The decision to put Horace Traubel’s With Walt Whitman in Camden (1906-1996) on the web would seem to be an easy one to make. Customarily mined by scholars for those scattered comments, recorded by Traubel as the words of the poet, relevant to whatever project the critic has underway, it is an ideal candidate for introduction into the kind of research environment that computers and the web have made possible: mass or focused word-searching of long or difficult-to-locate texts. Then again, the folks at the Walt Whitman Archive had reason to be trepidatious: the obstacles of encoding roughly 5000 pages of text, scanning scores of images, copyediting the entire work, marking it with machine-readable code, and obtaining permissions to reproduce the most recently published volumes were daunting, and kept the project on the back burner for a long time. In 2002, the decision was made to go ahead with the project; having myself learned how to encode digital texts by transcribing and tagging pages upon pages of contemporary reviews of Whitman’s poetry, I thought I knew full well what we were undertaking when the directors of the Whitman Archive asked me to be With Walt Whitman in Camden’s editor. 
But as often happens in the preparation of a digital edition, the material task of coding the text has thrown into relief a host of editorial issues—some uniquely visible within the context of scholarly editing thanks to the position of Traubel’s text in the Whitman canon, and some that plague most digital editions at the moment, but are particularly pronounced in the case of With Walt Whitman in Camden. We have quickly found that the digital edition of Traubel’s conversations with Whitman fascinatingly both reframes questions about intentionality and authorship long familiar in the field of textual theory and raises new questions about the capacity of our current standards to capture the structure, both material and ideal, of Traubel’s document. This brief essay will sketch out first some of the advantages of an online edition of With Walt Whitman in Camden, then turn to some examples of what transferring the work to electronic form has begun to teach us about Traubel’s text and about digital editing technologies. 
Thought about in material terms, With Walt Whitman in Camden is both a multimedia extravaganza and an open vault for the study of literature and culture of the nineteenth century. As Gary Schmidgall has pointed out, Whitman was willing to converse privately about issues he “would never have dared to voice in public or in print.” When he discovered that Traubel was taking notes on their talks, he agreed to allow the exercise on the condition that Traubel “include all the hells and damns.” Thus we read Traubel’s report on Whitman’s candid expressions about everything from politics (Benjamin Harrison, he thinks, is a “shit-ass”) to publishing (“What a cute—devilishly cute—lot the publishing wolves are”) to Traubel himself (“I’m afraid sometimes you’re a little too much inclined to the didactic”). Traubel’s text contains a host of embedded letters by an astounding range of nineteenth-century Anglo-American cultural figures, from Bronson Alcott to Bram Stoker. With the visitors to his Camden home, Whitman discusses international literary figures, science, sexuality, race, United States imperialism, art; the Whitman of With Walt Whitman in Camden is a kind of cultural clearinghouse, a vortex for debates about literature and spirituality with no parallel in American letters. 
With Walt Whitman in Camden also contains a range of other materials, visual and textual. When hyperlinked with resources that are already extant or that will be available in the future manifestations of the Whitman Archive, a rich intertextual picture will be refracted through Traubel’s work. The last volume ends, for example, with a transcription and a reproduction of one of Whitman’s poetry manuscripts, “A Thought of Columbus.” This document will eventually appear in the Poetry Manuscripts section of the Whitman Archive, which will offer transcriptions, annotations, and digital images of the original manuscripts. The physical, archival location of the original document will also be displayed by way of an electronic unified finding aid currently under development. As it happens, Traubel also published this image in the periodical Once A Week in 1892—a very different context from that of With Walt Whitman in Camden; this image will appear in the periodicals section of the Archive. The same is true of the portraits of Whitman included in Traubel’s book. Facing page 544 of volume 3, for example, is a famous image captioned, “WALT WHITMAN AND HIS REBEL SOLDIER FRIEND, PETE DOYLE (1889).” The photograph appears in the Whitman Archive’s “Images of Whitman” resource, in the 1860s section, with annotations that expand on the origins and significance of the image; all such images can be cross-referenced from With Walt Whitman in Camden. 
Although Traubel’s book is a fascinating combination of image and text, the print edition of With Walt Whitman in Camden is notoriously clumsy for researchers to use. It is hard to find: all but the last two volumes are out of print; volume four is extremely rare. Few libraries, even research libraries, contain all nine volumes. If one is lucky enough to have a complete set available, inadequate indexing remains an obstacle. The first few volumes index proper names only. Though the later volumes are indexed more thoroughly, there is neither a compound index nor a concordance of the work. Finally, in part because Traubel declared in volume one that he had chosen deliberately to present the compendium “as it was originally written,” without improving either “the formal grace of the recital” or elaborating on its contents, subsequent editors have eschewed annotation, even as general knowledge of some of the people and events of Whitman’s time dwindled with the 90 years it took to print the text. Our digital edition of With Walt Whitman in Camden will combine fully searchable electronic text of the nine volumes with digital images of the ephemera that were reproduced in the original books. The edition will be expandable to include annotations and an introduction to the work sketching out the literary and material contexts from which Traubel’s text emerged, while giving a sense of the ways in which the use of it in Whitman studies is both powerful and problematic.
The benefits extend inward to the world of humanities computing, as well. Our edition joins a range of other digital projects that grapple with questions about how to encode prose texts with difficult attribution issues and unusual layout features. With Walt Whitman in Camden presents fundamental questions about authority and textuality. For starters, the work is haunted by those questions of intelligibility and transcription shared by any oral history, and particularly by those made before the advent of sound recording. But there are broader questions, because what was generated between the stages of Traubel’s experience with Whitman, his notes, the transcriptions of them, and the final printed book was a complicated collaboration. One even hesitates to use that term: did Whitman know they were “collaborating”—that is, was it clear to him what Traubel was planning to do? What could it mean for one to collaborate against one’s will or without one’s knowledge (given that Whitman was not alive for the production of the published work)? Furthermore, despite Traubel’s insistence in the note “To Readers” that prefaces the first volume that he had “done nothing negatively to disguise any poverty in the portrait and nothing affirmatively to falsely enrich it,” critics have observed that, as Schmidgall puts it, “the proprieties are never breached” in With Walt Whitman in Camden, particularly when it comes to the question of Whitman’s homosexuality (xiii). The texts and letters from Whitman’s home that Traubel embeds in his account are not chosen utterly at random, nor are all of those he took from Whitman included.
The absence of transgressions of politeness notwithstanding, Traubel’s project was controversial for some people in Whitman’s circle of friends and allies, and it has been for critics ever since. These issues gesture toward the question of how accurate Traubel was in capturing Whitman’s “voice.” Critical opinions have differed. Traubel says that Whitman asked him to depict himself in his full range of expression, regardless of genteel literary convention. But are the damns and hells in With Walt Whitman in Camden Whitman’s—or are they Traubel’s version of Whitman? The same could be said of larger thematic concerns—of Traubel’s reports of their discussions of politics, sexuality, or race, for example. And finally, the moments of antagonism between Traubel and Whitman that are depicted in With Walt Whitman in Camden transmit ambiguous signals: on the one hand, that Traubel is willing to report Whitman’s now-famous exhortation that he “[b]e radical—be radical—be not too damned radical” suggests Traubel’s evenhandedness. But on the other hand, this moment reminds us of the fact that Traubel was more politically radical than Whitman, and that he used Whitman’s words, once the poet had passed on, to underwrite his socialist work in journals like The Conservator. 
This mysterious passage between Traubel and Whitman, then—that always-difficult space of translation and transformation—brings us to a tricky stand: the “authenticity” of With Walt Whitman in Camden cannot be established in any straightforward way, yet because most of the book consists of transcriptions of Whitman’s conversations or of his correspondence, it generates critical interest principally inasmuch as it is considered an extension of his literary oeuvre. Traubel’s text has been crucial for Whitman researchers since its publication; in important biographies of Whitman by Jerome Loving, David Reynolds, and Justin Kaplan, references to With Walt Whitman in Camden constitute between 10% and 20% of the total citations (126 of 1001 citations in Kaplan; 236 of 2321 in Reynolds; 209 of 1014 in Loving).  Shimmering at the heart of the critical study of Whitman, then, is a text that calls attention to its own problematic status as “evidence,” and perhaps thus to the difficulty of the concept of evidence more generally. While there is something deliciously Whitmanian about this—it is the editorial equivalent of the move Whitman makes to explode the contiguity of creation and ownership when he writes that “every atom belonging to me as good belongs to you”—it is worth pointing out that Whitman often works against the grain of uncertainty in With Walt Whitman in Camden, not merely by giving Traubel material documents that have survived to us, but by referring to them positivistically as “data” on more than one occasion.
With Walt Whitman in Camden’s first sentence issues a two-fold challenge to potential future editors. “My story,” Traubel announces in “To Readers,” “is left as it was originally written.” The first challenge is to leave the text as it appears. In a series of arguments that weave in and out of defenses of “spontaneity” and editorial nonchalance (“I have had no disposition since to do what I had no time to do then”) Traubel spins the web of illusion that is familiar to us from theoretical arguments in the critical study of bibliography: that he has given “the observer every privilege of vision,” that the story has told itself.  Yet whether or not we take seriously the magical picture of pure reflection of Whitman’s “words,” “manner,” and “faith,” the editorial imperative is clear: leave in the mistakes, the oddities, the “imperfection.” Digital presentation allows us to correct irregularities, for example in spelling, silently (so that word searches will function according to readerly expectations of the moment), and to give the user options about whether or not to display such corrections or other editorial apparatus as she reads.
But the second challenge is generic: Traubel calls With Walt Whitman in Camden a “story.” The next time he names it explicitly, he gives it a more objective state, referring to it as a “record,” and then ups the ante to “recital,” as if he were merely repeating something someone else wrote. By the end of the first paragraph, Traubel’s mediation has completely evaporated—the material record is simultaneous with ineffable time: “I do not want to reshape those years.” For Traubel, the generic instability of With Walt Whitman in Camden was a virtue, and the illusion, if not the fact, of the absence of mediation or editorial touch was a key component of unsettling the monumentality that helps establish the aura of authenticity in biographical texts. Certainly as readers we are not obliged to decide whether the text is a biography, a fiction, a record, or an oral history, or whether it functions within some other genre. But when the text takes a digital form, its editors must take a position, because structurally, the individual sections or inclusions signify differently, stand in a different hierarchical relationship to “the author” or to “fact” in these different genres. At the same time, because With Walt Whitman in Camden is being published in the context of the Whitman Archive, to be properly searchable the text as a whole must declare at least a provisional relationship to Whitman’s oeuvre. We have decided to call it a “mediated” Whitman text, and to list Traubel as its “author,” even though in “To Readers,” Traubel positions his text as not-very-mediated, and himself as less an author than an editor. This decision is partly pragmatic, preserving the functionality of the Archive and its focus on Whitman as a major author, and partly theoretical: we seek by designating Traubel as “author” to push ourselves, as we create the code by which computers will “read” the work, to preserve its constitutive tensions. 
Two examples of dilemmas we have encountered in moving from the printed text to an electronic rendition illustrate the fascinating constellation of issues Traubel’s text suggests. First, a note about what you will see below: in <angle brackets> are the “tags” that we use to denote particular features of a text. These tags have rules; they exist in a hierarchy that computers understand, so that features of the text can be located in relation to a signifying architecture: a quotation, for example, appears within a chapter that appears within the body of the book that appears within the text as a whole. The language we use to choose these tags and this hierarchy is called Extensible Markup Language (XML), and the rules are for the most part set by an editing-standards collaborative called the Text Encoding Initiative (TEI).  Some archivists refer to XML as the “acid-free paper of the digital age” because it is platform-independent and non-proprietary; the long-term viability and searchability of the works we edit thus mandate that the entire Whitman Archive be XML-encoded. XML encoding allows us to offer searches of With Walt Whitman in Camden that are based on content, but also ones based on structural elements.  In my examples below, it is this ability to search for structures that is significant. As users of the print volumes of With Walt Whitman in Camden know, the design of its text and image layout, its journal-based format, and its embedding of a range of texts are crucial structural elements for apprehending the way Traubel was trying to make Walt Whitman understood.
A first example emerges out of the particular field of literary and social discourse that Whitman and Traubel shared. Both men were trained printers, and their work together on the production of November Boughs and other texts, depicted in Traubel’s record, shows an active, sophisticated sense of the importance of controlling the material presentation of a book. Traubel’s involvement in the arts and crafts revival of the fin-de-siécle is well-known, and Whitman famously exerted vigorous control over the appearance of his own texts, from typeface, paper, and binding choices to the layout of words on the page. The physical text of With Walt Whitman in Camden shows much the same kind of attention; the physical contiguity of William Bentley’s production of the last two volumes suggests the continuing power of Traubel’s original design.
The digital edition of With Walt Whitman in Camden could not, of course, pretend to capture the full complexity of the physical text. TEI provides a powerful range of ways of encoding physical features from binding and paper choice to textual structures and images. Still, because XML has a hierarchical structure, it has difficulty suggesting that individual elements of text could be considered as elements in more than one hierarchy. (Particularly challenging, for example, has been the encoding of the marginal glosses that fill the first edition.) Most of the time, there are easy ways to work around this problem, but even the TEI guidelines, which contain an entire chapter on the issue, confess that “non-nesting information poses fundamental problems for any encoding scheme, and it must be stated at the outset that no solution has yet been suggested” that handles multiple hierarchies elegantly. 
With Walt Whitman in Camden contains a persistent feature that, in the light of the inability to capture its multiplicity within the hierarchies set up by TEI, emerges into critical view in ways that might have remained invisible without the pressure of thinking about an electronic edition. Traubel formats the letters embedded in his text in a more or less standard way, with an important exception. Here is a letter from John Addington Symonds to Whitman, for example, with the passage as presented in the print version first, followed by the XML-encoded version:
GAIS, SWITZERLAND, June 13, 1875.
My dear Sir. I was very much delighted some weeks ago to receive a copy of the New Republic with a little memorandum in your handwriting. […] (WWWC I: 203)
<q who="John Addington Symonds" type="written">
<text id="anc.00043" type="letter">
<addrLine>Gais, Switzerland, <date value="1875-06-13">June 13, 1875.</date>
<salute><hi rend="italic">My dear Sir.</hi></salute>
<p>I was very much delighted some weeks ago to receive a copy of the New Republic with a little memorandum in your handwriting. […]
Most of the markup above is cognate: each letter we encode has a unique identification number; we designate the beginning of the body of the letter so that it will display properly as a letter; label the address and standardize the date so that they can be searched separately. The salute is also indicated as a separate element: if they wish to, readers of the digital edition will be able to summon all of the salutes used in letters to Whitman—perhaps to trace patterns in forms of address, say, of comparative familiarity in greetings to the poet. But notice that the <salute> falls inside the opening material, but outside the first paragraph (designated by the tag <p>). In the original, the salute is italicized, but run into the first paragraph line, religiously. By italicizing the salute, Traubel sets it apart as a separate structural entity; but running it into the first paragraph, instead of giving it its own line, he suggests a collapse of the content of the letter—its unique qualities—with its more formal ones. The informal and the prescribed, then, run together, are made ambiguous in the hierarchy of understanding Whitman’s correspondence. While it is possible, by manipulating the interface we use to translate this XML markup into readable text in a web browser, to depict the text more or less as it appears in the print version of With Walt Whitman in Camden, it is not possible to encode structurally the violation of typographical hierarchy represented in Traubel’s choice to render Whitman’s correspondence in this way.  The particular social issues raised in Whitman’s poetry—and indeed, underwriting Traubel’s intimate attachment to him and by extension With Walt Whitman in Camden itself—make a difference here. For a writer deeply interested in the question of friendship, and in integrating friendship with the formal and public structures of politics, this issue looms larger than it would in another author’s work.
The second example gestures to larger editorial questions about how we understand authorship, and how we represent the relationship between the “creator” of a text and the work we read. Here is how we might mark up a typical oral exchange between Whitman and Traubel (the <q> tag designates a quotation):
Applying this scheme to the entire work allows searchers to find all passages spoken or written by Whitman or by any other known speaker or writer whose words appear as quoted verbatim in Traubel’s record. The procedure seems fairly straightforward—until we ponder what to do with a moment like this one:
Again, strictly speaking, this is an easy call for the tagger: “Walt Whitman” makes a declaration, “Horace Traubel” responds, and “Walt Whitman” accedes to the modification. But when we look closely at the mechanics of the exchange, it becomes more nebulous. Whitman begins, in fact, by quoting an aphorism. Aphorisms cannot, by nature, be attributed to any speaker, so we are no doubt justified in simply designating Whitman as the speaker, rather than qualifying this further by saying he is here also quoting someone else.  Traubel’s fragmentary response does not really stand on its own without Whitman’s precedent; it in fact implies it; since he records only his response, however, we can fairly designate “Horace Traubel” as the speaker. But Traubel calls attention to the fact that Whitman’s follow up, “repeating the phrase after me,” quotes Traubel himself. Whitman’s final phrase, “they are best able to stand it,” then, is in fact a collaborative construction between himself and Traubel, built upon and modifying an implied and unattributable aphorism. TEI does not have a tag for this sort of thing, and as With Walt Whitman in Camden develops and Traubel becomes more adept at anticipating Whitman’s speech, even interrupting the poet to finish his sentences at times, instances of this kind of discursive production proliferate. In content—the passage quoted above, in which Whitman and Traubel negotiate the “communism” of the production of November Boughs—and in structure, With Walt Whitman in Camden elegantly demonstrates the unresolveable problem at the heart of our ongoing interest in Traubel’s text: its status as an authorial object whose author is indistinct yet whose value for establishing the sense of the “author Whitman” makes it inescapable. In our current markup scheme, when handling quotations, we have decided to prioritize the utterance, but in this case and in the case of the <salute> above, we will be pursuing ways of capturing more of the complexity of these features of the text, which may mean changing some standards.
To conclude, I return to the question of what sort of text Traubel’s volumes can be said to constitute, and how re-presenting them affects the work. The publication of Schmidgall’s Intimate with Walt in 2001 raised once again the question of the status of With Walt Whitman in Camden in the Whitman “canon.” Schmidgall, like Walter Teller in his Walt Whitman’s Camden Conversations (1973), grappled with the question of distillation: how to select from Traubel’s text in a way that preserves its breadth but adds drama, in a readable length. Teller chose only those passages that “appealed” to him, then reproduced them in alphabetically ordered thematic groups. Intimate with Walt features much more intervention and explanation by its editor, but is again organized along thematic lines. These editions necessarily dispense with crucial structuring features of the text: repetition, chronological order, the quotidian mode, the back-and-forth of image, quotation, description, and conversation. They may as a result encourage the reader to take Whitman’s statements, a priori, as “authentic,” by removing the repeated self-authentication scheme that Traubel wrote into each entry (including for example the time of his visit, recording of people present, occasional weather conditions, and Whitman’s mood). While Traubel no doubt intended this contextual information in part to validate his reporting, it also repeatedly calls attention to his role as mediator. 
These efforts at condensation, while solving the problem of making With Walt Whitman in Camden readable, throw into relief the question of textual authenticity. The play of provocation and appropriation that constitutes Traubel’s record, the offering of evidence on a day-to-day basis, the listing of visitors received and letters written—all these contribute to the verifiability of Traubel’s experience with Whitman. While never quite giving us the reassurance that we have in front of us the true spoken words of the poet, these inclusions, tedious though they can at times be to read in the text’s original form, make an argument not unlike that of the photographic detail of the trompe l’oeil still-life paintings that were so popular in Whitman’s day: they stake a claim for veracity, for faithfulness. Writing the life of a man for whom the experience of camaraderie was the foundation of democratic sympathy and, by extension, trust, Traubel would have felt tremendous pressure to use every means of engendering a sense of faithfulness in his work. The complete presentation of the work, and the inclusion of its facsimiles, pictures, and portraits, is crucial to conveying the complex game of authority-establishment that constitutes a major drama of With Walt Whitman in Camden. It is significant that both Teller and Schmidgall suggest that, in the latter’s words, “there is no substitute for the experience of reading all from beginning to end” (viii). Presentation on the web will allow this completeness and a new flexibility: the search and retrieval capabilities of the digital With Walt Whitman in Camden will permit users to create their own selections and reframings, putting the entire text, with images, at their disposal. The task that remains, then, is to embrace the challenges, like those detailed here, that digital representation presents to us, so that we can create editions that make the story of these challenges available to readers.
 The author would like to thank the people who have worked toward this edition in multiform ways. These include, at Duke University, Allison Dushane, Socorro Finn, Leigh Spoon, the staff of the Center for Instructional Technology, Paul Conway, the staff of Rare Books and Special Collections, and David Ferreiro; at the Walt Whitman Archive, Kenneth Price, Ed Folsom, Brett Barney, and Brian Pytlik Zillig; and at W.L. Bentley publishing, William Bentley.
 For the online edition of volume one of With Walt Whitman in Camden, see http://www.whitmanarchive.org/disciples, under the Horace Traubel section.
 Gary Schmidgall, Intimate with Walt: Selections from Whitman’s Conversations with Horace Traubel, 1882-1892 (Iowa City: University of Iowa Press, 2001), viii; Horace Traubel, With Walt Whitman in Camden, Vol. 1 (Boston: Small, Maynard and Co., 1906).
 The “A Thought of Columbus” manuscript is in the Charles Feinberg Collection at the Library of Congress. For manuscripts and the finding aid, see http://www.whitmanarchive.org/manuscripts; for images, see http://www.whitmanarchive.org/gallery.
 See Bryan Garman, “‘Heroic Spiritual Grandfather’: Whitman, Sexuality, and the American Left, 1890-1940,” American Quarterly 52 (March 2000); 90-126.
 Jerome Loving, Walt Whitman: The Song of Himself (Berkeley: University of California Press, 2000); David Reynolds, Walt Whitman’s America: A Cultural Biography (New York: Knopf, 1995); Justin Kaplan, Walt Whitman: A Life (New York: Simon & Schuster, 1980).
 See for example the discussion of intentionality and meaning in D.F. McKenzie, Bibliography and the Sociology of Texts (Cambridge: Cambridge University Press, 1999), 31-76.
 Traubel, WWWC, 1:vii-ix.
 XML is itself a derivation of Standard Generalized Markup Language, known as SGML. The set of rules invoked by any particular SGML-based document is called a Document Type Definition (DTD); the Whitman Archive has defined its own DTD, based on TEI standards. For a more detailed introduction to XML in the specific context of text encoding, see the “A Gentle Introduction to XML” at the TEI website, http://www.tei-c.org.
 “Platform-independent” means the code will work on most operating systems (Unix, Windows, Macintosh, Linux, et al); “non-proprietary” means that the code is not exclusively owned by any entity or company that could restrict rights to use it.
 C.M. Sperberg-McQueen and L. Burnard, eds., TEI P4: Guidelines for Electronic Text Encoding and Interchange, Text Encoding Initiative Consortium (XML Version: Oxford, Providence, Charlottesville, Bergen, 2002): Chapter 31, “Multiple Hierarchies.”
 In SGML there is a strategy, using the CONCUR feature, that would address our dilemma; it is unfortunately unavailable in XML. An option in XML is to make redundant content with different markup at equivalent levels in the hierarchy; this requires a much bigger file and makes updating the text extremely time-consuming in the case of a very long work like WWWC.
 Though we can be reasonably certain, based on the pattern of conversational reportage that Traubel deploys, that this first quotation is Whitman speaking, it is possible that the situation is even more ambiguous: this quotation might be reproducing that thing that Joseph Cook “has been saying” that irritated Whitman. Again, beyond designating the speaker as unknown or ambiguous, at this point it would not be possible to present the two known possible attributions of the quotation as equivalents.
 Schmidgall, Intimate with Walt; Walter Teller, Walt Whitman’s Camden Conversations (New Brunswick: Rutgers University Press, 1973), 6.