 | |  |
Notes
This paper is based on the preface to my forthcoming A Concordance to the Poems of Emily Dickinson (Ithaca, N.Y.: Cornell University Press).
"Problems in the Making of Computer Concordances," XV, 1-14. See also the next paper in this volume (pp. 15-31) "Electronic Computers and Elizabethan Texts" by Ephim G. Fogel.
The distinction between a word-index and a concordance is often blurred—as in the definition of a concordance given by Webster's Third New International Dictionary: "an alphabetical verbal index showing the places in the text of a book or in the works of an author where each principal word may be found often with its immediate context. . . ." The last part of the definition equivocates on the key issues; the definition also fails to note the existence of concordances to languages. For a clear distinction between word-indexes and concordances see Roberto Busa's introduction to Varia Specimina Concordantiarum of Aquinas's liturgical hymns (Milan, 1951). According to Father Busa a word-index is "a list which gives for each entry the numerical listing of quotations only" whereas "when under each word all the lines that contain this word are transcribed, one by one, we have the 'concordance'" (p. 8).
The most famous example of this is to be found in "I taste a liquor never brewed—" (poem #214) in which the variant last line, "Leaning against the—Sun—" adds a perfectly appropriate concluding image that is absent from the original line, "From Manzanilla come!" which, in effect, adds only a rum name.
Johnson's own one-volume edition, The Complete Poems of Emily Dickinson (1960), unfortunately also does this. In his preface Johnson says he has adopted only those variants underlined by the poet, but his practice here is inconsistent.
Edith Perry Stamm, for example, in her article "Emily Dickinson: Poetry and Punctuation," Saturday Review, March 30, 1963, pp. 26-27, 74, criticizes Johnson for not recognizing that the dashes are elocution marks to guide the reciter of the poems. Her theory is easily disposed of simply by trying to read facsimiles of the poems according to these supposed marks.
As a supplement to Johnson's edition a facsimile edition of at least the best poems would be welcome. With the current techniques of electroprinting, however, the individual student of Emily Dickinson can make a good beginning for himself by reproducing the facsimiles in Johnson's edition, in Bolts of Melody, and in Charles R. Anderson's Emily Dickinson's Poetry (N.Y. 1960).
The ordinary print wheels used by the IBM "printer" consist only of periods, parentheses, hyphens, and commas for punctuation marks. The Arnold concordance, which was printed before the costly print wheels could be purchased, was done without punctuation, but the Yeats concordance is fully punctuated.
But variant spellings that perhaps indicated different words—"straight" and "strait", for example—were included.
Three exceptions were made in applying this principle. Published variants for poems #59 and #160 were included because they appear to derive from manuscripts now lost. And the published second stanza of #57 was retained because it poetically complements the stanza of a poem whose manuscript had been torn off at the top. Poems in Emily Dickinson's canon that exist only in their published forms were, of course, included as such.
Johnson often notes in his edition the nature and location of variants in another version of the poem—but not always. Consequently it was necessary in preparing the concordance to collate all variant versions with the main text.
This classification of variants had to ignore whether the variants were from single worksheet or semifinal drafts; from worksheet, semifinal, and fair copies of the same poem; or from differing fair copies, semifinal drafts, and worksheet drafts. To indicate the provenance of each variant in the concordance would have entailed a bewildering system of signs; the only sensible alternative seemed to be to combine them, regardless of their sources, and to refer the user of the concordance to Johnson's edition for their origins.
One result of combining variants was that words repeated in different manuscripts were not repeated as such in the concordance. The concordance could not represent within the limits of a single line the stages of composition beyond the including of different variants after the word used in the earliest finished version. Thus, for example, in poem #329 ("So glad we are—a Stranger'd deem") there are two manuscripts, the fair copy represented in large type and a semifinal draft from which it was redacted. In the semifinal draft the last line of the poem, "Could not decide between—", is given as "Could not discern between—" and then variant choices for "discern" are listed as "conclude" and "decide". In giving this line in the concordance as COULD NOT DECIDE [DISCERN / CONCLUDE] BETWEEN it seemed unnecessary to repeat "decide" in the brackets to indicate that the word had also been a variant choice at one stage of the poem's composition.
An example of a variant in lining not handled in the concordance also occurs in this line—or in these lines, to be exact. In the worksheet draft of the poem the line is given as two lines: "I perish to behold" and "Another such a might". If the lining of the worksheet draft were kept in the fair copy, the variants would have been handled as variant lines. But as the concordance has to follow the lining of the main texts in Johnson's edition, what was a variant line in an earlier draft is a variant phrase in the fair copy and in the concordance.
The eccentricities of Emily Dickinson's punctuation were, of course, followed in the concordance.
See, for example, poem #533 ("Two Butterflies went out at Noon—") where a fair copy was later turned into a worksheet draft. Johnson's own ordering and lining of the variants in this manuscript have been questioned by William H. Matchett, PMLA, LXXVII (1962), 436-441. See also poem #1591, a worksheet draft so rough that Johnson attempted no reconstruction; in preparing it for the concordance, all that could be done was to group together the phrases that appeared to be variants of one another.
Nine of these poems—216, 433, 494, 824, 1213, 1282, 1357, 1358, and 1627—are labelled as versions I and II or earlier and later versions; poem #148 simply has an "or" separating the two versions given in a single manuscript. Johnson's editorial practice in giving double versions of these poems seems inconsistent, for in some of them he seems to have abandoned his procedure of giving principal representation to the earliest version of the best manuscript. In other cases, the minor differences between the double versions hardly make the distinction worthwhile, unless it was also to be applied to more radical differences between two or more versions of a poem given only a single representation in the edition. There was, at any rate, no way of including the double versions as such in the concordance without modifying in some way Johnson's numbering of the poems in the canon.
The list of corrections also adds dashes at the end of #290, l. 4 and #299, l. 4 of the copy sent to Susan Dickinson.
There are also some very minor errors in the line numbering used to key variants to poems. These errors and their corrections are as follows:
- # 532, for 20] read 19]
- # 577, " 25] " 26]
- " " 26] " 27]
- " " 27] " 28]
- " " 28] " 29]
- #1479 " 7-8] " 7-8
- #1508 " 15] " 11]
- #1646 " 6. " 5.
#331 is an early variant of #342, #992 is an adaptation of #937, and #1616 is a slight adaptation of #1525. See NEQ, pp. 242-243.
The version found by Leyda is given in Anderson's Emily Dickinson's Poetry, pp. 324-325; it has one new variant—"candles" for the "candle" given in the Norcross transcript—and this new variant was included in the concordance.
Johnson's list of these in Poems, III, 1206, gives only twenty-four, but his edition shows additional titles for #1 (Valentine Week") and #1545 ("Diagnosis of the Bible, by a Boy—").
The table will also incorporate the changes in order and date made after the text of the poems went to press and given by Johnson in his preface, Poems, I, lxv.
See, for example, Frederick Mosteller's and David L. Wallace's use of "while", "whilst", "upon", and "enough" in analyzing the authorship of the Federalist papers, "Notes on an Authorship Problem," Symposium of Digital Computers, Annals of the Computation Laboratory of Harvard University, XXXI (1962), 163-197.
It should be possible, for instance, to instruct a computer to search the tape and print out all of Emily Dickinson's variant words and lines together with their poem numbers and consequently their approximate dates. Such information could be of considerable value in studying the composition of her poems.
The computer is really working with numbers in binary arithmetric rather than letters, and the words are alphabetized by comparing the numbers of the letters of a word with another word, arranging the two words in numerical order, and then proceeding to another word. If this operation suggests the way an idiot might alphabetize something, it is important to remember that the computer is a highspeed idiot, capable of comparing some 42,000 numbers a second.
| | |