Some Principles for Scholarly
Editions of Nineteenth-Century American Authors
by
Fredson Bowers
[*]
The first problem that faces any editor of a text from the nineteenth
century, or earlier, is whether to modernize. For nineteenth-century
American books there is only one answer: no gain results from
modernizing, and much is lost that is characteristic of the author. One may
safely say that nothing in the spelling, punctuation, capitalization,
word-division, or paragraphing of nineteenth-century books is likely to
cause a presentday reader any difficulty, whereas an attempt at
modernization is certain to destroy a number of the values of the original.
Every reason exists to preserve these classic texts in as close a form as
possible to the authors' intentions, to the extent that the surviving
documents for each individual work permit of such reconstruction. Indeed,
one may flatly assert that any text that is modernized can never pretend to
be scholarly, no matter at what audience it is aimed.
The second problem is whether to edit the text critically or to content
oneself with a reprint of some single document. Again, an argument cannot
really exist in favor of a mere reprint, no matter how neatly such a
procedure enables an editor to dodge his basic responsibility. It is probably
safe to say that no nineteenth-century text of any length exists that is not in
need of some correction, and possibly even of revisory emendation. Once
an editor tinkers in any way with his original, he has entered upon the
province of critical editing; and he had better go the whole way and be
consistent than dip his big toe in the water and then draw back in alarm lest
he suddenly find himself out of his depth.
The first step in critical editing is the so-called establishment of the
text. The first step in this process is the determination of the
exact forms of the early documents in which the text is preserved and of the
facts about their relationship to one another. That is, the early editions
within an author's lifetime, and within a sufficient time after his death to
give the opportunity for testamentary documents to be produced, must be
collated and the authoritative editions isolated. An authoritative edition is
one set directly from manuscript, or a later edition that contains corrections
or revisions that proceeded from the author. Authority divides itself
between
the words as meaningful units (i.e., the substantives) and the accidentals,
that is, the forms that the words take in respect to spelling, punctuation,
capitalization, and division. In this question the theory of copy-text
proposed by Sir Walter Greg rules supreme. Greg distinguished between the
authority of the substantives and of the forms, or accidentals, assumed by
these substantives. If only the first edition, set from manuscript, has
authority, as being the closest in each of these two respects to the author's
lost manuscript, then both authorities are combined in one edition. On the
other hand, a revised edition may alter the authority of some of the
substantives; but the transmission of the author's accidentals through the
hands, and mind, of still another compositor destroys the authority of these
features of the first edition, set from manuscript. An eclectic text must be
constructed which combines the superior authority of most of the words in
the revised edition with
the superior authority of the forms of words in the first edition.
The determination of authority is not always easy in a later edition.
For example, in Hawthorne's Scarlet Letter three editions
(i.e.,
three different typesettings) were made during his lifetime. In the second
edition, set and printed in 1850 within two months of the first, 226 pages
were completely reset, but 96 pages were printed from the standing type of
the first edition. In the 226 pages of the resetting occur 62 variants from the
first edition, of which three are corrections of first-edition typographical
errors and four are typographical errors in the second edition. Twelve
different words (i.e., variant substantive readings) appear in these pages,
and there are 43 changes in spelling, capitalization, and
word-division.
Of more import, in the 96 pages of standing type, someone ordered
eight variants, of which three are spelling, four are punctuation, and one is
division. Here if anywhere the author's intentions would be visible if he had
ordered these changes in standing type; but an editor will find no clearcut
authority in the changes, and indeed some evidence that at least two of them
go contrary to Hawthorne's observed characteristics. Once these variants
are rejected as non-authorial, therefore, the conclusion must be drawn that
Hawthorne did not supervise the production of the second edition and hence
no revisions can be accepted from the reset type-pages, although a few
corrections will prove useful.
Nor does the third edition, the last in Hawthorne's lifetime, yield any
readings other than a continuation of corruption, and some necessary but
obvious correction. None of the 37 additional alterations in the words seems
to have any chance of being an authorial revision, and most are clearly
errors.
In these circumstances, the editor is forced back to the first edition
as the sole authority. But the question then arises, what is the specific
authority of each page of this first edition, for it is possible for copies to
vary because of changes made during the course of printing. Mechanical
collation on the Hinman Machine of eight copies of The Scarlet
Letter discloses
four differences in readings, but all of these seem to have resulted from
type being loosened during the course of printing so that the progression is
from correctness to error. However, unless an editor had established the
correct readings where these errors exist, he would wrongly have imputed
the errors to the first edition; and it is possible that he might have emended
differently from the original reading. For example, because an exclamation
point dropped out very early in the printing of page 228, no edition before
the Centenary recovered this original authoritative punctuation, for all
editors were content to follow the second-edition comma that the later
compositor inserted when he came to the blank space in his copy.
The collation of multiple copies reveals other possibilities for
variation. In The Scarlet Letter, interestingly enough,
economy
of printing led the printer of the first edition to typeset the last two text
pages in duplicate. Fortunately no differences appear in these two settings,
but the possibility of variation is always present. For example, the new
Preface to the second edition was also set in duplicate, and here one
typesetting has a comma that appears to be authoritative, whereas the other
omits it. An editor who neglected to collate a number of copies might have
reprinted arbitrarily from the wrong typesetting and thus, even though in a
small matter, have departed from Hawthorne's intention.
Even if the first edition were printed from plates, machine collation
is necessary to discover concealed printings within the so-called first
edition, for the possibility exists that plates may be altered between
impressions. For example, the Ticknor and Field cost books list four
printings from plates in 1851 of The House of the Seven
Gables, and one printing in 1852. Since no copy of an 1852 printing
has turned up, it seems clear that one of the unidentified 1851 printings
represents the fifth impression with a title-page date unchanged. No book
collector or librarian has the least idea which printing his precious
first-edition copy represents, but the Hinman Machine discloses the order
by combining the evidence of type batter with the evidence for resetting of
damaged plates as well as various mendings. It is pure luck that these
extensive plate repairs were carried out without producing any changes in
the text to baffle the non-bibliographical editor, and it is clear
that Hawthorne (if he saw any errors) ordered no revisions between these
printings.
But not all changes made in plates from printing to printing are so
respectful of the text. For instance, the third edition of The Scarlet
Letter was printed from plates in 1850 and these plates remained in
use at least as late as 1886. In the course of the various repairs made in this
interval, five different words got altered so that the text of the final
printings from these plates differs from that of the initial printings in this
respect as well as in dozens and dozens of punctuation marks worn off or
quite altered by batter. These changes have no authority, but it is clear that
Hawthorne himself made some alterations in the plates for one of the later
printings of The Marble Faun, revisions of which an editor
must
take account if he knows about them.
When an author's manuscript is preserved, this has paramount
authority, of course. Yet the fallacy is still maintained that since the first
edition was proofread by the author, it must represent his final intentions
and hence should be chosen as copy-text. Practical experience shows the
contrary. When one collates the manuscript of The House of the
Seven
Gables against the first printed edition, one finds an average of ten
to fifteen differences per page between the manuscript and the print, many
of them consistent alterations from the manuscript system of punctuation,
capitalization, spelling, and word-division. It would be ridiculous to argue
that Hawthorne made approximately three to four thousand small changes
in proof, and then wrote the manuscript of The Blithedale
Romance according to the same system as the manuscript of the
Seven Gables, a system that he had rejected in proof.
A close study of the several thousand variants in Seven
Gables demonstrates that almost every one can be attributed to the
printer. That Hawthorne passed them in proof is indisputable, but that they
differ from what he wrote in the manuscript and manifestly preferred is also
indisputable. Thus the editor must choose the manuscript as his major
authority, correcting from the first edition only what are positive errors in
the accidentals of the manuscript.
However, when words differ in the print from the manuscript, as they
do a certain number of times, the question of authority arises. Any
difference in words can arise only by reason of printer's error that
Hawthorne did not catch in proof, or by reason of changes that Hawthorne
himself made in the lost proof-sheets. Each variant, thus, becomes an
editorial responsibility, to be adjudicated on the evidence available. In
The Blithedale Romance Hawthorne can be assigned
twenty-four
of the verbal proof-changes between manuscript and first edition. The
printer is responsible, fairly clearly, for the remaining seven of the
thirty-one differences in wording.
Here we encounter the theory of a critical edition. Obviously, an
editor cannot simply reprint the manuscript, and he must substitute for its
readings any words that he believes Hawthorne changed in proof. Once
more, if one argues why not reprint the first edition and be done with it
— then two questions of evidence are pertinent. First, in reprinting
the
first edition of The Blithedale Romance, one would be
attributing to Hawthorne seven words that are actually printer's errors.
Secondly, if an author's habits of expression go beyond words and into the
forms that these take, together with the punctuation that helps to shape the
relationships of these words, then one is foolish to prefer a printing-house
style to the author's style. This distinction is not theory, but fact.
Hawthorne's punctuation, for example, is much more meaningful in respect
to emphasis and to delicate matters of parenthesis and subordination than
is the printing-house style in which Seven
Gables and The Blithedale Romance appeared. In each
book, the real flavor of Hawthorne, cumulatively developing in several
thousand small distinctions, can be found only in the manuscript.
Sometimes the Greg formula that authors' substantive revisions in
later editions must always be followed, when identified, but that the best
authority for the accidentals remains the edition set directly from
manuscript, produces some complexity, and the result will agree in a
number of details with no preserved document, even though it will
represent the nearest approximation in every respect of the author's final
intentions. An eclectic editor must be prepared for any eventuality. In
The Marble Faun, for example, Hawthorne seems to have
proofread the original English edition rather carelessly, because he made a
number of corrections, and a few revisions, in the first American edition,
which was set up from the English sheets. Fortunately, we have the
manuscript preserved (though lost sight of for many years). Hence an editor
will base his edition on the manuscript copy-text but will substitute, first,
any words from the English first edition that he thinks are authorial
proof changes, and then, in addition, any variant readings from the first
American edition that appear to him, also, to represent the author's
alterations, and then, in addition, the authoritative Hawthorne revisions and
corrections made later in the first-edition plates. This is indeed an eclectic
text, but the unique results will be closer to Hawthorne's characteristics
than any single preserved document, not only in respect to the insertion of
Hawthorne's final intentions within the framework of his original
accidentals but also in respect to the scholarly refusal to reprint errors from
either the first English or the first American edition any more than the
rejected readings of the manuscript.
Granted that an editor has established a critical text that will stand up
under the most searching investigation of scholarship, then what will
scholars want from his apparatus? First, a list of the internal variants in the
first and in any other authoritative edition as revealed by the collation of a
number of copies of each, preferably on the Hinman Machine. Secondly,
a complete list of all editorial changes in the selected copy-text. These
changes comprise corrections and revisions admitted from later editions as
well as the editor's own alterations. For the sake of the record, the editor
should list the earliest edition from which he draws any alteration. Textual
notes should discuss briefly any arguable emendations, or failure to
emend.
The next item is the Historical Collation. This should contain all the
substantive alterations from the established edited text found in a group of
significant later editions. An edition is to be defined as any new typesetting.
Obviously, any edition within the author's lifetime may be significant and
must be collated. Thereafter the editor's discretion may enter. Usually it is
important to select editions that have been influential in the formation of the
text, or that have been commonly used by critics. For instance, the decision
was made by the Centenary editors of Hawthorne to confine the Historical
Collation largely to the Boston collected editions published in the Ticknor
and Field line to Houghton Mifflin, as well as any separate editions
published within the author's lifetime. For The Scarlet Letter,
therefore, the Centenary Edition records the readings from the second
edition of 1850, the third edition of 1850, the Little Classics edition of
1875, the Riverside of 1883, and ends with the Autograph of 1900.
Included always are the first English editions in case any authoritative
changes were made in the copy sent abroad, and usually any modern edition
that has been freshly edited in fact instead of in theory.
To insure accuracy, the sets of plates are taken as representing the
various editions, and the earliest and latest printings from each edition-set
of plates have been collated on the Hinman Machine and their variants
recorded in the Centenary apparatus. All printings from plates within the
author's lifetime have also been collated whenever variants appeared
between the first and last impressions of any set of plates instituted before
his death.
Although this Historical Collation is chiefly a record of the corruption
of the text, it serves as a useful object lesson in the untrustworthy nature
of various commonly esteemed editions. More important, however, this list
insures that all cards are on the table. If any collated edition has authority
not recognized by the editor, the critic will find the record of its variants
and all the evidence on which, throughout, the editor made up his mind
about the details of the text.
When a manuscript is preserved, an important separate list will
contain a record of all the rejected readings and revisions during the process
of inscription. Moreover, the variants in any preserved proof-sheets should
be recorded with the same scrupulousness and for the same critical
purpose.
So far as I know, a problem that no editor has faced concerns the
word-division whenever a compound in the copy-text is divided at the end
of one line and the start of the next. The exact form of all such compounds
must be settled so that the edited text will contain that one that is
characteristic of the author. Since editorial judgment is sometimes involved
in this process, a list of such divided possible compounds should be
provided. Correspondingly, the modern printer will divide a number of
compounds so that a reader will not always know the exact form in the
original. A second section of the compound list should note the copy-text
reading in all such cases.
The amount of collating and checking in such an edition as has been
outlined is very heavy indeed; but only this editorial process scrupulously
carried out will produce editions of American classics that will stand the
test of time and, heaven willing, need never be edited again from the
ground up. When scholars editing American literature will bring to their
task the careful effort that has been established as necessary for English
Renaissance texts, say, then the editing of American texts will become a
respectable occupation at long last, and not a piece of hack work for the
paperbacks.
Notes
