University of Virginia Library

Establishing a Text: The Emily Dickinson Papers
by
Thomas H. Johnson [*]

It was fifteen years ago this month that I first saw the manuscript volume which Edward Taylor had assembled during his long life as pastor in Westfield, Massachusetts. He had labeled it his "Poetical Works," and bequeathed it at the time of his death in 1729 to his descendants with the injunction that the poems should never be published. At the time, I thought the problems ahead were very challenging. And so, no doubt, they were. First of all, would scholars and lay critics share my belief that Taylor deserved to be published? Would the Corporation of Yale University—the legal heirs—grant permission in the light of Taylor's prohibition? Could one learn to read Taylor's handwriting? Most important, would the pattern of a poet emerge? As I say, this was several years ago, and seems at this remove, when all the questions are answered, a simple story. For there was but one heir: a notable institution that seeks advancement of learning. There was but one manuscript: a 400-page volume assembled by the author in orderly fashion. The handwriting was never, or almost never, an obstacle. And there was no copyright problem at all.

With Emily Dickinson I again feel challenged, and all the problems seem even more complicated. Perhaps they are. Or perhaps I'm beginning again. In either event, I accept the challenge and enjoy the fun. Though Dickinson never requested that her poetry remain unpublished, the fact is that at the time of her death in 1886 her poems were still in manuscript. The mere handful published before then either were issued surreptitiously or were anonymous. She bequeathed them, along with her other effects, to her sister Lavinia, who with passionate singleness of purpose determined that they should be known, since to her, at least, Emily was a poet. Had the poems been left untouched until 1950, then to be transferred intact to a

learned institution, the parallel with Taylor would be reasonably obvious. At this point the analogy breaks down.

All who have had access to material touching upon Emily Dickinson's life and writing agree—I think without exception—that she knew during her twenties that she was uncommonly gifted; that by the time she was thirty-one, when she sought advice from Thomas Wentworth Higginson in the now famous letters written in April, 1862, she did indeed crave assurance regarding a talent which at times literally overwhelmed her. "Alone I cannot be," she wrote at this time in a remarkable poem:

How was this gift to be shared? She must have been somewhat prepared to accept Higginson's hesitation to advise publication, for it matched the opinion of other writers and critics whom she knew and respected, gentlemen who knew something of her writing, notably Samuel Bowles and Josiah Gilbert Holland. She should delay submitting any letters to the world, they all told her in effect, until she had learned "control." Since she could no more "control" the quality of the hosts who visited her than she could alter her wren-like size, she must therefore, in her own lifetime at any rate, sublimate her desire for public recognition, however compelling the wish for it may have been. That the longing was present seems beyond doubt. In the first place, it was within the months just preceding her first letter to Higginson that she began to make copies of her poems, presumably from original drafts that were then destroyed. These early original drafts do not now exist and there is no record that they survived her. Perhaps she disposed of them as, one by one, she transcribed them into packets or fascicles.

Written fair, these assemblages of anywhere from ten to twenty transcribed verses are the gatherings, threaded loosely at the spine, which she laid away in her cherry bureau. The number of poems so transcribed during the years 1861-62-63 is very large indeed. On the evidence of stationery and of handwriting—and much will be said about both later—this large body of verse, fully half of all her extant poems, belongs in these three years. A packet or two may have been completed as early as 1858, but there are none at all for the years preceding. In fact there is not to my knowledge a single poem which in its present transcription can be safely assigned an earlier date.

The conclusion therefore is inevitable either that she wrote verses at white heat for a period of three or four years at the turn of the decade as

she passed her thirtieth year, or that the packets assembled at that time represent fair copies made from drafts which had been accumulating for the previous eight or ten years, roughly from the time she was twenty. Unless the originals, which certainly must have existed, are recoverable—and there is no reason to think they are—the issue can never be resolved by direct evidence. It is outside the province of this paper to try conclusions, and the process of editing the Dickinson papers has not yet advanced to the point where final speculations are warranted. I cite the problem, however, as a major example of the nature of the work involved in establishing a provenance of Emily Dickinson's poetry and the order of its composition. The task is further complicated by the fact that she seems to have had no discoverable design in the order of the packets or of the poems within a packet, and that since her death the packets have been handled, separated, and reassembled by several hands. The possibility of confusing the original order—if one did in fact exist—has been multiplied by each handling. If these packets represent Emily Dickinson's copying of her earlier drafts, and if the earlier drafts are irrecoverable, then an exact chronology of actual composition can never be established. A terminus ad quem date must therefore, certainly for the most part, be all that we can expect for the poems which we conjecture she composed in her twenties.

So far I have emphasized Emily Dickinson's compulsion to write poetry and to preserve it as she wrote it, with its half-rimes, broken stanzas, unexpected figures, even though the advisers she consulted counseled greater smoothness and regularity, with fewer oddities of thought. I have pointed out the difficulties that beset an editor of her poems who attempts to establish chronology, at least down to 1863 when she seems to have been convinced that publication in her own lifetime was out of the question. At about this time she began to develop a new medium for her verse, and a way of sharing it with her friends. She incorporated whole poems or parts of them in almost every letter she wrote. Such had not been markedly true earlier. She had always been a notable letter writer, from the time that her earliest extant letter, written to her brother Austin when she was eleven, shadowed forth the sensitive perceptiveness and originality which are her special genius. By the time she was thirty her contact with the world, except for the few members of her immediate family, was almost solely through correspondence. This she now freighted with poetry. Complete poems might thus be conveyed. Or it might be but two lines, or a quatrain from some longer poem recently transcribed into a packet, but here adapted by the change of a pronoun to apply to a particular recipient or situation. Sometimes the poems are archly disguised as prose.

From an editorial point of view such readings must be treated as variants. Often a letter is solely a poem; it usually is without salutation, but is signed "Emily." The text of one such was recently acquired by the Library of Congress. It consists of two quatrains, and the handwriting as well as the content suggest that it was inspired by a homesick mood during her enforced stay in Cambridge where in 1864 she underwent eye treatment. It is written in pencil, because during these months the physician forbade her to use a pen.

The poem was first published in the 1894 edition of the Letters among those, Mrs. Todd notes, "sent to the Hollands at various times." Presumably all the Holland letters published in 1894 were returned to Mrs. Holland and subsequently destroyed, but the poem as it appears in the 1931 edition of the Letters is printed as two quatrains (not as in 1894 as a single eight-line stanza), and offers an alternative reading for the fourth line. Presumably therefore Mrs. Todd had seen another version before she brought out the 1931 edition.

She sent this letter-poem to some member of her family in Amherst, probably to Sue, her brother Austin's wife. It presents a textual problem that is not unusual, and one that is especially interesting because it may never be satisfactorily solved. The Library of Congress holograph is the only manuscript version I know to be extant. It differs from the published versions chiefly in its fourth line which reads "Is easy, possibly." ("An Emigrant to be / In a Metropolis of Homes / Is easy, possibly"—not "Is common possibility.") And it offers no alternative reading.

At this point an editorial footnote is demanded, and unless further information alters the conclusion, it will go something like this: "The letters to Dr. and Mrs. J. G. Holland, published in 1894, were returned to Mrs. Holland after they were transcribed, and presumably were destroyed. Access to a second manuscript, with an alternate reading offered for the fourth line, must have occurred before 1931. The single-stanza version Dickinson sent the Hollands probably substituted common for easy, and may have shown no stanza division. Such variants in

wording and stanzaic form are not exceptional when the same letter-poem is written to more than one correspondent. Unless the hypothetical second manuscript proves to be extant, and can confirm the reading of the 1931 text, one queries the substitution of possibility for possibly. Is it a copyist's or printer's error? It renders the line meaningless and destroys the meter."

The purpose of any editorial task is to establish as definitive a text as possible, and to give it a chronology. For Emily Dickinson, it will be to edit all poems and letters known to exist, to give them order, and to place in context, as far as possible, those which, once published, were ultimately destroyed. Of the near 2000 poems known to have been written—most of them already published—a very large number exist in holograph. Of the some 1200 letters, a much greater percentage are still unpublished; and of those in print, a great many were lost or destroyed subsequent to publication. For example, Emily Dickinson's correspondence with her cousins, Fanny and Lou Norcross, was fairly voluminous. It covered many years and was particularly intimate on a domestic level. The spinster cousins, shy by nature, permitted use of the letters for the 1894 edition with the proviso that every personal allusion to them or to others be deleted. On that basis the letters appeared, so blue-penciled as at times to be fairly uninterpretable. They were then returned to the sisters, and in the manner deemed proper for families of their status, were destroyed at the time of their death. Much that would help interpret other letters, or give better insight into Emily herself, is thus irrecoverable. The Norcross letters, and others which suffered a like fate, will continue to tantalize, never explain. But the large number of letters that do still exist are not only intrinsically important (for Emily Dickinson takes rank as a letter writer) but are of tremendous importance in understanding her verse and helping to give it a chronology.

Probably no part of the editing of the Dickinson papers is more exacting than that which involves establishing chronology. Emily dated nothing after 1850, and after her death Lavinia with characteristic thoroughness burned all letters written to Emily. Three possibilities are open for assigning dates to the letters. First there is internal evidence. If it is direct, it is conclusive: "Father died a year ago today." But it often is circumstantial. "Perhaps the flowers wilted," she says in a letter written presumably in late summer, 1880, "because they did not like the Pelham water." The Amherst water supply was first piped in from Pelham earlier in that summer. Second, there is the handwriting. And finally there is the paper that she used. The degree to which these latter two possibilities can be effective is proportionate to the mass of manuscripts which can be studied. For Dickinson the volume is large. I do not hesitate to conjecture

that her manuscripts can be dated within a twelvemonth, when the handwriting is checked against the evidence of the paper groups. I will shortly discuss paper groups. Let me say something here about Emily Dickinson's handwriting.

A scholar's first reaction to placing dependence upon handwriting as a guide to dating a manuscript is apprehensive skepticism. Of course handwriting changes, but can the changes be interpreted in a chronology? I knew at the start that I would be compelled to seek an answer to the question, but I began with no assurance I would be led to conclusions which could be accepted. There are too many human factors. People are subject to moods which handwriting reflects. You write one way when you are rested, or another when you are in haste. The size of letters may vary with the size of the sheet written on, depending on how much you wish to say in a given space. The form and shape of the characters depend also upon the implement you use, as well as the quality of the stationery. A sharp, stiff pen gives results very different from those formed by a stubby pencil. A glossy ledger sheet permits a movement of arm, hand, and fingers that rough, resistant linen will not allow. Absorbent foolscap is something different still. Finally, I thought, let anyone allow a manuscript of his own, which he has no recollection of ever writing, to be placed before him for dating solely on the basis of the calligraphy; how close can he come? Within five years? within ten? Perhaps not even that. And if not, the use of handwriting will be more of a hazard than a help in attempting to establish a chronology for a poet, where the purpose is no less than to discover the growth of an artist.

Yet when I had begun the task, which I knew was necessary to undertake even though the results after many tedious and discouraging hours might only lead into a blind alley, I found trustworthy patterns emerging. And now, after a year of intensive study of the Dickinson handwriting, covering the span of her years, I feel confident that great reliability can be placed upon a chronology that derives from it. Bear in mind that we are not working in a vacuum. Though no poem is ever dated, and no letter dated after 1850, there are many, many letters that can be assigned exact dates, either on a given day, or within a week or a month. For our purposes, that must be considered as close as we can ever get. Next, there is always the check we can apply by arranging manuscripts into paper groups. For the undated documents these three forms of evidence are the only possible ones which can be used. In many instances, that is, where no internal evidence is possible and where the paper is unidentifiable, the handwriting alone must furnish the clue. In such cases the degree of reliability is lessened. Yet it often happens that a poem, identifiable only by handwriting, is exactly duplicated in

another manuscript that can be precisely dated. These cross-checks appear with enough frequency to buttress the confidence I have in our procedure. Let me now explain the way the Dickinson handwriting can be made to tell its story.

First of all, I repeat that the effectiveness with which any handwriting can be used is proportionate to the mass and coverage available. We have both here. And obviously one proceeds from the known to the unknown. Setting the poetry aside to begin with, let us consider only the letters to which dates can be assigned through internal evidence. (And here may I digress long enough to say that photostatic reproductions for such work are essential. They are expendable. They can be handled, arranged, cut, and pasted onto charts for purposes of comparing formation of letter, line slants, the length of ascenders and descenders of g's, f's, p's, and so on; and in Emily Dickinson's case they tellingly reveal the story of her linked and unlinked letters.)

My preliminary conclusion, reached after a tentative examination of all datable letters, is one I still hold: That it will be possible to assign a given year to any manuscript of sufficient length from evidence of handwriting alone. Such a calendar year must be assumed to extend a few months backward or forward. But with those reservations in mind, a chronology is possible within fairly limited periods in which manuscripts—and this is important—may be given their relative association. The second conclusion was that all writing in ink could be judged by commonly applicable rules, but that another set of rules applies to the writing in pencil. I prepared a loose-leaf notebook which would allow incorporation of specimens of four pieces of writing from each year of undated material—from 1850 to 1886. The plan was to place a photostat of two handwriting samples in ink and two in pencil, each representative of different periods in each year, chosen always from manuscripts that could be dated by internal evidence. The more samples per page the better, for thus one could determine whether variations were trivial or significant. Some pages still remain blank. For instance, there are several samples in pencil of writing in the latter part of the year 1864 and early 1865, but none in ink. Since these are the months she was under orders not to use a pen, it is unlikely that she would have done so. By extension, and because I believe she followed her doctor's orders, I doubt whether any document in ink will be found in that period. Similarly for the period after 1878. Why this abrupt adaptation?

One cannot pursue a study of the handwriting without considering the problem of her eyes and the general state of her health. Why did the character of the letters balloon so in the 'seventies, and the letters unlink? Emily Dickinson's death was due to some form of nephritis. It may

never be possible to establish the exact nature of the affliction, though it was then diagnosed simply as Bright's disease. But there are hypotheses that cannot be overlooked, even though they may ultimately remain suggestive only, for lack of sufficient proof. One not uncommon cause of a fatal nephritis is now known to be a deepseated streptococcic infection that becomes recurrent. The evidence for such in Emily Dickinson's case is striking. She was withdrawn from Mount Holyoke very suddenly and against her wish in the spring of 1848 by her parents, who sent Austin to fetch her home quite unceremoniously. They had heard indirectly that she was suffering from a severe throat infection that her letters home had tried to belittle. The recurrences were so continual through the next spring and summer that the family effectively prevailed in their demand that she give up thought of returning to the seminary. The letters she wrote during the 'fifties indicate her susceptibility to nose and throat colds. Injury to the kidneys was perhaps begun, though probably not known. It happens that this could have a great deal to do with her handwriting.

Such organic injury, where it does exist, can in its early stages produce a partial blindness or periods of visual aberration that doom the victim to impaired sight. With Emily Dickinson it certainly had progressed far enough by 1864 to compel her trips to Boston, to doctor's orders that she use a soft pencil in place of a fine-pointed pen, and to an acceptance—later modified—of darkened rooms. As years pass, her neuroticism becomes clear enough, but the degree of its basis in a physical handicap is not clear and may never be established. By 1867 her writing certainly has increased in size and the letters within words are broken to a point that one reckons, not in terms of linked words or syllables, but in terms of those that are unlinked, so general has become the separation of letters. The process was so steadily continuous that by 1875 only an occasional of, th, or Mr. remained fastened. She was still writing most of her letters in ink, however. In this year the size of the letters she formed with her pen reached a maximum. By 1878 she evidently found it necessary to forego ink altogether, and no part of any word is linked after that year. But the size of letter decreased when she used only a pencil. From her nervous collapse in the early 'eighties till her death in 1886 the progressive changes become increasingly marked.

I have inquired of physicians whether this increased size and gradual unlinking could have relationship to her vision and fatality. The evidence at the moment is inconclusive. But the bearing of her medical history upon the problem of her handwriting cannot be ignored, and the ultimate verdict will certainly include a study of her health.

There are two areas into which handwriting changes fall: the general

and the particular. The general includes characteristics involving five to eight-year periods. During the 1850's all words are linked, the writing is small and flowing, marked by a roundness that is almost copybook in its style. In these years she often signs her name "Emilie". The unlinking of letters begins in the early 'sixties and is well along to completion ten years later. The letters become progressively more angular, and the pages show a sense of drive and forcefulness. The "ie" for "Emily" is permanently discarded after 1861. In the 'seventies the size of the letters reaches a maximum, the use of a pencil is more frequent, and in the later years of the decade supplants the pen altogether.

Particular changes are far more complicated and overlapping. Charts made of each individual capital letter for each successive year tell a story quite easy to follow, for each year some capital letters undergo striking changes not found in other letters. The same phenomena are observable in lower-case letters, with the further advantage that the alterations in the small letters can be subdivided according to their use initially, medially, and finally. For example, in 1859 her final "d" is formed by an upswept ascender curving right. In the years preceding and following it sweeps left. From '55 to '58 the small "h" used initially is hooked to the left. The list of these minute variations through the years is very extensive, and each serves as a reference to and a check upon the others. Given a sufficient number of words in any document, it is possible to assign a probable date within comparatively narrow limits.

Since I have never tried before to study handwriting with the purpose in mind of dating a manuscript by characteristic changes in it, I cannot say whether all handwriting is so subtly variable. The progressive unlinking of letters, in this case over an eighteen-year period, is enormously useful, and made possible by a physiological change that perhaps is comparatively rare. In her case I believe it theoretically possible, if enough manuscript existed, and if each manuscript used enough letters in sufficient combination, to track down dates of composition within the limits of a given week. But the quantity of manuscripts is wanting, and even if they existed it would take years to compute and equate the frequencies without the aid of an electric eye and a robot tabulator. Our laborsome method of charts must remain but a very rough estimate where handwriting alone is the clue. At best, it must allow a margin of error measured in serveral months. At worst, as in instances where the documents are so brief that tell-tale combinations of letters do not appear, the probable limits of error can extend through two to three years. It is at this point particularly that we turn to paper groups for aid.

Emily Dickinson was fastidious in her stationery selection. Whether it was wove or laid paper, it is almost invariably of an identifiable quality,

and though over the years there is a great variety of it, when the types have been identified and grouped, a distinct paper pattern emerges. I suggest that for the moment we consider the rule, not the exception. It is true that she had several paper types at her desk at one time, and that she might on occasion use a sheet in 1875 left over from a batch presumably exhausted in the 'fifties. Such exceptions are easy to spot, and when they occur they rule out the aid which paper groups often can give.

The task of arranging paper into groups is tedious in the extreme, and has not yet been completed. It involves identifying whenever possible (and it is possible 95 per cent of the time) every paper type that she used, measuring it, matching the groups, and sub-dividing them according to their millimeter measurements. Since the identification of a document by its group is desirable as a check on the handwriting, this work must be pursued with no reference whatever to handwriting or other identifying quality.

Much of the laid paper she used was embossed with, say, a head of Minerva, or a capitol dome, or a basket of fruit, or the name of a stationer or manufacturer. The wove paper is generally watermarked and often dated. Let us suppose that all Weston's Linen 1868 paper has been identified and grouped, and that it totals 75 items. Millimeter measurements show that it falls into three sub-groups. That is to say, the sheets have all been folded once, to make four pages, with the spine vertical on the left, the way boxed stationery usually is sold. When we open it out for measurement, we find that the four outer corners for the entire quantity are identical in size. But the middle vertical measurements—the part that becomes the spine when folded—are in three distinct divisions: group A measures 180 mm; group B, 183 mm; group C, 177 mm. These groups, then, were trimmed folded, and represent three distinct batches. Let us say that the number of sheets in each group is roughly even, about 25 in each. Turn now to absolutely identifiable documents in each group. There are five in group A—all in the year 1872; ten in group B—covering 1874-75; eight in group C—all 1877-78. Turn back to group A. Of the remaining items in it, 20 had been dated about 1872-73 on the basis of handwriting; and three are unmistakably later. The remaining two are too brief to guess at closely from handwriting. They are clearly not later than 1876 nor earlier than '73. All that can be said of them is that they belong with group A, and fall within a pattern common to that batch.

Now group B, where the ten identifiable items fall within the years 1874-75. Let us suppose that 12 of the 15 items remaining have been assigned dates, on the basis of handwriting, ranging from 1874 to '78. The remaining three had been classified as doubtful, with a tentative range of dates between 1872-74. I think it reasonable at this point to

assume 1872 is too early a date for any paper in group B. It seems more likely that group A was purchased from the stationer and used sometime before group B; that group B, the same kind of stationery but from a different batch, was purchased later; that therefore the evidence of the group as a whole points to a date not earlier than 1874 for the three doubtful items.

Similarly in group C, if but one item were conjectured to fall before 1877, I should think the burden of convincing us of the possibility would be more difficult, once the pattern of the group is shown to extend forward from that year, not back of it.

For purposes of clarity, the illustration has been simplified, but it is representative of the use to which paper groups may be put. Evidence from it can never be final, nor can it be used except in conjunction with other evidence. But where it exists, it must be considered.

I think it important to underline at this point that I am presenting problems, not reaching conclusions. To establish a text for Emily Dickinson, particularly of her poems, in terms of exact chronology will never be possible, for all the reasons that have been set forth. The evidence must derive solely from handwriting and from paper groups, and whether taken separately or together, they cannot do more than suggest areas of time. The most that can be done is to give the physical evidence in the greatest possible detail.

It remains finally to say something about the text of the poetry. There would obviously be no problem if every poem were left in a single and final version. But such is not the case. At a rough estimate, I should say that 80 per cent of the poems exist in finished versions, copied fair, with no alternate lines or variant readings suggested. Of the remaining 20 per cent—some 400 poems—she offers variant readings, for the most part with no indication which reading she might ultimately have selected in a final version. Any editorial choice therefore becomes impossible in a definitive edition, since it can represent only an editorial preference.

In one instance I thought she herself had provided a solution. One of the poems which she copied into a packet had several suggested readings for eight different words in the course of the five stanzas, but with no indication of her choice. (Sometimes, though infrequently, her choice is indicated by an underlining of the word she prefers, or by a deletion of the one she has rejected.) Then I found the same poem included in a letter to Higginson with choices made in every instance. Here, then, seemed proof that she had established her final version. But in another letter to another correspondent, written at substantially the same time, she has included the same poem—also evidently a final version—wherein she adopted six of the choices made in the Higginson letter, but

selected two from among her variants in the remaining instances. If any conclusion is to be drawn from this citation, it would seem to be that there are no final versions of the poems for which she allowed alternate readings to stand in the packets.

Later when anthologists are compiling selections for lay readers, they may do what they wish about selecting the "appropriate" version. The text for my edition in preparation must simply record the lines as she wrote them, noting the alternates and the variant readings. It would be misleading to give the impression that many such exist. Even of the 300 or 400 poems that show them, by far the largest part offer alternates for but one or two words in any poem.

Some dozen or so work sheets exist: those scraps of paper which represent the first draft of a poem—often potentially a great one. They are the jottings which throw brilliant flashes upon the creative spirit in travail. Such is the poem "Two butterflies went out at noon," which has been reproduced in facsimile in Bolts of Melody. So far as I know, it never proceeded beyond the work-sheet stage. What shape she might finally have given it! But we sit, helpless to gain her insight, knowing only that such speculations "tease us past thought, as doth eternity."

Here then are the problems. Our tools are method only. As we crave solutions, daydreaming what we might learn from Dickinson herself if we could visit the Stygian world and hold parley with her Shade, I am reminded of a story told me some years ago of William Lyon Phelps. "Will Rhett Butler return to Scarlett O'Hara?" a student queried after finishing Gone with the Wind. "That is an interesting question," he answered. "I am dining with Margaret Mitchell tonight, and I will try to bring you back the answer." "And what did she say?" the student asked next day: "What was Rhett's choice?" "She said," replied the professor, "that she hadn't the slightest idea."

Notes