Considering that virtually none of these authors received a formal education, the degree of literacy found in the slave narratives is quite remarkable. It is little wonder that questions of authorship arose. Nevertheless, as scholars such as John W. Blassingame, Jean Fagan Yellin, and William L. Andrews have shown in great detail, the fugitive slaves were by and large the authors of their own tales, even if the editorial hand of an abolitionist corrected grammar or reshaped the flow of the narrative.
This is why Hannah Crafts’s narrative, if authenticated, would have such great historical importance: to be able to study a manuscript written by a black woman or man, unedited, unaffected, unglossed, unaided by even the most well-intentioned or unobtrusive editorial hand, would help a new generation of scholars to gain access to the mind of a slave in an unmediated fashion heretofore not possible. Between us and them, between a twenty-first-century readership and the pre-edited consciousness of even one fugitive slave, often stands an editorial apparatus reflective of an abolitionist ideology, to some degree or another; here, on the other hand, perhaps for the first time, we could experience a pristine encounter. This is not to imply that the “written by himself” or “herself” subtitles to so many of the slave narratives should be questioned: it is only to say that never before have we been absolutely certain that we have enjoyed the pleasure of reading a text in the exact order of wording in which a fugitive slave constructed it.
Nickell points to Crafts’s use of polysyllables—words such as magnanimity, obsequious, and vicissitudes—as proof that Crafts was not “an unread person.” Simultaneously, he continues, Crafts’s misspellings are legion: “incumber” for encumber, “benumed” for benumbed, “meloncholy” for melancholy, “your” for you’re. The curious combination of these two tendencies, moreover, is still another sign of the autodidact, “consistent with someone who struggled to learn.” Crafts’s progress from slavery to freedom overlaps precisely with her progress from “illiterate slave girl to keeper of ‘a school for colored children.’” Her references to Byron, to “the law of the Medes and Persians,” and the “lip of Heraclitus”—as well as her biblical epigraphs and other allusions—suggest the eclectic reading habits of a highly motivated person devouring the arbitrary selections in a small library in a middle-class, mid-century American home. Remarkably, Wheeler left a listing of the books in his private library, to which Crafts ostensibly would have had access. A list of these titles, compiled by Bryan Sinche, appears in Appendix C. In other words, Hannah Crafts wrote what she read, as is abundantly obvious from her uses of conventions from gothic and sentimental novels. In fact, no similar blend of genres exists in the antebellum tradition of African American writing.
Dorothy Porter’s letter to Emily Driscoll in 1951 had referred to Crafts’s text as a “manuscript novel” and as a “fictionalized personal narrative.” Even without researching Crafts’s life or any of the details of her narrative, it is obvious that, however true might have been the events upon which the episodes in her tale are based, Crafts sought to record her story squarely within the extremely popular tradition of the sentimental novel, replete with gothic elements.
If all of this were true, however—and all of these fictional elements are to be found in The Bondwoman’s Narrative—then how could I ever find Hannah Crafts? That is to say, if her tale is a fiction, how could I verify that she had once been a slave, and was a fugitive, as her subtitle claims her to be, “recently escaped from North Carolina”? If I were lucky enough to find a black woman living in New Jersey (where she claims to be teaching “colored” children at novel’s end) named Hannah Crafts—which I had become increasingly skeptical about being able to do, because of the text’s references to Ellen Craft’s cross-dressing, possibly pointing to “Crafts” as a protective pseudonym—how could I ever verify her claim to be an escaped slave? In other words, it occurred to me as I read Dr. Joe Nickell’s amazingly detailed report that I possessed a manuscript that was written sometime between 1853 and 1861, that read like a novel despite its title and its internal claims to be a slave narrative, and that was in all probability written by a black woman who might not ever be found, which seemed to be the way that Hannah Crafts had wanted it. Nevertheless, this quasi-gothic, sentimental slave narrative—no matter how fictionalized I found it to read—rang true at times, especially in her account of the master-slave power relation; her depictions of life in Virginia, North Carolina, and Washington; and, as Wyatt Houston Day had suggested, her various passages about routes and methods of escape adopted by fugitive slaves. How was I to proceed with the search for Hannah Crafts?
As a rule, novels do not depict actual people by their real names. Slave narratives, by contrast, tend to depict all—or almost all—of their characters by their real names, to help to establish the veracity of the author’s experiences with and indictment of the brutal excess implicit in the life of a slave. I write “almost all” because of an occasional change of name to protect the narrator’s modesty or those who might be harmed back on the plantation by the revelation of the author’s identity. Harriet Jacobs became “Linda Brent” and altered the names of characters, in large part because of her revelations about selecting a white lover out of wedlock and bearing his children. And indeed, Dickensian names such as the overseer in Frederick Douglass’s Narrative, aptly named Mr. Severe, seem a bit too good to be true. (Actually, the overseer’s name was Sevier, but Douglass’s tale is so chock full of detail that an occasional allegorically named character is a relief!) But as a rule, fictions of slavery—whether Uncle Tom’s Cabin or Mattie Griffith’s Autobiography—tend not to contain characters named after the author’s actual contemporaries, people who lived and breathed. (A historical novel like Frederick Douglass’s The Heroic Slave is an obvious exception.) If I could find Crafts’s characters in historical records, then, the possibility existed that she had known them as a slave.
I wrote earlier that I was pursuing the authentication of The Bondwoman’s Narrative using two separate procedures. One was the scientific dating of the manuscript, using sophisticated techniques that could ascertain the approximate date of paper, ink, writing style, type of pen, even the use of thimbles to affix paste wafers, and the other mysterious processes that Dr. Joe Nickell used to date the manuscript between 1853 and 1861. The second method on which I had embarked simultaneously was the exploration of census indexes and records, using research tools developed by the Mormon Family Library, especially its Accelerated Indexing System (AIS), which is an alphabetical listing of the names recorded in each federal census since 1790.
I became familiar with this index when researching the identity of Harriet E. Wilson. When I found Mrs. Wilson’s residence in Boston in 1860, using the Boston City Directory for that year (essentially the predecessor of a telephone book, without, of course, telephone numbers), I thought that it would be a straightforward matter to find Mrs. Wilson through AIS, but she was not listed. My colleague at Yale, the great historian John W. Blassingame, encouraged me to examine the actual manuscript record of the 1860 census for the street on which Wilson was reported to be living. Reluctantly, I agreed to do so, asking a research assistant to travel to the Boston Public Library, where the manuscript was held. I presumed that Mrs. Wilson had moved, or died, or been away from home on the day that the census taker knocked on her door. My research assistant, to her astonishment and my own, found that the bottom of the page on which Wilson’s name appeared had been folded under. The photographer who had made the microfilm on which the AIS index was based had not realized this, hence lopping off an entire section of that neighborhood. Had not Blassingame insisted that I pursue my research to its original source, I could never have established Harriet E. Wilson’s racial identity.
Numerous problems obtain with census records, not the least of which are human error, poor spelling, phonetic spelling, and the fact that some people will lie about their birth dates or birthplace, their ethnic identity, or their level of literacy. Not everyone
wants to be located, locatable, or identified, especially if she has a reason for which to forge a new identity. Many former slaves never could be certain of their birth dates in the first place, and some even shifted this date (usually forward) decade by decade as a researcher tracks them through each successive census. Spellings can also be quite arbitrary, necessitating a broad approach to an array of phonetic possibilities for one’s subject. Crafts, for example, can be written as Krafts, Croft, Kroft, Craff, etc. Census records can be a blessing for researchers, but they cannot be used uncritically. Just as important, indexes of census records are not entirely accurate, as I discovered when I used my great-great-grandmother Jane Gates as a control for the 1860, 1870, and 1880 census indexes, since I possessed copies of her listings in those records, which our family had made ourselves at the relevant county courthouses. Nevertheless, she did not appear in the AIS index. Electronic indexes—on CD-ROM and on-line, none of which existed, of course, when I went in search of Harriet Wilson—can be enormous time-savers but can never replace examination of an actual document. Human error in the replication of such an enormous database as the U.S. federal censuses is inevitable.
All of these caveats notwithstanding, I embarked upon a systematic examination of census records, using the Internet and a most efficient researcher at the Mormon Family Library in Salt Lake City, Tim Bingaman. Tim Bingaman was a godsend, not only because of his good humor and expertise with databases but because my travel was still restricted on account of my recuperation from hip-replacement surgery. I would phone Tim and request a search of this source or that, and back—by phone, fax, or mail—would come the result. Eventually, the search for Hannah Crafts would involve several archives: the Mormon Family History Library in Salt Lake City, the Library of Congress, the National Archives, Harvard, and the University of North Carolina, as well as family genealogy web sites and CD-ROM indexes, including the records of the Freedman’s Bank, recently published by the Mormon Family History Library.
I began my research by compiling a list of all the proper names of the characters who appear in Hannah Crafts’s manuscript. As I have written, if I could find at least one actual person named among her characters, then it would be clear that Crafts based her novel on some aspect of her own experience; that the novel was, to some extent, autobiographical; and that she, quite probably, knew the institution of slavery personally and may even have been a slave herself. The question would be one of degree.
By my count, thirty-one characters appear by name in The Bondwoman’s Narrative. At least two characters—Mr. Trappe, Hannah’s mulatto mistress’s torturer, and Mr. Saddler, a slave trader—were certainly named allegorically. Then, too, the slaves listed by first name would be extremely difficult, if not impossible, to find, since the slave censuses listed slaves by age and gender under the name of slave owner rather than by the name of the slave. So I set the names of slaves such as Catherine, Lizzy, Bill, Jacob, Charlotte, and Jane aside. Then I began to pursue each name in alphabetical order, using the 1840, 1850, 1860, 1870, and 1880 federal census indexes.
I can only summarize the mountain of research that these searches produced and the frustrations, false starts, and dead ends attendant upon this kind of research. Let me say that the peaks and valleys of exhilaration and frustration when pursuing this sort of research are extreme, and not for those easily discouraged. Finding Harriet Wilson, as difficult as it was, by contrast was much simpler because it was more localized, confined initially to Boston. In this search, we cast our net wide, of necessity.
I was convinced that the success of the historical search for Crafts’s characters would turn on locating her masters, especially Sir Clifford de Vincent. Either in Virginia or in the United Kingdom, I believed, Sir Clifford would be found. When he was not, I began to despair that Crafts’s tale was entirely fabricated, or at least she had changed the names of all her characters—just as Harriet Wilson had done—and that this avenue of research would lead to a dead end. Even worse, no one named Hannah Crafts was listed in any census indexes that we initially searched. Still, she had located the first part of her narrative in Milton, Virginia—and the Milton that is found in Charles City County, on a bend in the James River, southeast of Richmond, fit her description of the region so very well. An extended, alphabetical search just might yield some clue about the real identity of perhaps one or two of her characters. So, rather than abandon this aspect of the search, I pressed on.
The first indication of a name in a census matching a name in the text was that of Charles Henry, the second of the novel’s three characters having a first and last name. (Sir Clifford was the first.) Two Charles Henrys are listed in the 1850 Virginia census, and one in the 1860 census. “Charley” Henry, in the novel, was the son of “Mr. and Mrs. Henry,” Hannah Crafts’s kind new masters.
Crafts’s characters Mr. and Mrs. Cosgrove, who took possession of the Lindendale plantation after the death of Mr. Vincent, would be difficult to trace, given the absence of a first name of either. But the 1840 Virginia census lists one Cosgrove, the 1850 Virginia census lists three Cosgroves, while the 1860s census lists four, all living in various parts of Virginia.
These similarities in surname were obviously too vague to be of much use, given the absence of first names. Only geographical proximity could help connect them in some way. The first promising association came with the location of Frederick Hawkins, the novel’s third character with two names. The 1810 and 1820 Virginia censuses list a Frederick Hawkins living in Dinwiddie County. No Frederick Hawkins appears, however, in the censuses between 1830 and 1850. The distance between Milton and the closest northwest boundary of Dinwiddie County is about thirty kilometers, or 18.6 miles. When we recall that Hannah and her first mistress, the tragic mulatto, became lost on their way to Milton, it is at the home of Frederick Hawkins that they arrive. This was a very promising lead, seemingly too much of a correspondence to be entirely coincidental.
Once I had a location for Frederick Hawkins in Dinwiddie County, I could then return to the Virginia census listings in search of the Vincents, the Henrys, and the Cosgroves, to see if any lived near either Milton in Charles City County or in a nearby county, such as Dinwiddie. Nathan Vincent and Elisa Vincent lived in Dinwiddie County in 1830. Edward Vincent, Joseph Vincent, and William Vincent lived in Henrico County in 1840. In 1850 Nathan Vincent lived in Dinwiddie County and Jacob Vincent lived in Henrico County. Thomas Cosgrove lived in Henrico County in 1840, John Cosgrove lived there in 1850, and Frank Cosgrove lived there in 1860. Twenty kilometers separate Milton from the southeast border of Henrico County. Similarly, seven Henrys are listed as living in Henrico County in 1850, and one John H. Henry is even listed as a Presbyterian clergyman, age thirty-three, born in New York and living in Stafford County, which is eighty miles from Milton. It seemed possible to me that the Cosgrove, Henry, and Vincent families in the novel were named after these families living relatively close to Milton. The names of these characters, like the name of Frederick Hawkins, do not seem to have been arbitrary; the fact that the surnames of these characters matched real people who lived so closely together in one section of Virginia suggested that it was at least possible that Hannah Crafts had named her characters after people she had known in Virginia as a slave.
I wrote above that I had been pursuing Hannah Crafts along two parallel research paths. While I awaited the scientific analysis of the manuscript itself, I was gathering raw data from a variety of archives and sources. If Hannah Crafts had drawn upon her own experiences as a slave in Virginia and North Carolina as the basis of the events depicted in her novel, then sooner or later these two paths of research would have to overlap, or mirror each other, in their findings. Despite this expectation, nothing prepared me for the fascinating manner in which this mirroring would occur.
Joe Nickell had suggested near the end of his report that he felt that “the novel may be based on actual experiences.” Why did he think this possible? Because of Crafts’s peculiar han
dling of two of the characters’ names:
There are changes that may be due to fictionalization of real persons or events, such as the change of “Charlotte” to “Susan” [pp. 47 and 48]. More telling, perhaps, is the fact that the name “Wheeler” in the narrative was first written cryptically, for example as “Mr. Wh—r” and “Mrs. Wh—r,” but then later was overwritten with the missing “eele” in each case to complete the name [pp. 152–156].20
What these manuscript changes imply is that Hannah Crafts most probably knew the Wheelers and that Wheeler was their actual name. Even more surprising is the fact that she has disguised their name initially, and then filled it in later, suggesting that the reasons she had wanted to veil their identity no longer obtained when she decided to fill in their names. Moreover, it is clear that she wanted to leave no doubt about the Wheelers’ historical identity, about who they actually were.
The Henry Louis Gates, Jr. Reader Page 13