Editorial Policies and Methods for Preparing Texts


There are many sites on the Internet featuring some of Poe's works, usually the better known poems (such as “The Raven”) and tales (such as “The Fall of the House of Usher”), and several sites which purport to offer Poe's “Complete” works — none of these sites, however, can accurately make this boast. At best, they might offer all of Poe's important poems and/or tales, often without clearly identifying the actual source of their texts. (Most of these sites have taken their material from the “Raven” edition published by Collier & Sons in five volumes in 1903, though this source is rarely acknowledged. Although a good popular edition, those texts do contain errors; they also embody editing which does not necessarily reflect Poe's original texts. Several errors created by the early Griswold edition of Poe's works find their way uncorrected into the 1903 “Raven” edition, such as the minor but  unwanted “s” on “mortals” in the fifth stanza of “The Raven.” More importantly, many of these texts are offered without the kind of careful proofreading necessary to correct problems with scanning and typing. Many sites, indeed, use texts containing not only the same version but also the identical typographical errors, indicating that the easy road of copying has widely been taken rather than the difficult path of scanning and proofreading.) Wherever possible, only the original sources have been used for texts on this website, retaining Poe's punctuation and spellings.

This site is intended not only to cover the well-trodden ground, but to fill in the considerable gaps between what Poe wrote and what is easily available to most readers. Even for the tales and poems, Poe wrote a great deal more than the handful of items read so regularly. Here, one may read Poe's final tale, the unfinished “The Lighthouse.” One may also find the poem “Deep in Earth” and the fragmentary “The Beloved Physician.” In addition, here are selections from Poe's essays, literary criticism and such miscellaneous writings as the “Doings of Gotham” letters, the “Marginalia” and Poe's introductory material for his “Conchologist's First Book.” It is hoped that by providing free access to a broader selection of Poe's works, it will quickly become evident that he has been misjudged, based on too narrow a reading of items that reveal but one expression of his genius.

Another goal of this site is to provide bibliographical information on the various printings, revisions, and reprintings of Poe's writings, chiefly during his lifetime. Whenever possible, we will also offer the actual texts for first and other important or significantly revised versions. In this regard, we see Poe as a conscious artist, creating with effort and careful attention to details. Although the casual reader may be quite content with reading a single version of a tale or poem, such an approach inherently brings with it several problems. One issue which is all too frequently glossed over is the fact that Poe wrote and rewrote many of his works, sometimes several times, often making significant changes. Many readers are entirely unaware of the fact that a version of one of Poe's works, conveniently gathered and printed in a book, is, in most cases, a reading copy, editorially derived from one or more of these variants. Another problem with presenting a single text is that it is necessarily removed from the original contexts to a degree which places each item in a kind of absolute isolation, or an unintended juxtiposition. How is one to respond to “A Predicament” without its proceeding “How to Write a Blackwood Article”? How is one to recognize the satire or parody in “The Visionary” or “Metzengerstein” without understanding its original place in Poe's unpublished Tales of the Folio Club? And Poe's “Sonnet — to Science” may take on a slightly different meaning as the introductory poem to “Al Aarraf” in both the 1829 and 1831 editions of Poe's poems.

For criticism, some effort has been made to distinguish the substance of the items. A review is considered to be a formal comment on the merits of the work, even if much of the review is merely material quoted from the work under review. A notice is a short comment on the work, usually no more than a few sentences, with some indication concernings its merits, literary or as a product. An announcement is merely a comment that a work has or is about to be released, without any indication in regard to merits.


Types of Texts:

People read texts for reasons as diverse as light entertainment and deep scholarly study. In many cases, particularly for texts which have been presented in multiple authorized forms, as is true for many of Poe's writings, it is inherently impossible to present a single text which adequately meets such a broad range of uses. Fortunately, on a website, it is possible to do something that would generally be impractical in a printed edition, namely to provide full multiple texts. Such texts may, for the sake of convenience, be grouped into the following categories:

  • Reading and Reference Texts — Most casual readers are looking only for one text of a work, a text which reasonably represents the intentions of the author in a format which serves the needs of a general audience. Exactly how such a text should be prepared is a matter of much debate in bibliographical circles. Some readers find notes, annotations, word definitions, and varorium lists of variations between texts to be useful, while others find the apparatus for such extentions of the text to be intrusive and disruptive to the reading process. For this reason, there should, perhaps, be a distinction between a reading text (presenting the text alone) and a reference text (presenting the text and some combination of extensions). At present, no formal policy has been determined for this website, although it will be a long-term goal.
  • Historical Texts — Many texts for a single work exist in multiple forms. Poe's tale “MS. Found in a Bottle,” for example, was printed in no fewer than 5 versions during Poe's lifetime, and “The Raven” was printed and reprinted literally dozens of times, evolving through at least 18 versions that are presumed to have authorial sanction. These forms may include manuscripts, printed text, and revised proofs. A desirable distinction between these texts is those which carry at least some authority from the author, and those which are merely reprints.
  • Comparative and Study Texts — For readers interested in the evolving history of a text, a useful tool is a presentation that points out the various changes made during the history of that text. Such a comparison may be made between two selected texts, or between a larger set of texts. At present, the current scheme for comparative texts is an experiment, and under development. All changes between the texts noted are included, both in substantives and accidentals. Thus, the attempt has been made to capture all differences in phrasing, spelling, formatting (italics), and punctuation. A list of the texts included in the comparison is given at the top of the page, each with a short-hand notation identifying that text. Changes in the texts are given between double gull brackets (“{{” and “}}”), and in blue. The brackets and the text notations (the year and text number from the list at the top of the page, such as “1843-01:”) are given in bold, with multiple date notations relating to a single change being listed and separated by semi-colons. Sets of related changes are given in a contiguous stream, divided by double slashes (“//”). Using a few examples from “The Tell-Tale Heart” are representative, in the excerpt “I arose {{1843-01: , }} and argued . . .” the comma appears only in the 1843 text from the Pioneer. The rest of the phrase appears in all versions. In the excerpt “But anything {{1845-02; 1850-03: was }} better . . . ,” the word “was” does not appear in the 1843 text in the Pioneer, but was added to the 1845 text for the Broadway Journal and retained in Griswold's text of 1850. In the excerpt, “yes, it was this! {{1843-01; 1845-02: He had the eye // 1850-03: One of his eyes resembled that }} of a vulture . . . ,” both the 1843 text in the Pioneer and the 1845 text in the Broadway Journal read “He had the eye of a vulture . . .,” while the 1850 text printed by Griswold gives it “One of his eyes resembled that of a vulture . . . .”
    A study text takes advantage of essentially the same concept as a comparative text, except that the study text documents a single text with manuscript changes, such as the Duane copies of the Southern Literary Messenger, the J. L. Graham copies of Tales and The Raven and Other Poems, or the draft manuscript of “Murders in the Rue Morgue.”

In addition to such texts, the bibliographical entries will list scholarly and noteworthy texts (also, technically, categorized as historical texts), and miscellaneous texts and special versions (such as translations, radio shows, and film adaptations).


A Few Words on Method:

Generally, printed text is scanned on an HP Scanjet. Optical Character Recognition (OCR) is performed with Scansoft OmniPage11.0 and saved as a simple ASCII text file. This intermediary file is imported into Corel WordPerfect 10.0, where the text is examined for obvious errors, many of which are conveniently highlighted by WordPerfect's spell checker. Manuscript material, which cannot be interpreted by OCR, and text from sources which, for one reason or another (such as small or esoteric fonts, or poor quality printing) cannot be scanned, is entered by hand. Variants are usually created from a basic version of the text, modified as required by direct comparison to the target version. Greek and Hebrew text, which does not use standard character sets, may be provided as an image, scanned and manipulated under Adobe's PhotoShop 6.0. Foreign languages often use characters which are not part of the standard language character sets supported on all computer operating systems. We are still evaluating the best way to address this problem.) Illustrations originally published with the text are also scanned and processed under Adobe Photoshop 6.0.


Since OCR and spell checkers are notoriously unreliable, and inherently limited by their mechanical nature, text must ultimately be verified manually. OCR, for example, routinely misreads “hath” as “bath” and “thine” as “shine,” especially when the source font is smaller than 12-point. Spell checkers are inadequate for Poe's historical spellings and frequent coining of words. Thus, in the end, there is no way around manually proofing each text — and for historical editions, proofing each text against the original. This proofing must be done with great care and attention — and, for the best results, multiple times. At the end of each text, a proofreading mark is maintained to keep track of its status. The mark is always the last text on the page, in a very small font. As an example: “[S:0 - Works, 1850]” indicates that the text has not been proofread, and that the source for the text was the original 1850 edition of Poe's works. Although all texts are at least minimally reviewed during the process of formatting, this is not considered to be sufficient care to qualify as proofing. “[S:1]” indicates that it has been proofread once, and “[S:2 - MS, 1842]” indicates that it has been proofread twice, using the original 1842 manuscript as the original text. If no mark is present, the page should be presumed as not having been proofread.


For historical texts, every attempt will be made to honor the original formatting as much as practical given the limitations of HTML. Included in such considerations are spelling, punctuation, capitalization, and the use of italics. Some suggestion of the relative size of type, but the use of very small fonts has been avoided. (Some note will be given for texts where there are altered.) Pagination will be noted, giving page and column indications where approrpriate. These indications will be given in square brackets, and in bold. Footnotes have been moved to the end of the text, to avoid breaking the flow of paragraphs merely for the sake of dogmatically reproducing an artifact of the printed form. (The location of the footnote in the original will be noted.) End-line hyphenation will be eliminated for words which would, normally or in Poe's general usage, appear without hyphen if they occurred in the middle of a line. (At some point, we will note end-line hyphenation at the end of the text.) End-line hyphenation which occurs across pages will be handled in a similar manner, so that the whole word appears on the page where it began.





[S:0 - JAS] - Edgar Allan Poe Society of Baltimore - Works - Editorial Policies and Methods