Autobiography of Mark Twain (173 page)

Typed by Hobby between early August 1906 and late January 1907, TS3 comprises fewer than one hundred and fifty pages. Prepared as printer’s copy for six installments in the
North American Review
, it reproduces the first section of “Scraps from My Autobiography. From Chapter IX,” one Florentine Dictation (“Robert Louis Stevenson and Thomas Bailey Aldrich”), and excerpts from the Autobiographical Dictations through 21 May 1906. It consists of four batches, each beginning with page 1. Three of the batches include the text for a single
Review
installment (NAR 2, NAR 3, and NAR 16), and one batch encompasses three installments (NAR 4, NAR 5, and NAR 6). TS3 was typed primarily from TS1, incorporating its revisions, and includes further changes made to accommodate magazine publication.

Above is a diagram of the textual relationships between TS1, TS2, TS3, and TS4 of the 1906–9 Autobiographical Dictations and the twenty-five NAR installments that published excerpts from them (and from one other typescript). The “Early Attempt” preface is not present in either TS2 (now incomplete) or TS4 (complete), but both typescripts include “My Autobiography [Random Extracts from It],” followed by the second three-part preface (“The Latest Attempt,” “The Final [and Right] Plan,” and “Preface. As from the Grave”) and four Florentine
Dictations. TS3 was typed from the revised TS1 for all but one installment, NAR 16, which was typed from the revised TS2. The first and second parts of “Scraps from My Autobiography. From Chapter IX” were published in NAR installments 2 and 17, respectively, typeset directly from a typescript made by Jean Clemens in 1902. See the Appendix “Previous Publication” (pp. 366–67) for a list of the contents of each NAR installment.

Throughout the pages of TS1, TS2, and TS3 there are revisions, corrections, and editorial instructions in two colors of ink, in lead pencil, and in blue, purple, and red pencil. These were made not only by Clemens himself, but also by the following people: his typists, chiefly Hobby; his secretary, Isabel V. Lyon; Albert Bigelow Paine; George Harvey and David Munro, editors for the
North American Review;
Clara Clemens Gabrilowitsch; Paine’s successor as literary executor, Bernard DeVoto; and DeVoto’s assistant, Rosamond Hart Chapman. There are also specific instructions to omit this or that passage, each signed “ABP” by Paine and (though also in Paine’s hand) “CG” for Clara Clemens Gabrilowitsch, who along with Paine had been charged with deciding which of her father’s papers to publish.

Paine felt free to alter the typescripts (and some manuscripts) by writing his changes on them, and even to hand some of them to the printer to set up his 1924 edition. Markings by Paine and DeVoto have been especially problematic. Paine’s handwriting can be difficult to distinguish from Clemens’s, especially in small verbal changes or punctuation. But he clearly renumbered many pages, styled the texts for his publications, annotated and “corrected” them in large and small ways, and occasionally scissored out passages he intended to suppress. Paine’s blue crayon printer’s-copy page numbers and his typesetter’s galley numbers (in plain lead pencil) for his 1924 edition are scattered throughout, and many typescript pages are smudged with printer’s ink and pierced by a spindle hole, both signs that they literally served as setting copy. In preparing copy for
Mark Twain in Eruption
, DeVoto also wrote (in pencil) on the original typescript pages. He inscribed editorial notes to himself, to Rosamond Chapman, and to his typist, Henry Beck; he struck through whole sections of text; and he was so irritated by the typed punctuation that he canceled much of it, penciling through the offending marks so emphatically that it is sometimes difficult to recover the original reading.

This edition of the
Autobiography of Mark Twain
offers the reader an unmodernized, critically constructed text, both of the preliminary manuscripts and dictations and of the final text that Clemens intended his “heirs and assigns” to publish after his death. The editorial construction adheres to his intention as it is manifest in the most authoritative documents available, or can be reliably inferred from them, and aims at presenting the texts exactly as he would have published them, so far as that is possible—that is, as they were when he ceased to
make changes in them. Except for the revisions the author made for magazine publication (discussed below), all of his revisions and corrections are adopted, whether inscribed on a surviving typescript or detected by collation when the revised typescript is missing. Every decision to adopt (or not) is reported in the Textual Commentaries at
MTPO
, which also record every alteration that the editors have made in the source texts.
²

Many of the changes that Clemens made on TS1, TS2, and TS3 were aimed at shortening, taming, or softening the texts selected for publication in the
North American Review
. These changes are
not
accepted into the edited text, on the grounds that Clemens was clear that they were temporary concessions to propriety, not permanent alterations to the text. Other changes were corrections or revisions made for purely literary reasons, and these
are
adopted. Very occasionally the two kinds of changes are intermixed, and in those cases we err on the side of caution, retaining the original uncensored version.

Wherever works like the
Autobiography
were created solely by dictation, they pose all the usual problems of textual transmission plus some additional ones not native to manuscripts in the author’s hand. In a dictated text, unless the author has specified more than words while or before dictating (“use a semicolon not a period after that word”), the punctuation, spelling, emphasis, paragraphing, and many other small details simply
never existed
—only the inflections, gestures, and pauses of the author speaking and the grammatical structure of his sentences. It makes no sense to say that the author “intended” to spell a word in a certain way, since in speaking the word he may not have been thinking of any particular spelling at all. For some kinds of punctuation, like end-of-sentence periods and question marks, the speaker is more constrained by “rules” and probably comes a little closer to actually intending terminal punctuation, whereas the intended placement of commas, semicolons, colons, dashes, and so forth is less clear.

Of course Clemens did not dictate his autobiography in order to produce a text
without
punctuation. Unlike a public speech, where the authorially intended form is actually an oral performance, dictation was intended to result in a written record, in this case a double-spaced typescript that could be reviewed and corrected and ultimately published in the normal way. So to what extent should we accept the spelling, punctuation, and other details as typed by the stenographer who, we must assume, produced such details without specific instructions from the author? Clemens’s review and correction of the typescript made from Hobby’s stenographic notes (TS1) is some assurance that whatever he found wrong or misleading he corrected, but such assurance only goes so far, and his ability or willingness to scrutinize the transcript for such details was limited.

A dictating author and his stenographer collaborate to produce a text, and their respective contributions cannot easily be pried apart after the fact. To take a simple example, nowhere do Hobby’s typescripts record the kind of hesitation, reiteration, and self-correction that must have occurred even in Clemens’s slow and deliberate speech. By mutual though tacit agreement, such things were no doubt omitted, or silently repaired and smoothed over by the stenographer. Fortunately, in the case of Hobby, the collaboration was highly satisfactory to Clemens. Her stenographic notes are presumed lost, but the accuracy of her work can to some extent be judged from the resulting typescripts. They are double spaced (leaving room for revision and correction) and unfailingly neat; the rare typing errors are discreetly erased and corrected by her, with occasional doubtful spellings likewise identified by a lightly penciled question mark. And the number of corrections (as distinct from revisions) inscribed by Clemens is very small. There is some evidence that he trained Hobby to punctuate as he liked. Twenty months after dictation began, a journalist who visited Clemens reported that he “dictates slowly, using the semicolon mark, of which he is particularly fond, as frequently as possible. When the copy is handed to him by the stenographer it is almost always ready for the press, so few are the corrections to be made.”
³
It is easy to find passages in the dictations that exemplify this pattern, in which the typist used semicolons where full stops would, to an uninstructed listener, seem the more natural punctuation.
⁴
There is other evidence that Clemens actually dictated punctuation and other details. Almost forty years after his death, Lyon remembered that “Paine used to say when he was dictating he’d walk slowly up & down and say ‘period’ or ‘paragraph.’” And in 1908, when Hobby left Clemens’s employment and he had to break in a new stenographer, Mary Louise Howden, he took even more explicit control of these details. Howden herself recalled in 1925 that Clemens “put in the punctuation himself. His stenographer was never allowed to add so much as a comma.”
⁵

Whether Clemens literally expressed the punctuation, described his preferences to Hobby, or expected her to learn his style more or less osmotically makes little difference: her punctuation of the typescripts is remarkably close to the use patterns found in Clemens’s holograph manuscripts. We know Hobby did learn from his corrections on the typescripts, eventually spelling “Twichell” and “Susy” correctly, for instance. Inevitably, when expanding her shorthand she sometimes mistyped slightly unusual words: “silver boring” for “silver bearing,” “visited” for “billeted,” and “driveling” for “drizzling.”
⁶
But the author and the stenographer were remarkably well attuned to one another. As a result, Hobby’s spelling, punctuation, paragraphing, and so forth on TS1 (made directly from her notes), as well as her rare corrections of these details on any subsequent typescript, whether marked in her hand or introduced
while typing, command assent. So, if TS1 lacks aparagraph break which was then supplied by Hobby when she created TS2, we adopt her change as a correction of the original typescript. On the other hand, when TS2 shows a change in wording not initiated by Clemens on TS1, the change is
not
accepted unless it is a necessary correction, one that would have been made by the editors whether or not Hobby made it. Most such small verbal differences between typescripts were obviously inadvertent.

Because Clemens never prepared any of the texts for actual typesetting and publication, it is not always possible to follow his instructions. For instance, to the dictation of 20 February 1906 Clemens added a note: “Insert, here my account of the ‘Hornet’ disaster, published in the ‘Century’ about 1898 as being a chapter from my Autobiography”—a reference to “My Debut as a Literary Person.”
⁷
But if that bald instruction were carried out, the resulting text would be both self-contradictory, because it would still include the remark “I will go no further with the subject now,” and deeply puzzling, because it would contain a very long and rather irrelevant digression. Similarly, he noted in the dictation of 12 January 1906, “(Here paste in the proceedings of the Birthday Banquet)”—referring to the thirty-two-page illustrated issue of
Harper’s Weekly
commemorating his seventieth birthday.
⁸
The length and nature of this publication make it impractical to carry out his instruction in a printed volume. Indeed both instructions are more plausibly construed as instructions to himself rather than to his editors. In such cases, Clemens’s intention can be described and—wherever possible—the reader directed to the relevant text. In the case of “My Debut,” the text is already included with the preliminary manuscripts and dictations in this volume. The impracticality of including the
Harper’s Weekly
issue is overcome by directing the reader to a scanned copy at
MTPO
, and the elastic boundaries of the website may be used for other, similar cases, the rationale for which is always explained in the Textual Commentary. In general, such rough edges are an inevitable part of works that were not set into type and published by the author. On the other hand, if informal remarks can be rendered intelligibly in their own right (“I will ask Miss Lyon to see—but I will go on and dictate the dream now”) they are included in the edited text, even though the author would doubtless have removed them had he carried out the revision he planned.
⁹