[page 37:]

THE TEXT

The text used for this edition is that of the 1838 edition published by Harper and Brothers, the first complete edition and the only one authorized by Poe. The collation for the two installments published in the Southern Literary Messenger (January and February 1837, 3:13-16, 109-16) will cover both substantives and accidentals. The two parts were presented by the journal in an indifferent and careless fashion — a fact which makes it seem unlikely that even the first benefitted from the personal attention of Poe. He must have had a chance to proofread and edit it, since the magazine appeared at the beginning of each month and the lack of a December issue afforded an extra interval in which to prepare and correct copy for January. However, in addition to part one of Pym, Poe published his poems “Zante” and “Bridal Ballad” and critiques of books by Bryant, Tucker, Irving, Reynolds, and Anthon — a sizable amount of material to check over at a time when he was not well. He apparently did not know of the impending separation, often wrongly thought to have been initiated by himself. Thomas Willis White, enjoining secrecy, wrote to Beverley Tucker on December 27, 1836, of his intention to give Poe notice “in a week or so.” The confrontation must have come soon thereafter since White's statement at the end of the January number speaks of Poe's having “retired” from the “editorial department” as of January 3.(1) White was entirely responsible for publishing the second installment and may have been disinclined to treat the material seriously; his letter of January 19 to Tucker ends with a postscript: “Poe feels his situation at last — I see but little of him — but I hear a great deal about him and from him. I am tired out with hard work” (Jackson, pp. 111-12). The result of his “hard work” was the February issue planned for January 22 and issued on January 25. To Poe himself, White wrote a letter on January 17 proving clearly Poe's small role in planning or arranging any aspect of the second installment, which began at a suitable narrative break (the end of 2.3): “If it be possible, without breaking in on my previous arrangements, I will get [page 38:] more than the first portion of Pym in — though I much fear that will be impossible. ... You are certainly as well aware as I am that the last $20 I advanced to you was in consideration of what you were to write for me by the piece” (Harrison, Works, 17:42). White did indeed double the allotment of Pym, which he disliked both for its content and its expense. In an undated letter he designated it as “stuff” for which Poe outrageously “demanded three dollars a page.”(2) It was White's habit to expect a supply of contributions gratis. We cannot tell whether the publisher had used up his tolerance, money, or space. Certainly the abrupt break at the end of the second installment (4.3) makes it likely that Poe had given him the rest of chapter 4 and probably 5 because of the narrative continuity. On the other hand, Poe may have preferred to save the unpublished text for the full-length book then possibly being envisioned according to the advice of the reader for Harper and Brothers, John Inman, and of J. K. Paulding. There must have been material to show as a basis for the contract, presumably signed in February or March in view of the registration in June 1837 (see Fig. 4) and the announcement of the intended publication in the May 1837 Knickerbocker Magazine.(3)

The two installments, the second being twice the length of the first, show no significant differences in the proportionate numbers of typographical errors (ten in all). In the 1838 text all are corrected but in later sections five errors of punctuation appear, as well as a misspelling, which is corrected through a bracketed, inserted letter (see p. 50 for the full list). There are two apparent or arguable misprints: “under weigh” (2.9) and “followed” without a sequent “it” (3.4), neither of which has been changed in modern editions. Aside from the changes in wording demanded by his alteration of the ages of the lads, the date of their sailing, and the story line (see 1.1) Poe apparently wished to improve the accuracy of his language, although later distractions must have prevented him from adequately revising the text beyond 4.3. Salutary was his shift from “steerage” to “forecastle,” although incorporated only in 2.5, from “ground” to “floor” (3.8), and from “speech” to “voice” (3.8). Most probably [page 39:] it was he, not the copy editor, who made a few rhetorical improvements, as in “might be expected” (1.10) and “prying” for “prizing” (2.17). The majority of changes are in the direction of more modern spelling and punctuation. Compounds are now hyphenated, such as “sea-biscuit,” “rough-looking,” and “hiding-place,” while many previously hyphenated compounds or separated terms are made single words: “everything,” “southwest,” “ringbolt,” and “seaworthy.” There are exceptions, such as “under tone” and “mean time,” both allowed then. Forms are systematically changed, arguing either a house style different from the Messenger's or a genuine change in Poe's practice: Adverbs in “-wards” drop the “s,” as in “afterward”; almost all nouns ending in “-or” are accorded the British spelling of “-our” as in “neighbour” and “endeavour”; and “farther” becomes “further,” but not always, as in the title. The changes in punctuation are probably the most numerous (about 175 out of the 300) and serve to tighten the narrative somewhat, for prudently used semicolons replace many of the end-of-clause dashes. Many superfluous commas are deleted, especially before both parentheses of a set. Occasionally spellings seem purely arbitrary, such as “despatch” for “dispatch,” but are allowable variations at the time. Notes are provided for such variants as “lilach,” “deposite,” and “villany”; likewise for apparent errors in incorrectly copied material (20.4). The numerous grammatically weak constructions have been left unchanged, despite the widespread tendency of editors to follow the Griswold text in correcting a presumably wrong agreement of subject and verb (see no. 43 in Table 2, pp. 46-47). One clear error, in the accenting of “ricocheting” (22.5), may represent Poe's careless habit in this regard rather than the typographical blunder (see 22.SA).

In presenting this text of Poe's longest work of fiction, I have inserted minimal editorial alterations, in fact, a total of only five — for accidentals. Inconsistent forms, such as those which follow, have been retained since all were acceptable in his day or were variant spellings of names. (Similarly, forms which may vary among the three works in this volume have been retained.) There is no need to discuss whether or not they represent an indifferent attitude toward his own book, his habits of composition at that period, the slackness of typesetters working on a piece of fiction (unlikely for the reputable firm of the Harpers), or haste in the last stages of book production. The materials for further examination and speculation — by no means complete — include these instances (with no designating quotation marks): amidships (9.5) / amid-ships (6:12); Bennet's Islet (17.11) / Bennett's Islet (24.3, 24.5); boarding nettings (22.6) / boarding-nettings (18.8, 20.10); farther (title, 21.4, 22.5, 23.7) / [page 40:] further (21.7, 23 bis.4); head sail (6.3) / headsail (14.6); unconceivable (10.1) / inconceivable (10.4, 14.5, 23 bis.10); jollyboat (6.4, 8.7, 8.10, 14.19) / jolly-boat (1.8) / jolly (1.9); Kerguelen's Land (14.3, 14.10, 14.19, 16.6) / Kerguelen's Island (14.7, 14.10); lookout (14.6, 15.10) / look-out (1.7, 6.12, 7.1); Pole (Note.3) / pole (16.1, 16.3, 16.8, 17.8); around (2.14, 5.2, 22.13, 24.12) / round (5.2, 22.4, 23.9, 23 bis.2, 23 bis.3, 23 bis.10); sea-boat (14.1) / seaboat (14.6); ship furniture (6.5) / ship-furniture (2.7, 2.16, 4.6); single reefed (17.2) / double-reefed (7.5); Vredenburgh, (1.1, 2.2, 17.6) / Vredenburg (2.5); Weddel (15.9) / Weddell (16.5, 16.10); whaling vessel (3.3, 4.1, 6.5 n.) / whaling-vessel (2.5). It is clear that the Harpers neglected to impose regularity upon these and other alternate forms throughout Pym. This possibly accords with the editor's failure to revise the numbering of the chapters when a new “chapter XXIII” was inserted by Poe, presumably in the last stages of typesetting. On the other hand, the small number of outright typographical errors may indicate that the firm gave Poe's book fairly careful attention.

The same certainly cannot be said of the staff of the J. S. Redfield company of New York City, which helped Rufus W. Griswold to publish The Works of Edgar Allan Poe in four volumes, the last of which, containing Pym, was issued in the spring of 1856. Griswold obviously had no manuscript or corrected copy from the hands or the effects of Poe, who had not even changed the misnumbered final chapters. The importance of this text can be gauged from the eighteen reprints, of 1857 through 1884, under the imprint of Redfield or of W. J. Widdleton, W. C. Bush, or A. C. Armstrong.(4) Even wider currency has been given to the errors and arbitrary changes in the Griswold edition through its use by almost every major edition of the novel since the 1880s.

The table of “cruxes” that follows (pp. 46-47) will enable the reader to judge the merit of many of these changes in the Griswold edition. In addition, given below is a series of items in that text demonstrating the careless and negligent preparation of the printer's copy-text which, presumably, showed no author's alterations. Since all copy was destroyed, we can only infer that Poe was not responsible for alterations; in view of his expressed opinion of the work, this inference seems warranted. The variations from the 1838 text were numerous — in spelling, punctuation, and wording although almost entirely unimportant and probably, [page 41:] for the most part, accidental. Of the outright errors in the text of the Harpers (see p. 50) the first six, being obvious, were corrected. It should be noted that after the edition of 1856, several of the stereotype plates were altered to correct patent errors in the Griswold “first” edition. (e.g., those in 7.1, 7.10, 13.2). In the pairs given here, the first member is from the 1838 text, the second from that of 1856 (with no designating quotation marks): as we were fast / and we were fast (1.4); which floated / that floated. (1.10); rope wrapped / rope wound (1.11); iron-bound box / iron-boundbox (2.7); until / [nail (4.1); the mean while / he mean while (4.2); cook's gang / cock's gang (6.11); Peters's design / Peter's design (7.1) there were / there were (7.3); eleven / elevn (7.10); pantaloons' / pantaloon's (9.5); Peters's grasp / Peter's grasp (12.19); through into our / throughintoour (13.2); period / pesiod (13.12); period / perid (13.17); sprung a leak / sprunk aleak (13.23); words. / words: (13.23); absolutely / absolutoly (14.2); constructed / constructed (14.13); runs parallel / run's parallel (16.3); date. / date: (16.6); entry. / entry: (16.7); 47° 31’ E. / 47° 13’ E. (16.9); longitude / long. (17.8); comprehend / comprehend (20.1); Harris, and / Harris, (20.8); believe / believed (21.2); demolished /demolished (22.7); efficient-not less/ efficient not less (24.1); days after / days afer (24.2); upon the summit / up on the summit (24.5); handkerchief / hankerchief (24.7); THE END. / omitted (end of book). Looking at these many changes, we find few that clearly improve the text, while many impair the reading. None of them requires the involvement of the author.

In addition, the spellings of numerous words were deliberately changed in the 1856 text, either by Griswold himself or, more likely, by the Redfield copy-editor or compositor. In general, a more modern, less British style was favored. For example, words in “-our” dropped the “u” as in the following (the basic 1838 form is here given): behaviour, colour, demeanour, endeavour, favour, flavour, labour, harbour, honour, humour, neighbour, vapour, and vigour; similarly, afterward, connexion, despatch, endite, imbitter, macaroni, moveable, recognise, steril, stupified, survivers, and visiters. And, finally, hyphens were either inserted or removed freely, arbitrarily, and inconsistently in many words, such as aftersails, bearskins, boarding nettings, fireplace, fish scales, half-deck, New-York, packet book, seacoast, sea elephant, seashore, seaswallow, seagull, shellfish, ship-furniture, singular-looking, soapstone, Southwest, stateroom, storeroom, street-robber, and whaleboat. Neither the hand of Poe nor that of Griswold seems to be operative in these many alterations from the 1838 text, which must be accepted as solely authoritative.

Yet in the United States the 1856 edition has continued to exert its [page 42:] influence through a succession of purportedly accurate and sound texts. In 1895 George Woodberry and E. C. Stedman produced the first American collected edition having any serviceable notes and critical apparatus (BB.19, 30, 50, 57). Unfortunately, they accepted Griswold's text as authoritative and also asserted their right to correct “errors ... to accord with later usage and taste” (l:ix). For a few glaring mistakes, the editors go back to the 1838 edition, but they produce a few new ones, such as “parrallel” (14.5) in their beautifully formatted book. This type of irresponsibility is true also for the “definitive” edition of 1902 edited by James A. Harrison, the most widely consulted and quoted text. Despite his claim to be using the 1838 edition on his title page and notes of collation (3:329), Harrison basically follows the Griswold text. The words in “-or” consistently follow the pattern of 1838, but the, hyphens are often those of the 1856 edition, while “undertone” reverts to the SLM form; allowed variants, such as “lilach,” are silently corrected. With a kind of ingenuity, R. A. Stewart, who prepared the collations, manages to omit enough of the changes to give the false impression that he has really traced all the variants of 1837-1838, a claim disproved by a glance at my list of cruxes for the first four chapters. Perhaps Harrison regarded Pym as a minor component in the Poe canon, speaking of it as “made up of equal ingredients of Poe, ‘The Ancient Mariner,’ and Benjamin Morrell's ‘Narrative’ ... [,] full of graphic horror and exquisite though terrible landscape-painting” (Works, 1:133).

The deleterious influence of Griswold's edition has extended to several more recent and more accessible texts. The widely used Complete Poems and Stories of 1946, edited by Quinn and O‘Neill, purports to reproduce the 1838 “first printing” (p. 1085) and yet includes many of the Griswold edition changes along with several 1838 readings. In the field of the multicopy paperback book we find the Griswold or the Harrison Pym in W. H. Auden's popular compilation of 1950 (BB.70) and Edward H. Davidson's Selected Writings of 1956 (BB.71). Decidedly better but not unflawed is the 1960 edition prepared by Sidney Kaplan with a stimulating introduction (BB.72). This purports to be “a faithful reprinting of the first edition,” but even the full title comes from Harrison's alteration (3:329), and numerous unnecessary, silent corrections are made. Richard Wilbur's edition of 1973 follows Griswold's save for the correction of obvious typographical mistakes. Most recently, the edition of Harold Beaver (1975; M.36) follows S. Kaplan's general approach, adhering in the main to the 1838 text but introducing over twenty changes.(5) [page 43:]

In England the London branch of Wiley and Putnam rushed the book into print in October 1838, pirating the text from one of the hundred copies that Harper and Brothers sent over. There was no arrangement between the two firms, contrary to the view of John W. Robertson and George Woodberry.(6) Years later, George P. Putnam reported that his father showed Daniel Appleton, recently arrived from Connecticut, a copy of Pym as “an American contribution to geographical science,” which they might “reprint ... for the benefit of Mr. Bull” and of their own pocket, of course. Appleton took a half interest in the venture.(7) Wiley and Putnam saw no reason to preserve the text inviolate; the title was abridged and also recast, with the last events taking place in the 84th parallel (see Fig. 2). To attain greater credibility the publishers omitted the last paragraph about the white giant and inserted a note to the Preface:

As an unnecessary epitome of the last two paragraphs of the Preface and the first three of the “Note” or afterword, it underscores the firm's fear that an American tall story might be summarily rejected by the British, despite its cargo of “.science.” Other minor changes in the text were made, most of them intended to standardize the language for the public: “skewed,” “half-bred Indian,” “paralysed,” and “befel.” Curiously, in “boldly” of 3.8 it anticipated the pointless change of the 1856 edition. The few Harpers' typographical blunders were corrected and the numbers of the final chapters were made sequential, but with the last, extra one dropped in favor of a rule printed between the two chapters. In regard to the cruxes listed in Table 2 below it was the closest to Poe's text of the early reprints and was the basis for the first British editions: the 1841 cheap “Novel Newspaper” version by Cunningham (see [page 44:] Fig, 3) and the adaptation of the first half published for juvenile readers in 1857 by Beeton's Boys' Own Magazine, in twelve installments (see M.3). The latter preluded numerous French adaptations of the whole book for youth. The title of the next British edition signifies a bid for a similar audience: The Wonderful Adventures of Arthur Gordon Pym (1861).

The major reprint of the book in England was that issued by the Poe enthusiast, John H. Ingram, in his four-volume edition of The Works in 1874-1875 (M.6). It appeared in many reprintings and via many different publishers, including a few in New York City. It is strange that Ingram adopted Griswold's text, apparently unexamined, as his own, while vehemently critical of Griswold's orientation to Poe (see Temple Bar, June 1874, 41:375). Of approximately fifty changes and errors introduced, Ingram retained half and corrected half, and introduced two errors of his own in addition to a few British spellings or variants, such as “galvanised” (22.11).

Since the translation by Charles Baudelaire has been the basis for almost all the editions of Pym in foreign tongues — numbering over a hundred in addition to those in French — it is proper to inquire about Baudelaire's source-text. There seems no recognition that a problem exists. In the “definitive” 1934 edition of Baudelaire's works, Jacques Crépet asserts that the 1856 text displays only punctuation variations from that of 1838; yet he is not impelled to compare these two in the face of unreconciled differences between the three texts prepared by Baudelaire, on the one hand, and the two American editions, on the other. Briefly — Baudelaire spoke about translating the book on May 7, 1856, one month after the edition appeared in New York and optimistically hoped to publish the first installment in July 1856 (Crépet, pp. 246-47). He may have been relying, for the first four chapters, upon the set of the Messenger furnished him by William W. Mann, the Paris correspondent of the magazine, or upon an 1838 London or American edition.(8) It is conceivable that by the end of 1856 Baudelaire managed to procure a copy of Griswold's fourth volume, even though it had taken him over two years to procure the first volumes of 1850. There are several phrases in the Moniteur Universel translation (February 25-April [page 45:] 18, 1857) that could come only from the 1838 edition, including the full descriptive title which never appears in Griswold's fourth volume.(9) Yet in the list of cruxes for Baudelaire's versions, the majority seem to stem from the 1856 text; of course, a translation often conceals specific details of origin. Subsequently there were translations by Charles Simond, Armand Juin, and Armand Masson (K.7-9), but it may be assumed that almost all non-French translations, until very recently, have been based upon the sensitive version of Charles Baudelaire.

[page 48:]

END-OF-LINE HYPHENATIONS IN THE 1838 COPY-TEXT

The following compounds, hyphenated or fused, are hyphenated at the end of the line in the copy-text. The form in which they have been transcribed in this text and listed below represents an inference derived from current usage, other appearances, or parallels in this or other texts by Poe.

1.3 mainsail (sic in SLM)

2.4 good-for-nothing

2.8 unlooked-for

2.14 street-robber. (street robber in SLM)

2.16 iron-bound (sic in SLM)

3.1 taper-wax (sic in SLM)

3.3 sea-biscuit (without hyphen in SLM)

3.4 forefinger (sic in 24.10 but with hyphen in SLM)

4.4 bloodthirsty (sic in 20.9)

4.4 hair-like

5.7 forecastle

6.4 shifting-boards (sic in SLM)

6.4 Firefly

6.4 hatchway (sic in 8.7)

6.9 hiding-place (sic in 6.7)

7.5 storm-staysails

7.10 sea-burial

8.12 weather-lanyards

8.14 foremast

8.16 stern-post (sic elsewhere)

9.8 companion-hatch

10.5 bowsprit

10.7 sea-stores

11.2 storeroom (sic elsewhere)

11.14 seaweed

12.7 fellow-sufferers

12.19 slip-knot

13.9 olive-jars

14.1 arm-chests (sic in source)

14.2 tinder-works (sic in source)

14.2 crockery-ware (sic in source)

14.8 southeast

14.12 breakbones (sic in source) [page 49:]

15.4 northeast

15.7 sugar-cane (sic in source)

16.2 northwest

17.2 foresail

19.3 elephantfish (cf. parrotfish in 19.3)

19.6 goodwill

20.12 firearms

21.5 and 7 soapstone

21.7 rock-blaster

22.8 hiding-place (sic in 2.8)

23.1 penthouse (possible, but no basis for preference)

24.3 southeast

END-OF-LINE HYPHENATIONS IN THE PRESENT EDITION

The compounds or possible compounds in the list below are given in the form to which they have been resolved according to the 1838 edition of Pym. The four occurrences of line-end placement in both editions are resolved in the manner indicated in the parenthesis.

Locus

1.1 sloop-fashion

1.4 bilge-water

1.7 whaling-ship

1.8 jolly-boat

1.8 copper-fastened

2.1 somewhat

2.5 somewhat

2.5 bulkheads

2.12 overboard

3.8 ham-skin

4.3 forehead

4.4 fur-trader

4.6 ship-furniture

5.7 readjust

6.2 hogsheads

6.4 hatchway (sic in 8.7)

6.13 hiding-place

7.5 foresail

8.13 foremast

9.1 starboard

9.8 companion-hatch (sic in 12.13,13.3)

10.4 thanksgiving [page 50:]

13.6 main-chains

13.3 northwesterly

13.15 rebound

13.18 bedclothes

14.1 arm-chests (sic in the source)

14.2 tinder-works (sic in the source)

14.3 twenty-ninth

14.5 contradistinction

14.11 seagulls

14.18 well-regulated

17.10 masthead

20.9 leavestaking

20.11 seaman

22.4 flatboats

23 bis.l overarching

23 bis.3 soapstone

24.7 coat-pocket

24.14 pathway

Note.8 swift-flying

THE TREATMENT OF ERRORS IN THE COPY-TEXT

Of the nine clear or probable errors in the text of 1838, five involving punctuation have been silently corrected from the following original forms:

seventeenth eighteenth (12.9)

28 N., (13.23)

13 W., (13.23) 31 W., (14.3)

frame, Their (18.3; discussed in 18.3A, para. 2)

The remaining ones are discussed in the commentary notes as indicated:

followed (3.411)

weigh (2.911)

ricochêting (22.5A)

scarely (23 bis.7A)

No other changes have been editorially introduced. Unusual but allowed forms, such as “villany,” “stupified,” and “lilach,” are discussed in the related commentary notes. All variants are recorded for both substantives and accidentals. Inconsistent forms are retained, unchanged.

[[Footnotes]]

[The following footnote appears at the bottom of page 37:]

1.  For the letter see David K. Jackson, Poe and the Southern Literary Messenger (Richmond, 1934), pp. 109-110. Poe's editorial reference in the Broadway Journal, 1:183, to his “secession” is scarcely credible. See, however, the ambiguous matter of Poe's leaving to accept Dr. Hawks' invitation to work for the New York Review above (p. 24).

[The following footnote appears at the bottom of page 38:]

2.  Cited without any source only by Mary E. Phillips, Edgar Allan Poe (Philadelphia, 1926), 1:547. See also White's letter to Tucker of January 24, 1837, expressing contempt for Pym, cited by William Doyle Hull, “A Canon of the Critical Works of Poe” (Ph. D. diss., University of Virginia, Charlottesville, 1941), p. 65.

3.  For the Knickerbocker notice, 9:520, see J. V. Ridgely and I. S. Haverstick, TSLL, Spring 1966, 7:68, and Alexander Hammond, ATQ, Winter 1978, 37:9-20. Through the courtesy of the Manuscript Collection of Columbia University, I have checked all the Harper archives, including Demarest's contract ledgers, from 1817 on and all the notes left by Eugene Exman, only to find no trace of the Pym contract.

[The following footnote appears at the bottom of page 40:]

4.  For details of these and other editions (and translations) of Pym, numbering 320 items, see my bibliography in ATQ, Winter 1978, 37:93-110. These two are items BB.2 and 4. Parenthetical references to this list are made merely by capital letters and numerals hereafter.

[The following footnote appears at the bottom of page 42, running to the bottom of page 43:]

5.  H. Beaver claims (p. 32) to follow the 1838 edition “as edited by ... Griswold [page 43:] without substantive amendment of either spelling or punctuation” — a rather paradoxical copy-text.

[The following footnotes appear at the bottom of page 43:]

6.  Robertson, Bibliography of ... Poe (San Francisco, 1934), p. 170, and Works (1914), 4:349-50.

7.  For a brief discussion of Putnam's memories about Pym, both true and false, see Studies in American Fiction, Spring 1974, 2:46-49.

[The following footnotes appear at the bottom of page 44:]

8.  In this set Baudelaire must have found John Daniel's March 1850 SLM article on Poe which he translated and presented as his own in the Revue de Paris (March and April 1852). The included death-ship sequence (10.3-6) has wrongly implied Baudelaire's using the whole of Pym at the time. His relying on Daniel has been noted in the Yale French Review, 1953, pp. 65-69, and in Seven Tales by ... Poe (New York, 1971), Introduction, by W. T. Bandy.

[The following footnotes appear at the bottom of page 45:]

9.  This last point is noted by Le Dantec in his Pléiade edition of Poe's works, Oeuvres en prose, p. 1100, with reference to the use of the 1838 edition, but he fails to collate variants for proof that Griswold's text was also used. W. T. Bandy has given me proof that the full title was used also when it was published in Le Moniteur Universel of Paris in February, March and April 1857 (26 installments). It is noteworthy that this at once publicized the work throughout Europe so that the Mechitarist Friars' Publishing House of Vienna later in 1857 published an abridged version (96 pp.) translated into Armenian by Father E. Chakejian, based on the French version. This book presumably to be used as a monastic schoolboy text or, possibly, as an example of an interesting book of modern exploration — one of the favorite areas for the Friars' published translations-was therefore the first foreign version of Pym in book-format, preceding Baudelaire's volume of 1858 (see ATF, 1978, 37:96-97).