[page 45:]

Poe Texts in Transition

Benjamin Franklin Fisher IV, editor. Poe the Craftsman: The Changing Fiction. Essays in Honor of Richard P. Benton and Maureen Cobb Mabbott. The Library Chronicle (University of Pennsylvania), 41, No. I (Spring 1976), [1]-99.

The University of Pennsylvania Library Chronicle, which in recent years has presented Professor Fisher’s extensive textual essays on “The Assignation,” now devotes an entire issue, under his general editorship, to six studies of Poe’s short-story texts from genetic or evolutionary perspectives. A briskly written introduction, “Poe and the Art of the Well Wrought Tale,” announces the organizing purpose of the collection (“to serve Poe studies by treating textual history and evolving versions of the tales”) and summarizes the state of textual criticism on Poe. Its endnotes included, this opening piece by Fisher constitutes a useful descriptive bibliography of textual scholarship and the often divergent critical conclusions based on textual evidence. It is followed by the showpiece of the collection, Alexander Hammond’s “Edgar Allan Poe’s Tales of the Folio Club: The Evolution of a Lost Book,” and by shorter essays on Poe’s revisions in “Bon-Bon” (by James W. Christie), “Silence” (by Fisher), “William Wilson” (by Marc L. Rovner), “The Murders in the Rue Morgue” (by Joel K. Asarch), and “The Mystery of Marie Roget” (by Richard Fusco). As the editor remarks, each contributor “presents material according to individual preference, being restricted only by the requirement of textual history or Poe’s compositional habits.” Their differences of approach result in part from the available documents. Hammond, Fisher, and Asarch employ Poe’s literary manuscripts; the others confine themselves to printed versions because the only autograph matter, Poe’s handwritten revisions in printed copy, is not instrumental to their purposes. The first three essays make thoughtful use of sources and analogues in the writings of Poe’s contemporaries and in earlier literature; the last three discuss Poe’s intentions with little or no reference to other writers. Clarification of textual history is the main concern of only one piece, Hammond’s, while the others are inquiries into the meanings of Poe’s tales, drawing upon the evidence of revisions to challenge rival interpretations or to show that Poe’s conception of a given work or genre altered. Common to the last five essays is the not unreasonable conviction that Poe’s artistry matured in the successive versions of his texts. But two of these essays strain, in my view, to extract much significance from little revisionary data. In their quality of scholarly craftsmanship, the essays comprising Poe the Craftsman are not equally admirable; but the weaknesses of the least successful pieces as well as the strengths of the soundest ones will afford instruction to textual scholars.

Hammond’s monograph, which occupies a third of the volume, extends his “Preliminary Notes” on the 1833 version of the abortive Folio Club project (Poe Studies, 5 [1972], 25-32) by examining its design and evolution from 1831, when it seems to have been tentatively conceived, to 1836, when Poe abandoned it. The first major section of the study culminates in a conjectural reconstruction [column 2:] of the eleven-story collection, in symposium format, that Poe offered to the New-England Magazine in March 1833. Drawing upon the physical and textual evidence of two surviving manuscript leaves and upon the word-counts of the probable versions of the Folio Club tale-texts, Hammond prepares a model of the hand-lettered, stitched booklet from which they were salvaged — a manuscript Poe most likely assembled in 1833. The model’s features include the order of the tales and their fictional tellers, the “burlesque criticism” (now lost) occurring between recitations, and the probable versions of all the story texts, some of which had been published in 1832 bur would be reprinted, with extensive variants, in 1835 and 1836. In the second section, Hammond studies the fortunes of the collection from mid-1833, when Poe risked six manuscript tales made up in booklet form on the Baltimore Saturday Visiter’s short-story competition, to the autumn of 1835, when his hopes of marketing Tales of the Folio Club through Carey, Lea, & Blanchard collapsed. Featured here is a record of Poe’s attempt to publish the book by subscription immediately after his contest victory was announced, plus a derailed chronology of dealings between Poe and the novelist John P. Kennedy; Kennedy and Henry C. Carey; Carey and Eliza Leslie (editor of The Gift, who selected a Folio Club story for inclusion in that annual for the 1835-36 holiday season); Kennedy and T. W. White (owner of the Southern Literary Messenger); and Poe and White. The focus remains on the Folio Club collection, which Kennedy induced Carey to accept, late in 1833, for publication. Carey procrastinated, rejected Poe’s November 1834 request for an advance, and advised him to seek publicity through separate printings of the tales. On Kennedy’s recommendation in April 1835, White employed Poe as a regular contributor. Besides writing new fictions for the Southern Literary Messenger — tales that would take their place in later, expanded conceptions of the Folio Club volume — Poe eventually provided nine of the original eleven for reprinting or first appearance in the Messenger. In April or May 1835, he retrieved the bulk of the manuscript from Henry Carey; parts of it remained in The Gift office and would be recovered only the next February. A sixteen-story Folio Club volume was announced in the August Messenger, agreeably to Poe’s July request that White puff the project. In September Poe told Kennedy that White himself was willing to print the book, asking whether Carey, Lea, & Blanchard would consent to be nominal publisher. Carey received this suggestion with astonishment and scorn, dashing Poe’s prospects conclusively. Hammond’s third section describes “the probable makeup of the sixteen- and seventeen-tale Folio Club collections and Poe’s efforts to publish them during 1836.” At the instance of James Kirke Paulding, Harper & Brothers read the collection but rejected it in March. In September Poe approached the Philadelphia publisher Harrison Hall with a proposal for a seventeen-tale collection; simultaneously or very soon afterward, the New York branch of the British firm Saunders and Otley read an unfinished collection. Hammond infers that this version, like the one proposed to Hall, lacked the revisions and additions to the symposium framework which an expansion from eleven stories to sixteen or seventeen would have required, at a time when Poe was heavily engaged at other tasks. [page 46:]

The study illustrates literary detective work of the first order. Worthy of emulation, as well as applause, are Hammond’s acknowledgment of the complex and diverse intentions of the individual tales; his close, subtle, and responsible reasoning on matters of textual transmission and biography; his judicious use of previous Poe scholarship and standard references; and his enterprise in locating and correlating documentary material that illuminates this crucial phase of Poe’s career. One error and a few gaps in the argument might be noted. The reference to “October 12, 1835” — a misprint for “October 12, 1833” — in the July 20, 1835, entry of the chronology may temporarily confound the reader of so circumstantial an account. Hammond estimates that the lost holograph of “MS. Found in a Bottle” occupied ten pages of the 1833 Folio Club gathering (the number of words in the tale divided by the number of words on a page of the “Siope” manuscript). This estimate is pretty drastically out of line with the implied size of the “Bottle” manuscript selected by Miss Leslie from Carey’s Folio Club copy in November or December 1834. If, as Kennedy thought when reporting the transaction to Poe, the price was “a dollar a page,” the manuscript numbered fifteen pages. (In The Gift, the story took up twenty-one printed pages.) Again, since the surviving leaf of Folio Club prologue, paginated 9 and 10 points to eight pages of fore-matter, we must assume that eight pages followed the end of “Siope” (paginated 61 and 62) in the gathering Hammond describes: the two extant leaves were originally conjugate. But eight pages would not suffice for “MS. Found in a Bottle” at a minimum length of nine or ten pages, much less for that tale plus some pages of “burlesque criticism” on “Siope.” Conversely, eight pages would be too many for the commentary on “Siope” alone, if criticism of the previous three tales totalled only about eleven pages. These considerations do not seriously jeopardize Hammond’s model or even force him to abandon the inference that “MS. Found in a Bottle” was the next tale after “Siope.” The anomaly might be explained by supposing that the “Bottle” text bridged two gatherings, or that separate leaves were added near the end of the first gathering, or that after the critical debate on “Siope” a few leaves were left blank or even cut away. To acknowledge these possibilities would be to render the argument more complete, though neither more nor less persuasive.

As Christie observes in “Poe’s ‘Diabolical’ Humor: Revisions in ‘Bon-Bon,’” the earliest tales are the ones which underwent most frequent and most drastic revision. The evolution of “Bon-Bon” — “The Bargain Lost” in the first printed version — affords a clear view of Poe’s “success with comedy and his transformation into a competent craftsman in short fiction” Christie does justice to many aspects of the evolving tale, studying especially the changes between the 1832 (Philadelphia Saturday Courier) and 1835 (Southern Literary Messenger) versions. These aspects include the Folio Club context, the influence of popular devil-literature and Blackwood’s mannerisms, Poe’s curbing of runaway precocity, the impressive tightening of the thematic, dramatic, and symbolic structure, and the connections between Poe’s comic writing and his serious exercises in Gothic terror. Like the urbane devil of the later forms of the tale, Christie displays good taste in [column 2:] matters intellectual, and considerable resistance to scholarly intoxication.

Fisher (“The Power of Words in Poe’s ‘Silence’ ”) concedes that the original “Siope — A Fable,” published in The Baltimore Book in 1837, is a “pastiche . . . with obvious dependence upon sources.” But the chief thrust of his analysis is to refute G. R. Thompson’s reading of the tale as a parody of the Disraeli-Bulwer-De Quincey school of “transcendental” orientalia. Even the early version reflects a complexity and ambiguity of intention, and Fisher takes pains to show that successive revisions of the work, from manuscript to The Baltimore Book to Tales of the Grotesque and Arabesque (1840) and Broadway Journal (1845), strengthen its serious implications and mute its satiric overtones. Near the beginning of his essay, which, like Christie’s, employs explication among its many approaches, Fisher transcribes the fragmentary fair-copy manuscript (1833?). A photo-facsimile of the two pages accompanies the transcript. Regrettably, Fisher copies without explanation a group of features which strike me as nonauthorial: vertical lines dividing the text into segments of one hundred words. Some readers will also be puzzled by Fisher’s omission from the transcript of another feature of the photocopy, the number “71” at the conclusion of the text. Like the vertical lines to which it is related (I find that it corresponds to the number of words beyond the last full hundred), it should have been explained in a note or in the body of the essay. Another shortcoming is the misquotation in Fisher’s analysis (p. 65) of MS “Demon looked” as “Demon, and looked”; the formal transcript, however, is correct. Evidently this error is the printer’s for the phrase “Demon, and looked” occurs two lines below. Since it garbles the grammar of the quotation in full, the blunder should have been detected in proofreading, even without reference to the manuscript. Like two other contributors, Fisher places an exaggerated interpretive value on the variations of accidentals in printed texts; unlike them, he does allow (in a note) that certain accidental features in one printing may be attributable to the typesetters. The strength of Fisher’s argument lies in his thorough research. He pays heed to the Folio Club origins of “Siope,” notes the presence of other exotic items in The Baltimore Book, reports the admiring early reviews of the tale, traces its analogues of style and substance in Poe’s own poetry (with a credit to Hammond), in the Bible, in Moore, and in Byron. He is concerned with the genre of the piece, and, of course, with the textual changes it experienced. These investigations coalesce nicely to validate Fisher’s reading of “Silence” as a psychological fable, an “intense drama of the self.”

In “What William Wilson Knew: Poe’s Dramatization of an Errant Mind,” Rovner maintains that Poe plays ironically on Wilson’s perceptiveness about physical qualities — his “eye for detail” — versus his blindness to moral qualities and relationships. These latter, of course, are paralleled and symbolized for the reader by the physical details themselves. Rovner pursues this thesis by showing that Poe’s revisions between the 1839 and 1845 (Broadway Journal) versions of the tale add specificity to Wilson’s descriptions of persons, places, and objects. Parts of the analysis are independent of revisionary data; but the brightest spot, Rovner’s remarks on Wilson’s preoccupation with [page 47:] his acquaintances’ clothes, finds ample support in the changes at the end of the tale. Rovner contends that Wilson’s opening comments express not guilt (and the ethical awareness that guilt would presuppose) but a desire for sympathy consistent with his “moral obtuseness” — a contention which similarly harmonizes with Poe’s modifications of the second paragraph.

Unfortunately, Rovner’s technique of comparison often requires the reader to consult the two texts of “William Wilson” (or Harrison’s text and R. A. Stewart’s collation) to ascertain exactly what textual changes took place: some quotations are abbreviated, and hence perhaps misleading. A misquotation, “for me” instead of “From me,” will disturb even readers unfamiliar with the text, since it nearly inverts Poe’s metaphor and Rovner’s demonstration. One wonders about the choice of Burton’s (October 1839) as the “original” printed form and of the Broadway lournal as the “final version.” According to the Burton’s credit line, “William Wilson” is reprinted from The Gift. Heartman and Canny identify the printing in that annual for 1840 as the first appearance of the tale. (Hammond’s documentation reveals that by 11 September 1835 The Gift for 1836, containing “MS. Found in a Bottle,” was already in circulation.) If Rovner has determined, by collation or documentary evidence, that the Burton’s printing indeed antedated The Gift, he should not keep such high cards hidden in his sleeves. Furthermore, he makes no mention of the reprint in the Philadelphia Spirit of the Times for 5-8 September 1845, the last lifetime appearance according to Heartman and Canny. Its authority and textual relationship to the earlier printings deserve to be specified, however briefly.

Four variant texts of “The Murders in the Rue Morgue” are compared by Asarch in “A Telling Tale: Poe’s Revisions in ‘The Murders . . .’” — the surviving manuscript, Graham’s for April 1841, the 1843 Prose Romances of Edgar A. Poe pamphlet printing, and the 1845 Tales (Asarch acknowledges and reports, in a note, Poe’s few and small autograph revisions of this last form in the Lorimer Graham copy). The successive modifications “affect overt meaning, structure, and underlying implications,” Asarch maintains, and they result in a final version “more credible, though more grotesque,” than its precursors. His case for Poe’s assiduous improvements between the manuscript and Graham’s, and between Graham’s and the pamphlet, relies heavily on the citation of accidental changes. In the first transition, “Poe corrected the punctuation and included approximately twenty-five additional commas.” Although Poe, who joined Graham’s staff in April 1841, may have had a direct hand in the production of the issue, he cannot safely be credited for the tale’s new accidentals. (The Graham’s variant substantives, on the other hand, can be assigned to Poe with almost total confidence.) Asarch’s remarks on the Prose Romances text implies that all its differences from Graham’s are in the area of punctuation. These innovations are said to improve the tale noticeably: again Asarch attributes them unhesitatingly to the author. That each version is “more fully punctuated than the last” would hardly prove, to researchers conversant with current theories of copy-text and with nineteenth-century publishing practices, that Poe involved himself purposively and constantly in the revision of accidentals. Rather, it would illustrate the truism that successive copy [column 2:] editors and compositors impose successive layers of styling upon reprint texts. The notion that all variant features in reprintings over which Poe exercised a measure of control should be accepted as authorial squares neither with modern views of textual transmission nor with actual collation of the texts. Even when Poe had the Broadway Journal and edited sole proprietorship of all copy save, perhaps, the musical items, compositors’ errors in accidentals and possibly unauthorized substantives turn up in the reprints of his tales (for example, “beleiving” for “believing” and “irrevocably” for “irrecoverably” in “The Spectacles,” Broadway Journal, 22 November 1845, reprinted from the Dollar Newspaper). Can the possibility of nonauthorial changes in punctuation, spelling, and capitalization then be excluded? Asarch’s illustrations of added commas in the Tales version of “Murders” do not exemplify changes in meaning, modifications of tone, or sharpenings of the artistic design. Poe’s concern that “For the want of merely a comma, it often occurs that an axiom appears a paradox, or that a sarcasm is converted into a sermonoid” finds no confirmation in these optional pointings. We might remember that Herman Melville, a writer far less fastidious than Poe about the printing of his works, complained to Dix & Edwards about proof-copy for The Piazza Tales: “There seems to have been a surprising profusion of commas in these proofs. I have struck them out pretty much . . . .” And evidently his instructions were ignored (24 March 1856, and note, in the Davis and Gilman edition of Melville’s letters).

While as a rule Asarch’s responses to alterations in wording are more reliable, sometimes his reasoning seems slippery. He asserts that Poe’s excision of the detail of Madame L’Espanaye’s head rolling to some distance from the trunk marks the achievement of “additional shock value and greater drama” — an enlarged emphasis on “the horror of the scene.” More conventionally, he then notes that the substitution of “I felt a creeping of the flesh” for “I shuddered” displays “a further movement toward the grotesque.” The presupposition of Poe’s design to make the tale “more grotesque, and less everyday” can become a Procrustean bed.

How ironic that the critic should call the orangutan’s owner a “soldier” in a paragraph on Poe’s increasing valuation of clarity and precision! If this lapse suggests a careless reading of the tale, certain others betray relaxed standards for the accuracy of typed copy or impatience with proofreading chores during production. A page and line citation which should read “175,29” is given as “125,29.” In a quotation from the body of the tale, “back and front room” becomes “back and front rooms”; and the tale’s epigraph from Urn-Burial is deprived of a word, “himself,” and of italic type for “all.” Poe’s “taxed” and “tasked” become, in an endnote, “tax” and “task.” Finally, it seems that Asarch has confused William H. Graham, the publisher of Prose Romances, with George R. Graham, the publisher of Graham’s Magazine.

Fusco’s analysis of textual changes between the first printing of “The Mystery of Marie Roget” (Snowden’s, November 1842-February 1843) and the reprint in Tales (1845) seeks to prove that the versions manifest different literary personalities and different theories of reasoning. In the interval Poe’s “admiration for the ratiocination theme altered”; he was less eager in 1845 than in 1842-43 to [page 48:] display his intellectual superiority to his readers. Due largely to Poe’s reverses in public ratiocination (cryptography, the Barnaby Rudge prediction, the actual solution of the Mary Rogers case), he not only introduced qualifying changes into the Tales text of “Marie Roget” but skewed his ratiocinative fiction in two new directions. One was parody of causal reasoning, in “‘Thou Art the Man!’” The other, culminating triumphantly in “The Purloined Letter,” was the deductive solution of a purely invented crime. If “In 1842 he was interested in applying a theory of detection to the real world, . . . by 1845 he no longer cared about that theory”; “he retreated from reality and reverted to a more purely abstract approach,” where “the method mattered more . . . than the result.” I am sorry to say that this promising hypothesis is inadequately elaborated and defended. Fusco fails to mention that “The Murders in the Rue Morgue” itself (1841) involved a fanciful outrage. And despite the role of physical clues in Dupin’s discovery of an ape as the killer and of the means of ingress and egress, deduction through imaginative identification is the burden of the opening section on the mental powers and is Dupin’s means of producing the sailor to corroborate his explanation. To assert that Poe ceased to be interested in the application of ratiocinative techniques to the “real world” also seems to disregard the preoccupations of Eureka.

Fusco’s case about Poe’s growing respect for his audience depends heavily on the increased conventionalism of spelling, word-division, and punctuation in “Marie Roget.” But we are not shown that the Snowden’s forms (or, for that matter, the revised 1845 forms) are authorial. Concerning the divided terms “any thing,” “no body,” and “for ever” in the magazine print, Fusco declares, “To see, or pronounce, these words separately requires more effort, which gives them special emphasis; this emphasis was diluted deliberately in 1845.” Before arguing such a point, the textual critic should ascertain Poe’s own indisputable practice in manuscripts circa 1842, or exact transcripts. Next he should seek out occurrences of the questionable features in Snowden’s other contents during the winter of 1842-43. Poe’s manuscript habits in 1845 and the printing conventions of Wiley and Putnam ought to be investigated in the same manner. I suspect that house-styling accounts for such variations as “any thing”/”anything,” for the alteration of “visitor” to “visiter,” and for a share of the “improved” 1845 punctuation.

Numerous misquotations from his sources mar Fusco’s work. Checking only quotations from the Letters and the 1845 Tales text (I used the Lorimer Graham copy, at the University of Texas), I found a dismaying total of seven inaccuracies in the seven pages of essay text. Fusco changes “express purpose” to “expressed purpose” and silently omits a phrase — “the supposititious” — in Poe’s letter to P. P. Cooke, while in quotations from “Marie Roget” he gives “under the pretense” for “under presence,” “have availed himself and had he been on the spot” for “have availed himself had he been upon the spot” (Poe altered “upon” to “on” in pencil in the Lorimer Graham copy; elsewhere, Fusco calls attention to a different substantive revision, but here he cites the 1845 print without qualification), and “purposes of elopement” for “purpose of elopement.” At least one detail from Snowden’s — a semicolon after an [column 2:] introductory phrase — seems to be in error, but I have not compared it with the original. No responsibility of the textual scholar is more urgent than the accuracy of his transcriptions and quotations. In this respect Fusco’s piece is not only disappointing in itself but gives an unhappy conclusion to the volume as a whole.

Poe the Craftsman contains much solid historical scholarship and some exciting investigations of Poe’s intentions and maturing artistry. At the same time it reveals that textual criticism of Poe has not yet matured, has not sufficiently involved itself in the theoretical and methodological advances of the past thirty years. The declining reliability of accidentals as a text proceeds from finished manuscript to first printing to revised reprint, the susceptibility of even some classes of substantives to corruption during reprinting, the need to consider whether the author proofread, could have proofread, or could not have proofread any given printed version, the value of naming the exact specimen of a book or magazine used in collation, the necessity of multiple proof-readings by the scholar — these are a few of the concerns of textual editing and textual criticism pertinent to the study of any modern author. Poe is no exception.

Joseph J. Moldenhauer, University of Texas, Austin