The Ghostly Presence of James Hemings

As indicated by Jefferson’s request—to “send for” Hemings, and “tell him” he would be glad to receive him—Hemings was rarely, if ever, someone to whom Jefferson directly wrote. There are additional letters in The Papers of Thomas Jefferson Digital Edition that refer to Hemings, however, and these too can be identified by searching the archive’s editorial notes as previously described. But the list of letters that results from such a search does little more than to reinscribe the absence of James Hemings in the Jefferson archive as a whole. The author or addressee of each letter appears in bold red type: “To Paul Bentalou, 25 August 1786,” “From Philip Mazzei, 17 April 1787,” while James Hemings, the subject of the search, is relegated to smaller type, often encased in brackets, for Hemings was most often referred to by first name alone, most likely, as Lucia Stanton points out, to “preserve conscience and principle by increasing the social distance between master and slave” (84). Rather than reveal his role in crafting Jefferson’s notion of republican taste, a contribution that would clearly justify his presence in the Jefferson archive, the format of this list of search results reinforces the transactional nature of the system that placed him, like Harriet Jacobs, as discussed in chapter 4, outside the realm of humanity altogether.

But a rank-ordered list is not the only way in which search results can be presented. Consider, instead, a visualization of those same letters that I created, which dramatically shifts the archival frame. I see this shift as enabling a focus on the “surface of things” (Foucault, 58). This phrase, borrowed from Michel Foucault, is central to Stephen Best and Sharon Marcus’s formulation of “surface reading,” a set of critical practices that emphasizes attending to the materiality of the text and the structure of its language, as well as to our own affective or ethical response to the work. This perspective, Best and Marcus believe, can counter the symptomatic reading practices that insist on excavating deeper meaning and exhuming hidden truths. Surface reading, they explain, enables scholars to see shadows in the archive as “presences, not absences, and let ghosts be ghosts, instead of saying what they are ghosts of” (13). For Best and Marcus, as for many scholars of slavery, the ghost functions as an additional figure of absence. In its liminal status, it represents the condition of social death experienced by the enslaved. In its shadowy form, it captures a sense of what is palpable, yet cannot be fully grasped. In its lingering presence, it conjures a sense of the haunting of the present by the past. In its critical contribution, it gestures toward a textual plane that “insists on being looked at rather than [one that] we must train ourselves to see through” (9).

The figure of the ghost, like the notion of the surface—or, for that matter, like the illusory experience of eating that I have explored throughout this book—suggests something readily perceptible but not easily understood. Indeed, there are times when absences in the archive must linger, and the example of the absences associated with James Hemings’s life, and his tragic death, is one such time. We cannot gain access to his inner life, nor should we necessarily continue to seek to do so. And while we might consider how his life might be reimagined in the present, as Colson Whitehead did for Harriet Jacobs, as discussed in chapter 4, we might learn more if we begin by asking not only what, but also how we have come to know.

The visualization in Figure 13 represents one way in which we might begin by asking how we have come to know. I created this image using Protovis, a JavaScript-based tool kit for data visualization developed by the Stanford Visualization Group.13 Protovis facilitates a range of formats for visualizing social network data, including the arc diagram. Unlike the force-directed layouts more commonly employed to visualize network data, the arc diagram clearly identifies each individual, or “node,” in the network, but foregrounds the connections between nodes—or “edges” in network terminology—through the arcs that dominate the image.14 I generated the underlying data by searching the archive’s content and editorial notes for letters that concerned James Hemings, using the expanded search features described earlier in this chapter. After compiling the search data in a spreadsheet, I then wrote a script in the Python programming language to convert the spreadsheet to the JavaScript Object Notation (JSON) format required by Protovis. This process involved identifying each correspondent that mentioned Hemings as a unique node; identifying each additional person with whom that individual had corresponded about Hemings; and then calculating the number of letters each pair of correspondents had exchanged. Even at this level, the level of the archive’s surface, the process of enumerating the letters that mention Hemings begins to illuminate his presence in the archive, suggesting how correspondence networks such as these can provide a means of visually acknowledging the archive’s ghosts.

One notable feature of the arc diagram is that it allows clusters of nodes to be arranged into groups. In this case, I grouped the people who corresponded about James Hemings according to their relationship to Jefferson. Reading from left to right, the diagram lists Jefferson and his family, his political correspondents, his Virginia friends and colleagues, his plantation overseers and free plantation staff, his enslaved plantation staff, and finally, people who do not fall into any of those categories, or about whom we have little or no biographical information. An arc connecting two names indicates correspondence between them, and the width of the arc indicates the frequency with which they corresponded. Because this data is derived from Jefferson’s personal archive, all of the arcs, as expected, connect to him. The widest arcs link Jefferson with Nicholas Lewis, Jefferson’s neighbor in Virginia; George Jefferson, Thomas Jefferson’s Virginia agent (although apparently not a close family relation); and Richard Richardson, who worked as a plantation overseer at Monticello. Presumably, Jefferson corresponded with each of these men about the materials and services required for Hemings to create his artful cookery for the plantation’s residents and guests. In this way, the surface view of Jefferson’s correspondence also acknowledges the reach of Hemings’s cooking—centered in the kitchen, but extending across Monticello in the ingredients he purchased, the dinners he prepared, and the politics he subsequently influenced through the flavors of his food.15

However, the fourth wide arc in this diagram, the arc that connects Jefferson to Evans, cannot be linked to Hemings’s culinary labor. This is an insight that the archive’s surface view makes visible in a way that a listing of the same results does not. As previously noted, William Evans, by his location at the Indian Queen, served as a nodal point in the more material, and hence more easily preserved, network of print. For this reason, Evans’s presence in the Jefferson archive is more readily discerned. In contrast to the return of a “Name” search for James Hemings, chillingly void, a “Name” search for William Evans yields a chain of correspondence through which additional details of Hemings’s eventual fate can be discerned. An examination of this correspondence makes evident that Hemings had been involved in negotiations for employment with Jefferson well before Jefferson sought Evans’s help. Having spent the first twenty-five years of his life in bondage, Hemings understood the importance of defining the terms of his employment in advance. As evinced by a letter written to Jefferson from another acquaintance, Francis Say, dated one day after Jefferson issued his request to Evans, Hemings had already requested that Jefferson “send him a few lines of engagement and on what conditions and what wages [Jefferson] would please to give him” (Papers, 33:53). Further specifying that the offer should be in Jefferson’s “own hand wreiting [sic],” Hemings demonstrates his own awareness of the power of print—and, in particular, the power of Jefferson’s personal hand, as president-elect—to stand in for the de jure agreement that his status as a black man, even free, precluded him from ever wielding to its full effect (Papers, 33:53).

For reasons unknown, Jefferson failed to comply with this request. The next letter in the archive is from Evans to Jefferson and suggests Hemings’s confident tone. Although we do not know what Hemings actually said, Evans reports to Jefferson, “The answer he returned me, was, that he would not go [to Washington] untill [sic] you should write to himself” (Papers, 33:91). Here, we receive a powerful confirmation of Hemings’s literacy, his business acumen, and his determined stance. Despite its importance, however, this letter does not appear in the results of a keyword search for James Hemings, as the editors have not added his name to the notes associated with this letter. Whether or not Evans influenced the outcome of this situation, the Jefferson archive also does not say. Hemings never became the chef at the White House. An eight-month gap in the correspondence between Jefferson and Evans ensues. The subsequent, and final, exchange in the archive, from November 1801, constitutes the entirety of the evidence that documents the circumstances of Hemings’s suicide.

The ghost of James Hemings need not stand for something, as Best and Marcus caution. To be quite certain, the ghost of James Hemings means enough. And while we, as scholars, might seek to know more about Hemings’s life, his story is one that is impossible to retrieve (Jefferson, Papers, 36:20). To recall the words of Hartman, every story that takes shape in the archive of slavery is “predicated upon impossibility—listening for the unsaid, translating misconstrued words, and refashioning disfigured lives—and intent on achieving an impossible goal: redressing the violence that produced numbers, ciphers, and fragments” (2–3). Thus even as we consider the information we might gain from the “numbers, ciphers, and fragments” in Jefferson’s correspondence, transformed from absence into presence through computational means, we are reminded, with the foreknowledge of Hemings’s suicide, of how little of his life— and not only his cooking—we will ever truly know.

Visualizing Absence in the Archive of Slavery

Is it possible to visualize the impossibility of retrieving knowledge about Hemings’s life story? Even more fundamentally, is this a task that should be undertaken at all? As I have suggested throughout this book, I believe—following Hartman, Fuentes, and others—that the answer is, ultimately, yes. The stories of those like James Hemings are precisely what enable us to identify, in locations ranging from Jefferson’s dinner table to the inn at the sign of the Indian Queen, the forms of cultural expression and of aesthetic taste that are too often underacknowledged in larger narratives of the nation’s cultural foundation. More than that, the archival traces that document this sense of taste—precisely because of the impossibility of their full recovery, and the impossibility of complete redress—are what enable us to better comprehend the significance of the absences that structure the archive of slavery, as they do the archive of the United States as a whole.

But how, then, to pursue this task? The unlikely confluence of an archive always already deformed and a methodological school associated with the digital humanities that makes use of the term “deformative criticism” suggests one approach. In Reading Machines, Stephen Ramsay describes how the process of “deliberately and literally” altering the “graphic and semantic codes” of a text through computational means—what he calls the digital “deformance” of the text—results in a “critical self-consciousness that is difficult to achieve otherwise.”16 According to Ramsay, this “critical self-consciousness,” by which he means a deliberate form of subjective engagement with the text, allows the “liberation of the potentialities of meaning.”17 In the case of James Hemings, however, this subjective engagement exposes the impossibilities of meaning. The result becomes, instead, an image that holds open the space of absence, enabling those who view it to contemplate, and make meaning from, what knowledge remains undisclosed.

In this image of absence, pictured on the next page, I sought to dismantle the letter as the unit of the archive. Rather than privilege the relationships between letter writers, I examined each word of content on an equal plane. To begin, I obtained the letters included in The Papers of Thomas Jefferson Digital Edition in XML form from the University of Virginia Press.18 I then extracted the content of the letters from the XML files.19 Next, I employed a technique from the field of computational linguistics called “named entity recognition” (NER), which is used to automatically identify, or “recognize,” the words in a document, or set of documents, that represent the names of things, such as people or places. For instance, the mention of “James” in the phrase, “my former servant James” would be identified as a person, in contrast to how the editors of the Digital Edition only identified the people to whom Jefferson wrote letters, or from whom he received same. With the help of the NER software (I used the implementation included in the Stanford CoreNLP tool kit), I was able to automatically identify most of the references to people mentioned by name in the contents of the twenty-five thousand letters that had been digitized at the time I conducted the initial analysis for this project.

For the purposes of this visualization, however, I limited my scope to the fifty-one letters that the editors of The Papers of Thomas Jefferson identified as including references to Hemings or to a member of the Hemings family. To this more selective corpus, I added the seven letters I discovered through additional search techniques, including a letter that refers to what Jefferson came to call the “tragical end of James Hemmings” (Papers, 36:20).20 I again employed NER to identify the people named in the letters, and then wrote a Python script to parse the output into human-readable form. The result was a list of names, which I then reviewed by hand in order to eliminate the discernable errors and duplicates, such as the fact that “Hemings” was sometimes spelled with one “m” and sometimes with two, as in the line quoted just above. The fact that Jefferson almost always used diminutives when referring to the men and women he enslaved also contributed to the complexity of the data under analysis. James Hemings, for example, was referred to as Jamey, Jim, and even Gimmé while in France.21 After resolving such discrepancies to the best of my knowledge, I then wrote a second script, also in Python, in order to determine which names appeared together in the same letter. Finally, I formatted these relationships to be displayed in the arc diagram as shown.

Because this visualization shows the relationships among people mentioned in the letters, rather than the people to whom Jefferson wrote (or from whom he received letters), what emerges into view is evidence of the complexity of the relations among individuals, and across social groups. Significantly, the arcs that link Jefferson to the men and women he enslaved are much more prominent than those that link him to his family members and friends, suggesting the degree to which Jefferson relied on his enslaved plantation staff to implement his various directives. One can imagine that these directives included the purchase of provisions for his table, seeds for his farm and gardens, and other supplies that supported his project of producing republican taste. In this way, the visualization conjures a sense of the scope of Jefferson’s dependence on the men and women he enslaved in order to advance this project, even as it cannot re-create what these people said in their conversations with Jefferson or with each other, where they went in order to conduct their required transactions, or how they truly lived their everyday lives.

The multiple, overlapping arcs that comprise this visualization also prompt further consideration about the multiple networks of power embedded in the Jefferson archive. There is evidence, of course, of the chokehold of slavery, that “encapsulation” of capitalism that, as Paul Gilroy has demonstrated, “provided the foundations for a distinctive network of economic, social, and political relations” that persist to this day (55). But the arcs that link Hemings and his family to the other enslaved men and women on the plantation also provide a visual marker of the economic, social, and political networks that were sustained through systems of communication that “passed below the radar,” as Ivy Wilson has observed, and therefore are far more difficult to perceive in the archive today (29). In other words, because the relationships that are visualized are derived from the names mentioned in the letters, but do not correspond to any specific relationship beyond appearing together in the same letter, they are more abstract than, for instance, the correspondence network depicted in Figure 13, which shows the documented relationships among those who wrote to and received letters from Jefferson. After all, goods bartered or exchanged leave no financial record, news communicated orally leaves no written trace, and political rhetoric articulated in the vernacular leaves no tangible ideology, so these actions and ideas can never be as clearly documented in a textual archive as can writers and recipients of letters. This image thus helps to conjure a sense of the other powerful networks that are contained, if not explicitly documented, within The Papers of Thomas Jefferson.

To return to the documents in the archive with this image in view fundamentally shifts our focus. As confirmation, consider this letter from Jefferson to Evans, written on November 1, 1801, the first in the trail of correspondence to reveal his awareness of Hemings’s death: “A report has come here through some connection of one of my servants that James Hemings my former cook has committed an act of suicide. As this whether true or founded will give uneasiness to his friends, will you be so good as to ascertain the truth & communicate it to me” (Papers, 35:542). This letter endures in the archive as an emblem of the “precarious lives which are visible only in the moment of their disappearance,” as Hartman eloquently asserts (12). Notably, this letter, which is the first entry to appear in the results of a search for “James Hemings” in the Digital Edition, is one of only two documents in the archive that refer to Hemings by both first and last name. The letter is also significant for the oral “report” that it documents, the reference to the “connection” of one of Jefferson’s “servants,” and the mention of the “friends” who uneasily await confirmation of this news. Jefferson’s language thus points to Wilson’s below-the-radar networks of communication, as well as to the social networks that supported Hemings, and the circulation of subjects—Hemings once among them—who moved apart from the plantation world that Jefferson sought to control. To visualize this movement within the archive, rather than to represent the archive as static or fixed, resists what Best has described as the “logic and ethic of recovery” that reinscribes bodies and voices as lost (157). This image of absence, instead, challenges us as scholars to make the unrecorded stories that we detect—about eating as about life—expand with motion and meaning.

The Long Arc of Visual Display

At a time when the use of data visualization is becoming increasingly prevalent both in popular culture and in scholarly work, we must also, necessarily, recall the long, fraught history of visual display. It is not without irony to observe that this history passes directly through Jefferson and the way in which he utilized his own graphical displays of information, in the form of charts, lists, diagrams, and tables, to advance his empirical worldview. As I. Bernard Cohen explains, the “inductive” approach to knowledge favored by Jefferson and many others at the time “implied an experiential test of knowledge or of system, the same kind of criterion of truth that in the sciences had become Newton’s ‘Proof by Experiments,’ or a reliance on critical observations.”22 This reliance on “critical observations” in turn derived from the Lockean belief that the creation of knowledge begins with sense perception, the same belief that undergirded the philosophy of taste. When applied to the sense of sight, this belief occasioned the emergence of additional ideologies, as well as a new form of scientific expression, a form that could more effectively convey the “factual” nature of the phenomena observed.

Jefferson forged his approach to visual knowledge-making at the College of William and Mary, where he studied with the Scottish mathematician and natural philosopher William Small. In his autobiography, Jefferson cites Small as his most significant mentor. “From his conversation,” Jefferson recalls, “I got my first views of the expansion of science and of the system of things in which we are placed” (Memoir, 1:2). Jefferson also notes that Small returned to Europe, although he does not comment on Small’s subsequent career. But only a decade later, in the 1770s, Small would go on to train the young William Playfair, the Scottish political economist now viewed as the leading progenitor of modern data visualization.23 Playfair employed painstakingly composed charts and graphs—many the first of their kind—in order to advance his economic and political arguments about the British Empire. In “Exports & Imports to and from All North America,” pictured in Figure 15, Playfair effectively demonstrates the impact of the American Revolution on Great Britain’s balance of trade. Unlike Jefferson, he was not certain that revolution, at home or abroad, would result in any positive effect. As he explains in the preface to the third edition of The Commercial and Political Atlas, published in 1801, “A great change is now operating in Europe, and . . . it is impossible to guess in what it will most likely terminate” (iii–iv). Although he feared that the new century might be defined by “war and contention,” he agreed with Jefferson about one thing: that the visual format of his charts and tables would ensure that the underlying data would be understood and remembered for generations to come (iv). “On inspecting any one of these charts attentively,” he pronounces in the introduction, “a sufficiently distinct impression will be made, to remain unimpaired for a time, and the idea which does remain will be simple and complete” (xiv).

Jefferson demonstrates a similar desire to present an idea that remains “simple and complete” in his Notes on the State of Virginia, first discussed in chapter 1. The Notes constituted Jefferson’s extended response to the Comte de Buffon’s theory of New World inferiority, or “degeneration,” as he termed it. Widely considered the most famous example of this form of scientific expression (and scientific racism) in the United States, the Notes includes, for example, tables comparing the size of animals in Europe and America, listings of indigenous American vegetables, and an extensive catalog of Virginian birds. As Bruce Dain observes, Jefferson’s visual presentation of these “supposedly unvarnished facts,” without recourse to analysis or explanation, was intended to “testif[y] that Buffon’s idea of the inferiority of New World nature was absurd, an instance of prejudice and over-theoretical imagination running away with the facts” (28). In Jefferson’s mind, as in Playfair’s, the visual presentation of evidence aligned it more closely with an inductive methodology, and bolstered belief in the factual basis of what had been observed firsthand.

The implications of the visual rhetoric of the Notes also extend from Jefferson’s desire to assert the unequivocal nature of the evidence presented, to a parallel attempt—informed by the discourse of taste—to enforce a unanimity of response among the book’s citizen readers. Christopher Looby, in his work on the political dimensions of taxonomic natural history, draws upon moral sense theory (if not the discourse of taste directly) in his argument about how the preponderance of “graphical, two-dimensional” modes of presentation in the Notes was deliberately “intended to foster” a “uniformity of sentiments and conceptions” among those who read the book (265). Because the nation’s democratic governance relied upon the citizens themselves to make appropriate political decisions, it was of crucial importance to Jefferson, as we learned in chapter 1, that these citizens learn to cultivate a uniform set of behaviors and beliefs. Thus in his graphical mode of presentation, as in the table comparing the quadrupeds of Europe and America pictured in Figure 16, we see how Jefferson promotes a form of political control beyond the sense of taste; this is one enforced through his visual display.

But Jefferson had no public audience in mind when he traced the columns, rows, and rule lines in the small, leather-bound volume that he called his “Farm-book,” pictured in Figure 17. In the Farm-book, Jefferson recorded the names, birth dates (when known), familial relationships, present locations, and countries of origin of the men, women, and children he enslaved. In the representation of this information about the people of Monticello in diagrams that resemble the charts and tables of the Notes, Jefferson enacts a far more odious form of subjugation and control: his reduction of persons to objects, and stories to names. In contrast to the story of James Hemings, told through a combination of presences and absences in the Jefferson archive, the single line in the Farm-book that fixes James Hemings—“Jemmy. 1765.”—serves as a reminder of the violence that can be enacted through visual display. “There is no bloodless data in slavery’s archive,” as Jessica Marie Johnson reminds us. “Data is the evidence of terror, and the idea of data as fundamental and objective . . . obscures rather than reveals the scene of the crime” (70). The “crime” to which Johnson refers is the first enunciative act of enslavement: of transforming a rich human life into salable property by recording that person’s identifying information as data in a book. In this context, the data of James Hemings in the Farm-book conjures a cautionary tale of its own: a reminder to examine the underlying assumptions and biases embedded in the research methods, the technical structures, and the methods of presentation that we, as scholars of that archive, employ.

In a now canonical essay of the digital humanities, “Humanities Approaches to Graphical Display,” Johanna Drucker cautions that scholars in the field must resist the “intellectual Trojan horse” of data visualization, in which “assumptions about what constitutes information . . . are cloaked in a rhetoric taken wholesale from the techniques of the empirical sciences that conceals their epistemological biases under a guise of familiarity” (para. 1). By these “techniques,” Drucker refers to the panoply of line charts and bar charts that appear in contemporary scientific publications, which can be traced directly back to Thomas Jefferson, as we have just learned. But in the more specific case of how Jefferson records the life of James Hemings in his Farm-book, we should take heed to recall the “epistemological biases” of Enlightenment humanism itself. One by-product of the belief in the human capacity for reason—and, in particular, of the ability to transform sensory experience into knowledge—is the assumption that anything observable carries the status of a fact.24 We see this very perversion of observational science in Jefferson’s own lack of understanding of his range of scientific, social, and personal biases. By recording Hemings as “data” in his Farm-book, Jefferson supposed that Hemings might become an object of empirical knowledge, one not only controlled but also understood through visible, visualizable facts.

In this way, the Farm-book calls into question the positivist rhetoric so often associated with contemporary data visualization, rhetoric that derives from Jefferson and his age. It is no coincidence that critics most often point to Jefferson’s racial taxonomies, as articulated in the Notes, as evidence of the limits of his empirical science. Certainly, as Timothy Sweet has suggested, Jefferson’s assessment that “the races of black and red men . . . have never yet been viewed by us as subjects of natural history” indicates how Jefferson fails to “reflect critically on his own process of data-gathering and inference, [and] on the larger implications of the paradigm in which he work[ed]” (110). Following Foucault, Sweet cites these lines as an instance of the epistemological “gap in the Enlightenment scientific paradigm” that prompted the emergence of the modern human sciences (110). Thus when Drucker contends, in her essay on visualization, that the “humanistic concept of knowledge depends upon the interplay between a situated and circumstantial viewer and the objects or experiences under examination and interpretation,” we might more precisely identify the gap made manifest by Jefferson’s unreflective scientific racism as the one that, heeding Drucker, we as scholars must seek to close.

Jefferson’s epistemology of the visible—what I have defined as the tripartite relation that he posits between the observable, the visualizable, and the truth—also subtends his conception of race. Recall from chapter 1 how Jefferson understood black and white people as “distinct” racial groups (Notes, 270). Also recall from that chapter how Jefferson sought to reserve certain internal capabilities—namely, the capability to cultivate good taste—for white people alone. In his racial taxonomies, we can see an additional attempt to identify certain external features—visible features—that might allow him to continue to enforce his damaging racial divide.

The visualizations of James Hemings’s traces in the archive that I have presented in this chapter seek to reveal the “grid of control” that consigned him first to social and then corporeal death (Chun, 56).25 They also seek to reveal a “lived social reality” rich with community, kinship, and support (57). In the context of an archive—and an ideology—that effaces these relations, I have sought to use what Wendy H. K. Chun describes as “the technology of race” against itself (40). By deforming the archive through computational and visual means, I have sought to reveal some of the possibilities of recognition that The Papers of Thomas Jefferson itself resists. I have also endeavored to expose the impossibilities of recognition—and of cognition—that remain essential to our understanding of the archive of slavery today.

Culinary Labor, Digital Work, and the Archive of the Early United States

Jefferson’s “emancipation agreement” with James Hemings, introduced in chapter 1, is another document that, we now know, Jefferson penned in his special ink, encased in his imported paper, copied in his copying press, and then placed in his personal archive to preserve.26 In that document, the second of the two in the entire Jefferson archive that refer to Hemings by his full name, Jefferson insists that Hemings train another person “to be a good cook” before he can be freed (Papers, 27:119). With this stipulation, Jefferson offers enduring textual evidence of Hemings’s culinary expertise. The fact that Jefferson decided to preserve this document in his personal archive reveals how he was, at times, required to recognize—if not ever to redress—the flawed logic that suggested Hemings should be reduced to data in order for his labor to be seen. In sharp contrast to the entry for James Hemings in Jefferson’s Farm-book, which is a distillation of stolen labor, and life, of the highest degree, the emancipation agreement with Hemings identifies his labor as an “art”—indeed as techne—the precise form of applied, experiential knowledge that Jefferson himself most esteemed.

As exemplified by the copying press that he not only utilized but also designed, Jefferson particularly admired the “mechanic arts,” as technical knowledge was then described, and saw such arts as intimately related to his empirical worldview (Marx, 3). And yet, Jefferson’s supposition that if Hemings were to simply train a replacement cook then his absence would not be felt at Monticello reveals an additional limitation of his observing eye: his lack of awareness that there were aspects of Heming’s culinary work that he was unable to perceive. In the agreement, Jefferson does not acknowledge the intellectual aspects of Heming’s cookery, such as his ability to select the particular foodstuffs that would represent Monticello’s unique terroir or to combine flavors that would best please the palates of Jefferson and his guests. Neither does Jefferson register the affective impact of Heming’s cooking—the work of influencing, through Hemings’s specific methods of preparation and presentation, the development of Jefferson’s own conception of republican taste. The condition of chattel slavery of course fundamentally precludes any equivalence between Hemings’s culinary labor and work today. But it remains instructive to consider how the dimensions of Hemings’s techne that transcend the visible might, in turn, help to illuminate the invisible aspects of digital labor in the present.

With this notion of invisible labor in mind, let us return to the Jefferson archive as we most often encounter it today, in The Papers of Thomas Jefferson Digital Edition. When we are reading a letter from that archive online, as I described with respect to Jefferson’s letter to Evans that began this chapter, we see only the final result of the myriad forms of labor that led to the archive’s digital instantiation. Like the “artful cookery” for which James Hemings was renowned, we must work backward from the finished product in order to identify the work—intellectual and affective as much as technical—that went into the archive’s finished form.27 More specifically, we must consider the processes that contribute to creating a digital archive, those that, like a single dish of Hemings’s creation, involve much larger networks of people, resources, and ideas. There is, for example, the process of transcribing each letter in the Jefferson archive from his original handwriting to the structured XML that underlies the contents of the Digital Edition. This was likely a process that took place over decades, and was enacted by numerous people in multiple roles: first the editors of the print edition (or more likely, their research assistants), who transcribed the manuscripts into the text printed in each book; then those who took that text and transformed it from something like a Microsoft Word document into a plain text file, which could then be further manipulated as data; and then those who inserted the XML tags that would mark each section of each letter, and each notable feature, so that they could be formatted on the screen. Each of these portions of the process drew upon different sets of skills: first, the ability to decipher eighteenth-century penmanship; then, the capacity to translate machine-readable text across multiple file formats; and then, the knowledge of XML and related encoding standards, as well as the ability to contribute to a technical project team, to name only a few. Like the people and their skills who contributed to any single one of Hemings’s culinary confections, we cannot know each of their names, or the details of the roles that they played. But a consideration of the processes that we know to have contributed to the final product, as well as the skills that were required for each phase, can do much to acknowledge the otherwise invisible labor that contributes to the information in the archive that we are able to see, and learn from, today.28

We are not able to see any of James Hemings’s “artful cookery,” of course. Nor can we ever hope to taste it. But there exists one artifact that gets us closer to the labor, both visible and invisible, that Hemings contributed to each dish that he put on Jefferson’s table. It is a recipe that appears in a cookbook kept by Virginia Jefferson Randolph Trist, Thomas Jefferson’s granddaughter; one of three hundred she transcribed from a wide range of sources. This particular recipe is labeled “Snow Eggs,” and it describes the process of making an elaborate meringue. The meringue requires ten eggs, the yolks separated from the whites. The whites are whipped, infused with powdered sugar and an additional flavor—either orange flower or rose water “if you prefer.” The mixture is then poached in a milk bath, “spoonful by spoonful,” yielding a set of oval-shaped meringues that resemble the eggs from which they came. The remaining milk bath is then cooked down into a custard, “according to the thickness you wish to give” it. The custard—what we would call today a sauce anglaise—is then strained and poured over the meringues, which then rise to the top. “A little wine stirred in is a great improvement,” we are told. The recipe concludes with a line of attribution, “James, cook at Monticello.”

This “James” is, of course, James Hemings. While Jefferson himself transcribed several recipes for some of his (presumably) favorite dishes, some of which were likely dictated to him by Hemings, this is the only known recipe, in any archive, to be explicitly credited to Hemings himself.29 That it exists not in the Jefferson archive, but in a set of papers housed at the University of Virginia Library that span the years 1825 to 1936, points to how Hemings’s own culinary legacy, as much as the republican taste that his cuisine enabled, extended far beyond the end of his tragically abbreviated life. The number of advanced techniques involved in the recipe—the whipping of the egg whites into an airy cream; the poaching of the spoonfuls of cream back into egg shape; and the precision required in the cooking of the custard—each attest to the high degree of Hemings’s culinary training and skill. In addition, the numerous references to choices that were required to be made in the midst of the cooking process—the flavor of the egg mixture, the texture of the custard, and the addition (or not) of wine—underscore the various forms of tacit knowledge, invisible to the eye, that Hemings also possessed to the highest degree.

There is something affirming about the circular yellow and brown stains that dot the recipe, suggesting that this particular cooking process was followed by Trist or someone in her kitchen, and therefore further suggesting that Hemings’s influence could be experienced, to some degree, decades after the last meal that Hemings himself ever prepared. And yet the idea that Jefferson’s granddaughter—and not any of Hemings’s kin—would be the one to perform this embodied incantation of Hemings’s “artful cookery” produces an equal-but-opposite sense of unease. We do not have equivalent records that document the Hemings family, of course. But the affective dimensions of this lack, whether experienced as unease, or silence, or shadow, or absence—as I have named it in this chapter—also holds value. For it is through these unsettling absences that the most expansive version of the archive of the early United States emerges into view. This is an archive that encompasses impossibility, and that knowingly depends upon the interplay between scholar and text, as well as between archival technologies and archivists. An emphasis on eating offers one entry point into this expanded archive, but there exist many more. What unites these multiple points of access is an understanding of the archive—of the early United States, or of any other domain—not as a neutral repository of knowledge, but instead as a tool for exposing the limits of our knowledge. It is only by acknowledging these limits that we can, at last, begin to see.