[page xxvi, continued:]

II. PLAN OF THIS EDITION

CRITERIA FOR SELECTION

It might be possible to show from Poe's own remarks in his critical writings that he thought a tale ought to be what we now usually call a short story. But he was not rigidly consistent in classifying his own works, and he estimated their number in 1844 and 1845 in a way that makes me conclude with Woodberry that Poe counted as tales what others might call essays.* Hence I have established a category of “Tales and Sketches,” and include in it shorter prose articles by Poe in which there is an element of fictional [page xxvii:] narration.† The principle governing my decisions is a desire to include all the articles that involve narration. Extreme brevity does not lead to exclusion. I have included nine definitely new pieces, over Harrison's collection, most of them short. They include, however, Poe's latest and unfinished story, “The Light-House.”‡

The canon of Poe's imaginative prose is not hard to establish. Practically everything collected here is either his acknowledged work, or established by incontrovertible evidence. Two or three sketches are accepted with slight reservations, but only one, the strange story called “A Dream” (1831), can be classed as of highly doubtful authorship, and even this one cannot be firmly dismissed. It may be added that we have no reason to think any tale composed by Poe in maturity has been lost.

SOURCES OF THE TEXTS

(For details see Sources Collated, at the end of Volume III)

The sources for the texts of Poe's short fiction are of several kinds. In addition to nine complete manuscripts and three fragments that survive, there are authorized first printings in monthly magazines, weeklies, daily newspapers, and gift books known as annuals. There are also reprints authorized by Poe which show revisions. No proofsheets survive.

Three collections of the shorter fiction here issued during the lifetime of the writer. They are: (1) Tales of the Grotesque and Arabesque (Philadelphia, 1840), issued late in 1839 in two volumes containing twenty-five stories; (2) The Prose Romances (Philadelphia, 1843), of which only the first number, containing “The Murders in the Rue Morgue” and “The Man that was Used Up,” appeared; and (3) Tales (New York, 1845). This last is a selection of a dozen stories made by E. A. Duyckinck. [page xxviii:]

In addition, we have copies of several volumes in which Poe made manuscript changes, usually with a view to future reprintings:

The first of these is called PHANTASY-PIECES. It originally consisted of the two-volume set, Tales of the Grotesque and Arabesque, for which Poe made a new title page and table of contents in manuscript and in which he indicated numerous emendations, some of them abortive. Only the first volume survives; it was found in Poe's trunk after his death. The second volume has disappeared. I suspect that it was broken up and used as copy by Griswold's printers, and was the source of the Works texts of “Metzengerstein” and “Hans Pfaall.” There is another copy of Tales of the Grotesque and Arabesque, given by Poe to the Misses Pedder, in which the author added a manuscript footnote.

Of the Tales (1845) there is the celebrated J. Lorimer Graham copy (bound up with The Raven and Other Poems), with penciled manuscript revisions. Poe kept the book with him, even in his final summer, and hence the changes can hardly be dated individually, since all were not necessarily made at one time.

Manuscript changes were also made in what is known as the Bishop Hurst copy of Eureka (1848).

There are two bound files of periodicals containing Poe's manuscript alterations. One is the Duane set of the Southern Literary Messenger with a few changes, meant as printer's copy for the collected edition of 1840. The other file, the Broadway Journal, was given by Poe to Mrs. Helen Whitman, with numerous markings in pencil. Most of them are acknowledgments of his authorship of unsigned articles, but they include a few important verbal revisions.

Finally, there is the Rufus W. Griswold edition of the Works of the Late Edgar Allan Poe (1850 and 1856). The versions there are usually to be accepted. Careful study of the variants shows that Griswold had for some tales obviously superior readings, improvements that must have come from Poe. For these it may be assumed that Griswold had clippings with revisions in Poe's hand.§ Even when Griswold's edition shows but one or two changes, the presumption [page xxix:] is in their favor, and when the last periodical text is verbally the same as Griswold's, nothing is gained by insisting on a return to the former. Hence, although the merits of each text are considered individually, Griswold's Works of ... Edgar Allan Poe (Volumes I and II, 1850; and Volume IV, 1856) is accepted for the majority. But some of Poe's latest changes were not included in Griswold's texts. He did not use the J. Lorimer Graham copy of Tales (1845), undoubtedly because he did not receive it until after he had sent an uncorrected copy of the book to the printer. Neither did he have access to the file of the Broadway Journal Poe gave Mrs. Whitman in 1848. Our texts have incorporated these changes.

ARRANGEMENT

The tales in Volumes II and III are arranged chronologically. The precise order in which Poe wrote them cannot be completely determined; but the approximate date of the original version, as well as the exact date of the earliest publication, is now known for all the stories except one. The exception is “Why the Little Frenchman Wears His Hand in a Sling,” which presumably first appeared between 1837 and 1839 in a periodical not yet discovered. The advantages of arrangement by date are very great to any serious student. He cannot fail to see development, for both Poe's method and his style changed markedly with the years, as did his opinions.

The order of the present edition differs to some extent from Harrison's because of the discoveries made since his great work appeared in 1902. More than fifty additional texts are collated here; by far the largest number are manuscripts and early printings. At one time or another I have seen all the printed originals described, and every manuscript known to be extant.

DISCUSSION

The discussion of the individual tales follows the plan set forth in my Preface to the edition in Volume I. The most pertinent passages are restated here:

The introductory note to each item gives the history of composition and an account of Poe's major sources if they are certainly [page xxx:] known or plausibly suggested. The history of publication and the list of texts which follows include mention of all authorized versions of the item now known. In the case of articles which neither Poe nor his literary executor collected, record is made of first publication in periodicals and of first inclusion in books.

Purely aesthetic criticism is deliberately kept to a minimum. It is usually confined to pointing out the merits of widely recognized masterpieces and to evaluation of pieces less well known. Something is said of the estimates of his work (not always favorable) that stem from Poe himself, but very little of his vast influence on his contemporaries and on those who have come after him. Mention is made, however, of all known reprints in America and in Britain before 1850, and of translations made on the Continent during Poe's lifetime.

In the notes I give, as far as I can, the sources of Poe's direct and indirect quotations, and explanations of his references and allusions, that may not be clear to a reader of today. Cruces are discussed, but record is usually omitted of views outmoded in the light of present knowledge and of explanations withdrawn or abandoned by their proponents.

Credit for their discoveries is given to my predecessors, I hope with some thoroughness. I am also indebted to my students and friends for unprinted suggestions given to me directly, which I have tried to acknowledge and for which I express my sincere thanks.

VARIANT READINGS

My policy is of extreme respect for the text. I correct nothing verbal in Poe's texts except obvious misprints. Fashions in spelling changed while Poe was writing and some of his forms look curious to our eyes, but spellings of his time have been allowed to stand, and, if they deserve it, receive comment in my notes.

In a few cases it has seemed desirable to reprint in toto the earliest version of a tale as well as the final text.

All verbal variants in the authorized texts are recorded, but only such variations in accidentals as seem of any significance. Differences in spelling, if the forms used were generally accepted in Poe's day, are not recorded, but typographical errors in all texts have been listed among the variants. [page xxxi:]

Changes in italicization are recorded.

Punctuation variants are not recorded, with the following exceptions:

1. Poe's manuscript changes in the J. Lorimer Graham copy of Tales (1845) are recorded.

2. Complete punctuation changes of the seven texts of “Ligeia” are recorded, as an example of Poe's work in revising his tales.

3. Poe's manuscript changes in the punctuation of PHANTASY-PIECES are summarized but not recorded, with the exception of those in “Ligeia,” as mentioned above.*

4. Differences in paragraphing are not recorded, except in the case of surviving manuscripts, and in “Ligeia.”

5. Hyphenations in the manuscripts are recorded if they differ from the printed texts. Only in the manuscripts can one be sure that the word compounds originated with Poe.

The following kinds of emendation have been made:

French. Verbally Poe's French is allowed to stand, but errors in accents and spelling are corrected — from other texts, if possible; if not, editorially. The incorrect forms in Poe's texts are recorded in the variants.

Greek. Errors in spelling are corrected, but accents are used only if they appear in the original.

Punctuation. In some texts emendations in punctuation seem required. Actual errors are corrected, generally from first printings, but editorially if necessary. The errors are recorded in the variants.

Two kinds of changes have been made silently. (1) Hyphenations within a particular tale have been made consistent according to the form apparently favored by Poe. (2) In the case of texts from the Broadway Journal, there are a number of occasions when, in order to fill out a line to the right-hand margin, the printer inserted a dash after the period following a sentence. A few of these dashes were carried over to Griswold's edition of the Works, where they [page xxxii:] often fall in the middle of a line. All these “printer's dashes” have been silently eliminated. They have nothing to do with Poe's punctuation, although Harrison and other editors have reproduced them in their texts from the Broadway Journal.

ABBREVIATIONS

AL    American Literature

DAB    Dictionary of American Biography

DNB    Dictionary of National Biography

MLN    Modern Language Notes

OED    Oxford English Dictionary

PMLA    Publications of the Modern Language Association

SLM    Southern Literary Messenger

[[Footnotes]]

[The following footnote appears at the bottom of page xxvi:]

*  Woodberry, Life (1909) II, 405. Poe called “Mesmeric Revelation” a “tale” in a letter of May 28, 1844, to Lowell; an “article” in a letter to George Bush, January 4, 1845; and an “essay” in a letter to Griswold, February 24, 1845.

[The following footnotes appear at the bottom of page xxvii:]

†  For the purposes of presentation, it is convenient to group together the longer stories in Volume IV of this edition. The Narrative of Arthur Gordon Pym (1838) is of book length; “The Journal of Julius Rodman” (1840) was meant to be, as was, apparently, “Hans Pfaall” (1835) at first. The pieces are closely akin in manner, although the two earlier involve the impossible while the other does not.

‡  Besides “The Light-House,” the new pieces are “A Dream,” “Instinct vs Reason,” “Cabs,” “Moving Chapters,” “Desultory Notes on Cats,” “”The Swiss Bell-Ringers,” “Theatrical Rats,” and “A Reviewer Reviewed.” I also include a number of pieces collected by Harrison but placed elsewhere than among the tales.

[The following footnote appears at the bottom of page xxviii:]

§  There is preserved an example of Poe's correcting in this way one of his printed pieces. See “Poe's Revision of Marginalia;” in Ex Libris (January 1940), a quarterly leaflet issued by the friends of the library, The Johns Hopkins University.

[The following footnote appears at the bottom of page xxxi:]

*  The punctuation changes in PHANTASY-PIECES are particularly interesting, however, because we are certain that Poe made them. For this reason these changes have been summarized and the summary placed under the list of texts of each tale. In the twelve tales involved, 1832-1839, Poe was most concerned with his use of dashes, industriously eliminating a large number in favor of commas and semicolons. In 1848 he was accusing printers of doing this, to the vexation of the writer. See “Marginalia,” number 197, Graham's Magazine, February 1848, p. 130.