A fascinating literary mystery surfaced in 2001 when the distinguished scholar Henry Louis Gates Jr.—chair of the Department of Afro-American Studies at Harvard University—purchased at auction a 300-page holographic manuscript (one written in the author’s own handwriting; see figure 3.1). Its title page read, “The Bondwoman’s Narrative / By Hannah Crafts / A Fugitive Slave / Recently Escaped from North Carolina.” Gates (2002a, xii) immediately recognized that “if the author was black, then this ‘fictionalized slave narrative’—an autobiographical novel apparently based upon a female fugitive slave’s life in bondage in North Carolina and her escape to freedom in the North—would be a major discovery, possibly the first novel written by a black woman and definitely the first written by a woman who had been a slave.”* Beyond the excitement, of course, was the question: was The Bondwoman’s Narrative authentic?
To address that crucial question, I was commissioned to examine the manuscript by Laurence J. Kirshbaum, chairman of Time Warner Trade Publishing, which was planning to publish the novel. I had been recommended by document expert Kenneth Rendell (see chapter 2). In a letter dated April 24, 2001, Kirshbaum referred to Professor Gates’s finding that the manuscript appeared to date from around 1855. “We need your investigative expertise to authenticate this date,” wrote Kirshbaum. I had subsequent discussions with him, Gates, and Rendell, who had taken a preliminary look at the manuscript. We agreed that I would make a detailed examination of the manuscript’s physical makeup as well as any internal evidence relating to its date and authorship. Meanwhile, Gates and other researchers were attempting to verify the historical existence of a Hannah Crafts.
*The first novel published by a black woman was Our Nig (1859), written by Harriet E. Wilson, who had been born free in the northern United States.
Figure 3.1. Titled The Bondwoman’s Narrative, a 300-page manuscript could be the earliest novel written by a black woman.
Subsequently hand-carried to my lab by courier, the manuscript began to give up its secrets and slowly yield clues about its mystery author. For six weeks I subjected the ink, paper, and other traces to a battery of tests—stereomicroscopic (see figure 3.2), chemical, and spectral (ultraviolet). At night, I read a typescript of the novel, looking for clues in the text.
Of my subsequent findings, Gates (2002a, xxx) wrote, “Nothing in my experience as a graduate student of English literature or a professor of literature for the past twenty-five years had prepared me for the depth of detail of the results of Nickell’s examination, nor for the sheer beauty of the rigors of his procedures and the subtleties of his conclusions.”
Gates published my report as an appendix to The Bondwoman’s Narrative (Nickell 2002). It treats the following aspects: provenance, paper, ink, pen, handwriting, erasures and corrections, binding, and textual matters. My findings are summarized here.
Figure 3.2. The author examines the alleged slave-written narrative using a stereomicroscope. (Photograph by Thomas Flynn, Center for Inquiry)
Professor Gates provided some letters regarding the known provenance of the manuscript. It could be reliably traced back to a 1948 catalog from Emily Driscoll, who operated an autograph and book business on Fifth Avenue in New York City. Driscoll subsequently sold the Hannah Crafts narrative for $85 to Dorothy Porter of Howard University. Driscoll told Porter she had “bought it from a scout in the trade” and that all she knew of its prior history “was that he came upon it in Jersey!” (Driscoll 1951). This information showed that the manuscript’s provenance was incomplete but not suspicious, being traceable back half a century, to a time when it was not so valued and therefore not such a target for forgery.
Mrs. Porter’s typed record described the work as a “Manuscript Novel” and a “fictionalized personal narrative” that had been “written in a worn copy book” (Nickell 1993c). Actually, it was penned on stationery (subsequently bound) that consisted of folio sheets (sheets of paper folded in half and thus having two leaves and four pages). The manuscript consisted of four different types of stationery sheets, two of which bore embossed stationers’ crests (designs impressed in the upper-left corner of letter folios beginning in the late 1830s). These crests were from the Southworth paper manufacturing company, one of which appeared in examples from 1856 and 1860 (as shown in a catalog I compiled of more than 200 such marks [Nickell 1993c]).
As part of my detailed examination, I studied the paper’s rag content (Nickell 2002, 291):
Stereomicroscopic examination of the surface of the various pages reveals the presence of bits of thread, occasionally still colored red, blue, etc., indicating the paper pulp was not bleached but was made largely of white cloth. I obtained some small slivers of paper from frayed outer edges (slivers that were about to become dislodged in any case), moistened a sliver with distilled water and teased it apart on a microscope slide, stained it with Herzberg stain [see Nickell 1996a, 178–79], and observed the fibers microscopically. I identified rag—linen and cotton—fibers (the latter with their characteristic twist) but found no evidence of ground wood pulp (first successfully commercially produced in North America in 1867).
Among other analyses, I employed transmitted light (backlighting), which showed that the wove paper lacked a watermark. However, I was excited to discover a type of “accidental” watermark consisting of translucent stitch-type markings running across some leaves; I recognized them as having been produced by the seam of the wire belt of an early paper machine.
The ink proved to be ordinary iron-gall ink. To learn this, I used a nondestructive procedure I had developed with forensic analyst John F. Fischer. Whereas some document examiners either make tests on the document itself or remove samples of ink by scraping with a scalpel or punching out tiny disks of inked paper, we devised a technique that is much less destructive. As I reported (Nickell 2002, 294):
In this procedure, a small piece of chromatography paper is moistened with distilled water, placed over a heavy ink stroke, and rubbed with a blunt instrument using moderate pressure, by which means a small trace of ink is lifted onto the paper. Such samples were taken randomly from several locations throughout the manuscript. The chemical tests were then conducted on the chromatography paper.
Two reagents were used. First, hydrochloric acid was applied, which produced a light yellow color typical of an iron-gall ink that lacks a provisional colorant. (A blue reaction would have indicated a colorant such as indigo; red would have indicated logwood.) This was followed by potassium ferrocyanide which yielded a Prussian-blue color, thus proving the presence of iron and indicating an iron-gallotannate ink. This type of ink was the most common in use during the middle of the nineteenth century.
Rarely in the manuscript, the ink was blotted with the application of sand, but never with blotting paper.
Significantly, examination with ultraviolet light revealed that there were numerous instances of “ghost writing”—mirror-image, fluorescing traces of writing caused by ink corrosion from the facing page. Its presence here was in marked contrast to its absence in the fake Jack the Ripper diary (discussed in the previous chapter), and this was one of several signs of age in the manuscript.
That The Bondwoman’s Narrative was written with a dip pen was evidenced by sequences of writing that became progressively lighter and then abruptly dark again, as the writer recharged the pen. Also, sometimes the pen strokes became finer, indicating a shift from a blunt to a sharp nib—a characteristic of the quill pen. Stereomicroscopic inspection showed the absence of nib tracks and other features that confirmed the use of quills (which were largely abandoned by the end of the Civil War). I determined that standard goose quills had been used, rather than the crow quills employed to produce the tiny script sometimes affected by Victorian ladies as an expression of femininity (see Nickell 1990, 3–4).
Figure 3.3. The Bondwoman’s Narrative, purportedly by one “Hannah Crafts,” is in a modified round-hand style of script, consistent with its purported 1850s authorship.
The handwriting throughout the narrative was of a class succeeding American round hand (circa 1700–1840), called modified round hand (see figure 3.3). It dominated from about 1840 to 1865 (before being supplanted by the Spencerian system of 1865–1890). Of course, a handwriting cannot be precisely dated by style, since people tend to continue their manner of writing into old age. Because the writing materials of the manuscript indicated it had been written in the 1850s, the absence of archaic forms (such as the long s) suggested to me that the writer was relatively young. The penmanship was of a quality I described as “serviceable” (neither elegant nor untutored), and though it was a natural, genuine handwriting, it had been penned relatively slowly, as if to render it legible. The punctuation was eccentric (lacking periods, having apostrophes and quotation marks on the baseline like commas).
The author used a wide variety of correction and revision techniques, ranging from making strike-outs with the pen and providing careted insertions (using an inverted v to indicate placement) to scraping with an ink-eraser knife. Sometimes, while wet, the ink had been wiped off with the little finger (the direction indicating the right hand; see figure 3.4), a common practice of the quill pen era. (The later steel pen dug into the paper too much and eliminated the practice [Nickell 1991].) More extensive revisions had been made by covering the old text with slips of paper bearing the new text; these were affixed with halves of moistened vermilion-colored paste wafers (see figure 3.5). To make a better bond, the pasted areas were impressed not with the usual seal-like device bearing a waffle iron–like pattern but apparently with a thimble—one of several indicators that the author was probably a woman.
Since the manuscript was composed not in a blank book but on sheets of stationery, it needed to be bound. Evidence of pinholes (see figure 3.6) indicated that it had originally and amateurishly been sewn with needle and thread, apparently by the author (just as Emily Dickinson did with her poems [Shurr 1983]); the author then numbered the pages. Much later (as indicated by soiling and abrading of the manuscript’s original first and last pages), the manuscript was professionally bound with a black cloth-covered pasteboard binding (Nickell 2002). This appeared to date from sometime after 1880 (Rendell 2001).
In short, the author’s writing accoutrements included quill pens, a penknife, a container of iron-gall ink, folded stationery sheets (including those produced by the Southworth paper company), an ink-eraser knife, a sander (filled with common sand for blotting ink), a box of vermilion paste wafers, a pair of small scissors, and possibly a paper knife (to slit an occasional folded sheet into two leaves). These items could have been kept in a combination writing and sewing box that included a thimble (used to press down the wafers to make a better bond) and needle and thread (to sew the pages into a book). All these materials were consistent with the 1850s.
Figure 3.4. Evidence that ink was wiped off by the little finger before drying—a common practice for making corrections in the quill pen era—is consistent with the period in question (and indicative of a right-handed author).
Figure 3.5. To correct the manuscript, its author used vermilion paste wafers to attach slips of paper over the replaced text.
Figure 3.6. Evidence of pinholes reveals that the manuscript had been bound in an amateurish manner using needle and thread.
The text of The Bondwoman’s Narrative offered further information about the author and the date of composition. The narrative was not that of an unread person, despite many misspellings: meloncholy for melancholy (59), benumed for benumbed (285), and your for you’re (109). Polysyllabic words such as magnanimity (94), demoniacal (127), and ascertained (230) flowed from the author’s pen. The admixture of good vocabulary skills and poor spelling suggested that the writer had struggled to learn to write. The readability level of the novel—determined by applying a common formula (Bovée and Thill 1989)—was high. The following passage, describing the protagonist’s visit to a gallery of ancestral portraits in the stately mansion Lindendale (18), was assessed at the eleventh-grade level:
Though filled with superstitious awe I was in no haste to leave the room; for there surrounded by mysteryous associations I seemed suddenly to have grown old, to have entered a new world of thoughts, and feelings and sentiments I was not a slave with these pictured memorials of the past They could not enforce drudgery, or condemn me on account of my color to a life of servitude As their companion I could think and speculate In their presence my mind seemed to run riotous and exult in its freedom as a rattional being, and one destined for something higher and better than this world can afford.
The author’s erudition was even more impressive, referring to “the laws of the Medes and Persians” (23), suggesting appearances that “were enough to have provoked a smile on the lips of Heraclitus” (173), and speaking of “the meaning of nature’s various hieroglyphical symbols” (252).
I found many indications that the work was a novel, despite the author’s protestations in the preface that “being the truth it makes no pretensions to romance.” There were, for instance, Gothic elements, including the shadowy gloom of Lindendale, a curse-containing “legend of the Linden,” and suggestions of various supernatural elements. The novel’s lengthy exchanges of dialogue, as well as elaborate scene-setting descriptions and other conventions, stood in contrast to the true slave narrative of Frederick Douglass (1845).
Yet there were indications that the novel may have been based on certain actual experiences. For example, one clue in the manuscript led to a North Carolina slaveholding politician. I noticed that the name “Wheeler” in the novel had often been underlined, and when I inspected closer using the stereomicroscope, I discovered that the word had originally been written as if to conceal the identity—for instance, “Mr. Wh____r” and “Mrs. Wh____r”—and later filled in with the missing letters to complete the name. I related this to Professor Gates, who learned that the reference was to an actual North Carolina slave owner, John Hill Wheeler (Gates 2002b, 331–36).
There were many literary influences in The Bondwoman’s Narrative. As discovered by astute Princeton graduate student Hollis Robbins (who went on to become director of the Black Periodic Literature Project at the W. E. B. DuBois Institute at Harvard), there were many outright borrowings from Charles Dickens, in particular from Bleak House, which was published in the United States in 1853, following serialization in 1852 and 1853. The following two examples show nearly verbatim borrowing (Gates 2002b):
Bondswoman’s Narrative |
Bleak House |
Gloom everywhere. Gloom up the Potomac; where it rolls among meadows no longer green, and by splendid country seats. Gloom down the Potomac where it washes the sides of huge warships. Gloom on the marshes, the fields and heights. Gloom settling steadily down over the sumptuous habitations of the rich, and creeping through the cellars of the poor. Gloom arresting the steps of grave and reverend Senators; for with fog, and drizzle, and sleety driving mist the night has come at least two hours before its time. . . . |
Fog everywhere. Fog up the river, where it flows among green aits and meadows; fog down the river, where it rolls defiled among the tiers of shipping, and the waterside pollutions of a great (and dirty) city. Fog on the Essex marshes, fog on the Kentish heights. Fog creeping into the cabooses of brigs; fog lying out on the yards, and hovering in the rigging of great ships; fog drooping on the gunwales of barges and small boats. Fog in the eyes and throats of ancient Greenwich pensioners. . . . Most of the shops lighted two hours before their time. . . . (chapter 1) |
It is a stretch of imagination to say that by night they contained a swarm of misery, that crowds of foul existence crawled in out of gaps in walls and boards, or coiled themselves to sleep on nauseous heaps of straw fetid with human perspiration and where the rain drips in, and the damp airs of midnight fatch and carry malignant fevers . . . |
Now, these tumbling tenements contain, by night, a swarm of misery. As on the ruined human wretch, vermin parasites appear, so, these ruined shelters have bred a crowd of foul existence that crawls in and out of gaps in walls and boards; coils itself to sleep, in maggot numbers, where the rain drips in; and comes and goes, fetching and carrying fever. . . . (chapter 16) |
Throughout The Bondwoman’s Narrative, numerous distinctive words and phrases had date significance. I found that selected words from the text were correct for the mid-nineteenth century. In addition, historical references rang true, as in the mention of “vagabond Irishmen” (248), which may have been prompted by the increase in immigration resulting from the great Irish famine of 1846–1847. One very specific indicator was mention of “the equestrian statue of Jackson” in Washington (246); that sculpture was completed in 1853, thus providing a date before which the manuscript could not have been written. As to the latest date the novel might have been written, I reasoned as follows (Nickell 2002, 307–8):
Throughout the narrative, references to slavery are in the present tense (as in the Preface’s mention of “that institution whose curse rests” over the nation). This would make no sense if written after the war. Neither would the author’s claim to being “A Fugitive Slave” who had “Recently Escaped from North Carolina.” Mentions of “a deed of manumission” (p. 53), “a slave state” (p. 104), and “an Abolitionist” (p. 202) are all correct for the pre–Civil War period. To have omitted any mention of secession or the outbreak of war itself would have been counterproductive if written after 1861. Following the war, the story would have seemed passé, perhaps thus helping to explain why it went unpublished.
I concluded that, considering all the evidence, a date of 1853–1861 was indicated, supported by consistent evidence from the writing materials.
The link between the narrative and John Hill Wheeler meant that either Hannah Crafts had been an escaped slave of Wheeler’s or she had created a persona as such. Gates (2002a) demonstrated that the novel’s mention of the Wheelers’ runaway slave “Jane” was surely based on the real-life escape of Wheeler’s slave Jane Johnson. She and her two young sons had traveled with Wheeler to Philadelphia, where they were assisted in their escape by black Underground Railroad activist William Still and white abolitionist Passmore Williamson. Jane subsequently faded into obscurity (Gates 2002a, xlvi–lvi; Flynn 2004, 372–73).
In 2004 Gates and Robbins published In Search of Hannah Crafts, their edited collection of critical essays on the narrative. Some of those fascinating essays proposed “suspects” in the quest for the identity of the author. I was asked to contribute a chapter assessing each possible candidate in light of my original findings and, conversely, to reconsider my profile of the author in light of the new evidence (Nickell 2004).
Gates himself conducted an extensive search for Hannah Crafts. He found a black Hannah Kraft in Maryland, but she was too young and happened to be illiterate. He also found a Maria H. Crafts in New Orleans, but her signature did not match the “Hannah Crafts” on the title page. Expert genealogical sleuth Katherine E. Flynn (2004) uncovered two more possibilities—a white Boston teacher and her mother—but their handwriting also failed to match.
Two other candidates for authorship mentioned by Gates were revisited by Flynn and distinguished English professor Nina Baym. Flynn (2004) reopened the case on Jane Johnson, whose candidacy Gates (2002c) had rejected because she had apparently been illiterate, having signed court documents relating to her escape not with her signature but with her “mark.” I cannot do justice here to Flynn’s excellent and important research, but she found several obstacles to Jane’s being Hannah. For instance, according to Jane’s co-rescuer, William Still, “she has never been allowed to read” (quoted in Gates 2002a, xlix), which contradicted the Narrative author’s claim that she had been taught reading and writing as a child by an elderly white couple. Moreover, although Jane Johnson commendably became literate five years after her escape, it seems unlikely that she could have progressed from total illiteracy to writing a novel of the Narrative’s quality in so short a time. Thus, there are many reasons to doubt that Johnson was Hannah (Nickell 2004, 411–12).
Baym (2004) also returned to a candidate rejected by Gates: a free black woman from New Jersey named Hannah Vincent. However, Vincent remains only a possible author until stronger evidence—such as handwriting—confirms or disproves her authorship.
We must remember that the burden of proof is on the proponent; it is not up to others to disprove a claim. We must also guard against wishful thinking, being careful to follow the evidence and not get ahead of it.
Here is my updated profile of Hannah Crafts: She was a talented female writer who had read much imaginative literature and probably produced other literary works, such as poems or short prose pieces, both before and after The Bondwoman’s Narrative. She was relatively young when she composed the novel, which was written between 1855 (when Wheeler’s slave Jane Johnson escaped) and 1861 (the advent of secession and the Civil War). Her identity as an African American slave rings true, but as Baym (2004, 316) observed, that might be because “the account rings imaginatively true” (emphasis added). Whatever her race, it appears that she struggled to become educated. That, along with some indications of frugality (she recycled paper for correction slips and used quill pens rather than the more expensive steel variety), suggests that she was not a middle-class Victorian lady. She was surely a Christian and likely a Methodist, just as Hannah was portrayed in the novel. We have a copious quantity of her handwriting and may someday discover her true identity.
