The Vance

First chapters


The Foreverness
Message Board

The Cosmopolis




Wilver the Water-walker said thoughtfully: “I would resign my important position in an instant to rejoin the old troupe! What do you say, Master Zamp? Why should we not revive the glorious old times?”
“Never be guided by sentimentality!” advised Zamp. “I advise you all to remain with Garth Ashgale, whose terms of employment are stable; as I recall he discharges only the notoriously incompetent.”
“He can also be a difficult task-master,” grumbled Wilver. “He wants me to perform my act without glass stilts, which is difficult.”
“We all have similar problems with Ashgale,” said Thymas. “For instance, in his production The Extraordinary Dream of Countess Ursula Gandolf and I must simulate strange animals in questionable poses.”
(From ‘The Magnificent Showboats…’, p.130—all page references are to VIE volumes.)

Strange Animals In Questionable Poses
(The Reconstruction of the V.I.E. Texts)
by Alun Hughes

During rehearsals Zamp attempted to simplify and modernize certain obscure phrases, and again found himself in controversy with Gassoon, who insisted on fidelity to the original. “All very well,” cried Zamp, “but speech is spoken that it may be understood. Why present a drama which simply bewilders everyone?”
“Your mind lacks poetry,” Gassoon responded sharply. “Can you not imagine a drama of hints and dreams which totally transcends the animal titillations and spasms and hooting sounds upon which your reputation is based?…Authenticity must be our watchword.”
(From ‘The Magnificent Showboats…’, p.121)

One of the aspects of the VIE will be the restoration of Vance’s works to the state intended by their author…as far as possible all ‘editorial improvements’ will be eliminated. (From a Statement of VIE Intention.)

It is hardly a controversial stance. The “intentionalist” approach to textual criticism has been the dominant twentieth-century approach. The author’s intention is paramount; all changes from the author’s original are corruptions; the ideal is the restoration of the author’s uncorrupted original.

In some senses, the task is not too difficult. The texts exist in a relatively small number of versions; there were few interventions in proof; we need not worry about variants within editions; the language presents few difficulties. On the other hand, standard techniques of textual criticism have not conventionally been applied to 1950s pulp magazine publications; the notion of the author “seeing his work through the press” hardly applies; the author is living, but cannot be expected to be fully familiar with texts he wrote fifty years ago. Electronic texts exist for the later novels, but in an electronic environment the problems of textual genealogy and version control can sometimes be exacerbated rather than simplified. For example, we had electronic texts for ‘Ecce and Old Earth’ and ‘Throy’, but in multiple versions whose relationship to each other and to the printouts and published editions required considerable analysis. The computer files were seldom final, with revisions often made to printouts. Because printouts might be made more than once and revised separately, the sequential and cumulative textual development of the earlier texts was replaced by a set of variants that were related in rather more complex ways.

A fairly common view is that the definitive text is the last published during the author’s lifetime. This relies on a degree of belief in the author’s willingness and ability to see each edition through the publication process that is manifestly not sustainable in the context of late twentieth-century genre publishing. As a general rule, it is probably safer to think of the definitive text as the last for which the author retained demonstrable responsibility.

In practice, this means that the first edition is likely to be the most definitive, because later ones will be more or less corrupt, unless Jack revised or otherwise took responsibility for a later text.

However, as a principle this needs to be interpreted carefully. An interesting problem is raised by ‘Guyal of Sfere’. This story, originally part of ‘The Dying Earth’ (published in the VIE as ‘Mazirian the Magician’) and written around 1944, was revised by Jack for its re-publication in the collection ‘Eight Fantasms and Magics’ in 1969. The revision was quite drastic, shortening the story appreciably and clearly affecting its colour and texture. It must represent the author’s final intention—but that presents the textual editor with two problems. First—frankly—not all would recognise it as an improvement on the original. To some eyes it represents an uncomfortable compromise between styles separated by over twenty years. Second, should we publish ‘Mazirian the Magician’ with one part heavily revised and now stylistically inconsistent with the rest? Our conclusion was to treat ‘Mazirian’ as the “text” in question, allowing us legitimately to include the original ‘Guyal’, with the revised version considered as a separate text and published in the VIE as such.

One could cite further examples. The early stories ‘The World-Thinker’ and ‘I’ll Build Your Dream Castle’ were significantly revised by Jack for their book publication many years later in ‘Lost Moons’. In his introduction to that volume, Jack says:

Two [stories]…were so embarrassing that I rewrote a few stand-out passages, a lick-and-a-promise operation rather like putting rouge on a corpse.

In fact Jack’s revisions were rather more substantial than he suggests, though I would not quibble over his description of their effect. Some would argue that ‘The World-Thinker’ has a kind of historical importance as Jack’s first published story (though not his first written, or sold) and as such the original version has an interest of its own.

Although we had little difficulty with the concept of restoring Jack’s original writing by eliminating editorial intervention, we had far more difficulty with the idea of ‘final intention’. Should we accept as final intention those changes made reluctantly, and to ensure a sale, at editorial request? Many of the revised short stories, for example those revised (along with ‘Guyal of Sfere’) for ‘Eight Fantasms and Magics’, read as neither a product of their original time nor of the time of their revision, and we have often taken a view that a text appropriate to the context of its publication in a particular collection may not be the most suitable text for the VIE. But we have not been consistent in the way that we have treated these revisions and it would be difficult for us to articulate a single credible rationale: rather, we propose to make it clear, in each case, what the treatment of the text has been. Thus, ‘The World-Thinker’, in its revised (1981) version, and ‘I’ll Build Your Dream Castle’ in its original version, appear in the main volumes, with the substantially different version of the latter, published in ‘Great Stories of Space Travel’, 1963, appearing separately (in this volume), and the not substantially different original of the former not included. Crusade to Maxus was heavily revised for its re-publication in ‘The Augmented Agent’ in 1986, with editorial changes as well as the author’s; the VIE version has elements of both, with a major episode from the original reinstated. ‘Noise’, ‘The New Prime’ and ‘The Men Return’ similarly represent a set of judgements as between the originals and the 1969 revisions in ‘Eight Fantasms and Magics’, but take as their bases the originals. In the same volume ‘Cil’ was changed very slightly from the first version published only three years earlier, and in the VIE the story appears in its context in ‘Cugel the Clever’ (originally published as ‘The Eyes of the Overworld’) in its original form. Telek contains many of the textual changes in the revised version but restores some major excisions. ‘When the Five Moons Rise’ largely restores the original, but the differences are very minor in any case. ‘The Miracle Workers’ generally follows the revised text, though again the level of revision is fairly low.

Where the author is known to have re-edited his text, usually parsimoniously and with the aim of eliminating the occasional awkwardness or mistake in the previously-published text, we have had no difficulty in adopting the revised text. This applies to ‘Abercrombie Station’ and ‘The Moon Moth’ as published in ‘The Best of Jack Vance’, but of others in that collection, ‘Ullward’s Retreat’ is restored based on its appearance in Galaxy (1968), ‘Sail 25’ based on ‘Future Tense’ (1964), ‘The Last Castle’ on the typescript prepared for the Tor 1989 publication, and ‘Rumfuddle’ based on the typescript submitted for ‘Three Trips in Time and Space’ (1973). Throughout, we have been concerned to ensure that the texts presented in the VIE are authentic Vance, and especially where later versions are known to have been the subject of editorial intervention, we have often preferred earlier versions. This is also the case with three of the stories in ‘Future Tense’, including ‘Ullward’s Retreat’ but also ‘The Gift of Gab’ and ‘Dodkin’s Job’. Although the author is known to have revised these, mainly by deletion, there are also substantial editorial changes. However, ‘Sail 25’ was considerably added to by the author for ‘Future Tense’, and we have therefore preferred that version, though we have also attempted as far as possible to eliminate editorial intervention.

In the case of ‘Ullward’s Retreat’, the various versions in ‘Future Tense’, ‘The Best of Jack Vance’, and ‘Dust of Far Suns’ differ largely in accidentals, that is, particularly in punctuation, and we have had to make judgement calls in differentiating between authorial and editorial intervention. The VIE text takes elements from each.

The stories collected in ‘The Dark Side of the Moon’ and ‘The Augmented Agent’ (1986) were heavily edited in an attempt to modernise the texts. With the exception of ‘Crusade to Maxus’, revised by the author, and the removal of the last line in ‘DP!’ (see below), we have not used these texts.

Towards the end of the process of writing a book or story—probably after the end as he saw it—Jack would often comply with a publisher’s request for changes, but without enthusiasm, and minimally, making the fewest changes to the text that he could consistent with satisfying the publisher’s requirement. In the case of the text for which we have the fullest evidence—‘Emphyrio’—including correspondence with the publishers, this is very clear. In ‘Gold and Iron’, the ‘happy’ ending has been disowned by the author, to the extent of suspecting (we should remember, fifty years after its composition) that it had been added by another hand. However, there is not much doubt that the offending material is by Vance himself, and a reasonable explanation is that the happy ending was required by the publisher and provided by the author with a minimum of enthusiasm and a maximum of despatch. For the VIE appearance, the text is based on the original, and suspect, appearance in ‘Space Stories’ but, in order to respect the author’s intention, a slightly revised ending approved by the author has been incorporated. In contrast, in its first appearance (‘Avon Science Fiction and Fantasy Reader’, 1953), a grotesque final line was added to ‘DP!’ by its editor (“But I am seized by an irresistible urge to ‘tell-off’ a rotten, inhumane world…”). This was removed by the author himself for its re-publication in ‘The Dark Side of the Moon’ in 1986.

It is not, however, always necessary to presume that the author’s intention is what he actually wrote. To take an example from ‘Wyst’: it is clear from the typescript evidence that at a fairly late stage Jack decided to change ‘Alastor Agency’, which appeared numerous times in the book, to ‘Alastor Centrality’. But he missed some of the occurrences, leaving an inconsistency which I would find it impossible to argue that he intended. We standardized on ‘Centrality’. However, in other instances of inconsistency, especially where the evidence was less clear-cut, we have not attempted to second-guess the author. In the same text, what is apparently the same beer is referred to twice as ‘Dark Wort’ and twice as ‘Dankwort’. Neither wishing to toss a coin nor wishing to suppress any of the richness of the text, we have left both versions in. Careful readers will notice that we have done the same with the two versions of the letter from Aleida Gluster to the Connatic that appear in chapters 1 and 9. That was the subject of considerable debate within the Textual Integrity team, and it would be fair to say that we are still not unanimously sure that we were right. Many of our decisions have necessarily been judgement calls, and it would be unreasonable to think that none of our decisions could have been taken differently.

There is evidence in correspondence that Jack accepted sensible editorial intervention in a number of works, for example ‘The Man in the Cage’ and some of the editorialising in the Joe Bain books. One school of thought prefers the first printed edition as “copy-text” (the preferred reading) on the grounds that the author can be expected to have concurred with the editor’s or proofreader’s emendations—after all, did he not check the proofs? Another prefers the manuscript, if available, on the grounds that it is simply not a safe assumption that the author checked, corrected and approved the proofs.

In Jack’s case, the evidence suggests that he gave proofs comparatively little attention, and sometimes none at all. Jack’s editors and proofreaders have, at various times, “normalised” spelling and hyphenation; corrected clear and trivial spelling errors; corrected what they thought were errors but were not; changed what would conventionally be seen as bad grammar; clarified awkward phrasing; killed elegant phrasing; replaced recondite words with commonplace; arbitrarily changed punctuation, word choice, and even chapter order.

In ‘Wyst’, Jack’s ‘pastes’ became ‘pastas’ and in ‘Magnificent Showboats’, ‘cast a vague loom’ became ‘loomed vaguely’. In ‘Madouc’, Casmir ‘made shift to ease himself upon the quivering body of one of his pretty pageboys’ except in the Ace edition, where we have ‘pretty pages’. In some rather overenthusiastic editorial updating in the stories in ‘The Dark Side of the Moon’, ‘visiphone’ becomes ‘videophone’, ‘dome’ (of a spacesuit) becomes ‘faceplate’, ‘airplane’ becomes ‘flier’, ‘teleview’ becomes ‘screen’ and ‘platoscope’ becomes ‘instrument’, then, rather desperately, ‘digital enlarger’. Editors were also often unkind to Jack’s coined words. In ‘The Killing Machine’, a ‘sumptuous flying car, decorated in the most elaborate fashion, with colored lumes’ becomes decorated with ‘plumes’ to please only the computer spellchecker.

Editorial intervention in the pulp magazines typically consisted of cutting to length; changes to punctuation to speed the pace, eliminating colons and semicolons and shortening sentences; changes to individual word choice, often substituting a common word for an unusual one. Scene breaks would be inserted and many new paragraph breaks added to suit the columnar format. Major structural alterations to stories were rare.

In the book publications, there is a considerable contrast between the rough-and-ready editing of the high-volume publishers such as Ace, and that of hardcover publishers of greater reputation, such as Doubleday and Bobbs-Merrill. The level of intervention could be high or low in each case, but in the former, one would be likely to see very many small changes to word choice and punctuation, where in the latter (and especially with the mysteries) there might be fewer but individually more substantial changes, often but not invariably negotiated with the author. Not surprisingly, the mysteries have been more carefully edited for internal consistency, not a normal Vancian strength. We have seldom intervened to improve consistency in the texts; generally, we have limited ourselves to implementing naming consistency where we believe that to have been the author’s intention, along with a small number of changes to achieve narrative sense, for example in ‘Durdane’. There are major internal inconsistencies in ‘Magnificent Showboats’, which we have left as an exercise for the reader, and also in ‘The Green Pearl’. This is from Steve Sherman’s textual integrity notes on this text:

An interesting aspect of this text is that there are rather a lot of continuity problems. For example, at one point Tatzel clasps her arms around her knees—at a time when she has a broken leg. When Melancthe first comes to Trilda, Shimrod takes her cloak. A page and a half later he takes her cloak again—and her gown comes off. Where in Suldrun’s Garden the term of a Ska slave is thirty years, in this text it is twenty.

The largest continuity problem has to do with the Rote of Persilian. At first mention, Casmir is reflecting on the matter and recollects only vaguely the words used to foretell the ascent of Suldrun’s son to the Throne Evandig. The second mention is in a dialogue with Tamurello, and now Casmir remembers the exact words of the Rote. But of course in Suldrun’s Garden the actual source of the prophecy about Suldrun’s son is Desmëi.

 Some editorial changes are perfectly understandable: for example, all editions of ‘The Dying Earth’ after the first transpose the first two chapters, presenting them in chronological sequence. But the first Hillman edition is true to the author’s intent and in ‘Mazirian the Magician’ we have restored the author’s preferred title and chapter order. In ‘Suldrun’s Garden’, Chapters 25 and 26 appear in different orders in different editions. The typescript evidence we have to hand indicates that the correct order is as in the Underwood-Miller edition, which again is not perhaps the natural chronological order but represents the author’s intention.

So, do editorial changes have any textual validity for the purposes of the VIE? It is possible, and reasonable, to argue that the author, even if he did not check his proofs, nonetheless expected his work to be changed by his publishers. Some of this work may legitimately be seen as restoring the author’s intention—that is, we presume, for example, that the author did not intend crass mistakes in spelling or grammar. If so, we should take advantage of that proofreader’s work. That does not absolve us of the responsibility of restoring, as far as possible, the author’s intention.

Of over 400 editorial proofreader’s emendations in ‘Wyst’, we accepted fewer than twenty, or about 5%. In addition, we have corrected clear errors which the proofreader missed—perhaps another five changes. The harder TI decisions have tended to be the judgement calls; this is why we had a review mechanism in place that allowed these decisions to be tested at a level approved by Jack and Norma themselves when necessary.

The varied nature of the textual evidence that we have for Jack’s work means that we have not been able to define a single detailed methodological approach that we can apply to each text. Sometimes there is nothing beyond a single published text (‘Vandals of the Void’, ‘Bad Ronald’); or there are multiple drafts and setting copy (‘Emphyrio’, ‘Tschai’, ‘Magnificent Showboats’); divergent published texts with no available manuscript evidence (‘The Languages of Pao’, ‘The Dragon Masters’, ‘Marune’, ‘Trullion’, ‘The Star King’); published texts with setting copy (‘Nopalgarth’, ‘The Last Castle’); published texts with typescript versions that are probably final (‘The Dark Ocean’, ‘The Houses of Iszm’, ‘Durdane’); published texts with typescript versions that are known not to be final (‘The Man in the Cage’); electronic files, which have their own problems of interpretation, along with printouts and published texts (‘Madouc’, ‘Ecce and Old Earth’).

Each case has needed its own approach, but the overarching textual principles are constant. My own rule-of-thumb was “authentic, but not painfully authentic”—this seemed to strike a chord amongst Textual Integrity workers, and served us well through many difficult decisions, as we attempted to restore Jack’s intentions but not his slips of the pen.

In assessing the evidence for each text in the VIE, we tried to examine each textual version that was likely to be of any value. We have not examined every re-publication of every text, so, for example, while the version of ‘Abercrombie Station’ published in ‘The Best of Jack Vance’ is a valuable text slightly revised by the author, we took it as a fairly safe bet that the version published in ‘The Science Fiction Weight-Loss Book’ is of no textual significance.

However, we made no assumption that later editions of the same work, even by the same publisher, are textually identical. For example, the trade and mass-market paperback editions of ‘Madouc’ published by Grafton are different.

From the guidance notes to Textual Integrity workers:

In assessing the manuscript material, it’s useful to know how the “typical” Jack story or novel was put together. Let’s take a hypothetical “mid-period” novel (actually it’s ‘Emphyrio’). Jack would write an extensive outline—up to a third the length of the final work—initially in his own handwriting, then typed up by Norma. The novel itself would then begin, again as a handwritten draft, in Jack’s characteristic hand, with many crossings-out, interpolations, and doodlings; there might be a second handwritten draft. Norma would type this up and Jack would revise it again, usually extensively. There might be a second typewritten draft which Jack would again revise, this time probably rather less, and Norma would type it up again. The final version of the typescript might or might not contain further handwritten emendations. A typescript with more than the occasional handwritten change is probably not the final one. 

Norma Vance, on the editorial process:

In the case of both ‘Throy’ (published by Tor) and ‘Madouc’ (by Berkley), I think Tim Underwood, who published his limited editions first, shared his edited disks with the two publishers just for letting him do the limited editions. But just yesterday I looked into Tor’s ‘Ports of Cal’ to verify the name of the world from which the Flauts came and discovered a very annoying error. It seems that the Tor editor corrected the sentence “You will hear few complaints.” to “You will hear a few complaints.” I looked into the Underwood version which read as it should. It frightens me to think what else I might find, but thank goodness for the VIE!

When I finished typing a story on my computer, I then printed one copy. Next, I read the printout and made corrections and if John [Vance] had time he would read the MS and usually would find more things to correct. We made a list of the errors with page numbers; then I went back to my computer and corrected all the pages on my disks, after which I printed four copies. I sent one to Tim, one to our publisher and one to our friend Bill Schulz and kept one for myself. After a time Bill Schulz would return his copy with penciled-in suggestions which we could either reject or accept as improvement which Jack often did. Back to my computer with the improvements. Next I would tell Tim what changes we had from Bill so that he could incorporate them into his copy-edited MS. Eventually a copy of Tim’s MS arrived which we would check for accuracy. In the meantime Jack could still be making changes due to queries from Tor (or Berkley, as the case might be), or possibly due to his own dissatisfaction with some passage or other. And then there were more changes for me to make. I reprinted only the pages with corrections or rewritten passages to the number of MSS in existence. I removed and substituted the new pages in my copy and in copies I had made for England and France (who were always anxious to get started on their own publication), and mailed the substitutions to Tor (or Berkley) and to Tim Underwood.

Most of Jack’s most popular books have had multiple reprintings, sometimes by the original publisher and a new editor, or sometimes by new publishers and new editors. It’s no wonder to me that differences are being found in the various versions. Every different editor probably had a field day finding new things to adjust to his or her preference. Jack was losing more and more eyesight and could not be bothered with reading galleys. And as hard as I have tried to be careful, errors do happen with a work so complicated. I’ll take my share of the blame, but careless work by editors and publishers can take most of it. The VIE, I hasten to add, is beyond reproach.

Jack never or almost never made significant changes in proof, at least in the early and mid periods. This is, to some extent, why some of the texts are in the state they’re in, and one of the reasons why the VIE is necessary. There are unlikely, therefore, to be major author’s changes after the final version of the typescript, either second thoughts or restoration of text changed in the copy-editing process.

We were able to learn lessons from the way the manuscript materials have been treated that we could then transfer to the TI work on texts where no prepublication evidence is available. A trivial example is Jack’s consistent use of ‘blond’ for a male subject and ‘blonde’ for a female—proofreaders have inevitably preferred ‘blonde’.

A classic textual criticism problem arises with ‘The Languages of Pao’. The original version, published in Satellite Science Fiction in 1957, is considerably shorter than the subsequent book versions (Avalon, 1958, and Ace, 1966—very similar texts, with the Avalon edition containing a couple of paragraphs not in the Ace, and preferred). Did Jack expand the magazine version to book length, making cuts as well as he did so, or cut the book for the magazine appearance? Were both texts edited versions of a longer original? Each contains passages not found in the other. As a team we are not unanimously agreed on the history of this text, but we have chosen, with the author’s agreement, to produce a composite of the two, longer than either; this gets the reader the maximum Vance, but risks “reconstructing” an original that never was.

The only public collection of Vance manuscripts, typescripts, proofs and correspondence is at the Special Collections Department of the Mugar Memorial Library at Boston University. The Library has been a very great help to the VIE, especially in their willingness to support a very disparate and distributed community of workers on the project. The Vances also have manuscripts, typescripts, tearsheets, correspondence and other evidence. However, there are still rather too many texts where there is no evidence available other than the published texts. These include most of the pulp magazine stories, but also novels including ‘Trullion’, ‘Marune’, ‘Clarges’, ‘Big Planet’, ‘The Star King’, ‘Gold and Iron’, ‘Mazirian the Magician’, and ‘The Dragon Masters’. We were relatively unsuccessful in locating and getting access to manuscripts known to be in private hands.

One irritating characteristic of the evidence is that when there are alternate versions consisting of a magazine appearance and a book publication, we seldom have manuscript evidence. This is because the final typescript and its carbon would be sent one to the book publisher and the other to the magazine publisher, leaving no typescript in the Vances’ hands, and therefore much less likely to be preserved. However these are exactly the cases where we would have found manuscript evidence most valuable. For example, the texts of ‘The Star King’ published in Galaxy in 1963–4 and by Berkley in 1964 are surprisingly divergent, given the proximity of the publication dates. In such cases it is necessary to construct a hypothesis relating the different texts to each other in order to inform the textual reconstruction process. The publication dates are close enough together that it would not be safe to assume that the later publication represented a revision of the earlier, especially as the lead times for the two publishers might have been quite different.

In this case, there is little evidence available external to the texts. The Berkley publication references the earlier serialisation and indicates that the texts are different; there is some circumstantial and stylistic evidence that ‘The Star King’ was written some years before its first publication; interestingly, Attel Malagate is ‘Grendel the Monster’ in the serialisation and in the only extant typescript for ‘The Killing Machine’, published by Berkley later in 1964 with the name changed to Malagate but otherwise relatively faithful to the author’s typescript.

Perhaps what happened was something like this: Jack completed his original draft of ‘The Star King’ in around 1960, and the novel, like a number of others, spent several years in search of a publisher. By 1963 ‘The Dragon Masters’ had appeared to considerable acclaim and the novel was sold, possibly along with its sequel, ‘The Killing Machine’. It is quite possible that the book publisher delayed publication to allow time for the magazine serialisation to appear; we know this to have happened in other cases. That might explain the gap of only seven months between the publication of the two books, and why ‘The Killing Machine’, unlike its predecessor and successor, was never serialised.

It appears that the Galaxy version was cut to fit the space available, and it is reasonable to suppose that Jack did the cutting rather than that the work was editorial; it was normal to ask authors to cut to length, which they disliked, on the grounds that they were normally paid by the word, and the effect was therefore to commit them to extra work in order to receive less money. Here the hypothesis becomes totally unsupported by direct evidence, but one imagines Jack returning to a text written several years earlier, probably just after he had finished ‘The Killing Machine’, and feeling an urge to revise which he could not satisfy for the magazine version, but which might be a proposition quite agreeable to the book publisher.

Both texts—other editorial interventions apart—are therefore likely to be Jack’s work. Not all TI workers felt entirely convinced of that. In the book version, Chapter 10 contains a number of melodramatic additions in that section of narrative where Gersen encounters Dasce at Thumbnail Gulch—‘Gersen looked an implacable god, a god of vengeance’ and the like—which a number of us found difficult to accept as authentic Vance, and similar doubts were expressed by a number of VIE volunteers as to the authenticity of the changes to the treatment of Pallis Atwrode towards the end of the book. In the book version, Gersen arranges for her to be paid off from Malagate’s funds, which makes her ‘very cheerful’, while in the magazine version a tentative liaison between her and Gersen is suggested in the following paragraph (‘A year later Kirth Gersen returned alone to Teehalt’s planet …’) to have been, unsurprisingly, not a success. Readers of the VIE version will realise that we have not always accepted all the implications for the text even where we have agreed on the likelihood of a particular textual stemma; there has always had to be room for judgement, and the reader of the VIE is entitled to come to a view as to whether we have exercised that discretion wisely.

Some of the most useful evidence is the setting copy typescripts. These are the typescripts which were used in the production (typesetting) of the published edition, and they are especially useful because they contain copy-editor’s and proofreader’s changes marked up so that it is possible to see what’s going on and who changed what. These may contain markings by several different hands and in different colours, pen and pencil, and are therefore preferably consulted in the original rather than in photocopies. Few proofs exist but where they do (‘Emphyrio’) they add little to the evidence base. We restored these texts from setting copies: ‘Emphyrio’, ‘The Magnificent Showboats…’, ‘Nopalgarth’, ‘Tschai’, ‘The Last Castle’, ‘The Fox Valley Murders’, ‘The Pleasant Grove Murders’, ‘Wyst’, ‘Cugel the Clever’, ‘The Deadly Isles’, ‘The Houses of Iszm’, ‘The Palace of Love’, ‘Suldrun’s Garden’, ‘Night Lamp’.

Setting copies provide clear evidence of the level and type of editorial intervention in a number of texts. For example, the level of intervention in ‘The Pleasant Grove Murders’ is notably less than in ‘The Fox Valley Murders’. The Mugar collection includes a letter from Robert Ockene, Jack’s editor at Bobbs-Merrill, in which he concedes that his proofreaders went rather overboard on The Fox Valley Murders and promises to rein them in with the next book.

An unusual feature of the setting copy is that it contains handwritten dialogue between Norma and Jack that occurred during the revision process.

Text: Guy Benjamin was a civil engineer employed by the State Department of Highways.

Norma:   Works for private firm later.

Jack:       Changed jobs. This is reason for Benjamin’s locating in so out-of-the-way place as Pleasant Grove to begin with.

Text:       [Starr Shortridge] “Father, dear dear father! I’m nineteen years old.”

Norma:   Make her 14 to start then. She’s only 17.

Jack:       No.

Text:       [Rex Kelly says, of Caspar Hubman] “His wife is one of these arty types—weaves rugs and makes pots.They both drink martinis.”

Norma:   How would he know?

Jack:       [Strike] ‘They both drink martinis’.

Text:       [Joe, of Howard Griselda] “He’d love to find me sprawled out on a divan eating hashish.”

Norma:   opium?

Robert Ockene: Jack—This doesn’t bother me. Changing it would probably hurt it. Bob

Another advantage of setting copies is that they provide good evidence that the typescript in question was the author’s final draft. In a number of cases we have late typescripts or carbons, not setting copy, where it’s reasonable to suppose that its counterpart top-copy or carbon was the basis for the published edition. Typically these are fairly close textually to the published editions and have a small number of holograph amendments—usually there wouldn’t be a final, completely clean typescript, but any typescript with more than a few holograph changes per page would probably not be final. Texts restored from such typescript evidence included: ‘The Killing Machine’, ‘Durdane’, ‘Araminta Station’, ‘The Face’, ‘The Book of Dreams’, ‘Cugel: The Skybreak Spatterlight’, ‘The Green Pearl’, ‘Throy’, and ‘The Narrow Land’. In these cases close comparison with published texts might be necessary to glean evidence of late textual changes. For example, as previously noted, in ‘The Killing Machine’ typescript Attel Malagate is still Grendel the Monster, as he was in the magazine version of The Star King, but the change was clearly made very late in the life of the text, later than the version we have.

There is also a typescript of ‘Space Opera’ that looks like setting copy, and was clearly used during the editorial process, but which does not contain all the editorial emendations. This is what my notes made at the time that I first examined the typescript, referring to all the textual evidence, say:

(a) Synopsis—typescript with heavy holograph amendments

(b) Synopsis—typescript with minor holograph amendments, descendant

of (a), in agent’s folder (dated 3/10/64)

(c) Typescript with substantial holograph corrections

(d) Typescript, described as setting copy on the folder and as ‘Typescript with printer’s marks’ in the inventory.

However, the ‘printer’s marks’ (in red) are restricted to pages 1, 6, 36 & question marks on p23, 34, as well as some marks periodically through the text which might simply mark progress as the text was read; there are some pencil marks elsewhere in the text, typically reversing the order of single quotes and commas (also changing ‘which’ to ‘who’ on p39; also changes made by Jack in holograph (blue ink)).

Jack’s changes seem to have made it to the published text. This is clearly not setting copy as we normally understand it but does appear to be final draft, having the typical Vancian characteristics of clean(ish) typescript with modest holograph amendments.

The published text has clearly been copy-edited from this one; it contains typical editorial changes of punctuation, though the level of intervention is relatively minor.

Generally speaking, where we have final or almost-final typescripts, earlier versions have not provided much additional textual evidence, though in examples such as ‘The Man in the Cage’ where a partial earlier draft is all the manuscript evidence we have, it is useful though needs to be treated with care. However, on a significant number of occasions we have been able to resolve issues in the final text by reference to earlier versions, typically where they involve transcription errors. For example, in ‘Showboat World’ (Pyramid, 1975) the following passage appears:

The slave-dealer started to expostulate, but the magistrate said: “That is reasonable enough. Who would risk the consequences of fraud for a paltry few groats of iron?” (p. 164)

However, the typescript has ‘This reasonable enough’. Clearly it needed correction for publication, and not even the most conservative VIE editor would propose retaining the manuscript text. However, we were able to check the precursor typescript, where the reading is ‘This is reasonable enough’. The final typescript version is a transcription error from the earlier, and we are able to restore the correct reading.

Some apparently useful typescripts turned out not to be what we hoped. We were thrilled to discover a typescript of ‘Abercrombie Station’ in Oakland—it would be almost the earliest of any text that we had—but it turned out on examination to be retyped by Norma from the Ace edition of ‘Monsters in Orbit’, presumably for ‘The Best of Jack Vance’, where the few holograph changes are implemented. It therefore has value in determining the status of ‘The Best of Jack Vance’ text, but we were hoping for something else.

‘The Blue World’ is an interesting and unique case. This is the only Vance novel that is an expansion of a short story, a practice otherwise common in genre sf publishing of the time. The typescript in the Mugar collection, entitled ‘King Kragen’, is the full text of ‘The Blue World’, a part-carbon, part-top-copy typescript with holograph amendments varying from the minor to the very substantial in different parts of the text. It is typical of a pre-final version, i.e. there would normally be a clean copy of this version typed up for submission to the publisher. Jack would normally have made changes to that final version also, and a check of a number of passages against the published first edition shows some alterations that are typical copy-editor’s changes and some that are clearly Jack’s. Fortunately the extent of later editorial intervention is minor, and typically relating to accidentals.

One thing it demonstrates clearly is that the novel is indeed an expansion of the short story. For example, the character Semon Voidenvo (‘The Kragen’, Fantastic Stories 1964) becomes Semon Voiden in the early part of the ‘King Kragen’ typescript, later Semon Voiderweg, all changed in holograph to Semm Voiderveg, which is how the character is named in ‘The Blue World’.

Was this text any use as evidence for ‘The Kragen’? We needed to show that it bears a relation to the original Kragen manuscript and not, for example, that it is based on a retyping by Norma from Fantastic tear sheets.

In fact, some of the pages are indeed ‘The Kragen’ typescript; the starting point for this text was clearly the carbon copy of the original ‘Kragen’ typescript. That it was the original and not a re-type is fairly clear from examination, but is proved by (for example) page 49 (of the new text) where there is a holograph alteration by Jack which has made it into the Fantastic version. The original page numbers are also often visible, though crossed out and replaced by new ones.

What we have therefore are pages from the original ‘Kragen’ manuscript, in carbon, generally but not always with holograph amendments, interspersed with retyped new pages where the changes made are very substantial.

This is how the typescript is made up:

So for ‘The Kragen’ we have 46-and-a-bit typescript pages, or nearly half. The level of editorial intervention in the magazine version is relatively low. In the VIE, we have considered ‘The Kragen’ and ‘The Blue World’ as separate texts.

‘The Dragon Masters’ is an interesting example of a text where we have alternate published texts but no manuscript evidence. The two principal texts were published in Galaxy in 1962 and by Ace Books in 1963. It is reasonable to suppose that both texts derive from the same manuscript, though it is also of course possible that there were authorial revisions between versions submitted to the different publishers.

Each publisher edited the text in a different way—the Galaxy version is heavily edited for punctuation and sentence and paragraph breaks, whereas Ace retained more of the Vancian punctuation but were more willing to make arbitrary word changes.

The fact that there may be two different readings for the text with no manuscript evidence to choose between them does not mean that we have to toss a coin to arrive at the VIE version. It does mean, though, that we have to make reasoned judgement calls based on the evidence. Take the two different versions of the passage below:

Kergan Banbeck’s face became red. He half-turned before his men but, restraining his anger, spoke slowly and with careful clarity. “I have something you want. You have something I want. Let us trade.” (Galaxy)

Kergan Banbeck’s face became red; he half-turned before his men, but restraining his anger, spoke slowly and with careful clarity, “I have something you want. You have something I want. Let us trade.” (Ace)

Looking at the first difference, it is hard to imagine an editor changing an entirely conventional full stop to a semicolon—this would be very unusual, and the semicolon is typically Vancian. The second difference caused us some debate, but we concluded that the most likely explanation for the difference was that neither reflected the manuscript reading, where there was probably no comma in either location, but both editors had thought one necessary to balance the one after ‘anger’. In the case of the third difference, we have no clear hypothesis to inform us. The comma before the quotation marks is not typically Vancian (though a colon would be) and could be an error, but there is no evidence for that. We have had to make a judgement call as to which is the appropriate reading. The VIE text therefore becomes:

Kergan Banbeck’s face became red; he half-turned before his men but restraining his anger, spoke slowly and with careful clarity. “I have something you want. You have something I want. Let us trade.” (VIE)

What we were trying to do in this case was to establish what was likely to be the more authentic reading, rather than the ‘better’ one. It was not always possible, however, to base such decisions on clear evidence or even on credible hypotheses.

The titles for the texts as published in the VIE are the author’s own preferred titles, and in most cases were the working titles for the stories or were the titles under which they were originally submitted to publishers. A few special cases are worth noting. ‘Cugel: The Skybreak Spatterlight’ is a replacement for ‘Cugel’s Saga’, a title supplied by the publisher which the author never liked. It is, however, a new title constructed for the VIE. ‘The Telephone Was Ringing in the Dark’ never had a title; its VIE title is taken from the opening text of the typescript, and is approved by the author.

Volume 3 of the VIE, ‘Gadget Stories’, has a title which Jack has used to describe many of the stories in it:

‘Dream Castles’, ‘Sabotage on Sulfur Planet’, ‘Potters of Firsk’ (with its smarmy ending) came while I was trying to produce gadget stories. (‘Foreword and Cold Facts’, from Lost Moons, Underwood-Miller 1982)

‘The Wannek’, replacing ‘Servants of the Wankh’, is the author’s change eliminating the unfortunate (in the UK) connotations of the original. This also involved 189 individual changes in the text itself.

We do not claim the VIE as a landmark of textual restoration. We have found purity and indeed consistency of approach impossible to achieve. It would be hopeless to argue that we have never been swayed by our own preferences. Faced with choices and with no algorithmic means of determination, we have argued about which reading might be ‘better’ and would probably have done our readers a disservice had we not done so. Our decisions were often not unanimous and they were certainly not democratic.

Outsiders would have found the passion and effort expended arguing over the position of a comma, let alone a rearranged sentence or an altered phrase, bizarre. However, most of the variations between published texts and between texts and manuscripts are like this, and it is by the cumulation of such small-scale restorations that the Textual Integrity work has its effect.

In only a few cases have we differed significantly over methodology, where, curiously, we have tended to argue rather less vigorously. However, the underpinning principles and methods of the VIE have now been tested through several years of painstaking implementation, and though they may have evolved since we began with such naïvety five years ago, they remain consistent with our original principles, and have served us well.