University of Virginia Library

Bibliography and the Novelistic Fallacy
by
Bruce Harkness

It is a truth universally acknowledged, that a critic intent upon analysis and interpretation, must be in want of a good text.[1] It is also universally acknowledged that we live in an age of criticism, indeed of "new criticism"—which means that we as critics are dedicated to a very close reading of the text. Sometimes, it is true, that critical principle leads to abuses. The symbol-hunting, the ambiguity-spinning become wonders to behold. As one objector has put it, "nose to nose, the critic confronts writer and, astonished, discovers himself."[2] Nonetheless, the principle of close reading is held central by us all. Immediately that one contemplates novel criticism, however, an oddity appears: the last thing we find in a discussion by a new critic is some analysis of the actual text.

The modern critic is apt to be entirely indifferent to the textual problems of a novel. He is all too prone to examine rigorously a faulty text. As Gordon Ray and others have pointed out, even the Great Cham of British Criticism errs in this respect. F. R. Leavis defends the early Henry James in The Great Tradition: "Let me insist, then, at once, . . . that his [James's] 'first attempt at a novel,' Roderick Hudson (1874), in spite of its reputation, is a very distinguished book that deserves permanent currency—much more so than many novels passing as classics." Professor Ray adds that "Mr. Leavis goes on to quote three long paragraphs to illustrate the novel's 'sustained maturity of theme and treatment. . . .' These remarks are amply warranted by the passage that Mr. Leavis cites. But unhappily he has quoted, not the text of the first edition of 1877 [while carefully dating it from the time of composition to make it appear all the more precocious], which is simple enough, but that of the New york edition of 1907, revised in James's

intricate later manner. This leaves him," concludes Leavis's critic, "in the position of having proved at length what nobody would think of denying, that James's writing at the age of sixty-four has all the characteristics of maturity."[3]

Unhappily, few of us can afford to laugh at the poor new critic. We all know the truth that we must have a good text, but most of us do not act upon it. A commonplace? Yes, and unfortunately, I have only that commonplace to urge; but I claim good company. Jane Austen, with whom I started, recognized that Pride and Prejudice had no profoundly new meaning. She ironically developed upon commonplaces: don't act on first impressions; don't interfere in your best friend's love affair; don't ignore your younger daughters. My point is that, ironically, everyone ignores the bibliographical study of the novel. People who would consider it terribly bad form to slight the textual study of a play or poem—or even doggerel—commit bibliographical nonsense when handed a novel. It seems that the novel just doesn't count. A key error in many studies of the novel is simply this, that the novel is unconsciously considered a different order of thing from poetry—a poem's text must be approached seriously. I shall illustrate by mentioning the sins of editors, reprinters, publishers, scholars, and, alas, bibliographers. Then, after discussing a few of the many reasons for this bibliographical heresy, I shall turn to my main illustration of the need for textual bibliography, The Great Gatsby.

II

A list of representative errors, by no means exhaustive, by sound men whom I admire in all other respects will make clear how faulty the texts of novels are, and how little we care. A good editor has put The Nigger of the "Narcissus" in The Portable Conrad, an excellent volume the introductions to which contain some of the best Conrad criticism. But what, one may wonder, is the copy-text for The Nigger? A search through the book discloses two references, the less vague of which reads as follows: "It is from the editions published and copyrighted by the latter [Doubleday and Company] that the texts reproduced in this volume have been drawn" (p. 758).

After a spot of searching the reader can discover for himself that the copy-text for The Nigger of the "Narcissus" is not the collected English edition, which as is well known was Conrad's major concern. The copy-text was an early American publication, which Conrad

habitually did not supervize. The new critic immediately asks, does it make any difference?

The collected English edition was, as one might suspect with an author who was constantly revising, changed in many ways. This final version cuts down Conrad's intrusive "philosophizing," and corrects Donkin's cockney accent, among other shifts.[4] I yield to no man in my admiration for Conrad, but if he has a fault, it lies in that adjectival "philosophy" which is admired by some, charitably overlooked by others, and condemned by a few as pipe-sucking old seadog-talk. Surely the following, from the early part of Chapter Four, is inappropriate in the mouth of the sailor-narrator: "Through the perfect wisdom of its grace [the sea's] they [seamen] are not permitted to meditate at ease upon the complicated and acrid savour of existence, lest they should remember and, perchance, regret the reward of a cup of inspiring bitterness, tasted so often, and so often withdrawn before their stiffening but reluctant lips. They must without pause justify their lief. . . ." Most of this passage, and much similar sententiousness, were cut by Conrad from the collected English text; but they all stand in The Portable Conrad.

As for the class of books known loosely as "reprints," I suppose that no one expects a good text for twenty-five or thirty-five cents. These books I am not concerned with, but the more serious paperbacks, obviously intended for use in colleges, are sometimes faulty. For example, Rinehart Editions' copy of Pride and Prejudice reprints Chapman's excellent text—but suppresses the indication of three volume construction by numbering the chapters serially throughout.[5] Though three volumes are mentioned in the introduction, this misprinting of such a tightly constructed novel can only be regretted, for the effect on the college reader must be odd.

What of the publisher of more expensive novels? It can easily be seen that errors are not limited to the paperback field. Consider, for example, the one-volume Scribner edition of James's The Wings of the Dove, dated 1945 or 1946. Here is no scrimping for paperback costs, but the book is not what one would think. It is not a reprint of the famous New York edition; it is another, unacknowledged impression of the 1902 first American edition, dressed up with a new-set New York preface—an odd procedure the reason for which is not apparent. The publisher nowhere tells the reader that this is like some wines—an old text with a new preface. Yet one line of print would have made the matter clear. It is only by his own efforts of collation of the preface and the text itself that the reader knows where he is.[6]

To turn to the errors of scholarship, take F. O. Matthiessen's lengthy appreciation of Melville's phrase "soiled fish of the sea" in White Jacket. Melville's narrator says of himself, after he had fallen into the sea, "I wondered whether I was yet dead or still dying. But of a sudden some fashionless form brushed my side—some inert, soiled fish of the sea; the thrill of being alive again tingled. . . ." This section Matthiessen acclaims as being imagery of the "sort that was to become peculiarly Melville's . . . hardly anyone but Melville could have created the shudder that results from calling this frightening vagueness some 'soiled fish of the sea'!" Then follows a discussion of the metaphysical conceit and its moral and psychological implications.

As has been pointed out, the genius in this shuddering case of imagery is not Melville, who wrote coiled fish, not soiled fish. "Coiled fish" stands in the first editions of White Jacket, and to an unknown Constable printer should go the laurels for soiling the page with a typographical error.[7]

Matthiessen's error does not concern me now, but it does concern me that the scholar who first caught the mistake has a strange but perhaps understandable attitude toward textual matters. Recognizing that such an error "in the proper context" might have promulgated a "false conception," the scholar feels that the slip does not actually matter in Melville's case. Furthermore, he feels that Matthiessen's position is essentially sound — he was merely the victim of "an unlucky error." While sympathizing with common sense and professional etiquette, one may still wonder, however, how many such slips in illustration are allowable. Could the critic, if challenged, produce as many

sound illustrations as one would like? Does not Matthiessen, in his categorizing of conceits, virtually admit that this particular kind is rare in White Jacket?

When we look at the texts of novels from the other way, how many good editions of novelists do we have? How do they compare with the poets? We know a good bit about the bibliographies of Scott, Trollope, Meredith, but those of Dickens, Thackeray, Conrad, Hawthorne, and many more are completely out of date.[8] How many collected editions can be put on the same shelf with Chapman's 1923 Jane Austen? "We have virtually no edited texts of Victorian novelists," says Mrs. Tillotson in the introduction of Novels of the Eighteen-Forties (1954). How slowly we move, if at all.

Take Hardy for example. In 1946 Carl Weber said that "many scholars have apparently made no attempt to gain access to Hardy's definitive texts." In March, 1957, a scholar can complain that "As late as November 1956, sixty full years after the publication of the book, the only edition of Jude printed in the United States took no account of either of the two revisions which Hardy gave the novel. . . . The New Harper's Modern Classics edition . . . [however] is almost identical with that of the definitive 1912 'Wessex Edition.'"[9] One is hardly surprised that Professor Weber is the editor.

Sixty years is a long time, but American literature is no better off. Moby-Dick, our greatest novel, presents no problem of copy-text. Yet more than 100 years went by after publication before we had what a recent scholar called the "first serious reprint," by Hendricks House. Before that, the careful reader did not even know, for example, the punctuation of the famous "Know ye, now, Bulkington?" passage. But how good is this reprint? The same scholar — not the editor — asks us to consider it a definitive edition. His reasons? It contains only 108 compositor's errors and twenty silent emendations.[10] Would anyone make such a claim for a volume of poems?

So much for editors, publishers, scholars. The sins of the bibliographer are mainly those of omission. For well-known reasons he tends to slight 19th- and 20th-century books in general, and in consequence most novels.[11]

The critic therefore needs convincing that novels should be approached bibliographically. The critic appreciates the sullied-solid-sallied argument about Shakespeare, but not that of 108 typos for Moby-Dick. A false word in a sonnet may change a fifth of its meaning; the punctuation at the end of the "Ode on a Grecian Urn" can be considered crucial to the meaning of the whole poem; but who, the critic argues from bulk, can stand the prospect of collating 700 pages of Dickens to find a few dozen misplaced commas? Like the "soiled fish" reading of White Jacket, a few mistakes seriously damage neither novel nor criticism. They are swallowed up in the vast bulk of the novel, which by and large (and excepting a few well-known oddities such as Tender is the Night in which case one must be sure which text one is attacking) is decently printed and generally trustworthy. The critic feels that a mistake here or there in the text is immaterial. "It doesn't really alter my interpretation," is the standard phrase.

This attitude has long since been defeated by bibliographers for all genres except the novel. One wonders indeed, if the critic would be willing to make his plea more logical. Could not the attitude be extended to some formula for trustworthiness versus error? It ought not to be difficult to arrive at a proportion expressing the number of errors per page, exceeding which a novel could be condemned as poorly printed.

Amid bad reasoning, there is some truth to the critic's defence against bibliography. The argument can be shifted from the ground of a novel's size and a reader's energy to the aesthetic nature of the novel. The critic is certainly right in maintaining that novels are more loosely constructed, even the best of them, than poems or short stories. The effects of a novel are built through countless small touches, and the loss of one or two — whether by error in text or inattention in reading — is immaterial. Putting aside the counter claim that this truth is damaging to the critical and crucial premise of close reading, surely all is a matter of degree. And what is more, the theory applies mainly to character portrayal. If we fail to recognize Collins as a fawning ass on one page, we will certainly see him aright on another.

That much must be granted the critic. In other concerns, however, the novel may not be repetitive. To give just one illustration: F. Scott Fitzgerald's Last Tycoon as published in unfinished form contains a

boy whom the reader should compare to the "villain" of the piece, Brady (or Bradogue as he was called in an earlier draft). In Fitzgerald's directions to himself left in his MSS, he says "Dan [the boy] bears, in some form of speech, a faint resemblance to Bradogue. This must be subtly done and not look too much like a parable or moral lesson, still the impression must be conveyed, but be careful to convey it once and not rub it in. If the reader misses it, let it go — don't repeat."[12]

My last and painful reason why virtually no one is concerned with the texts of novels is this: most bibliographers are also university teachers and many of them suffer from schizophrenia. I do not refer to that familiar disease which makes us scholars by day and diaper washers by night, but that split in the man between Graduate Seminar number 520 in Bibliography and Freshman "Intro. to Fic.," 109. How many of us make bibliographical truths part of our daily lives or attempt to inspire our graduate students so to do? In this respect many bibliographers are like socialists and Christians: walking arguments for the weakness of the cause.

Let me give one or two illustrations from experience. Not very long ago I sat in a staff meeting while we worried over a sentence of Conrad's introduction to Victory in the Modern Library edition. The sentence contained the odd phrase "adaptable cloth," used about mankind. It made no sense until it was finally pointed out that "adap-table" was divided at the end of the line in both American collected edition and reprint — a domestically minded compositor was talking about a table cloth, while Conrad was saying that Man is "wonderfully adaptable both by his power of endurance and in his capacity for detachment." And our silly discussion had gone on despite long teaching, and one's natural suspicion of the cheaper reprints that perforce must be used in college classes.

More seriously, consider Dickens' Great Expectations, taught to freshmen at many universities, by staffs composed of men nearly all of whom have been required to "take" bibliography. Yet how many of these teachers have turned to the facts of serial publication to explain the figure of Orlick, extremely puzzling by critical standards alone? One immediately sees that Orlick's attack on Mrs. Joe, which ultimately causes her death, is used by Dickens to pep up a three instalment sequence the main purpose of which is simply to let Pip age. This sequence would have been too dull, too insistent on domestic scenes round the hearth while Pip gradually withdraws from Joe, were it not

for the Orlick subplot.[13] The novel apparently had to have thirty-six weekly units, and Dickens therefore could not simply skip this period of Pip's life. The figure of Orlick may not be critically acceptable, but he is at least understandable when one views him in the light of publishing history.

I am also indicting myself for not understanding this point; for it was not many months ago that I looked up the weekly issues of All the Year Round and now have far more detail than, as the saying goes, "the short space of this article will permit the discussion of." I was derelict in my duty partly because life is short and bibliography is long, but also partly because I unconsciously resented the editor of my paperback Great Expectations whose job I was having to do.

For I am more familiar with the schizophrenia than most people, though mine takes a different form. With critics I am apt to claim to be a bibliographer; among bibliographers, I proclaim myself a critic.

The critic, one must recognize, can argue on aesthetic grounds against working on the texts of novels. He can produce the tu quoque argument. And he can say that the bibliographer neglects what he is working on. Of 244 articles on textual bibliography in the Studies in

Bibliography list for 1954, only three were related tonovels.[14] "What has the bibliographer been doing?" asks the new critic.

It may be that under the aspect of eternity George Sandys' Ovid is more important than Conrad's Nostromo or Melville's Moby-Dick, but it would be hard to convince the novel critic of that.

III

For these reasons I have chosen F. Scott Fitzgerald's The Great Gatsby as my main illustration. It brings out nearly all my points: inconsistent editing, an unknown or unidentified text, a publisher who is good but vague, important errors in an important book, schizophrenia in the bibliographer-teacher. Not only is Gatsby a fine novel, but it is taught so often because it contains many of the basic themes of American literature: West versus East; the search for value; the American dream; crime and society; and in young Jim Gatz's "General Resolves," it even reaches back to Ben Franklin and Poor Richard.

How many know, however, what they have been teaching?

The Great Gatsby exists in print in three main versions: the first edition, beginning in April, 1925; a new edition in the volume with The Last Tycoon and certain stories, beginning in 1941; and a subedition of the latter text in the Modern Standard Authors series (Three Novels) together with Tender is the Night and The Last Tycoon, beginning in 1953.[15] Though Gatsby in the Three Novels version is another impression of The Last Tycoon plates I call it a sub-edition

because Gatsby's position is different, coming first in the volume, and there are many changes in the text.[16]

So far as I know, the only available information about the text of Gatsby is buried in the notes to Arthur Mizener's The Far Side of Paradise. Mizener says that Fitzgerald found a misprint in the first edition: the future Nick Carroway speaks of at the end of the novel should be "orgiastic," not "orgastic":[17]

Let us couple these comments with Matthew Bruccoli's interesting article on Fitzgerald's This Side of Paradise. Bruccoli is surprised that thirty-one errors are corrected in later impressions of the novel. He concludes that "the first printing was an inexcusably sloppy job," although Fitgzerald was himself in part responsible for the difficulty. We might infer two things, therefore: far fewer errors in Gatsby's first edition, and a correction of the word "eternal," in The Last Tycoon.

Not so. The correction to "external" is not made in the second impression of the first edition, nor in any impression of Last Tycoon (202.2, TN 38.2). Though there are only four changes from the first to second impression of the first edition, there are no less than twenty-seven changes between First and Last Tycoon. Between First and the 1953 Three Novels, there are more than 125 changes. Of these changes

about fifty are quite meaningless. They change "to-morrow" with a hyphen to "tomorrow," for example. Or they change "Beale Street Blues" to Beale Street Blues. This class of change will not be commented upon nor included in statistics, except to add that the publisher was not at all consistent in making such alterations.[19]

There are, in other words, 75 changes of moment between the first edition and Three Novels—forty-four more than in This Side of Paradise. Many of them are more important. Of the changes the August, 1925, first edition brought, the most important was the substitution of the word "echolalia" for "chatter" in the phrase "the chatter of the garden" (First 60 line 16).[20]

But we must remember that Last Tycoon and Three Novels are both posthumous, and that of the twenty-seven changes from First to LT, twelve are clearly errors, seven are dubious improvements, and only eight are clearly better readings. Of them all, the word "orgiastic," apparently, alone has the author's authority. What's more, the subedition Three Novels retains all but two of these bad changes. An example of an error begun in LT and continued in TN occurs on page 209.6 of First (296.8 of LT and 132.8 of TN). The sentence of Nick's, "It just shows you." is dropped from the text, thereby making the punctuation wrong and leading the reader to confuse speakers.

Between First and Three Novels the changes are of several kinds. In addition to the fifty or so "meaningless" changes, there are (a) fifteen changes of spelling, including six that change the meaning of a word and others that affect dialect; (b) seventeen changes in punctuation, including quotation marks, paragraph indication, and so on; (c) six incorrect omissions of a word or sentence or other details; (d) six proper deletions of a word or more; (e) thirty-one substantive changes —the substitution of a word or the addition of a phrase or sentence. For instance, Gatsby is transferred from the Sixteenth to the Seventh Infantry during the war. (See First 57.17, LT 201.12, TN 37.12.)

For when we turn to Three Novels we must move out of the camp

of strict bibliography into the field of its important ally, publishing history. Fitzgerald's own copy of the first impression, with pencilled notes in the margins, is now located at the Princeton University Library and was used to make the sub-edition.

Of the seventy-five changes between First and TN, thirty-eight are with Fitzgerald's sanction and thirty-seven are without. Most of the thirty-seven changes not recommended by Fitzgerald are "corrections" made by a publisher's staff editor or by Malcom Cowley, who supervized the sub-edition. However, some of this group are clearly errors, many of them having crept into the text by way of the Last Tycoon version. The noteworthy thing is that no reader knows the authority for any of the changes. The sub-edition itself does not even announce that it takes into account Fitzgerald's marginal comments—which, one would have supposed, would have been good business as well as good scholarship.

Furthermore, some of the thirty-eight "sanctioned" changes were only queried by Fitzgerald: no actual rewording was directed. An example is the phrase "lyric again in." Fitzgerald questioned "again" and the editor dropped it. But in five instances of Fitzgerald's questioning a word, no change was made—as, for example, Fitzgerald was unhappy to note that he had used the word "turbulent" twice in the first chapter.[21] There is also one instance in which Fitzgerald expressly asked for a change that was not made. At First, 50.1, Fitzgerald corrected "an amusement park" to "amusement parks," but the later version does not record the request (TN 32.32).

On the whole, one can say this, therefore: that about sixty of the changes from First to Three Novels are proper. That is, they either have the author's authority or are stylistic or grammatical improvements or are immaterial. I speak just now as a devil's advocate — a critic with a jaundiced eye toward bibliography. He would call the deletion of a comma from a short compound sentence "immaterial," though it was not done by the author.[22] I am trying, in other words, to make the text sound as good as I can. Problems arise, however, from the fact that awkward readings sometimes come from purely typographical errors, sometimes from editor's decision, and sometimes from Fitzgerald's own notes. Everyone would accept such changes as "an Adam study" for "an Adam's study," (First 110.26, LT 233.30, TN 69.30);

but by the same token few critics will be pleased by a Fitzgerald marginal correction reading "common knowledge to the turgid sub or suppressed journalism of 1902," instead of "common property of the turgid journalism of 1902" (First 120.11, TN 76.5).

We are left then with fifteen or sixteen errors begun or continued in Three Novels, errors which I trust even the newest of new critics would accept as having some degree of importance. That degree of course varies. The dedication "Once again, to Zelda," is left off, for example. Dialetical words are falsely made standard English, or half-doctored-up, as in this sentence where the word in First was "appendicitus": "You'd of thought she had my appendicitis out" (First 37.4, LT 188.37, TN 24.37). Sentences start without a capital[23] or end without a period[24] or are dropped altogether.[25] Quotation marks appear or disappear[26] and awkward readings come from nowhere. To illustrate that last: on page 149.10 of First Nick says that "the giant eyes of Doctor T. J. Eckleburg kept their vigil, but I perceived, after a moment, that other eyes were regarding us with peculiar intensity from less than twenty feet away." The eyes are Myrtle Wilson's but in Three Novels 95.1 (and LT 259.1) the sentence is confused when "the" is added without any reference and "from" and "with peculiar intensity" are dropped: "the giant eyes of Doctor T. J. Eckleburg kept their vigil, but I perceived, after a moment that the other eyes were regarding us less than twenty feet away" (italics added). Another dubious change is this: a joking slip or drunken mistake by Daisy is corrected — "Biloxi, Tennessee" becomes academically placed in its proper governmental locality.[27] One hardly needs to add that none of these changes have Fitzgerald's sanction.

The biggest errors, critically speaking, are ones that also occur in Last Tycoon. The principle of order in The Great Gatsby is a simple

one: Nick Carroway, the narrator, tells his story wildly out of chronological order, but in the order that he learned it — with one exception.[28] The first half of the book is concerned with the development of the outsiders' illusions about Jay Gatsby — he is "nephew to Von Hindenburg," and so on (TN 47). The second half is a penetration in depth of Gatsby's illusion itself. The shift in the theme of the book is marked by the one major sequence which Nick gives the reader out of the order in which he himself learned it. I refer to the Dan Cody episode from Gatsby's early days.[29]

Now the most important structural unit in the book below the chapter is the intra-chapter break signified by a white space left on the page.[30] In Last Tycoon and Three Novels four of these important indications of structure are suppressed.[31] Oddly enough, it is the one following the Dan Cody story that is the first one missing. The detail that divides the book into its two structural elements is botched.

In the Three Novels version of Gatsby, then, we have a book quite well printed — surprisingly so when we look at the galley proofs. They are filled with changes — with page after page added in long hand, with whole galleys deleted or rearranged. (I would estimate that one-fifth of the book was written after the galley stage.) And we have a book that tries to take into account the author's latest stylistic revisions. Unfortunately, it is also a book that has far too many errors.

Perhaps this is the place to mention the third Scribner edition of Gatsby, the paperback Student's Edition, which uses TN as copy text.

Have matters been improved? Some have, but more errors have been added. There are twelve changes from TN to SE: it makes two distinct improvements, including the replacement of the dedication; but it adds three places in which intra-chapter breaks are suppressed.[32] The other changes are "immaterial" typographical errors such as "turned to be," instead of "turned to me" (SE 71.17 and TN 54.20) and "police," instead of "police" (SE 27.27 and TN 22.19).

I hope it is clear, then, that Three Novels represents the best present text of Gatsby. No doubt it and the Student's Edition will be the ones most used in colleges for some time. It should also be clear that in Three Novels, we have this kind of book:

• 1. A book which nowhere gives the reader the authority for seventy-five changes, all of them posthumously printed.
• 2. One which fails to make use of all of Fitzgerald's corrections.
• 3. One which contains thirty-seven changes which Fitzgerald did not authorize — some of which are of most dubious value.
• 4. A book which contains at least fifteen quite bad readings, one of which is of the highest structural importance.

So, armed with this mixed blessing, or with the worse one of Last Tycoon, or worst of all, with a reprint by another publisher which has none of Fitzgerald's corrections and additions, many students unwittingly face the next semester with their prairie squints. Only a nonexistent, eclectic text, combining the best of the August, 1925 first edition and the Three Novels text of The Great Gatsby would be proper.[33]

Could we not as critics pay more attention to Bibliography, and we as Bibliographers to criticism? Can not we somehow insist that editing actually be done — instead of the practice of putting a fancy introduction on a poor text? Can not we have sound texts reproduced and publisher's history stated by the editor? Can not we know what it is we have in our hands? For it is simply a fallacy that the novel does not count.

Notes