Text: Burton R. Pollin, “Introduction,” Word Index to Poe’s Fiction, (1982), pp. i-xiv (This material is protected by copyright)


[page i:]


This volume was developed in response to an urgent need for an instrument to authenticate unsigned texts of Poe that he could have inserted into the journals that he edited during his energetic editorial career: Southern Literary Messenger, Burton’s Gentleman’s Magazine, Graham’s Magazine, The Mirror, and the Broadway Journal. In addition it can and, I trust, will serve many other purposes for students of the words of Edgar Allan Poe: the analysis of his style and vocabulary, his characteristic and underlying concepts, his broad range of allusion, and his language transformations. There was an obvious need for such a key to the integrated body of his fiction, to accompany the 1941 Bradford Booth Claude Jones concordance to the very small corpus of his poetry. The delay in preparing any concordance to Poe’s prose could be explained largely in terms of the extended period of waiting for a definitive edition of his fiction, long promised and long prepared by the late Professor T. O. Mabbott. Surely only Poe’s tales could comprise a truly significant and well defined corpus for such a source. The year 1978 saw the appearance of the Harvard edition, in two volumes, of the great bulk of his tales, posthumously completed through the admirable efforts of Maureen C. Mabbott and Eleanor D. Kewer. Proud to be involved in those volumes, as the Acknowledgments (I, v) indicate, I long cherished the idea of executing such a project upon the completion of the editing of the rest of Poe’s fiction (Arthur Gordon Pym, “Hans Pfaall,” and “Julius Rodman”). With the publication of this “third” volume of his works, under the imprint of G.K. Hall-Twayne, in August 1981, I looked forward to the editing of the rest of Poe’s prose works, but without the necessary concordance tool for authenticating his vocabulary and his stylistic and rhetorical usages. My own Dictionary of Names and Titles in Poe’s Collected Works (Da Capo, 1968) was useful for allusions, but it had to omit new works and variant forms not in the Harrison edition — the stipulated source. Likewise, my labors in the realm of “exceptional” or “creative” Poe vocabulary had served to produce Poe, Creator of Words (Baltimore: Enoch Pratt Library and Baltimore Poe Society, 1974; augmented, rev. ed. Bronxville: Nicholas Smith, 1981). But this was of limited scope compared to all of Poe’s non-created but very creative words in his full texts. (The number of words in his fiction, my computer printout shows, is 468, 688, which merge into a total of 25,907 entries or indexed words in the main text which follows.)

Before I could be definitive about the unsigned “Marginalia” and “Pinakidia” type of paragraphs strewn, I suspected, by Poe in the Messenger or other magazines, I needed to consult a concordance; similarly, for authenticating as Poe’s numerous unsigned and unattributed reviews, editorial miscellanies, obiter dicta, incidental remarks, footnotes, etc. in the Broadway Journal. Many difficulties of a practical nature immediately loomed: The text required three large volumes published by Harvard and Twayne, but a concordance furnishing the context of every word (or all [page ii:] save the high frequency auxiliary-type of words such as “the” or “is” or “and”) would require even more space than the three volumes. Who would publish the fruits of such an enterprise? Moreover, how long would hand-culling such a collection of words and their context require? — many years, at the least. Two major solutions at once sprang out of the grim necessities of publishing in these adverse times and out of my own experience: since the texts might be assumed to have been defined in the three volumes, why not produce a “word index” referring to the context in those volumes thereby saving a great deal of space? The index to Finnegans Wake by Clive Hart even calls itself a “concordance” although not giving the context to the reader directly. There are also other examples of concordance-locations without contexts: R. Garcia and J. Karabatosos, Concordance to the Short Fiction of D. H. Lawrence (1972), Leslie Hancock, Word Index to Joyce’s Portrait of the Artist (1967), and Gary Lane, Word Index to ... Dubliners (1971). I was not unfamiliar with concordance making, since Alice M. Pollin, my wife, had compiled concordances to the poetry of Eugenio Florit (New York University, 1968) and the plays and poems of Federico Garcia Lorca (Cornell Concordance Series, 1975). By my referring to but not giving the context (in the three volumes) a one-volume Index could be produced, if the proper format were devised.

The second aspect of the solution was to computerize the entire project, or at least the results of the editing and planning process for the sake of speed and potential accuracy conferred by the modern development of Babbage’s Calculating Machine, as Poe called it in “Scheherazade” with prescience and admiration (B1166/ 35). But several conditions had to be met: availability of expensive computer machine time, of a programmer, of human “inputters” who could type the data “onto” the disk and, above all, of a publisher willing to issue a large and possibly expensive book. In June 1981 these conditions were miraculously fulfilled, when I found myself holding final page proofs of the third volume of Poe’s fiction: called Collected Writings of Edgar Allan Poe: Volume I, The Imaginary Voyages (Boston: Twayne Publishers). Aware that the Gordian Press of New York, who had recently reissued Prescott’s Selections ... of Poe, also put out the reprint of the Ostrom Letters of ... Poe, I proposed a one-volume “Word Index to Poe’s Fiction.” The publisher, Mr. John Corta, liked the idea greatly. In rapid succession came the approval of Dr. Leon F. Landowitz, Director of the Computer Facilities of the Graduate Center of CUNY (42nd Street, New York City) and of Dr. Richard A. Styskal, Associate Dean of the Graduate School, CUNY. Dr. Landowitz assured me of the help of Mr. Larry Powell, of his organization, and of his staff. Eventually to be involved were also the work area and printer of the main computing facility of CUNY at 555 West 57th Street, New York City, where two and eventually three of the data inputters and I worked straight through the summer of 1981. The Hunter College Computer Facility also helped by [page iii:] providing me with a terminal where an efficient inputter spent many a night hour collecting data into the files of the project. (See the Acknowledgments at the end for more specific attributions.)

Various aspects of this Index might seem arbitrary and perhaps even incomprehensible if there were no description of the stages and circumstances of the development of the collection of the raw data for the text which had to be programed into the Index and its appendices. My three previous computer-aided texts (Godwin Criticism, the Poe Dictionary, and Music for Shelley’s Poetry) first involved the laborious key-punching of the data on IBM cards, with the consequent handling of these crude and precarious data-records. The development of word-processing instruments using a console screen set before the typewriter-like keyboard, with data files on central and safely stored disks, has greatly quickened and facilitated the collection and correction of considerable quantities of data — especially if one is willing to work almost around the clock. Having access to various terminals, for example, means that material can be inputted or stored and corrected during the night often far from the place where the central files or computer printers are kept. Thanks to the day and night functioning of the total computer system of the City University of New York, to the telephone-line linkages, and to accessibility to the central storage system, this Index was processed in all its stages in about six months. Almost incessant editing, proofreading, compilation of lists for correction, and consultation with the Graduate Center personnel were necessary. During the last three months, this Center chanced to install its own upper and lower case printer, so that time-consuming, inconvenient trips to the main facility (at 57th Street and 10th Avenue) were obviated, save for the final printout.

The first stage was the editing of the text of the three volumes (termed A, B, and C) after the Harvard University Press and the Twayne Publishers had both graciously granted their permission. My inserted markings next to and between the lines told the data inputters how to “type up” all the numbered lines. However. instead of numbers we used “T” and “M” and “V” for the words in Poe’s titles and mottoes and variants, the last being at the foot of the page in “A” and “B”. Ultimately, the locations would appear in this form: A0071/ 31 or B1137/ 28 or B806/ M, meaning Vol. A, p. 71, 1. 31; Vol. B, p. 1137, 1.28; Vol. B, p. 806, Motto. The fill-in zeroes reflect the fact that these locations were set by fixed columns at the head of the line, being extracted by the program through their position. While they could have been erased in the final printout text, before photographic copying of the text for publication, the uniformity of the columns improved both the general appearance and reading ease. With regard to copying out the text, we generally aimed at literal faithfulness, but only certain capitals and italics were to be preserved, even though both underlining and capitals were available to us in the CUNY printchain (first at [page iv:] 57th Street, later at 42nd Street). Proper nouns and the “I” (first person singular) of course retained the capital. There was a question about the numerous capitals used by Poe for rhetorical effect (Nature, Art, Death). Since almost every common noun that was occasionally capitalized would produce a separate entry in the list, there might be a considerable danger of swelling the number of entries beyond the scope of one volume or making it too bulky. Then too, the capitals might have been the reflection of an early editor’s preference, not of Poe’s deliberate wish as some of the variants might imply. Finally, the capitals in the titles, while deliberate on Poe’s part, sometimes doubled several lower-case words to no great advantage, especially since a word in the title of a tale automatically exposed its location through the addition of “T” to the page number. This would not prevail, to be sure, for titles of other persons’ works cited in the text by Poe, and some of these were arbitrarily lower-cased in my markings.

Finally, there was the large problem of capitals used to indicate the beginning of a sentence. Naturally, there was no purpose in signifying this purely grammatical function through a capitalized word (save for proper nouns which started sentences). All of these had to be lower-cased. On the other hand, it was my aim to cater to the habits of ordinary typists who knew the language of Wylbur sufficiently to input the text and make simple changes, such as corrections and insertions, and to recall lines from their disk numbers for subsequent alterations and corrections. Surely, I reasoned, it would be easier and quicker to type discursive prose with end punctuation uniformly reduced to periods. The programmer assured me that he would reduce initial capitals through his program and that we could add a special delimiter before every proper name (or I, first person) for retention of the capital. We therefore inserted a dollar sign ($) before all of these. The procedure was a mistake, I know now, for many of the dollar signs were inadvertently omitted by the typists; five months later I was still finding lower case words needing a restoration of their initial capital. On the other hand one of the typists (for a changing “crew” of four eventually produced the entire input text) casually lower-cased initial first person “I” and created a file of “i” which proved almost incorrigible in the long run. We also had enormous difficulties with capitals at the beginning of lines on the page through a gross misunderstanding of whether the program for “reading” final periods could operate from one line over to the next; this produced the need, eventually, for thousands of correctional changes through retrieval at the console, when it was found that the output Index contained numerous doublets of common words. We also found that the program for lower-casing unwanted capitals after periods was producing much trouble with initials in full proper names, including those enclosed in special symbols for reverse alphabetization (see below) and in abbreviations. One of the traces of this procedure is the presence in my Index of short abbreviations [page v:] printed without their period, such as “Mr”, “Mrs”, “Capt” and “Lieut”. In fact, we found it necessary eventually to devise a set of double delimeters (“plus” signs) around any words or abbreviations containing periods that we wished to keep in the index itself (see CHANGES ... below).

Another editing necessity concerned italics, which Poe employs with a great deal of freedom. Sometimes these are used for the titles of books, articles and quotations; often for emphasized words or phrases, but sometimes the title is in quotation marks and the author’s name is italicized; scientific names, in botany, for example, are inconsistently italicized. Since an italicized English word would often produce a doublet of an ordinary word, hence a new entry, I felt it better to keep doublets under one rubric or entry-name by canceling the italics, and the general rule that I followed was to use an underlining (italics in our Index) for all quoted foreign terms save when they were interjected into a passage of English prose. This led to few italics for the many French interjections of Madame Lalande, speaking passages of mongrel English in “The Spectacles.” Poe’s word coinages, as I indicate in Poe, Creator of Words (p. 16), were respectfully retained in almost every case, along with a few special words used for rhetorical emphasis. The method for designating italics in the inputting was somewhat awkward; since we early assumed that many whole phrases would need italics, we felt it best to enclose these phrases in two delimiters, to trigger the program into underlining all the individual words. The sign chosen was the percent sign, useless in our text for any other purpose. In practice this led to considerable trouble, largely through the machine’s perplexity whenever a space erroneously intervened between the word and the sign or whenever one of the two signs had been omitted. Yet typing-in a tagging symbol attached to each word would have led to equivalent trouble. The treatment of Greek words and of dialectal non-standard English forms also required special symbols or handling. Brackets eventually designated (for the program) all Greek words, which were printed out in the index with a “(Gr.)” after each of the separate words, while the created dialect or distorted forms in “Why the Little Frenchman Wears his Hand in a Sling” were designated by an increase sign (>). The computer cannot print accent marks for foreign tongues, all of which have to be dropped. This means also that homonyms which would be differentiated to the eye, such as the English article “a” and French “a” are con-corded together. In this case all of them are merged into one high-frequency rubric. The presence of homonyms in the Index must be realized. The user must be aware that “chest (17)” includes the word for “box” and also for “part of the body.”

From the beginning of our inputting the data (in July 1981) it seemed to me necessary to keep names meaningfully together. What would be the use of having the name of Martin Luther concorded under Martin and Luther? Hence the programmer and I resolved this in terms of a pair of delimeters that would enable the [page vi:] machine to extract the names and turn them around to index them on the surname, thus: “Luther, Martin”. This was done through the insertion of the signs for equals and plus before the name in the Index and a plus after the name: =+Martin Luther+ to be printed out in the index as “Luther, Martin”. Even the names of characters, I judged, would often be most useful to the reader, under the surname (cf. Miller), rather than as separate entities for the first and sometimes the middle names or initials. In all these cases the names were preserved as phrases. This led to some difficulties, to be sure. For example: “Ferdinand Fitzfossilus Feltspar” could not be printed out in the space allotted to the entries. In a few cases I had to revise Poe’s text slightly: e.g., this name became “Ferdinand F. Feltspar,” to fit into the space. The second name of “Napoleon Bonaparte Froissart” became “B.” This need for reduction applies to a few “words” humorously coined, such as “Alexander-the-Great-o-nopolis” (29 characters long), which became “Alexander-etc.-nopolis” — scarcely a desirable stratagem. In general, once our program functioned, the system proved feasible and, I trust, helpful to the reader. Originally I intended to keep all other meaningful phrases together, such as names of newspapers (New York Sun) or Poe’s titles (“MS. Found in a Bottle”) or special phrases often in foreign tongues (“comme it faut”). To do this required putting the words of an entire phrase on one line, that is ascribing the words entirely to the top line, on which the phrase began, rather than to two lines as in the book, since the index could properly print only one line number for each location. Moreover, the program for the Index could read each line as one unit but could not leap over from one line to the next. I spent the months of July and August 1981 editing the text into phrases and having the typists input it accordingly. This often entailed shifting a few words ending a phrase into the preceding or “upper” line to give the phrase only one line number, as was necessary according to space allowance. Then it appeared clear to me that I was really negating the basic nature of the entire “Word Index” by making it so often into a “phrase index.” For one thing it would be almost impossible to find a specific word buried in a long or even short phrase. For another thing, the programmer found it impracticable to format the text to accept phrases of more than 25 characters in view of the need for leaving space for a frequency parenthesis plus 8 spaces for at least one location on the same line along with spaces between the separate sections of the entry and its accompaniments, all of which was limited to only 42 spaces for each half of the page. Although originally we intended to have entries of two-lines, if need be, we finally ruled this plan out. I therefore had to cut back all the preedited and “inputted” lines to single word lengths. A large number of errors, however, would have ensued if we had shifted back to their original pages the many words at the end of phrases that had been transferred or rather given the preceding line number. Hence a certain number of indexed words will be “off” by one line number in the text from their [page vii:] stipulated location in the Index. For example, the phrase “Glory of the East” was given in the text as “Glory of the” on A0128/ 7 and “East” on A0128/ 8. I had first edited it into +Glory of the East+ to be printed as a phrase and ascribed to A0128/7. With the abandonment of the plus signs or of the phrase units system, the individual words were concorded, when the plus signs were removed by a program directive. However, the word “East,” having been assigned a new line number (7 in place of 8), retains this number still in my Index, although in the text it occurs at the start of the next line (line 8). This modification must be borne in mind when one consults the texts according to the number given in the Index. In short, if a word is supposed to be on a line and is not there, look at the beginning of the next line for its actual text location, always below the one given in the Index. Of course a very small number of words, almost all capitalized, are thus affected.

In order to preserve the meaningful unity of some short phrases, however, in an Index of individual words, I subsequently and at the end of the project resorted to the device of using a slash or virgule to keep together unitary phrases, such as “St. Louis” or “St. Pierre”. To the computer the slash counts as a regular alphanumeric character, equivalent to a letter or a number. Hence the words on both sides of the slash would be linked together as one unit. For example, “United/ States” as one parameter or word-unit is more significant and useful to the reader than would be the word “United” indexed with all other “United” words and, separately, ‘States.” It is true that the word “States” is eliminated from the Frequency List for capital “States” words, but the comprehensional value seemed to me to outweigh the merits of preserving the count, already somewhat tinctured by the variant forms of the same tale printed in the Harvard edition (see below). Other types of retained short phrases are magazine and newspaper titles (Edinburgh/ Review), a few exclamations (he/ he/ he/, since scattering the phrase would make “he” identical with the personal pronoun, and it would disappear into the high frequency words with no locations designated), proper names having two parts in the surname (De/ L’Omelette, de /-Stael, Le/ Brun), real or putative institutions (Park/ Theatre, States’/ College, North/ West Company, Folio/ Club, Royal/ Geographical/ Society), song titles (Judy/ O’Flannagan, Betty/ Martin), ship’s names (Jane/ Guy, Mary/ Pitts), and two-word unitary concepts (Post/ Office). A slash also prevents the erasure by the program of the period (see above); St. Andrews would be printed out as “St Andrews” and, if taken by itself, “St” would also look like the word for “Street” with the period erased.

There were several dangers in this method of using slashes for short phrases: first, of extending a phrase beyond the limits for the Index entries (25 characters) and second, of inserting too many confusing slashes that might interfere with ready reading. For example, it seemed to me preferable to have “Cape of Good Hope” as [page viii:] one unit, but this would require three slashes. In this instance. I relied on the reader’s intuition by yoking only “Good” and “Hope” together with one slash, producing “Good/ Hope.” In this matter of the slashes or the phrase units, I do not claim total consistency, as in “Rue/ Morgue.” It was indeed my expectation that the users of the Index might approach such phrases in various ways, e.g., through “Morgue” or through “Rue.” Hence, the inconsistency might serve the reader’s exact purposes, provided he eventually directs his attention to the different or possible elements. He must know that cross references are impossible in a computerized list.


Since the “text” used as the source of index words includes significantly different “variants” I must explain how the variants were edited and how differently the Harvard and the Twayne editions print them. All of Poe’s sketches, tales, and long narratives (i.e., Pym and “Julius Rodman”) were “concorded” with the exception of “A Dream” (A/ 6-9, discussed below). The two volumes of the Harvard edition, which were paged continuously, were assigned the letters A and B, while the Twayne edition was designated as C. The line count differs slightly, with 37 lines to a full page in the Harvard and 41 in the Twayne edition, requiring the use of two different line-rulers (provided with this Index). In the Harvard edition the variants are printed at the foot of the page, whereas in the Twayne edition they are all listed after the text of Pym (pp. 211-214) and of “Hans Pfaall” (pp. 436-449). There are none for “Julius Rodman,” and most of those for Pym concern accidentals. Typographical blunders are, of course, omitted, save that occasionally the Mabbott text fails to indicate when a variant is a blunder, as with “handsommest.” In several instances I have allowed this oversight to prevail, especially since contemporary forms may make difficult the matter of determining the allowable usage of the day. In Vol. C, I have discussed more fully in my notes the variations then permitted, but verbal variants (for successive printings) are multifold only for “Hans Pfaall,” not for Pym. Professor Mabbott has reprinted four of the tales which were only partially revised (“Loss of Breath,” “Bon-Bon,” “Lionizing,” and “Morella”); and for these “essential” variants, it would have been impossible to match or trace duplicate words in parallel passages as I had been doing for the footnoted variants. That procedure had involved my elaborately red-lining the footnotes, word by word, so that the inputting typists could copy out those words which were different from the material in the text above. But in these so freely varied four tales, no such checking for parallelism and identity of words could be done. Each tale was therefore concorded as a separate entity. It should be mentioned here that it is easy to find the variant words at the foot of the page save when the variant is embedded in a deleted paragraph as in “Philosophy of Furniture” (A/ 495/ V). In order to make it easier to trace the words in the very long detailed initial paragraph of “The Oval Portrait” (A/ 667), I have added putative line numbers [page ix:] for the 73 lines of the two-columned passage. The paucity of substantive variants for Pym and “Hans Pfaall” greatly eases finding the cited words on the pages totally devoted to variants in the Twayne edition (C).

For reasons given in his Introductions to the several articles, Professor Mabbott included as “Sketches” works of Poe which many would prefer to call “essays” (e.g., “The Folio Club” and “A Reviewer Reviewed” and “Autography”). For our concord-ing purposes these increases in the “data bank” enhance the worth of this Index. But I cannot accept as authentic the dream-story, called “A Dream” in the Philadelphia Saturday Evening Post of August 13, 1831, which Killis Campbell in 1917 suspected might be Poe’s work (p. A/ 5); and even Professor Mabbott cautiously states: “I feel a text of the story should be presented, with a caveat, as tentatively assigned to Poe.” There are really no positive reasons for ascribing it to Poe, who never claimed or even mentioned it. Were I to include all of its words in my list I should negate a major purpose of the Index, that is, to test such passages against the totality of Poe’s vocabulary. On these grounds, there are reasons to doubt Poe’s authorship very strongly. He does not use any of the following words in his fiction (and only five in his poems): physiologist (7/ 1); Nazarene (7/ 10) gurgle (7/ 15); Judaea (7/ 25); beauteous (7/ 28); crape (7/ 35); requiem (8/ 12); expiatory (8/ 11); candlestick (8/ 15); babe (8/ 22); matted (9/ 15); enrobed (8/ 26); light-shot (8/ 30); amaranth (9/ 3); grave-worm (9/ 14); parody (9/ 17). Since many of the situations in Poe concern death, torment, and guilt the absence of these connotative words among the others is a strong reason to reject the article. In addition, it is full of mawkish, sentimental, stereotyped, and fustian phrases that are inexpressive, tasteless, and unoriginal: p. 6 — “the slumbers of night”; the “God of Nature”; “the wild vagaries of the imagination”; 7 — “the bitterest woe that mortality ever felt”; “the mantle of clay”; life “shuddered to walk along”; “the chain of existence was broken, and a link dropped to eternity”; “a feeling of conscious pride ... this pride of the eastern world”; “the skirts of mortal vision”; “a perfect loveliness had thrown itself over animated nature”; “a mantle of crape” over the sun; each star “had ceased to twinkle”; “enveloped in the badge of mourning”; 8 — “they sang the hoarse requiem”; “the spirit of darkness spread his pinions”; “I could feel the flood of life slowly rolling back to its fountain”; “twas lost in utter darkness”; “the living fire of the candlestick”; “‘ twas still as the sepulchre”; “Nature mourned”; “the habiliments of sorrow” over the earth; “the sables of mourning”; “a column of light shot athwart the gloom”; “still awfully plain and distinct”; 9 — “amaranth which was wont to circle his brow”; “terror had tied up volition.” Surely this does not “eminate” (a contained spelling Poe never used) from Poe, who was even then writing those matchless satires and burlesques of the early tales of the Folio Club. It is not difficult to oust this from the canon where it is only “tentatively” assigned by Professors Campbell and Mabbott. [page x:]

The separate works which must be considered as sources of concordable words are eighty-three in the Harvard edition and three in the Twayne, two of which are novels (with 25 and 6 chapters). Several of the “sketches” in the Harvard Edition (volumes A and B) are brief newspaper pieces or short essays included for their fictive material (e.g., “Philosophy of Furniture”), but only the four-page “Dream” is excluded from this Index.


I have tried to be as scrupulous as possible in adhering to the “copytext” provided by the Harvard and Twayne editions of Poe’s fiction, even though this has produced some startling inconsistencies that look like errors in the lists of words that follow. Poe changed his spelling or punctuation habits from one period to another, thereby producing variations. It must be admitted, however, that he often wrote and published in a hurry; the evidence lies here. Even when one of his forms is incorrect, I have adhered to the text, although in every case I have checked into the source materials used by T. O. Mabbott or myself, to see what was actually printed. Very few changes have been needed: (A210/ 14) “wonderful” (“r” omitted); and “preeminently” (as in Griswold) for “preeminently” (A302/ 15). In one respect, however, I have had to supply an omission in the Harvard edition, the lack of a list of end-of-line hyphenations, for there was no way to verify whether the words with hyphens at the end of the lines coincided in spelling with the copytext being used save by going back to the source. It was necessary for me to indicate to my typists whether a word was to be inputted as a fused word or a hyphenated compound in every one of these. In a few cases, the text used by T. O. Mabbott itself showed a coinciding hyphen; for these, I consulted earlier texts and sought the most reasonable consensus according to the standard procedures in such cruxes. I can therefore claim that if one needs to know about hyphenations for the sake of exact quotations from the Harvard text, these indecisive appearances can be settled by consulting the form listed in my Index.

Some few changes in the cited texts were necessitated by the nature of the computer instrument and by the formatting of the text, in the two sets of four double columns, aiming to convey much, in little space. I have spoken about words that would not fit into the 25-character entry-span: others were “cock-a-doodle-de-d000000h” (B1021/ 12), “good-for-nothing-to-nobody” (B1372/ 2, B1374/ 15), and the two instances of “Alexander-the-great-onopolis” (B1369/ 19 and 20). These are all obvious to the reader, in any event. The oddities of the machine’s handling of alphabetization required some spelling changes. For one thing, the machine regards apostrophes as being of lower (or prior) order than any letter, so that in alphabetizing material, it inserts anything beginning with an apostrophe at the head of an alphabetized list: thus poeticisms or dialectical forms such as “’sembled” for resembled or “‘tis” and “‘twas” for “it is” and “it was” would begin the whole Index [page xi:] before letter “A” in a most confusing fashion. Similarly, Italian words, such as “ ‘huomo” and “‘io” go to the head of the list. My solution was to deprive eight such words of their apostrophe, have them appear in the order of their first letter (such as ‘t’ for “ ‘tis”) and afterwards add the apostrophe by hand. another problem was with the machine’s placing capitals after a lower-case letter. The programmer circumvented this process for a single capital letter beginning a doublet, such as “god” and “God,” but not for a capitalized digraph in words taken from the Greek. Hence, the name “AEschylus” in the Harvard text twice had to be inputted as “Aeschylus” to avoid its appearing after the word “azure” in the Index. It would have been desirable to circumvent the machine for other marks of punctuation, such as the apostrophe and the hyphen, but too much programming would have been involved; hence the user of the Index must remember, for example, that apostrophes and hyphens appear before letters, in spelling. The letter D file, for example, reads thus (each entry being separated by a semicolon in my list here with “skips” not being marked): D; d; d; D.C.; D.U.K.; D — ; d — d; d’autre; d’un; D’Alger; D’Avisson; dabble; dagger, etc. There is machine logic to this parade of forms and the pattern can be quickly grasped by the reader. Occasionally I substituted a full name for a letter designating a person previously mentioned (e.g., A735/ 36, “D.” became “Deluc,” a character in “Marie Roget”). This was to avoid multiplying the capital letters at the beginning and to render clear the meaning of the letter, which was, to Poe, a mere journalistic shorthand tolerated by editorial practice at the time. I did not do this consistently; hence, for example, the many instances of capital “P” ranging between B1032/ 20 to 1032/27. All represent “Mr. Poe,” who is called “myself” in “Mesmeric Revelation” (B1032/ 17). I introduced a few other changes in the text designed to aid the inputting or the reading by users. There is no degree sign on the computer printer; yet in Poe’s many nautical and geographical references this symbol was important. The inputters therefore used an “at” sign (@) for every degree sign, and these were programmed to appear as the word “degree.” Of course they became merged as high frequency entries with the word degree or degrees as used by Poe, so that we find a total number of 132 instances of “degree” and 134 of “degrees.” The dash presented its own problems since it was a mark of punctuation often at the end of a word, and by a sweeping program we had agreed to cancel end-punctuation early in the data-inputting. Hence, the inputters typed in two asterisks for Poe’s numerous dashes, these to be eventually converted in the Index itself to two hyphens, serving as a dash. Often, however, the asterisk was confused by one of my input typists as a plus, with disastrous results, in view of the use made of plus signs to enclose names and set phrases; hence I cancelled the period and the asterisks in several instances of “ — ult” (see A265: “your letter of the — ult.”) and there are 7 of the “ult” in the list. The period at the end of the word always created trouble, in view of the program for [page xii:] erasing periods at the end of sentences, and I decided to allow several common abbreviations to stand without any period: “Mr” and “Mrs”, “M” (for Monsieur), and “Capt”. It was theoretically possibly to “save” the punctuation at the end of a word, by inputting a pair of plus symbols around the word outside the period. Points of the compass, such as N. and S./ E., “No.” for “number,” “etc.” and “A.M.” were thus saved for selection. This proved extremely awkward for typing and led also to doublets of many abbreviations (such as Mr and Mr. or MSS and MSS.) which then had to be reconciled. Unfortunately, all corrections proved ultimately tedious and costly, to be made only on the input or original data, since we did not save a file of concorded output, to be corrected by itself. There were good reasons for this procedure, however, such as the fact that changes in any part of the index would also change frequencies and possibly the form and order of compound words, and both these categories formed separate appendices depending upon the original text for extraction and printing as a formatted string of entries.

There may be apparent errors, and despite a six month’s nightmare of round-the-clock proofreading, I feel certain that some real ones are present, but it is to be noted that Poe and the editors of the copytext which was chiefly used (the Griswold or Redfield edition) allowed many errors and inconsistencies to persist, which I hesitate to change. For example, in “X-ing a Paragrab” the joke lies in substituting an “x” for every “o” in the printed copy. But B1374/ 8 shows “your” as “yzur” rather than “yxur” as in the Griswold copytext. Neither T.O. Mabbott nor I have changed the “z” to an “x”. Nor have I changed Poe’s error of “Coleridegy” in A495/ V.

Poe’s misconceptions, errors, and inconsistencies are not to be corrected, I feel. Here are a few of the “variants” or alternate forms: “mavourneen” and “mavoureen”; “Capricornuto” and “Capricornutti”; “mizen-mast” and “mizzen-mast”; “bouquets” and “boquets”; “Bordeau” and Bourdeau”; “Edinburg” and “Edinburgh”; “innuendo” and “inuendoes”; “villany” and “villainy”; “De Stael,” “de Stael,” and “Stael.” While trying to preserve such distinctions, I have had to make changes in both the capitals and the italics, for many reasons. Accepting Poe’s arbitrary use of capitals in titles and for rhetorical effects might have threatened to swell the number of entries beyond the permissible size of one volume. This is also true of many instances of italics, sometimes corresponding to no usage of our day and possibly not the norms of his own (e.g., A18/ M: the author, not the Latin text, is in italics). It seemed most useful to take one of Poe’s most common reasons for using italics as my own fundamental method, serving as a way to indicate a foreign quotation to the reader; the Latin, French, German, Spanish, Italian words are therefore in italics, save when they are briefly interjected into a passage of English, with no sense of being quoted (see “The Spectacles”). Also, Poe’s regular habit, of italicizing what he thought to be his own word coinages, led me to italicize such words. Titles, unless in [page xiii:] a foreign tongue, of magazines and books are not italicized, nor do I represent by italics words and phrases in quotation marks. In editorially inputting both capitals and italics, I must claim a large area of discretion, however; likewise, in dropping the period in such abbreviations as these: Capt (Captain), Lieut, M (Monsieur), Mr, Mrs, and viz (viz.).


The frequency cut-off number proved one of the most crucial aspects of the entire Index, for the figure chosen and used in programming would determine the number of entries shown with all their locations, hence the length of the entire text. We considered as major determinants both the desirable size of the pages and the number of pages that could reasonably be contained within one volume in a book intended as a widely distributed tool for Poe studies. Our first cut-off was at the frequency of 41 occurrences, but Mr. Powell, the programmer, then had the happy thought of setting his program at 51, a felicitous conjectural thrust at half-a-hundred-plus-one. I kept in mind the virtual impossibility of a reader’s looking up more than fifty locations in the three source-volumes in the course of fulfilling a research project. If this were a concordance with the texts of each entry exhibited, the cut-off point might be in the hundreds (omitting the auxiliary verbs, conjunctions, prepositions, etc.) Until the costliness of such a long text (undoubtedly in several volumes) can be overcome, perhaps by using microforms, this Word Index may make its own contribution to studies of Poe’s style and vocabulary. Another partial solution to the lack of a text showing the multiple instances of interesting words beyond the level of 50 (such as “natural” and “opened” and “wilderness”) may be the production of a printout text with the program set for 200 or even 300 instances. This printout, perhaps produced also in multiple carbon copies, can be bound in cardboard folders and placed in several centers for Poe materials around the country, for the sake of students who may wish to look up the more than 50 uses of some of Poe’s words. Such centers might be in New York City, Baltimore or Washington, Bloomington, Iowa City, Austin, and Pasadena. This process will be subsequent to the publication of the bound volume. Similarly, the entire text of the Word Index has been placed on magnetic tape or disk pack, to be retained by the editor or the Graduate Center, City University of New York, for access in the future by data bank projects personnel, by Poe scholars, or by other institutions, who might even wish duplicates. There are many possibilities suggested by my own memory of the data-bank usage in Hitotsubashi University, Japan, of the tape of my computer-aided Godwin Criticism and in Oxford University of my Music for Shelley’s Poetry.

It should be noted that the book was inputted through the Wylbur on-line text-editing system at terminals in various locations in New York City. The computer was an IBM 3033 mainframe with a Digital Data Printer, upper and lower case, at the [page xiv:] Graduate Center, CUNY, with the final printing executed at the CUNY Computer Center at 57th Street, New York City.





[S:0 - WIPF, 1982] - Edgar Allan Poe Society of Baltimore - Works-Word Index to Poe's Fiction (B. R. Pollin) (Introduction)