railway tickets (though just what the cultural significance, if any, of numbers on American watches might be, then becomes the equivalent mystery). Similarly, the double-revision rendering of Mr Bankes’s mental image of Mrs Ramsay at the other end of the telephone would, in its turn, seem to be simply that: a sketch with variant brush-strokes of essentially the same picture:

(A1, 47; E1, 50)

The decision, however, about the degree of sympathy or antipathy with which to read, and make readable, Charles Tansley is inevitably a more serious matter. It makes a difference—while the alternatives clearly both grow out of the text’s potential—whether the New York edition reticently says of him:

then what they complained of about Charles Tansley was that until he had turned the whole thing round and made it somehow reflect himself and disparage them—he was not satisfied. And he would go to picture galleries they said and he would ask one, did one like his tie? God knows, said Rose, one did not. (A1, 16)

or whether the London edition sharply pronounces:

… and made it somehow reflect himself and disparage them, put them all on edge somehow with his acid way of peeling the flesh and blood off everything, he was not satisfied. And he would go to picture galleries, they said, and he would ask one … (E1, 18)

What is more: while the double-revision sketch of Mrs Ramsay at the other end of the telephone line was, either way, an invention made in the course of the revision, not an overwriting of antecedent text, the alternative renderings of Charles Tansley in the novel’s simultaneous public versions each differently modify one common antecedent passage in the proofs:

… and made it somehow reflect himself, and made them all feel in the wrong somehow—if it was fine well; then, the farmers, he would say, wanted rain—he was not satisfied. And he would go to picture galleries—could one imagine him looking at pictures?—and he would ask one …

(Proofs, 18)

What here may ultimately count for most, in critical terms, is our sharpened awareness of the play between the three variant characterisations that the text respectively realises—in the proofs, the American edition, and the British edition. What we also sense at times is an excitement at the possibilities—or indeed needs—for rewriting that the text opens up in the rereading towards revision. This is not only true when on occasion the reading of the proofs reveals a slip in stylistics that gets mended by one sure stroke, invariant in the two editions, as in the comment on Mr Carmichael’s success with the publication of his poems:

(A1, 202; E1, 208)

It is also exemplified in the double reading response to what the proof text conveys of Lily Briscoe’s sense of being caught up in one of those

(A1, 237; E1, 246)

This double rewriting of an initial attempt to articulate the dynamics of acquiring wisdom, and in the rewriting doubly to consider the reciprocity of experience and the mind, properly epitomises that very reciprocity of the reading and writing processes in revision that we are here discussing.

If in this instance the variation is tripolar, it remains bipolar elsewhere. Revising the text for one but not the other of the editions in preparation need not mean more than that both the proof reading unaltered and its revision qualify as equally valid public realisations of the text for To the Lighthouse. It makes little difference, for example, whether one reads ‘she sat in the window which opened on the terrace’ as the proofs and the New York edition agree (A1, 27), or simply ‘she sat in the window’ as the London edition is content to phrase it (E1, 29). Similarly, it is an even-handed alternative (in ‘Time Passes’) whether, among the ‘usual tokens of divine bounty’ which imagined visionaries pacing the beach might discern—such as ‘the sunset on the sea, the pallor of dawn, the moon rising, fishing-boats against the moon’—, the children, in whom this series of tokens culminates, should be engaged in one or two activities. The New York edition (again) follows the proofs in the doubling of ‘children making mud pies or pelting each other with handfuls of grass’ (A1, 201). According to the British text, they are merely ‘children pelting each other with handfuls of grass’ (E1, 207). The focus towards which this passage steers is anyhow not these tokens of divine bounty, but rather their antitheses of disharmony, ‘an ashen-coloured ship’ or ‘a purplish stain upon the bland surface of the sea’ signifying the upheavals in the order of nature in times of war.

Even more importantly, the time passing under the reign of war unhinges the world of the novel and disjoins the lives of its characters. The telling of Prue’s death in childbed gains poignancy from the juxtaposition of what ‘people said’ in variant response to it:

(A1, 199; E1, 205)

Throughout in the course of the revisions, one may observe several, and often subtle, rereadings of the characters. They are told with a greater reticence. While in ‘The Window’, for example, the London text goes with the proofs in tracing Mr Bankes’s thoughts as he watches Cam: ‘it would have been pleasant if Cam had stuck a flower in his coat or clambered over his shoulder, as over her father’s, to look at a picture of Vesuvius in eruption’ (E1, 40), the phrase ‘, as over her father’s,’ is deleted in the New York edition (A1, 37); as a result, Mr Bankes stands, subtly, further apart from the intimacies of the Ramsay family circle in the American text than he does in the British one. The critical question is not of course which reading is more appropriate or right. What is important to recognise is that the variation between the public texts is expressive of just those fluctuations of intimate familiarity and polite distance between the Ramsays and their guests that characterises the Lighthouse narrative as a whole. The inclusion of the phrase ‘as over her father’s’ is as fitting in terms of the novel’s patterns of meaning as is its exclusion.

In other instances, the focusing of character, combined sometimes with an increased reticence in character portrayal achieved through revision, is more noticeably a feature of the British edition. In ‘The Lighthouse’, for example, it is not ‘Just to please herself’ (A1, 281) that Cam would take a book from the shelf in the library, but, in the reading of the London text, ‘In a kind of trance’ (E1, 291). Just how his children see their father, and how the narrative itself sees Mr Ramsay, becomes increasingly important, especially towards the end of the novel. That ‘he was not vain, nor a tyrant (these were the things they hated him most for) and did not wish to make you pity him’ is what the proof text gives as Cam’s sense of him. The parenthesis ‘(these were the things they hated him most for)’ is identically deleted for both public texts. But thereafter, the British edition also ekes out his positive qualities. Cam now pronounces him ‘most lovable,’ ‘most wise;’ and she does so, as well, without adding that he ‘did not wish to make you pity him.’ Consequently, the variation pattern looks as follows:

(A1, 282; E1, 291)

It is in the course of the boat trip to the Lighthouse that Cam silently articulates these successively modulated feelings about her father. In terms of the construction of the novel, they echo and balance James’s fiercer thoughts of rejection. These at the same time, however, also undergo revision. The writing even at this juncture, late in the novel, engages in composing by radically recomposing the character of James. But we would not be able to appreciate this fully, were it not for the survival of the proofs. The published texts only sparsely shadow James’s inner turmoil that the proofs spell out in two adjacent passages. The first segment in question reads familiarly from the published texts as follows:

But he pulled himself up. Whenever he […] began

hearing the rustle of some one coming, the tinkle

of some one going, he became extremely sensitive

to the presence of whoever might be in the room.

It was his father now. The strain became acute.

(A1, 277; E1, 286)

Virtually identical in both public versions, this passage replaces a paragraph of palpably greater urgency in the proofs:

But he pulled himself up. Whenever he […] began

hearing the rustle of some one coming, the tinkle of

some one going, or that laugh which ended with three

separate “ahs”, each less than the last, like drops

wrung from the heart of merriment, it meant that

he was drawing near the thing he did not want to

think about (his mother), since it was terrible and

horrible to think of her with his father near; it

meant that something had started the sense of her,

as still by opening a drawer in a cupboard or

looking at a face—Rose’s for instance—through

one’s fingers one could recover her absolutely for

a moment. But it was horrible; the strain was acute.

(Proofs, 286–87)

Behind the published texts’ terse account of James’s deliberate ‘ceas[ing] to think’ under that strain, furthermore:

But all the time he thought of her, he was

conscious of his father following his thought,

shadowing it, making it shiver and falter.

At last he ceased to think; there he sat with

his hand on the tiller in the sun, staring at the

Lighthouse, […]

(A1, 278–79; E1, 288)

the proofs provide an extended account of James’s sense of his father after his mother’s death. This is how the deleted paragraphs read in the basic wording of the proof typesetting (a few revisional corrections marked in before these paragraphs were deleted wholesale indicate that attempts at retouching preceded their complete removal):

Now in London, now wherever they lived, they

were surrounded by distortions; lamentations;

and long speeches of violence; and old ladies

like Mrs. Beckwith being kind, and bald men

sipping tea and being clever while bread and

butter turned brown in the saucer, and there

one twiddled one’s thumbs in the heart of

unreality, sitting in the background on a stool,

and if in the middle of all this sighing and

being clever some one sneezed or a dog was sick,

nobody dared laugh. And the house grew darker,

he thought, and turned the colour of dusty plush,

and there were shrines in corners and nothing

could be moved, and nothing could be broken.

In the depths of the winter, or in those long

twilight months which seemed interminable, his

father, standing up very stiff and straight on a

platform in the city (to get there they must dine

early and drive eternally), proved conclusively

(but they could none of them listen) how there is

no God, one must be brave; for there is no God,

he said, while rows and rows of the ugliest people

in the world gaped up at him, in that greenish

hall, hung with brown pictures of great men. If

she had been there now, what would she have

done? he wondered. Laughed? Even she might

have found it difficult to tell the truth. He could

only see her twitching her cloak round her, feeling

the cold. But she was dead by that time. The

war was beginning. Andrew was killed. Prue

died. Still his father lectured. Even when his

hall was full of fog, and only sprinkled with

elderly women whose heads rose and fell, like

hens sipping, as they listened and wrote down,

about being brave, and there is no God, still he


Often they quarrelled among themselves

afterwards, what could one say to him? How

could one appease him? For he wanted praise.

He wanted sympathy. He wanted them to go with

him and listen to him, and to say how good it

was; how it was the greatest success. Rose said it,

forced herself to say it, but she said it wrongly

and he was angry; he was depressed. And James

himself wanted to say it, for he stood very

straight and very stiff, facing that dismal group

of people; one could not help admiring him;

liking him; as he stood there doggedly sticking

it out about God and being brave. So that sometimes

James would have liked to say it himself;

how he admired him; what a brain he had; and

would have done so, only his father found him

once with a book of his and sneered at him for

“it wasn’t the kind of thing to interest him”, he

said, whereupon James made a vow; he would

never praise his father as long as he lived.

There he sat with his hand on the tiller in the

sun, staring at the Lighthouse, […]

(Proofs, 288–90)

To have these passages of Cam and James in silent contemplation of their father preserved in the proofs permits us, in critical terms, to assess how discerningly the characters were adjusted in revision, and especially so, it seems, for the concluding sections of the novel. In terms of the present argument, of course, Cam’s warm views of her father, and James’s control of his inner conflict of feelings towards his mother and father, as they sail with Mr Ramsay to the Lighthouse, are identical in both public texts of the novel. The revisional changes take place between the proofs and the published text. Nonetheless, we have quoted these paragraphs, for they are perhaps the most remarkable passages of the book that were left behind at the pre-publication level of the text in the course of revision. Since they were so left behind, and hence are part neither of the British nor the American public text, it is true that they do not contribute to establishing the distinction between the simultaneous public versions of the novel. Yet comparing the paragraphs in proof against their revision as published helps us to appreciate the stringency of the text’s multiple options to mean.

At one of the novel’s crucial moments of composing and telling its characters, namely at the end of the ‘Window’ section, such variant options are (again) pursued concurrently. This is when, after a day of fluctuating between irritation and affection, Mr and Mrs Ramsay are granted their moment of intimacy together. On the level of the plot, it is also the moment after which Mrs Ramsay ceases to be the novel’s living centre. Every reader will remember the end of the ‘Window’ section, and in general terms also recall how the narrative here reaches its culmination. Specifically, however, the conclusion features significant versional differences that have seldom been highlighted as concurrent:

(A1, 185–86; E1, 190–91)

It is possible to construct a scenario of successive stages of revision, and even to base this in part on the real circumstances and time scheme of the preparation of the book for publication. By the evidence of the mark-up on the proofs for the New York edition, the first modification of the end was apparently the simple addition to the final paragraph of the line ‘She had not said it: yet he knew.’ This carried with it two consequences: it shifted the chapter’s final focus from Mrs Ramsay to Mr Ramsay; and it created a latent ambiguity: just what was it she had not said, yet he knew? Was it (as we were undoubtedly intended to understand) the words ‘I love you;’ or was it perhaps, and somehow confusingly, the sentence just actually spoken: ‘You won’t be able to go.’ Once the ambiguity was noted in the rereading, it was that sentence—voicing, as it did moreover, the Mr Ramsay note, rather than the Mrs Ramsay one—that was recognised as dispensible, and removed. Slightly modified, the sentence first added at the end of the paragraph to the New York proofs was moved into its place: ‘She had not said it; but he knew it.’ It refers unambiguously to her silent declaration of love. And with much more force than at the pre-revision stage in the proofs, the chapter ends again as it originally did: ‘And she looked at him smiling.5 For she had triumphed again.’

The final shaping of the text thus in the British edition all the more incisively marks the end as Mrs Ramsay’s end. An orthodox one-text edition of the novel would, on the foundation of this argument, recognise the American text as in a transitional state of revision and establish its critical text according to the British first edition. In the reality of the novel’s publication, however, the American and the British ending attain a simultaneous public presence. The critical insight to which this leads is that, according to the entire disposition of character and plot, either ending of the ‘Window’ section is a real and a valid textual option for To the Lighthouse. We build our field of critical understanding between the positions that these two endings mark concurrently. The challenge to critical editing is to support and make perceptively possible such a critical approach.

* * *

To the Lighthouse deserves a study edition that does justice to its first publication in simultaneous versions. Such an edition’s principles should be set out clearly and be consistently observed in establishing as well as in presenting the text. The principles should evolve out of careful text-critical investigations and aim at a concurrent presentation of the novel’s versional texts. To make such a presentation readable as well as usable will require explanatory and analytical notes.

Textually speaking, the point of departure for establishing a critical text of To the Lighthouse is the basic text layer of the first proofs, provided by the set of these proofs typeset by Clark & Clark of Edinburgh for the Hogarth Press in London that was sent from London as printer’s copy for the Harcourt, Brace & Company’s New York edition.6 These proofs are dated by date stamps between 31 January and 12 February 1927. The stages by which the novel’s text reached the first proofs cannot be fully recovered. The point of origin for the composition was the extant draft manuscript. It was begun on 6 August 1925, and finished (‘provisionally’, as a diary entry of 28 September 1926, comments) on 16 September 1926.7 Comparing the text in the proofs—set up from copy that must have reached Clark & Clark around mid-January 1927—with that of the first draft reveals that the book underwent extensive revision before it reached the proof stage. Evidence of this development is generally lacking. The book’s middle section, ‘Time Passes’, however, is documented at one intermediary point by a version of the chapter in typescript, datable to October 1926, from which, at that stage, a translation was made into French.8 Otherwise, all transitional documentation between draft and first proofs has been lost. But Virginia Woolf’s diary entry of 14 January 1927 describes how the physical side of the process of revising looked to her, and how she went about it: ‘Since October 25th I have been revising & retyping (some parts 3 times over).’ As a reference to her work on To the Lighthouse, this entry picks up on the more explicit note of 23 November 1926: ‘I am re-doing six pages of Lighthouse daily. This is not I think, so quick as Mrs D.: but then I find much of it very sketchy, & have to improvise on the typewriter. This I find much easier than re-writing in pen & ink.’

Woolf was observing working habits evident elsewhere (if, for example, the pattern of progressive composition and revision of Between the Acts offers a reliable analogy, as preserved in the originals at the Berg Collection in the New York Public Library). The typing up of first drafts (done in pen and ink) in the manner indicated resulted in typescripts that look as if they were carried forward on a wave: stretches of pages consecutively and singly numbered alternate with stretches of sheets typed for a second or third time with identical page numbers. At such wave-crest moments in the accumulating typescript, the identically numbered pages can be identified as first, second, or third typings of the same, progressively revised passages of text.9 From the cumulative revision typed out by Virginia Woolf herself, a further complete retyping was usually then prepared professionally. This is likely to have been, and in the case of To the Lighthouse must have been, done in close parallel with Woolf’s own typing. The diary entry for Friday 14 January looks back on work accomplished on Virginia Woolf’s part: ‘I have finished the final drudgery.’ Leonard is to be given the novel to read on the following Monday. That is, her professional typist may not have lagged behind her, if at all, by more than a day or two. Leonard in turn was not remiss in his reading: he pronounced the book a masterpiece by 23 January, and this tallies easily with the date stamp ‘31 January 1927’ on the first gathering of the first proofs set in Edinburgh.

Presumably the Woolfs sent an un-marked-up carbon of the professional typescript to Donald Brace before the New York publishers agreed to publish the book. Virginia Woolf notes on 12 February 1927 that Brace was less enthusiastic than he had been about Mrs Dalloway; but, she adds, his ‘opinions refer to the rough copy, unrevised.’ However, if the typescript read by Brace was unrevised, there is no indication that it differed from Clark & Clark’s printer’s copy (which, one may presume, would have been the top copy of the same typescript). Apparently, Woolf did not go over the professional typescript at all but expected to revise in print. The Edinburgh consignment of the first proofs was imminent when she recorded Brace’s reaction—or perhaps it had in part already begun to arrive, though the last proof gatherings, as we have mentioned, carry the date stamp of the very day of the diary entry, 12 February: a Saturday. Woolf goes on to describe the task awaiting her: ‘I have to read To the L. tomorrow & Monday, straight through in print; straight through, owing to my curious methods, for the first time. I want to read largely & freely once: then to niggle over details.’

This diary entry is quoted in the Shakespeare Head Press edition of To the Lighthouse, and indeed Susan Dick’s ‘Introduction’ to that edition also goes on to cite many, though not all, of the references in the diaries and letters to Woolf’s reading and revising the proofs. In terms of relating the progress of revision to the book’s production, however, further questions remain. Bibliographical facts and operations can be clarified that turn out to be not merely matters of book-making, but to relate to essentials of the text. One central question to be answered is just how the London edition of To the Lighthouse came to acquire its significant increase in unique changes over and above the state of shared revisions reached at the point when the extant set of proofs was sent off to New York, and thus also over and above the versional differences already thereby established. In terms of the routines of book production, one must assume that the further changes were made on revises. The routine of revises, however, is a stage that Susan Dick never allows for. Remaining either vague or silent on this issue, she nevertheless indirectly suggests that Virginia Woolf kept correcting the Hogarth Press exemplar of the one and only proof stage of which we still have a record, by way of its Harcourt, Brace fellow set. But actually, the extent and substance of revisions in the Hogarth Press edition going beyond the New York edition is so rich as to be more easily accounted for if we assume that it accumulated successively in the first proofs as well as in revises. As to the notion of revises itself, it is true that the term never occurs as such in Virginia Woolf’s diary entries. But this is probably because receiving and going through revises before giving the final go-ahead for printing was—and sometimes still is—a self-evident procedure in book production that hardly requires naming.

Unrecorded by Dick is a sentence from the diary entry of 5 March: ‘Finishing, correcting the last proofs that is to say, of a book is always a screw.’ According to Dick’s assumptions, what this would imply is that, while the extant first-proof set was sent to New York as Harcourt, Brace’s printer’s copy in late February, Virginia Woolf held on to the proofs for the London edition for another ten days or so, and continued to work on them until the early days of March. If so, she did not despatch the first proofs to Edinburgh until then. But it is equally possible that she returned the revised first proofs in their respective sets simultaneously to Harcourt, Brace for the New York edition and to Clark & Clark for the Hogarth Press publication. This would mean—the time schedule would have been tight, but not impossible—that on 5 March she was in fact reading revises and her diary would thus be speaking precisely and justly of ‘the last proofs.’

An intriguing piece in the puzzle of putting together this sequence of events is provided by the fact that, on or around 1 March, the concluding gatherings S to U(+) from another set of the first proofs were sent to New York (in two instalments). Why would this have been desirable, or necessary? Textually speaking, this belated consignment instructs the printer to delete the long passage of James’s thoughts about their father, and to work in the changes in Cam’s, as they make the novel’s culminating boat trip to the Lighthouse with him. In terms of character revision, as we have seen, these cuts and changes were momentous. In bibliographical terms, however, there was a thoroughly prosaic reason for the textual operation. It appears to have been suggested by a technical exigency. Its purpose, in terms of the London first edition, was to prevent an overflow of two pages of text into a new gathering and thus to contain the book within 320 pages, or a full twenty octavo (16-page) gatherings. It is quite conceivable that Clark & Clark warned Leonard and Virginia Woolf, as directors of the Hogarth Press, that to run the printing into another part-gathering with only two pages of text was extravagant. The point that it would mean a waste of paper, and an unnecessary cost, would certainly not have been lost on the careful Leonard Woolf. Such a warning, moreover, might have come with the return of the revises for final approval, when it was clear that the working-in of the revisions from the first proofs had not eliminated the overflow already apparent in those first proofs. The text changes that Virginia Woolf decided on in consequence were so significant—perhaps, in her opinion, even so happy—that it must have been important to her to incorporate them in the New York edition also, even though its typographical and bibliographical concerns were not affected. It would have been an advantage to take the sheets on which she marked up these further changes for New York from a remaining set of the first proofs, because the identical typesetting and pagination that the New York printers already had before them would help them to place these further instructions.

The inference, then, from the observable evidence is that the revises for the Hogarth Press edition were read in the first half of March 1927. Accordingly, the date entered on the draft manuscript: ‘finished March 16th 192<6>7’ carries full weight as witness of the fact. The diary entry of five days later, 21 March, seems consequently to have been written as Woolf wound up the process of ‘finishing, correcting the last proofs that is to say, of [her] book’:

Dear me, how lovely some parts of The Lighthouse are! Soft & pliable, & I think deep, & never a word wrong for a page at a time. This I feel about the dinner party, & the children in the boat; but not of Lily on the lawn. That I do not much like. But I like the end. (D-G 3, 132)

Receiving the revises with Virginia Woolf’s imprimatur a few days after 16 March gave Clark & Clark some six to seven weeks to incorporate the final corrections, to print the edition, to have it bound and to distribute it to the trade. To the Lighthouse appeared on 5 May, a date so significant that the Woolfs were surely aiming for it from the beginning of the entire production period, in January, on both sides of the Atlantic. 5 May was the day Julia Stephen died in 1895, and it is the epitaph for her mother that Virginia Woolf writes through Mrs Ramsay, the imaginative centre of To the Lighthouse.10

* * *

Ascertaining the recoverable facts and circumstances of the passage of To the Lighthouse from the first proofs through to the first editions, we have established the parameters for a critical edition in terms of the documents on which to base it. Textual criticism, in preparing the ground for the critical editing, has run half its course, and accomplished its first task. The second and remaining task lies in assessing the relative quality of the main witness texts themselves by applying text-related text-critical—that is, both bibliographical and critical—criteria and procedures. Would we be aiming for an orthodox critical edition on copy-text editing principles, a document text would also need to be selected, from among the extant witness texts, to serve as its copy-text. Since, however, our goal is rather to produce an edition that does justice to the existence of To the Lighthouse in two simultaneous public versions, the choice of a copy-text will not be a matter of overriding importance—though, as will be seen, the choice of a base text for the presentation of the two versions that we intend to propose will still require careful consideration.

For this enterprise, our first concern must be to establish each version on its own terms in a form of the highest possible authenticity. This means ascertaining how far, in every reading of words and punctuation, the witness for each of the novel’s versions—that is the British first edition, on the one hand, and the American first edition, on the other hand—provides an authorial as well as a non-corrupt version text. Assessing the record of each version text in its respective first edition, so as to gain each version’s genuine authorial text, means in turn stripping, from that record, the overlays both of house styling—most likely to occur in details of punctuation, spelling and typography—and of textual error.

Textual error may on occasion be inherited. It may have its origin in the proofs, from where it may descend to both, or only one, of the first editions. At 241.23, for instance, the proofs read ‘revivication’ and this goes unobserved in the British tradition of the text until the Hogarth Press Uniform Edition of 1930 eventually corrects it; the New York first edition, by contrast, immediately recognises this misprint and puts it right (A1, 233). Or the origin of a textual error may even lie in the typescript that served as printer’s copy for the Clark & Clark proofs. The proof phrasing of the second parenthetical passage in ‘Time Passes’, for example, looks deficient: ‘(One dark morning, Mr. Ramsay stumbling along a passage stretched his arms out, but Mrs. Ramsay having died rather suddenly the night before he stumbled along the passage stretching his arms out.)’ A clause to end the sentence seems to be missing. This may be the typesetter’s fault; or else, the typist already could have made the omission. Unfortunately, Woolf’s two efforts to correct and revise this passage, attempted separately for the two versions, did not in either case fully succeed in mending the phrasing.11

Where textual error has been introduced in either of the first editions, a positive check to detect it against the proofs is generally only possible, or at any rate is a great deal easier, for the New York edition. Since the extant proofs which provided its printer’s copy hold the record of every correction and revision actively made on them, any departure that the New York typesetters introduced is, if it is nothing else, house styling, mainly of punctuation and spelling; or it is a simple necessary correction of the proof text overlooked in the course of Woolf’s own correcting and revising. But if it is neither house styling nor a simple correction, a departure without instruction in the setting of the New York first edition from the first-proof text must be assumed to be an outright textual error. To establish a text of the highest possible authenticity for the American version of To the Lighthouse, one would eliminate such textual errors; probably accept the occasional correction of proof-text errors; and decide on a policy of how to deal with the house stylings.

A policy regarding the house stylings should relate to the general notion that the American and British editions make public simultaneous versions of the novel. One main reason that we have put forward for terming them simultaneous is the fact that the editions were published on the same day. One logical consequence of our basing the idea of versional simultaneity on the simultaneity of these acts of publication is to respect the American guise of the New York edition, and especially so in matters of spelling and punctuation. In other words: the American edition, in significant ways, and especially at the textual surface of the so-called accidentals, derives its versional individuality from its not fulfilling authorial intention in these matters. To respect its American guise is thus, in terms of underlying tenets of editorial theory, a gesture towards translating a social theory of editing into editorial practice. This might even stretch, if the production context warranted it, to accepting American idiomatic equivalents in words or phrases for English ones, even without the authority of an authorial instruction. The situation seems not to occur, however, in To the Lighthouse.12

The departures in the New York edition’s wording from the text set in the first proofs, or Virginia Woolf’s instructions for correction and revision on the extant exemplar of those proofs, can indeed with some confidence be identified as textual errors. The clearcut distinction is supported by the fact that, to the best of our knowledge, Woolf did not read revises on the American edition. She could thus not herself have made this change, for example, from ‘exaltation’ to ‘exultation,’

(A1, 166; E1, 171)

It is therefore entirely up to the editor to weigh the possibility of accepting ‘exultation’ as the American version-text’s emendation; or else to class it as a mistaken, if not unintelligent, guess at an intended reading, and consequently a textual error. Nor could Woolf have detected and amended mis-executions of her own revisional instructions that had resulted in textual errors in the American edition. One such instance is the strange compound ‘surface pool’ on A1, 266, where it is evident enough that the mark-up of the Harcourt, Brace printer’s copy was deficient, and the American text should really have read as the English one: ‘surface of the pool’ (E1, 276). Another case is the syntactic conundrum in the American edition: ‘the great in birth receiving from her, some half grudgingly, half respect’ (A1, 17). The proof mark-up presented a two-fold problem. Woolf had neglected to delete the second ‘half’ that she was replacing with ‘some’ and at the same time, she had placed that ‘some’ in an ambiguous position in the margin. The New York typesetters followed what they believed were their instructions, but the revised phrasing which the mark-up points to should be critically construed as ‘[…] receiving from her, half grudgingly, some respect.’

The British text, however, has: ‘[…] receiving from her, half grudging, some respect’ (E1, 19): for ‘half grudgingly’ it reads ‘half grudging.’ Is this a divergent revision, or is it a textual error in the British edition? Seeing that before revision, the proof text at this point read: ‘the great in birth coming to her mind now and then, half grudgingly, half respectfully’, is it safe to take the ‘grudging’ of the British text as an authentically revised reading? Or might this, in turn, be a textual error in the British edition caused by an accidental curtailing of the proof’s ‘grudgingly’ in the process of dovetailing the revision into the standing type? However, for the compound of changes before us, it would be text-critically unsound to accept one part—because it agrees with the evident revisional instructions for the American text—as a revision correctly carried out in print for the British edition, while rejecting another part—because it diverges from the American text—as a textual error. Grammatically, adverb and adjective are both possible in the given position, with only the subtlest shift in meaning, together with a perceptible modulation in rhythm and sound. The variance between ‘grudgingly’ and ‘grudging’ therefore constitutes a genuine versional difference between the American and the British text.

We have construed this example as a way of exploring how textual errors may be adjudicated in the British text. At the same time, what this experimental analysis highlights is the greater difficulty of isolating them there at all. We possess no pre-publication witness against which to check the performance of the Edinburgh typesetters. The typescript from which they set up the text has not survived, nor has the exemplar of the first proofs hitherto been traced that Virginia Woolf marked up for the British edition in parallel to that for the American edition. Nor, thirdly, do the revises seem to have been preserved on which, as we have suggested, a further round of revisions was entered in addition to those made on the first proofs. In general terms, however, this lack of documentation is to some extent offset by the fact that, contingent upon the routine procedures of book production itself, repeated rounds of correction were performed on the British text. In the course of the marking-up of the first proofs, not only was the text revised, but Clark & Clark’s typesetting was of course also corrected; and it was corrected a second time when the incorporation of the revisions from the first proofs was checked in the revises.

All this, naturally, did not guarantee an error-free text (typos like ‘revivication’ remained undetected), but the repeated working-over heightens our expectation that the British text ought generally to be sound. If textual errors persist, even though isolating them is trickier than in the American edition, the presumption is also that they are rarer. Nonetheless, we encounter, for instance, the phrasing at E1, 16.28–17.1: ‘[…] | like a Queen’s raising from the mud a beggar’s ‖ dirty foot and washing, when she thus admonished | […]’13 We would argue that these lines feature a textual error. To do so requires close bibliographical reasoning. The proofs have: ‘[…] | like a Queen’s raising from the mud to wash a ‖ beggar’s dirty foot, when she thus admonished | […]’ (cf. A1, 14). We assume that this should have been correctly revised to: ‘[…] | like a Queen’s raising from the mud and washing ‖ a beggar’s dirty foot, when she thus admonished | […]’ To perform the revision he was instructed to make, the compositor needed to remove line 16.28, which is the bottom line of page 16, and line 17.1 from the standing type of the adjacent pages. To replace them, he would have arranged two new lines of type, and he would have partly used the broken-up type from the removed lines to do so. In the course of the operation, he would have had three strings of type of approximately equal length to juggle with and to distribute over the end of the one and the beginning of the next new line of type; The three units were ‘dirty foot,’ (in standing type), ‘a beggar’s’ and ‘and washing’ and the latter two would appear to have been interchanged by mistake and put into each other’s intended positions. The first edition’s odd phrasing was the result.

The change effected produces a reading unique to the British edition. We believe it likely that this revision was made on the revises, and that this is the main reason that it was not identified as the textual error we assume it is, and thus not corrected before publication. In terms of the critical and bibliographic assessment of this passage and its textual error, the case would not be altered if the revision had been made on the first proofs. In this case, it is true that the revises would have offered an opportunity for checking that the revision had been correctly incorporated. Positive evidence, however, would be required that the opportunity was taken. It is a matter of principle of textual criticism and editing that a textual error passing under the authorial eye but left untouched in a round of correction cannot be considered as silently approved (and thus, by purely negative evidence, be regarded as no longer an error).

Oversights in correction, rather, are simply a fact of life in the transmission of texts. In the case of To the Lighthouse, indeed, these are not confined to easily detectable, and therefore in some sense trivial misprints. The persistent misnumbering of the sections in the final ‘Lighthouse’ division of the novel, for instance, is perhaps the most serious failure to correct that both criticism and editing have to contend with. Going back to the Edinburgh proofs, it most immediately affects the British tradition of the text, where the ‘Lighthouse’ division does not feature a section ‘2’ until the Everyman edition of 1938 marks an additional section half-way between ‘1’ and ‘3’ and calls it ‘2’. Whether this was inserted on authoritative instruction, however, is unknown. The American first edition, in contrast to both the Hogarth Press first and Uniform editions, notices straight away that section ‘1’ is irregularly followed immediately by a section ‘3’ in the first proofs, without an intervening section ‘2’. In restyling the section numbering from arabic to roman numerals, it simply calls ‘3’ ‘ii’ and renumbers all subsequent sections accordingly. Criticism may wish to resolve this crux by debating whether a segmentation into thirteen or fourteen parts is more appropriate to the ‘Lighthouse’ division of the book. What we would not wish to assume, however, is that Virginia Woolf, in failing to adjust the numbering sequence in the course of her several corrective rereadings of To the Lighthouse before publication, sanctioned the chapter’s lack of a section number, and/or a section, ‘2’.14

Despite a few textual errors and remaining misprints, the British edition of To the Lighthouse provides us with a thoroughly worked-over text. Its aggregate of correction and revision is appreciably higher than that of the American edition, and it remained under close authorial control throughout the period of the book’s production. As regards the accidentals of spelling and punctuation, the Edinburgh typesetters’ affinity to Virginia Woolf’s own styling was close, and their house styling would have been the natural—meaning: the conventional—extension of her own conventions as a writer, as well as an amateur tradeswoman in the composing room. (As an aside it may be noted how aware of the technical consequences of her revisional instructions Virginia Woolf often shows herself to be. She will, if she can, accommodate her changes to the typesetters’ need to shift type and typelines, helping them to limit the invasion into standing type to a minimum.) Both in terms of its textual substance, therefore, and of its British guise on the typographical surface, the Hogarth Press edition presents its own version of To the Lighthouse, distinctly individuated against the Harcourt, Brace edition. In the parlance of the intentionalist editors of old, it also, as it happens, provides the version of the text closest, overall, to the author’s intention.15

* * *

The editorial challenge of To the Lighthouse is to design an edition that will convey both the fluidity of the revisional finishing of the book and the versional distinctiveness and individuality of the simultaneous British and American editions. One way to accomplish these aims in book form would be a facing-page text edition. This would utilise the parallel presentation to make the versional difference readable by the horizontal comparison of the two texts in juxtaposition. Antecedent states of readings for either text, or both, would be footnoted at the bottom of the pages, as would any changes to the British text only—both corrections and revisions—that were made in British issues and editions after the first edition during Woolf’s lifetime. To each version, a small number of editorial emendations would lastly be necessary, and these, if not also footnoted, could be listed after the main facing-text block of the edition.

The essential aims of a two-version edition could, however, be accomplished less expansively in book form than through a relentless facing-page arrangement. The edition I am envisaging—while still featuring the three main presentational units described, namely the text page, its footnotes, and the appended matter—would challenge its readers by means of a dynamic and variable formatting of its continuous text pages. These would be more intricately arranged than plain reading-text pages. Mainly progressing as one text, and so emphasizing the novel’s forward flow, they would divide down the middle for passages or paragraphs where only a visualisation in parallel could adequately convey the divergence between the British and the American version. There would be two blocks of footnotes on the page. One would indicate the versional differences in any single words or phrases that were not displayed in parallel columns. The other would give the antecedent text from the proofs in cases where both versions were revised away from that text, whether identically or differently. Passages like the end of the novel’s ‘Window’ division might therefore be presented on the page in this edition as follows (where, at line 24, the text flow divides into two columns, the left-hand column represents the British text, and the right-hand column the American one):

Similarly, a passage in mid-paragraph where a couple of sentences were deleted in revision for the American edition only might be presented thus:

These proposals for an editorial presentation of To the Lighthouse in its two simultaneous versions should be understood as a first sketch only, open to reconsideration and modification. The choice to footnote the versional variant at line 11 of the first sample, for instance, has here been made mainly to illustrate that I would envisage using both parallel passaging and footnoting to communicate the versional differences. Yet, weighing the variation critically, the editor might equally decide to parallel ‘did not mind looking now, with him watching, at the Lighthouse. For’ against ‘remembered how beautiful it often is—the sea at night. But’ in the same way as the end of the section is paralleled; or as is the deletion in the American edition, which the second sample exemplifies, and where the absence of text is immediately apparent from the blank right-hand column.

The second footnote block to the first sample records, with reference to line 9, that the published texts agree in reading ‘do for him?’ against the proofs’ ‘do?’—this therefore is a revision common to both versions. Next, with reference to line 25 of the left-hand column, the antecedent reading of the proofs is also reported. A look at the right-hand column of the text block will confirm that the American edition transmits the proof reading unchanged (except for the removal of one hyphen). This footnote entry therefore underscores the way in which the left-hand column, as it stands, is unique to the British edition’s realisation of the novel’s text. Finally, the absence of footnoting in the case of the phrases, ‘She had not said it, but he knew it.’ of the British text, or ‘She had not said it: yet he knew.’ of the American one, implies that they have no antecedent reading in the proofs.

What the footnoting does not report at all is that, at line 9 ‘up’ and at line 19 ‘any thing’ in the first sample, the American text reads ‘up,’ and ‘anything’—that is to say, no space is given on the text pages to recording the variation in accidentals of the American edition. The text presentation in the edition here envisaged follows the accidence and styling of the British text, except in the right-hand (that is, the American-version) column where passages are displayed in parallel. This is where one might consider such an edition as in conflict with the rationale here developed for it. For after all, I explicitly argued above for the autonomy of the American version of To the Lighthouse and maintained that this was due not least to its independent styling. On the other hand, however, I emphasised for the British edition that it had received repeated rounds of correction and revision, and altogether more constant general attention from author, publisher and printer cooperating on the book over an extended timespan. It thus provides, by comparison, the more significantly and effectively worked-over of the novel’s two texts. This should justify a decision to choose the British text as reference and base for the edition proposed, even as that edition sets out to enable its users and readers to experience Virginia Woolf’s To the Lighthouse in the fluidity of its two simultaneous public versions of 1927.

* * *

A study edition of Virginia Woolf’s novel To the Lighthouse as we have outlined it is not limited, in conception, to what it proposes to accomplish technically, in terms of text analysis, editorial discrimination and a function-oriented presentational surface. A command of the methods of textual criticism and the observance of precision and accuracy in editing, as well as of user demands in terms of presentation, while necessary, is not in itself sufficient when textual criticism and editing are properly understood as belonging among the foundational disciplines of literary scholarship and criticism. As such, they must be assessed in terms of their achievement, and according to how far they enable, and how much they contribute to, the critical endeavour.

An edition allows us to experience a text materially. A text edition that is based on the text- and work-related records of composition, revision and transmission conveys the dynamic quality of the text’s materiality, which arises in its turn out of the generative potential of the creative processes of composition and variation. A text’s dynamics of composition and variation are themselves always already played out in a field of force between writing and reading. Not only, therefore, is there a need for editions that render a text’s range of variation accessible, in so far as it is possible for editions to do so, that is, as far as the texts and their variants survive in material records. It is equally essential that criticism in theory and method should recognise the nature of its own enterprise as a dialectic of reading and writing that constructs its discourse on the analogous discursive structure of its subject. The subject of literary criticism, the literary work, is always discursively structured in language. Its discursiveness plays itself out in that its text can always also be other. This is why revision is not accidental to it, but of its nature.16 The option of revision is always inherent—which is also exactly why the literary text and work are interpretable. But to recognise whether, when and how a text’s potential for revision has been realised, requires a material record of the processes of writing and reading through which works and their texts have been constituted, in form, in wording and in meaning.

The text of Virginia Woolf’s To the Lighthouse, as we have seen, can always also be other: the novel’s two simultaneous versions bear this out in very graphic terms. Thus, the character of Charles Tansley can be drawn both mildly ironically, and sharply; the children pictured as playing on the beach can be making mud pies as well as pelting each other with handfuls of grass, or they can only be pelting each other; perhaps, too, there can even be a ‘ring of exaltation and melancholy’ quite as much as a ‘ring of exultation, and melancholy’ in Mr Ramsay’s voice.17 Moreover, in addition to speaking in alternative ways, the text can always also be other in alternatives of speaking or being silent. Mrs Ramsay can say or not say, for instance: ‘You won’t be able to go.’ From exactly this passage, in which her saying or not saying these words distinguishes the simultaneous versions, we may find, indeed have found, a diachronic axis opening up into the text’s pre-publication state. On this axis, the material record allows us to register the alternatives of silence or wording as operating between the absence, in the proofs, of text corresponding to the phrases: ‘She had not said it; yet he knew./She had not said it, but he knew it.’ and the presence of these phrases in both the alternative wording and the alternative form (meaning here: the alternative placing) in the text’s published versions. As to wording versus silence, reticence versus explicitness, the reverse situation occurs, for instance, with respect to a James outspokenly resentful of his father in the proof text, and an emotionally controlled James in the published text.

The evidence from this textual and text-critical analysis, reflected on such terms, establishes a point of vantage for criticism. To the Lighthouse may be seen to spring from, and at the same time, through its composition and revision, to generate, oppositions of silence versus speaking. In general terms, it is true, this thematic as well as structural duality can be understood simply from the novel’s published text in either version. Insights can be deepened, furthermore, through a study of the material records as we have here described and analysed them. Yet it is ultimately only through an edition, we would argue, that the reader and the critic, freed from any interposed discourse, but with the edited material record set directly before them, will be able to experience for themselves the processes of writing and revision of the work and its texts, and to assess their significance. And this is ultimately why textual study and the editorial enterprise critically matter.18

1 Virginia Woolf, To the Lighthouse (London: Hogarth Press, 1927) (E1); Virginia Woolf, To the Lighthouse (New York: Harcourt, Brace & Co., 1927) (A1).

2 These are now housed in the Frances Hooper Collection, William Allan Neilson Library, Smith College, Northampton, Mass. The library’s courtesy in supporting the research for the present investigation is gratefully acknowledged.

3 It is item M31, in three parts, among the Virginia Woolf holdings in the Berg Collection, New York Public Library. It has been published in a transcript: Virginia Woolf, To the Lighthouse: The Original Holograph Draft, ed. by Susan Dick (Toronto: University of Toronto Press, 1982).

4 In the collation printouts used as illustrations in this article, the base text is that of the first proofs. It carries their page.line numbering, which is close, or identical, to the page.line numbering of the British first edition. The collated texts are from the American [us] and British [eng] first editions. These are subjoined line by line in parallel to a base text line, with their variants only printed out. Text identity in the collated witnesses is marked by ‘===’; absence of text in any member of the collation is indicated by blank space. Where the collation display progresses numbered line by numbered line without parallel collation lines, the text in all three witnesses is identical. The symbol combination ‘$_’ indicates a paragraph opening.

5 He is smiling, as well as she is, surely: the absence of a comma after ‘him’ has that double effect.

6 See above, note 2.

7 The Diary of Virginia Woolf. Volume 3: 1925–1930, ed. by Anne Olivier Bell, assisted by Andrew McNeillie (London: Penguin Press, 1982 (paperback); London: The Hogarth Press, 1980 (hardback)). This volume, in the Penguin edition, was referred to throughout for this article. The entry for 28 September 1926 extends over pages 111–12, the present citation is to be found on p. 111. For subsequent citations, the diary date given in the text should suffice.

8 An account of the arrangements for the translation into French is given in Virginia Woolf, To the Lighthouse, ed. by Susan Dick (The Shakespeare Head Press Edition) (Oxford: Blackwell Publishers, 1992); ‘Introduction’, p. xxviii; Appendix C of this edition (pp. 212–29) gives a transcript of the typescript itself.

9 For the surviving typescript pages of Pointz Hall—only late renamed Between the Acts—I have come to these conclusions by undoing—virtually, if not physically—their rearrangement undertaken by a former curator of the Berg Collection. Future research into Virginia Woolf’s working habits based systematically on the broad evidence in the archival holdings accessible on both sides of the Atlantic would seem greatly desirable.

10 In a letter to her sister Vanessa Bell of 15 May 1927—that is, ten days after the date of publication—Virginia Woolf knows and mentions the book’s exact number of pages: ‘Dearest, | No letter from you—But I see how it is— | Scene: after dinner: Nessa sewing: Duncan doing absolutely nothing. | Nessa: (throwing down her work) Christ! There’s the Lighthouse! I’ve only got to page 86 and I see there are 320. Now I cant write to Virginia because she’ll expect me to tell her what I think of it. | Duncan Well, I should just tell her that you think it a masterpiece. | Nessa But she’s sure to find out—They always do. She’ll want to know why I think its a masterpiece | Duncan Well Nessa, I’m afraid I cant help you, because I’ve only read 5 pages so far, and really I don’t see much prospect of doing much reading this month, or next month, or indeed before Christmas.’ In Congenial Spirits. The Selected Letters of Virginia Woolf. New Edition, ed. by Joanne Trautmann Banks (London: Pimlico, 2003), p. 224. Page 86 of the first edition, which is as far as Vanessa is supposed to have read, seems intriguingly significant. It speaks of Lily Briscoe: ‘She took up once more her old painting position with the dim eyes and the absent-minded manner […]ecoming once more under the power of that vision which she had seen clearly once and must now grope for among hedges and houses and mothers and children—her picture. It was a question, she remembered, how to connect this mass on the right hand with that on the left. … But the danger was that … the unity of the whole might be broken. She stopped … she took the canvas lightly off the easel.’ With what greater emblematic succinctness in a letter could Vanessa’s and Virginia’s sisterhood in art, together with the figurative signification of Lily Briscoe and her painting for the novel, be expressed?

11 A1, 194; E1, 199–200. The only genuinely critical attempt, to my knowledge, to emend this persistent textual error has been made by Stella McNichol in her 1992 Penguin Books edition of To the Lighthouse. Julia Briggs mentions the instance in her perceptive critique of Woolf textual criticism, ‘Between the Texts: Virginia Woolf’s Acts of Revision’, TEXT, 12 (1999), 143–65 (p. 153).

12 But it is a pervasive phenomenon in transatlantic double publishing generally. Fredson Bowers may be remembered by some as protesting vigorously against what he saw as a malpractice in publishing to adjust everyday idiom (‘gas’ for ‘petrol’ for example, or vice versa) when an English book was published in America, or an American one in England. What was anathema to the intentionalist could be taken in her stride by a post-intentionalist editor orientated, over and above intention, towards production and reception factors in the critical constitution of a text.

13 The line breaks (|) and the page break between pages 16 and 17 (‖) fall as indicated.

14 I argue the structural significance of 13 sections for ‘The Lighthouse’ below in ‘From Memory to Fiction’ on p. 290.

15 J. A. Lavin, ‘The First Editions of Virginia Woolf’s To the Lighthouse’, Proof, 2 (1972), 185–211, however, gave original currency to the opposite view. ‘[W]ithout having Woolf’s marked-up proofs to check against, [Lavin] mistakenly concluded that the American edition represented the latest and best state of the novel, “superior to the one published in England by Mrs Woolf’s own company” (187),’ comments Julia Briggs in ‘Between the Texts: Virginia Woolf’s Acts of Revision’, p. 151. This is a nobly reticent assessment. It is true that Lavin worked under the handicap of assuming the proofs to be lost; they were at the time still in private repository. Yet systematic and critical collation, combined with the use he did make of diary and letters, should by sheer force of text-critical logic have led Lavin to truer conclusions about the textual situation, the lack of the physical evidence from the proofs notwithstanding.

16 This is a central tenet in the laying of theoretical foundations for the study of revision by Roger Lüdeke in his Wi(e)derlesen. Revisionspraxis und Autorschaft bei Henry James (ZAA Studies: Language, Literature, Culture, no. 14) (Tübingen: Stauffenburg Verlag, 2002). Roger Lüdeke wrote the book as a dissertation under my direction, and he has taught me a wider understanding of authorial revision in literature.

17 It is true that we have assessed ‘exultation’ as a textual error in the American edition, for which reason it should not be accepted in a text-critically constituted text, conceived of as an authentic text, of the novel’s American version. Nonetheless, this reading is testimony to a discerning reading of the first-proof text, resulting in a change that was assumed to be corrective. Hence, also the variation of ‘exultation,’ against ‘exaltation’ goes to show that the text can always also be other.

18 Note in October 2017: This essay was written in the early 2000s. It conceives of the scholarly edition still essentially in book form. It may and should stand as it is, since books on many fronts continue to be the goal of the editorial endeavour, and legitimately so. But scholarly editing has equally meanwhile advanced far into the digital age. Personally, I would today reconceive an edition of To the Lighthouse as a digital edition and realise it in that new native environment of scholarly editions, even though under, in principle, the same theoretical, methodological and critical premises here laid out and argued.