INTRODUCTION  ❖  Editing a Fluid Text · Navigating the Typee Manuscript · Writing Typee · Scenes of Revision

Scenes of Revision: Expansion and Collaboration

At the end of June 1851—the same month in which he had confessed his unfoldings to Hawthorne—Melville was still deeply immersed in the final stages of Moby-Dick. Now he was writing to Hawthorne to tell him about his building projects on the house and farm, and how he was also building “shanties of chapters and essays” to flesh out Moby-Dick. By “shanties,” he at first had in mind something like the modest, improvised additions he was having done at Arrowhead, and this was in keeping with his notion that a book like Moby-Dick is always growing and therefore always incomplete, like Cologne’s great unfinished cathedral, an image he would use to conclude his comic chapter on “Cetology.” But discovering a pun, he also meant that the “chapters and essays” he was inserting into his narrative were like the shanties he and his former whaling mates had sung at sea. In all likelihood, Melville was using these chapters to expand the middle of his whale book where Ishmael departs from the action to become a kind of unsighted seer into the nature of whales and whaling. These chapters (what Howard Vincent called the novel’s “cetological center”) are lyrical and meditative yet jaunty and comic, and they allow Ishmael to sing out his presence; they build a voice that can speak for both Ahab and the crew, and make their humanities, failings, and obsessions seem natural. These late insertions into Moby-Dick take readers far afield, it would seem, but they serve an important narrative function. They are not just ramshackle shanties; they are songs.

Six years earlier Melville had also found himself wondering what to do about Typee. He had reached the end of his personal narrative of about eighteen chapters, and knew he did not not have enough text to market his book beside heftier and more factual Polynesian narratives. To make his slender volume longer, more “factual,” and seemingly more reliable, he would have to add to it more “chapters and essays.” But with Typee, his strategy was different from what he would do in expanding Moby-Dick. With the beginning, middle, and end of Typee already written out, and with Tommo’s voice already well established throughout, his first goal was simply to add filler concerning island life and custom, based initially on more of his personal experiences. But he soon realized his experience would not do the trick, and in researching various Polynesian resources for even more filler and fact, he began to react to what he was reading, and his filler turned into argument and digression. Whether he had intended it or not, his digressions modulated Tommo’s voice to reflect Melville’s own growing awareness of Pacific imperialism.

In the meditative chapters added to Moby-Dick, Melville allowed the voice of Ishmael to grow beyond anger and obsession. But in Typee, the digressions he inserted later in the process register more of Tommo’s anger and bafflement than meditation and resolution. Once Typee was published, reviewers complained of what they took to be the book’s “moral obtuseness,”1 and within two months Melville issued his revised American edition, which addressed the cause of that reaction by expurgating many of the digressions Melville had previously labored to add. As discussed in Melville Unfolding, Melville’s complicity in creating this version is evident in his attempt to convince British publisher John Murray to accept the expurgations for the sake of a smoother flowing narrative. “Such passages are altogether foreign to the adventure,” he wrote to Murray, “. . . I have merely removed passages which leave no gap” (Correspondence, 56). Melville’s claim suggests all the more that such “foreign” passages had not been part of the original narrative; they had been imposed on the narrative at some subsequent stage in the composition of the book. As grievous as the expurgations may be for modern readers, they brought Typee closer to the “adventure” Melville had originally written. The expurgated digressions, then, are clues for identifying the “chapters and essays” that Melville added to expand his narrative, and serve along with the manuscript itself as evidence for how Melville built Typee. The fuller story of this expansion process and Melville’s use of sources is related in Melville Unfolding. For our purposes here in constructing a revision narrative of the growth of Typee, it is enough to recount the salient features of Melville’s chapter expansion process as they relate to the stages of composition found in the manuscript.

Certain Hypotheticals: Splitting Chapters

To gain a clearer sense of how Melville converted his original eighteen-chapter personal narrative into the fuller, more digressive text he eventually published, we first need to think of his chapters as a series of more tightly bound anecdotes and vignettes that Melville would have had to cut apart and reposition in order to accommodate the newly inserted material. Some sense of these passages and their sequencing can be gathered from the subheadings for each chapter listed in the table of contents and at the beginning of each chapter in the print version of Typee. Melville may have used similar lists of topics as outlines to assist him in his initial composition, although no such documents have been found. Most certainly, the present subheading lists were devised either after the fact by Melville for his fair copy or by his publisher in preparation for publication.2 In either case, the contents lists designate segments of text within the printed chapters that may have first appeared in manuscript in a differently configured sequence in different versions of chapters. If we imagine that these text segments are like the tesserae of a mosaic, we can also imagine Melville pulling these pieces of text apart from each other and repositioning them to create new mosaics, or rather new chapters. Although none of the three surviving manuscript chapters reveals this kind of reconfiguring, some hint of it exists in the fact that manuscript chapter 12 exists on loose leaves, which may have been created when Melville began tearing it away from a composition booklet in order to insert a new chapter between chapter 12 and its formerly contiguous following chapter.

As none of these reconfigured chapters survives in manuscript, they are obviously hypothetical, but if such manuscripts were to be found, we could imagine they would necessarily exist in a cut-and-paste format: a combination of loose leaves drawn from the earliest of composed chapters affixed to large, pages-long insertions composed at a later date. What these cut-and-paste chapters might have looked like in manuscript is hinted at in the not-so-seamless texture of the print chapters themselves, which often exhibit abrupt or clumsy transitions from narrative to polemic digression and back again. And to go back further in time to envision the ur-manuscript chapters that preceded these cut-and-paste ensembles requires a further act of imagination in which we extract the digressive texts from narrative passages and speculatively reassemble the narratives into their earliest hypothetical form.

Stage II

The manipulations Melville made in Stage II of composition demonstrate this process (see the Growth Chart). If we look at the present print chapter 16, we find that it begins with Tommo’s “Melancholy condition” after Toby’s departure but ends with digressions on sewing and shaving. Similarly, print chapter 17 is a jumble of narrative and polemic text combining segments on Tommo’s “Improvement of Health and Spirits”; then a Montaignean complaint comparing “civilized and unenlightened People”; but also a segment on a “Skirmish” among the islanders; and then, too, material lifted from Stewart. Chances are the openings of both 16 and 17 were originally the two halves of an earlier, hypothetical ur-manuscript chapter we might call “manuscript chapter 13A/B.” This would have been a more coherent narrative chapter in which Tommo's “Melancholy condition” of print chapter 16 (let’s call it segment A) is followed up with “Improvement of Health and Spirits” of print chapter 17 (segment B). Chapter 13A/B was probably followed in manuscript by “manuscript chapter 14” now known as print chapter 18, the famous chapter relating Fayaway’s canoe ride and the appearance of Marnoo, the eventual agent of Tommo’s escape.

Like print chapters 16 and 17, print chapters 19 and 30 also seem to be odd mixtures of narrative and digression, even trivia. The opening to print chapter 19, entitled “Reflections on Marnoo” (segment A), is followed up with digressive anecdotes concerning tappa, popguns, and Marheyo’s shoe fetish; and chapter 30, devoted mostly to “Tattooing” (segment B), also contains seemingly random afterthoughts about taboo. What seems likely is that segments A and B in these chapters were originally conjoined in an ur-manuscript chapter 15 A/B that followed on the heels of ur-manuscript chapter 13A/B and manuscript chapter 14.

In looking at Melville’s text in this way, we begin to conceptualize more precisely the original terrain of the first version of Melville’s composition and the Stage II expansions that he might have been able to make on his own without resorting to outside source materials. Essentially, Melville split both ur-manuscript chapters 13A/B and 15A/B in two, stretched his memory to find material to augment each half, and created four different chapters (which eventually became numbered in print as chapters 16, 17, 19, and 30). This Stage II process increased the chapter count of Melville’s original narrative by two.

The narrative elements in the original configuration of Melville’s ur-manuscript chapters help us understand why Melville might have felt the need to make his preliminary expansions and how he performed them. The sequence of manuscript chapters 13A/B, 14, and 15A/B tightly and swiftly record Tommo’s post-Toby depression, his reversal via Fayaway to happier associations with the island, but then his anxiety over tattooing. In content and narrative arc, they lead directly to Typee’s concluding chapters (or what in Stage II are called ur-manuscript chapters 16, 17, and 18). These chapters, now known as print chapters 32, 33 and 34, relate in equally swift measure Tommo’s escape. Of course, Melville might have expanded his book by inventing conflicts and plot complications involving Tommo, Fayaway, and Marnoo, but the expectations of the travel genre obliged him to avoid such romancing; he had little recourse except to make Tommo’s stay in the valley seem longer simply by putting textual distance between the three key events in the arc of his already established narrative: Tommo's feelings of wretchedness at being a captive, his turn for the better, and then his turn for the worse. Essentially, Melville needed to pace his reversals more slowly.

What might have been the sequence of these hypothetical revisions? Manuscript evidence gives some clue, but first let’s consider the most likely sequence of revision. The first preliminary expansion probably involved “ur-manuscript chapter 15A/B.” Originally, this chapter marked the beginning of Tommo’s reversal in his affection for the Typees; it would have begun with (A) Tommo’s continued anxiety over the natives’ desire to keep him (now found in print chapter 19) and moved directly to (B) the natives’ attempt to tattoo him (now found in print chapter 30). But if Melville were to supply his readers with more detail on Typee life, he would have to sustain Tommo’s happier regard for the islanders and forestall his doubts about their keeping him. The problem with ur-manuscript chapter 15 A/B is that it would have rushed Tommo’s increasing wariness of his captors and made further digression on the Typees impossible. To slow things down, Melville would have split “manuscript chapter 15A/B” apart. To segment A, with its “Reflections on Marnoo” and its focus on the natives’ “strangest passions,” he would have added two freshly concocted anecdotes: Tommo’s manufacture of a popgun and Marheyo’s use of Tommo’s shoes as a necklace.3 He would also have added his description of tappa-making (derived from personal observation) at the end.4 The result would be the new and fuller chapter that would become print chapter 19. At this time, he would have also turned manuscript chapter 15’s original conclusion on tattooing (segment B) into print chapter 30 by adding various materials including his personal observations on taboo. By writing a separate chapter on Tattoo, he could stress more clearly the central idea that the threat of being tattooed “augmented [his] apprehension” (NN Typee, 220) and that it hastened his desire to leave. In subsequent stages, Melville would also insert more chapters between print chapters 19 and 30 to stall Tommo’s revelation and prolong his reversal even more.

With a similar revision strategy in mind, Melville probably turned next to his “manuscript chapter 13 A/B.” However, in this case his revisions required him also to tinker with one of the extant manuscript chapters. Here, as with “manuscript chapter 15A/B,” Melville would have divided his ur-manuscript chapter in two. To the first half on Tommo’s “Melancholy condition” (13A), he added the digressions on sewing and shaving to create print chapter 16. To the felicitous latter half, which first registers Tommo’s new “elasticity” and “altered frame of mind” (13B), he added the more serious concluding section on the Typees’ “Skirmish,” which while humorous nevertheless hints at troubles in paradise. Perhaps his writing out of the child’s play of the popgun anecdote used to expand “manuscript chapter 15A” into print chapter 19 actually triggered these deeper reflections on tribal militancy in the expansion of 13B into print chapter 17. One replays the other.

Material support for this second late addition on warfare comes in the form of an insertion slip that was attached to a leaf of manuscript chapter 10. Like the amusing popgun affair, Melville’s warfare anecdote in print chapter 17 comically reduces Typeean bloodshed to the loss of a thumbnail, but the “Skirmish” also includes three ominous gunshots and therefore demonstrates that the natives actually possess muskets. For Tommo, these gunshots disrupt the valley’s “general repose” and undermine any naïve notions that the Typees are “lambkins.” But where do these muskets come from? In describing the skirmish, Tommo observes the chiefs running off from the Ti, “grasping the muskets which were ranged against the bamboos” (NN Typee, 128). This print reference refers back to Melville’s earlier chapter describing the Ti, which includes a paragraph that registers, also in print, Tommo’s “surprise” at seeing “six muskets ranged against the bamboo” (92).

But significantly, the manuscript of this chapter makes no reference at all to muskets. Had the muskets been mentioned in manuscript, we could assume that their presence in Typee was a part of Melville’s original narrative and that the skirmish was also part of that original plan. But where the print version of Typee tells us we should find text about musketry in the manuscript, we find no text at all; hence the muskets and indeed the skirmish could not have been a part of Melville’s original narrative. What we do find on page 5 of manuscript chapter 10 are graphic insertion devices indicating that Melville had prepared additional text about musketry for insertion, and that he had put that text on a separate slip of paper (now lost) to be appended on manuscript page 5. (Pinholes in the leaf show where the insertion slip was once fastened.) Without a doubt, the reference to muskets in the skirmish scene of print chapter 15 triggered the addition of muskets to manuscript chapter 10. That is, in adding the skirmish episode, Melville realized that the “surprising” presence of musketry in the Ti could not be surprising at all because Tommo had already visited the Ti in manuscript chapter 10. Writing triggers memory once again, and in recollecting the skirmish Melville recalled the further detail of the Typees’ armory. That is, Melville returned to his earliest description of the Ti and added the muskets that the natives would later fire in the skirmish. Thus, we find traces of one revision (the added section on the skirmish), which no longer exists in a manuscript form, located in a corresponding revision (the insertion slip on muskets) of a manuscript that does exist.5

The expansions of hypothetical ur-manuscript chapters 13 A/B and 15 A/B constitute the main revision events of Stage II, and with these expansions completed, Melville had indeed created a longer book, but not long enough. In the revision process, he had begun to expose more of Typeean belligerence, thus heightening narrative tension and expanding to some extent his political awareness of warfare practices. The former naval seaman never saw action during his service, but a year and a half on a warship had shown him enough of muskets, and writing about the Typees’ armory and their maneuvers gave him an outlet for what he had kept inside about humanity, savage or civilized. At some point in the narrative, Tommo, too, would have to be made to realize that Typeean belligerence is not just an occasion for cultural comparison; it is a threat to his well-being. Even so, the comic treatment of the skirmish suggests that Melville needed to forestall his “apprehensions of evil”; he was not yet ready to give Tommo any real justification for escaping the Typee Valley. Melville’s humor, then, manipulates our reader response by minimizing violence even as it foreshadows worse to come.6

Stage III

With his narrative now more carefully paced and a revision strategy for stretching out the plot, Melville could turn to other forms of expansion, and in Stage III he added two full chapters of filler: a description of breadfruit (print chapter 15), which he inserted between manuscript chapters 12 and our hypothetical “manuscript chapter 13A/B,” and a similar digression (print chapter 28) on eating parties and raw fish, which includes references to the Ti and intimations of future cannibalism. The main argument for Melville’s having made these insertions in this stage is that they, too, seem to exhibit no debts to sources other than Melville’s personal experience. As already noted, we find material evidence of the insertion of the breadfruit chapter in that the manuscript’s loose leaves may have been created—that is, made “loose”—by Melville’s cutting them away from a stitched booklet in order to insert the breadfruit chapter between manuscript chapters 12 and 13 A/B. Melville’s expansions in Stages II and III range from politically, culturally, and narratorially relevant observations to the sexual insinuations of “Raw fish!”7 However, with a resulting chapter count of still only twenty-two, Melville’s need to expand further, if only to meet audience expectations of sufficient length, would have been impetus enough for him to search his mind for more material. And with his personal repertoire of anecdotes exhausted, he would have naturally turned at this point to source books. Stewart and Porter were inevitabilities.

Loose Leaves

The Typee manuscript consists of one booklet of ten leaves and six loose leaves, with a seventh (Leaf 12) currently missing. The text of the loose leaves flows directly out of the text that concludes the booklet so that there is no apparent interruption of composition from the booklet to the loose leaves. Moreover, the text in the loose leaves (and presumably the missing Leaf 12) also flows consistently throughout the leaves. In short, there is no break in the textual flow. Naturally, we want to know why, how, and when Melville composed, tore out, and reassembled these leaves.

The question of timing is a matter for speculation. The “neat” explanation might be that Melville created the loose leaves relatively early in the process before going on to make other, later changes, such as his insertion of paragraph marks, or his response to his brother’s editorial pencilings, or even his use of certain insertion slips. However, “messier” sequencings are equally possible. The creation of the loose leaves might have come after or concurrently with the pencilings and insertion slips. That is, the penciled suggestions may have triggered further revisions that eventuated in expansions that hypothetically led to the tearing out of the loose leaves; or the loose leaves might have been part of a process including the Fayaway insertion slip (now lost) on Leaf 14, perhaps as a preparation of the text for fair-copying. Another possibility is that the tearing of the loose leaves may have occurred in Stages II-V preceding the fair-copy stage of composition, a period in which Melville inserted additional chapters throughout his narrative. To my understanding, no material evidence exists to show conclusively when these loose leaves were created.

More, however, may be said about the how and why of their composition. One explanation is that they are the first six (actually seven when we count the missing Leaf 12) consecutive leaves of a second booklet, and that in removing for whatever reasons other later leaves (now lost) in the same booklet, Melville necessarily tore these now loose leaves apart from the second booklet. But while this may account for the first two loose leaves (which do seem to come from a single other booklet), it does not tell the whole story.

To understand that story, we need to recall some rudiments of bookmaking. Melville’s booklet consists of five sheets of folded paper, each inserted one inside the other, and all five stitched together, so that when opened at the middle, five leaves fall to the left of the stitching and five to right. If all six loose leaves were part of one such booklet, Leaves 11 through 15 would fall to the left-hand side of each stitched sheet. Moreover, the tears of the torn edge of each of these left-side loose leaves would match the tears on the edges of the corresponding right-hand leaves to which they were formerly attached (see Fig. 1). That is, like pieces in a puzzle, Leaf 11 would originally have been attached to what we would call “Leaf 20” (if it existed) and these two leaves together in their original untorn state would make up a single folded sheet forming the outermost sheet of this hypothetical single booklet. Similarly, Leaf 12 would have been attached to what we could designate “Leaf 19,” and Leaf 13 to a “Leaf 18.” Of course, leaves 20, 19 and 18 are now lost, but if this hypothetical booklet were real, Leaf 14 would have to have been attached to the actual surviving Leaf 17, and Leaf 15 to Leaf 16 (the two of which would constitute the innermost folded sheet). Accordingly, if the left-hand Leaves 11 through 15 were, in fact, originally attached to Leaves 16, 17, “18,” “19,” and “20,” then the torn edge of each leaf would correspond to the torn edge of its companion leaf. Of course, since the hypothetical “Leaves 18, 19, and 20” do not survive, we cannot explain the full story behind the actual loose Leaves 11 through 17, but we do have enough physical evidence to make one significant determination: the six extant leaves could not have come from a single hypothetical booklet.

Manuscript booklet

Fig. 1. Melville's manuscript booklets consisted of five sheets of paper folded and stitched in the fold to create ten leaves, for a total of twenty pages per booklet. Pictured here is the arrangement of loose Leaves 11–17 as they would appear in booklet form, if they all came from a single booklet. Note that Leaves 12 and those designated [MS] are missing. If the loose leaves did come from the same booklet, we would expect the torn edges of Leaves 14 and 15 to correspond to those of Leaves 17 and 16, respectively. Since they do not, nor does any torn edge of the extant loose leaves correspond to any other torn edge, we must conclude that the leaves come from at least two booklets.

This is certain because if these leaves did come from the same five-sheet, ten-leaf booklet, the torn edges of at least one leaf would correspond to the torn edge of one other leaf, but no such correspondence occurs. For instance, Leaves 14 and 15 would have to have interlocking tear marks with Leaves 17 and 16, respectively; but in fact the torn edge profiles of Leaves 14 through 17 are exactly the same rather than interlocking (see Fig. 2). That is, they appear to be consecutive leaves (all falling to the right of the stitching) that were grabbed together and torn from a booklet other than the one originally including leaves 11 through 13.

Loose leaves

Fig. 2. Loose Leaves 11 and 13 (Leaf [12] is missing) have unique torn edge profiles and may or may not come from the same booklet as loose Leaves 14–17. (Because Leaf 11’s profile indicates that it is an outermost leaf, it may have come from the same booklet as Leaves 14–17, although it would have to have been attached to a fifth, right-hand side leaf following Leaf 17.) Loose Leaves 14–17 exhibit the same torn edge profile, which indicates they are, in fact, consecutive leaves torn from the same booklet. Since the torn edge appears on the left of each leaf and goes beyond the crease of the original full-length sheets, we can assume that these leaves also came from the right side of each sheet or the latter half of the booklet.

We can deduce this for the following reason. When the loose leaves are laid on the table with their torn edges to the left, we can see that portions of those torn edges fold up (i.e., the tears cross over to the other side of the crease dividing the leaf from its formerly connected other half); thus, we can conclude that these leaves had to have been torn from the right side of a booklet. And since Leaves 11 and 13 do not share this tear profile, nor do they interlock with any of leaves 14 through 17 or with each other, we can also conclude that they were torn from a separate booklet.

Also relevant is the relative amount of revision evident on the loose leaves. On the one hand, two pages (MS 30 and 32) have two of the highest word count in the entire manuscript and hence the most revision per page (see “Writing Typee,Table 1). On the other hand, the rest of the loose leaves exhibit lower than average word counts, indicating that Melville probably performed more copying and less revision on each of these pages. Taken together, the loose leaves reflect extended passages of copying (suggesting fairly polished material) punctuated by brief but intense areas of fresh composition and subsequent revision. Chances are that Melville had worked over what became his loose-leaf materials considerably on separate pages, that he eventually copied the material into at least two booklets, that he then (for whatever reason) tore the relatively polished loose leaves away from those booklets, and that he subsequently revised and expanded the text of certain pages on those loose leaves by using insertion slips.

We can entertain at least two hypotheses to explain why Melville tore out his loose leaves: one involves the revisions relating to Fayaway and Toby; the other concerns the possible insertion of a digressive chapter. The two may work separately or together.

Fayaway and Toby

As it happens, manuscript chapter 12 (inscribed entirely on loose leaves) contains some of the most emotionally charged and sexually stirring passages in Typee. Here we find Tommo’s remorse, indignation, and guilt over the departure of Toby as well as his opening up to the nurturing of Fayaway, his ogling of young girls as they bathe, and his autoerotic description of Kory-Kory making fire. (See Melville Unfolding, chapters 7–9.) It is only natural that Melville’s most suggestive chapter is the scene of the kind of massive revision the loose leaves suggest. Indeed, the relative smoothness of composition evident in manuscript chapter 12’s loose leaves indicates that Melville may have performed his heaviest revisions on other loose leaves, which, having been torn away and discarded, were the textual precursors of his surviving loose leaves.

One possible scenario (see Fig. 3) is that Melville completed manuscript chapter 11 at the beginning of a second new booklet at the top of Leaf 11. (The fold and tear patterns on the edge of this particular leaf indicate that it is the left half of the outermost sheet of a booklet.) Having done with chapter 11, he then continued on with “Chapter Twelth” (as he calls it), writing on into this second booklet (which I shall designate B2), and on to a third (let’s call it B3). At this point Melville was most likely composing new text or at least creating considerable new material even as he copied from earlier leaves. In the process he found himself recollecting more clearly the aftermath of Toby’s departure, his own depression, Fayaway’s consolations, and the complex feelings of anger, guilt, despair, and passion he had experienced at that time. Quite possibly, he reached something of a block. Perhaps the words had originally flowed well enough onto booklet B2, but upon rereading he was not satisfied and began to revise. Typically, passages that he had begun to revise, then copy, further inspired him to compose new material, and as the writing block melted, he found himself expanding manuscript chapter 12 so that his drafting actually took him to the end of B2 and into the first half of B3.

Booklets 2 and 3

Fig. 3. One possible scenario for the arrangement of the loose leaves is that Melville began to compose his manuscript chapter 12 (now chapter 14 in print) in booklet B2 and wrote on, completing B 2 and continuing half way into B 3, expanding and continuously revising his narrative so much that he was forced to stop, remove these leaves, and recopy them onto the remainder of B 3. He tore away Leaves 11, 12, and 13 (probably individually) and preserved them. Then he copied text from the leaves marked “MS-A through L” onto blank Leaves 14–17, which he then tore away from his booklet in order to place them next to Leaves 11, 12, and 13. An alternate but not necessarily contradictory theory is that Melville’s manuscript chapter 12 (print chapter 14) was originally followed directly by a manuscript chapter 13, or what is now print chapter 16, and that at some late point in composing Typee, Melville decided to add a digressive chapter on breadfruit between these two manuscript chapters. In order to insert this digressive chapter (now print Ch. 15), Melville would have torn manuscript chapters 12 and 13 away from each other and placed the digression on breadfruit, or what is now print chapter 15, between them.

Of course, we do not know the precise events, but if this hypothesis holds, it follows that the later pages in B2 and the first half of B3 had become so heavily revised and overlong as to be virtually illegible in places. At some point, then, Melville found himself breaking up his booklets, preserving the first three leaves (11 through 13) of B2 and copying revisions of the remaining material from the rest of B2 and the first (left-hand) half of B3 onto the latter (right-hand) half of B3. Once he had completed B3 with his revised material, he tore out the more smoothly inscribed leaves that he wanted to keep, which we now call Leaves 14 through 17, and placed them beside Leaves 11 through 13. The fact that he was copying onto these leaves would account for the relatively small amount of revision we find on them, but the fact that Leaves 14, 15, and 16 bear the marks of insertion slips also attests that Melville had still more revising to do.

An Inserted Chapter

A second hypothesis (see also Fig. 3) proposes that Melville tore Leaves 14 through 17 away from the rest of B3 not in order to unite them with Leaves 11 through 13 but rather to separate them from the chapter that directly followed it. As it presently stands, Typee’s print chapter 14 (manuscript chapter 12) is followed by the digressive print chapter 15, which halts the narrative to discuss breadfruit. Print chapter 16 picks up the narrative, beginning as does chapter 14 with Tommo’s feelings of doubt about the Typees despite their persistent “kindness.” It continues to reflect on his own despondency (mentioned above) and the genial character of the natives. In fact, chapter 16 has more in common with chapter 14 than its present “breadfruit” predecessor, chapter 15. Chances are the digressive chapter 15 on breadfruit was composed and added in a later stage.

It is quite possible, then, that in its early transcription version manuscript chapter 12 (print chapter 14) was originally followed by what is now print chapter 16 (we might designate it the original “manuscript chapter 13”), and that in order to expand his book Melville, at some later point, inserted his chapter on breadfruit. To do this he would have had to tear out Leaves 14 through 17, which conclude MS chapter 12, and thereby separate them from the hypothetically named “manuscript chapter 13” (print chapter 16) in order to insert the new chapter on breadfruit, which we now call print chapter 15.

If this second hypothesis holds, then Melville may have torn his loose leaves apart during the later fair-copy stage of composition when (as discussed in Melville Unfolding, chapter 13) the writer made other chapter-length insertions into his text, including two chapters’ worth of insertions in the now-lost section of manuscript that precedes the surviving fragment. Of course, unlike our first hypothesis, this explanation by itself does not fully account for how Leaves 11 through 13 were separated from B2. Thus, the two hypotheses are not mutually exclusive, and a third possibility is that the two work together to explain the loose leaves. That is, at the time Melville saw fit to add a chapter on breadfruit, he was also reworking his material on Toby and Fayaway.

Whatever hypothesis works best to explain the loose leaves, Melville’s tearing of them and the texts associated with them was the product of a mode of revision that occurred late in the compositional process, and is a sign of important textual and psychological transformations.

Two Chapters More

A seemingly small but nevertheless intriguing fact concerning the Typee manuscript is that its chapters numbered “eleven” and “twelth” correspond to print chapters 13 and 14. This two-chapter discrepancy indicates that at some point after completing his “first draught,” Melville had to have added two chapters, or two chapters’ worth of material, to the first third of his narrative. When did Melville add the material, how did he do it, and why?

Although we do not have manuscripts showing the actual insertion of these two chapters, our observation that they were inserted is not hypothetical. Only a two-chapter insertion would make such a renumbering necessary. Of course, it is possible that in reviewing the early chapters of his narrative—chapters he would have originally numbered one through nine in manuscript—Melville may have found two that were simply too long and mechanically divided them into four shorter ones, thus converting nine original chapters into eleven without adding any new material at all. But the textual condition of the print version—the choppiness of the first chapters and the presence of roughly integrated source appropriations—indicates a far more complex scenario of expansion and revision. Moreover, data derived from the extant manuscript also reveals that Melville had to have inserted a dozen or so pages.

For the moment, however, let’s consider the timing of this stage of composition. My reasons for locating the two-chapter expansion of Stage IV at this particular moment in Melville’s revision process depend first upon the unlikelihood of their being added at any other later time. We know that Melville could not have renumbered his early manuscript chapters 1–9 after sending his fair copy overseas to publisher John Murray because of the circumstances related to a still later renumbering of chapters occasioned by the addition of material appearing in the latter half of Typee.8 In a letter to Murray dated December 6, 1845, Melville’s brother Gansevoort designated the new material as a revision of chapter 20 and two new chapters, 21 and 27. He also supplied Herman’s revised table of contents in which the author renumbered only the late chapters from “Chapter 20 on” (Howard, “Historical Note,” NN Typee, 280; Correspondence 31). Since these manuscript chapters 20, 21, and 27 correspond in number to their print equivalents, we know that Melville’s fair copy to which these later chapters were added must already have contained the two earlier chapters’ worth of material under discussion here , otherwise Melville would have had to have labeled his new chapters 20, 21, and 27 as chapters 18, 19, and 25. Thus, Melville did not add these two chapters at any time after Gansevoort handed the fair copy over to Murray in October 1845.

Could the two chapters have been added during the weeks of Gansevoort’s travel to London in August 1845? Evidence for this does not exist. In his surviving correspondence, Gansevoort indicates that around mid-September he had written a now-lost letter about Typee to Herman in response to a letter Herman had written (also lost) that had reached Gansevoort around September 1. Although Gansevoort does not report having received any new chapters in this mailing, this letter could have contained the two additional early chapters. Of course, Herman’s letter, which was sent by steamer, would have been mailed in early or mid-August. Since Melville obviously would not have had the material ready to give to Gansevoort in person before his departure on July 31, he would have had to have prepared the chapters to be added to his fair copy in the first two weeks of August. This would be scant time for such revisions, especially if any of the hypotheses about how these chapters may have been created (about which more later) hold true. Chances are Melville did the messy surgery himself before making a fair copy, and before entrusting that copy, with its two additional chapters, to his brother.

If we can reasonably locate this major two-chapter revision before the fair-copy stage, the next question is how Melville did it, or more specifically what materials he added to trigger a renumbering of his chapters.

Melville might have added two new chapters wholesale, or he might have expanded existing chapters, swelling them to the point that they needed to be cut into smaller consecutive chapters, or he might have performed both kinds of operation, adding a chapter wholesale and expanding another so that it would be divided into two new chapters. Each of these options is a likely scenario for Melville’s conversion of manuscript chapters 1–9 to print chapters 1–11.

To help focus the matter, we can also ask where in the text is Melville most likely to have made his additions? Of the now-existing material preceding the manuscript fragment (i.e., print chapters 1–11), one thing is certain. Chapters 5–11 follow one upon the other in brisk narrative progression. These involve Tommo's planning with Toby to jump ship, their five-day trek in the mountains, their entrance into the Typee valley, and the introduction of various native characters. Except for the insertion of the vocal telegraph image in print chapter 10 (see Melville Unfolding, chapter 10), there are no digressions on politics or religion in chapters 5–11, and the sequence of episodes is so tight that it seems highly unlikely that the story line could have made sense if any two of these chapters had not been present in Melville’s narrative from the start. The same thinking argues against Melville’s simply augmenting the adventure of print chapters 5–11 by swelling certain chapters so as to require new chapter divisions and numbers.

But print chapters 1–4 are quite different. In them we find several digressions on missionaries and Western incursions, a proleptic defense of Tommo’s decision to desert, anecdotes that jump around in time, a significant amount of material borrowed from sources, and even what might be taken as two separate beginnings to the book. In fact, until we meet Toby in chapter 4, there is little sense of a narrative, and a great deal that registers general discontent with French, British, and American authorities. These opening chapters are so “labored” that Leon Howard speculated long ago that they were “probably written, in their final form, only after a substantial portion of the later chapters had been composed” (Howard, Herman Melville, 94). The manuscript evidence of Melville’s renumbering of chapters (which Howard did not live to see) supports Howard's astute speculation. Exactly how and why Melville added to his first four chapters can be explained by their content. A brief summary of these chapters gives some sense of the odd textual diversity of Melville’s opening section.

Chapter 1 is a kind of grab bag consisting of bawdy religious and political commentary, its jaunty tone bristles with a sense of unjust deprivation and impatience to get ashore. The chapter opens with the stirring words “Six months at sea!” but ends a page later with a line of asterisks. (This section, which we may call 1A, appears in print as NN Typee 3.1–5.9, and takes up 67 print lines.) The narrative in Chapter 1 seems to begin again, with “Hurra, my lads! . . . we shape our course to the Marquesas!” This second opening (1B) is followed by details from “olden voyagers,” presumably Melville’s source books (NN Typee 5.11–6.10; 39 print lines). Chapter 1 ends with two ribald (and eventually expurgated) anecdotes about the islanders’ inspection of the missionary’s wife’s petticoats and their queen’s tattooed derriere (1C). Because these three bumptious, anecdotal, and somewhat polemical sections have no strong narrative interconnection, any of them could have been added to one another in several stages so as to create a new chapter.

Chapter 2 is an anomaly among the opening four chapters: it has a straightforward narrative progression. Beginning with a languorous description of the sea, it moves to a sighting of the island of Nuku Hiva, which the whaling ship Dolly circumnavigates; it describes the French gunboats in Nuku Hiva Bay, which the Dolly enters with the help of a drunken pilot, and it ends with visitations from first male and then female natives, who engage with the sailors in a night of debauchery. It is a coherent progression from an ocean picturesque to scenes of primitive and civilized sexuality. Only about one page of island facts, culled from Stewart, might be a candidate here as a late insertion (NN Typee 11.3–11.39; 36 print lines). Quite possibly, chapter 2 was Melville’s original opener.

Chapter 3 is the smallest chapter in the book and a digression upon French operations in Polynesia, ending with the anecdote of Mrs. Pritchard’s defiant flag-raising against the French invaders. Charles Anderson in Melville in the South Pacific does not indicate any debt here to sources, and Melville may have cobbled these materials together from tales he had gathered from his own experience. Melville may have inserted the chapter in its entirety (197 print lines), and some indication of its narratorial irrelevance is that it was removed wholesale in the American revised edition.

Chapter 4 may be divided into six parts:

  1. Tommo's defense of his decision to desert the Dolly (NN Typee 20.1–23.19; 117 print lines);
  2. his description of Nuku Hiva Bay, its inlets, and tribes drawn from Stewart (23.20–24.34; 54 print lines);
  3. two paragraphs on the Typees as “celebrated warriors,” possibly inspired by Porter (24.35–25.13; 18 print lines);
  4. a reference to the Katherine incident and other island treacheries, possibly from Olmstead, balanced by Porter’s treacheries as filtered through Stewart (25.14–26.28; 54 print lines);
  5. Tommo’s disquisition on the misapplication of the term “savage” inspired by both Stewart and Porter (26.29-27.26; 37 print lines); and
  6. a return to the causes of island hostilities and the parodic anecdote of the Glen of Tior (27.27–29.35; 87 print lines).

Melville may have built this chapter by adding any of these sections to any of the others.

It is anybody’s guess which of these text segments found in the four opening print chapters first appeared in Melville’s original narrative, and which were added later (and in what sequence). And we can construct any number of revision narratives to explain what might have happened. For instance, the opening section on the deprivations of sea life (1A) or Melville’s later defense of his apparent self-serving decision to jump ship (4A) may have been primary concerns for the writer and thus a part of the original genesis of his personal narrative; or they may have been afterthoughts and thus part of his Stage IV revision strategy. Conceivably, Melville might have had the materials he drew from sources in chapters 2 and 4 in front of him when he first sat down to write; or, as is more likely the case, he may have inserted those source materials much later. Until such time as the working draft manuscript for these first four chapters is discovered we have no direct material evidence to support any particular revision scenario, and it may seem that this necessarily dark stage in Melville’s revision of Typee is taking us far afield from the immediate applications of the manuscript fragment itself.

Indeed, in trying to build a revision scenario on the basis of the simple fact that Melville added two chapters, I must seem to resemble the scientists in Woody Allen’s Sleeper who upon the assassination of their Leader attempt to clone a new, full-bodied version of their recently exploded führer with that individual’s only surviving body part: his nose. But this should not prohibit us from making reasonable speculations about the growth of Typee, and in fact the manuscript fragment does provide material information that can help us delimit the speculation. This, I am sorry to say, will require some math.

One factor that distinguishes any one revision scenario from another is that each requires Melville to have composed varying amounts of additional text. At present, all we can know of these hypothetical additions is the amount of lines of text they ultimately generate in print. (See print line counts for each text segment mentioned above.) We know this because those print lines actually exist in the print text Melville first published. And we can indirectly measure whatever hypothetical additions a particular revision narrative requires simply by adding up their corresponding print lines.

Let us now assume that Melville composed his additions in the same fashion and at the same rate of writing as he composed the actual three-chapter manuscript fragment we have before us. With this manuscript text in one hand and its corresponding print lines in the other, we can derive a ratio of the amount of text Melville actually generated in manuscript that would yield the corresponding amount of text in the actual print lines he eventually published. That is, if we count all the words inscribed in manuscript chapters 10–12 (including, of course, all cancellations) and compare them to the words they actually became in print chapters 12–14, we could develop a formula for determining the probable amount of text Melville would have to have composed in order to add two chapters’ worth of text in his opening narrative, based upon the actual text he ended up printing.

Chances are the one or two mathematicians and theoretical physicists reading this have caught on to this formulation; even so, for the rest of the population whose eyes may be rolling out of exhaustion if not disbelief, consider this: (1) What we write in manuscript is never exactly the same as what appears in print, in terms of either content or quantity of words; (2) in the case of his three-chapter manuscript fragment, Melville inscribed more words than he eventually printed; (3) if that differential can be taken as a constant ratio, then we can reasonably assume in backwards fashion that any specific number of printed words must necessarily have derived from a corresponding larger number of manuscript words. As it turns out, the actual ratio of manuscript to print words is 1.2375. Put more directly, approximately 124 words in manuscript come to 100 words in print. All things being equal, we can also say that 100 print words represent 124 manuscript words. With the print words in chapters 1–4 easily countable before us, we can readily estimate the total number of manuscript words Melville might have inscribed.

But this bit of algebra alone does not solve the problem of what specific content Melville might have added to his initial narrative that caused him to bump up the number of chapters in this section by two. To determine this, we need to consider the material condition of Melville’s existing manuscript, in particular the fact that he inscribed his narrative on composition booklets.

As we know, the chapters in the Typee manuscript numbered 10–12 consist of one ten-page booklet and several loose leaves torn from similar booklets. Let us assume that in Stage I Melville initially inscribed his eighteen-chapter personal narrative on a set of such booklets, and that, accordingly, the original nine chapters preceding the three that we now have in manuscript were also inscribed on a discrete, whole number of booklets. That is, Melville did not at this stage tear apart booklets to reshuffle pages or use fractions of booklets, but filled each booklet with text (revising, of course, as he composed or proofread) and moved to the next booklet, filling it as well, and so on to the end of his narrative. It stands to reason, then, that whatever text Melville added to his first nine chapters to make his narrative grow to eleven would have to be text in excess of what he originally placed in the discrete whole number of booklets.

To clarify, let’s imagine Melville writing Typee. He picks up a booklet and begins to write something called chapter 1 by inscribing text, rereading, and revising a bit as he goes along. In the same booklet, he moves directly on to chapter 2, but filling the booklet before finishing his chapter, he selects a fresh booklet and continues his narrative, again filling each side of each page, again moving on to a new chapter, until a booklet is filled, and so on. And so on, into the booklet containing what he first called chapters 10, 11, and 12, and beyond that into booklets that complete his initial eighteen-chapter narrative. With this in mind, we can more concretely visualize a discrete set of booklets preceding chapters 10–12. And applying our mathematical ratio, we can further determine that this set consisted of five booklets.

Here is how I came up with the number five. From the extant manuscript, we know by direct word count that Melville filled the one surviving booklet with 7706 words. Through whatever acts of revision, these 7706 manuscript words eventually yielded 6227 words in the first British edition, which amounts to 518 print lines in the NN Typee edition. This gives us the ratio of manuscript words to print words of, once again, 1.2375. If we take the 518 print lines as a constant print equivalent for any manuscript booklet, and divide this figure into the actual print-line count of the first eleven print chapters, we derive a figure that represents the number of manuscript booklets Melville would have used in writing the text of his original manuscript chapters 1–9, plus the two chapters’ worth of text he added.

Here is how the division works out. Since the text of the print chapters 1–11 amounts to 2886 print lines, the number of booklets Melville would have had to have used in manuscript to write eleven chapters (2886 print lines ÷ 518 print lines per booklet) would have been five booklets, and a remainder of 296 print lines. In this calculation, the five booklets represent the original nine chapters Melville wrote in Stage I and the remainder of 296 print lines represents the additional text Melville would have added to create his two more chapters in Stage IV.

Assuming the constancy of our manuscript-to-print text ratio of 1.2375, we can now use that ratio to calculate an estimated manuscript length for those 296 print lines as being roughly equivalent to 4396 manuscript words. And given that the average word count for a manuscript page is 376, we can also determine that Melville would have composed some twelve additional manuscript pages beyond the five booklets’ worth of material that would constitute his original span of manuscript chapters from 1 to 9.9

These twelve manuscript pages, then, represent the amount of text that Melville added to his originally inscribed opening four chapters, the same pages that in turn led him to renumber his chapters. But since these pages do not survive, we must content ourselves by focusing on their 296 print-line equivalents in the first eleven chapters of Typee, or more precisely chapters 1–4. The question then becomes, which configuration of the sections in chapter 1–4 we have previously outlined comes within range of the 296 print lines of text?

Revision Segment Print Lines
Chapter 1: Complete 202
  1A: Opening 67
  1B: Opening 39
  1C: Queen’s Tattoo Joke 96
Chapter 2A: Stewart description 36
Chapter 3: Complete 197
Chapter 4: Complete 367
  4A: Self-defense 117
  4B: Stewart description 54
  4C: Warriors (Porter?) 18
  4D: Anti-Porter 54
  4E: Savage/Civilized (Stewart/Porter) 37
  4F: Glen of Tior 87

Table 3. Print-line lengths in chapters 1–4

With the print-line lengths of the dozen or so possible revision segments from the first four chapters in front of us (see Table 3), we can begin to determine the limits of possibility regarding the different revision scenarios that would amount to 296 print-lines that Melville may have added. For instance, the minimalist hypothesis that Melville simply added chapters 1 and 3 wholesale to increase his chapter count by two yields a total of 399 print lines. This is considerably more than the target number of 296 print lines; moreover, this view does not account for digressive segments in chapters 2 and 4 that clearly borrow from or are parallel to the Stewart and Porter sources that by themselves are most certainly the kind of late additions we expect to find in Stage IV.

A more complex set of scenarios requires us to see some of Melville’s original opening chapters as “ur-manuscript chapters” consisting of segments of the present print chapters, similar to those we find in Stage II. With this procedure in mind, one scenario might be that Melville’s first manuscript chapter was originally a combination of the two opener sections of print chapter 1 (1A and 1B) and all of chapter 2. In this case, one hypothesis would be that Melville added his ribald jokes (section 1C) and the Stewart description (2A) to this ur-manuscript chapter, separated the resultant overlong chapter into what are now print chapters 1 and 2, and then added chapter 3 wholesale. This would yield 329 print lines, which comes closer to our target of 296. But this hypothesis presumes that Melville’s original opening had its odd double opener (sections 1A and 1B) from the beginning. A more likely hypothesis is that Melville also added the second opener (1B) at the same time as he added the ribald 1C segment, but these additions (1B, 1C, and chapter 3) bring us to an excessive print-line count of 368.

Another scenario is that Melville originally conceived his ur-manuscript chapters 1 and 2 as they now exist in print, but that sections A, C, D, and E of chapter 4 were originally affixed to what is now the somewhat short print chapter 5, in which Tommo meets Toby and plans their getaway. We may call this “ur-manuscript chapter 4/5.” With this configuration in mind, we can hypothesize Melville adding the Stewart description (2A) to chapter 2, all of chapter 3, and either section 4B or section 4F to ur-manuscript chapter 4/5. This insertion would have caused Melville to divide the now-unwieldy ur-manuscript chapter 4/5 in two, separating the chapter 4 materials from chapter 5 to give us (along with the added chapter 3) the two extra chapters. The print-line yields for each case would be 287 (with 4B) and 320 (with 4F). These are attractive print-line counts, but the problem with either scenario is that sections 4B and 4F (which both draw from Stewart) would naturally go together, and yet to include both as a single insertion would give us an excessively high print-line yield of 374.

A third scenario draws upon the preceding two. Here, chapters 1 and 2 constitute a hypothetical ur-manuscript chapter 1/2 along with chapter 3 and ur-manuscript chapter 4/5. The number of hypotheses derivable from this configuration is large, and I shall pursue only one of the more promising ones. In this case, the original opener, manuscript chapter 1/2, consisted of two jaunty and ribald segments (1B and 1C) and all but segment 2A of chapter 2; and the manuscript chapter 4/5 combination consisted of Tommo’s self-defense (4A) and all of print chapter 5. Here, Melville would have added his new opener (1A), stressing shipboard deprivation, and the Stewart description (2A) to ur-manuscript chapter 1/2, and created out of this aggregate two separate chapters 1 and 2 to precede the already present chapter 3. Separating the self-defensive section 4A (which also echoes the deprivation theme of the newly added segment 1A) from ur-manuscript chapter 4/5 to create separate chapters 4 and 5, Melville might have added to segment 4A the three segments 4B, 4D, and 4F, which draw from Stewart, to flesh out what would become print chapter 4. This hypothesis has a preternaturally close print-line yield of 298, but it does not account for the two segments (4C and 4E) which draw upon Porter and amount to 55 lines. However, if Melville turned to Porter as a source only after reading Stewart (as seems to be the case), he may have added these Porter-related sections 4C and 4E well after he revised and renumbered his chapter for his fair copy, and as late as Stage VII when he was sending other Porter-related material, chapters 21 and 27, to Murray. (See Melville Unfolding, chapter 15.)

My purpose in subjecting you to this odd mix of speculation, ur-manuscript chapters, print lines, and arithmetic is to ground our thinking about the more plausible revision narratives of Melville’s late “translation version” of Typee in the material realities of the surviving manuscript. This is a good exercise in testing the limits of fluid-text analysis.

Gansevoort’s Penciled Revisions

As noted, some time after completing his manuscript (probably during Stage V), Melville submitted his work to a reader (probably his brother Gansevoort) for a vetting. The evidence of this discrete substage of revision is a set of faint pencilings on the manuscript so light as to be irreproducible on photocopy or microfilm and even unnoticeable on high-quality photographs. They can be seen in digital pictures but are best inspected directly on the manuscript itself. Most consist of nothing more than an elaborated dot placed in either the right or left margin. In some cases, words in the adjacent manuscript lines are underlined in pencil; and in two instances actual words are penciled in.

It is conceivable that Melville inscribed these markings himself. If so, he would have sat down to proofread, not at a writing desk with inkstand but in a comfortable armchair perhaps, and lightly marked in pencil certain problematic passages to be considered for a more thorough revision in ink at a later time. However, the light and tentative nature of these dots, lines, and words seems rather to indicate that they are the markings of someone licensed only to suggest, not make, changes. Thus, if Melville had made the pencilings himself, one might expect more forceful markings and more spontaneous insertions of words than we do in fact find. And since most of the suggestions are of a grammatical or mechanical nature, Melville would have most likely made the changes on the spot if he were vetting himself. But several factors strongly suggest instead that Melville gave his manuscript to a trusted family member to indicate problems he had with Herman’s text.

Gansevoort Melville is the most likely member. During the first half of 1845, Gansevoort was living at Astor House, the fashionable Manhattan hotel located on the site of the present Woolworth building just north of Greenwich Village. Herman was staying nearby with his other brother Allan at 7 Greenwich Street.10 Cooling his heels while eagerly awaiting a political appointment in the Polk administration for which he had vigorously campaigned, Gansevoort would have had time in his otherwise busy life to provide his editorial services. Since their father’s untimely death in 1832, Gansevoort had been a mentor to his siblings, and editing their writing was a function he had often performed before.

In the papers of Herman Melville’s sister Augusta Melville, we find several school composition books filled with essays on such varied topics as Washington, Robert Fulton and other heroes, “Mrs. [Felicia] Heman’s poetry,” the Sabbath, and the execution of Lady Jane Grey, all of which she wrote from January to June in 1836 while attending the Albany Female Academy. On one of her composition booklets she later wrote: “My composition with dear Gansevoort’s corrections (Albany),” and on her “Lady Jane” essay we find in Gansevoort’s hand the words “Corrected by GM,” with these final initials inscribed as an ornate circular seal.11 In 1836, Gansevoort was himself an exuberant stylist developing his own rhetorical skills; Augusta, a clearly talented but inexperienced writer in need of (and grateful for) fraternal “correction.” He was twenty years old, and she fifteen. Gansevoort’s markings of Augusta’s work are, therefore, far more extensive than the tentative pencilings he made nine years later on Herman’s draft of Typee. On Augusta’s drafts, Gansevoort confidently cancels her lines at will, supplying alternative wording for her. On the Typee draft, we find only three suggested words at two places, a small amount of intense cancellation, and the previously mentioned set of suggestive dots. Obviously, the twenty-five-year-old Herman did not require the kind of help the younger Augusta had needed and appreciated when she was fifteen; and Gansevoort, himself older and busier in 1845, would have had less time and less to correct in his brother’s more mature prose. Gansevoort could use his marginal dots (which do not appear on Augusta’s compositions) to signal problem areas that his talented younger brother could recognize and correct himself, once they were called to his attention.

Despite the absence of such dots on Augusta’s works, enough similarities exist between Gansevoort’s corrections of Augusta and the penciled corrections on Typee to identify Gansevoort as the one who vetted the Typee draft. In both cases we find wavy and looping cross-out lines to cancel multiple-line passages; in both, numerals (1, 2, etc.) are used to indicate the proper sequence of transposed words; and on Augusta’s essay, the penciled words “the” and “life” in Gansevoort’s hand resemble two of the three penciled words found on the Typee manuscript: “the” and “like” (RS10ms43 and RS11ms19).

There is good reason, then, to believe that Herman used Gansevoort as a collaborative reader of his manuscript. The twenty-eight marginal pencilings are invariably associated with a deliberate, responsive revision either in pencil or ink directly on the manuscript or with a change found in the first print editions. They fall into three categories relating to diction, sex, and Polynesia.


Seven revision sites involve a marginal dot with no further indication of what is to be revised, and yet a revision of some sort to an adjacent passage can be found in the first British edition. In most cases these nearly invisible penciled dots correspond to fairly mundane problems. For example, marginal dots appear beside lines containing the phrases “luxurious vegetation around” and “columnar rocks.” These words, unrevised in manuscript, later appear in the British edition as “luxuriant vegetation” and simply “rocks,” respectively (see RS10ms30e24 and RS10ms31e25). Presumably, Gansevoort was calling for more precise diction and image. The most curious instance of this kind of penciled-dot revision site occurs at RS11ms3, in which a still -undeciphered manuscript word (currently read as “chasten”) is underlined in pencil. Since the entire passage in which the word appears is absent from the British edition, we cannot be sure what Melville meant to say here, and apparently the pencil marking indicates that his brother could not make sense of it either.

Twelve of the sites with penciled dots in the margins do have written revisions in the adjacent manuscript line, some appearing in pencil and then finalized in ink, but most appearing in ink only. (Revisions of this sort appear most frequently on the loose-leaf pages of the manuscript.) One example demonstrates all three features in this category.

  1. My companion … did not appear by any means to relish it. [RS11ms191]
  2. My companion … did not appear by any means to relish […] this expression of my wishes.
  3. My companion … did not appear […] to relish the idea. [pencil revision]
  4. My companion … did not appear […] to relish idea. [ink revision]
  5. My companion … did not appear to relish the idea. [RS11e31]

Discussing Toby’s reaction to his plan for Toby to fetch medical supplies, Tommo initially states that his companion “at first did not appear by any means to relish it” (step 1). In an early proofreading phase of revision, Melville had canceled “it” in ink, replacing it with “this expression of my wishes” (step 2). But later, Gansevoort scribbled over “by any means to relish” and Melville’s inserted phrase “this expression of my wishes,” and then wrote below the scribbling near the previously canceled word “it” the replacement words “the idea,” also in pencil. He also placed a dot in the margin to indicate his suggested revision (step 3). Later, Melville returned to this revision site and in ink also canceled what his brother had earlier canceled in pencil, and inserted in ink the incomplete phrasing “to relish idea” to give “he did not appear to relish idea” (step 4). Presumably, in accepting his brother’s suggestion, Melville failed to notice his dropping of “the” in the intended phrase “to relish the idea.” But eventually, the phrase was corrected for the British edition (step 5).

A revision site such as this, with Gansevoort’s penciled dot and revision as well as Melville’s inked-over revision, is rare. Most of the eleven other revisions in this category are nothing more than a dot in the margin next to a line containing Melville’s inked revisions. The assumption is that Gansevoort did not feel the need to pencil in any specific revisions, and that his dot was enough to draw attention to an infelicity that his brother could fix in his own way. Or, perhaps, that the dot signaled a “talking point” for Gansevoort to remind him in a later conversation of what was needed. In one case, Melville has Toby say: “When the first stunning effect had subsided & left me restored to consciousness I perceived the three savages” (RS11ms93e126). Here, Gansevoort’s dot signals the problem of “left me restored to consciousness.” On the one hand, the phrasing, which would have us conceive of the subsidence of an effect as both a leaving and a restoring, is too abstract and finally inept. At the same time, this kind of convolution is not characteristic of Toby nor appropriate for the exciting narrative he is telling. One of Melville’s principal revision strategies was to clarify Toby’s voice (see Melville Unfolding, chapter 8), and in this site we see Gansevoort abetting that endeavor. And Melville’s later inked revision “corrects” the problem. He canceled “left me restored to consciousness” and inserted a new sentence down the left margin: “As soon as I regained my consciousness I perceived the three savages.” While typical revision, playing tentatively upon the margins of consciousness, is primarily grammatical, two other revisions involve sexual content.


In the highly sexualized passages that follow Toby’s second and successful attempt to leave, Melville had originally elaborated upon the native girls’ consoling body massage of Tommo by alluding to John Gay’s The Beggar’s Opera (1728).12 Gansevoort, however, felt the passage’s salacious details required censoring:

With Captain Macheath in the opera I could have sung “Thus I lay like a Turk with my doxies around,” for never certainly did effeminate ottoman in the innermost shrine of his serglio attended by lovlier houris with more excess of devotion than happened to me on these occasions I have mentioned. (MS12ms142e197)

Without supplying any marginal dot, Gansevoort marked out the clause referring to Macheath and his doxies with insistent looping lines. He also canceled the conjunction “for,” and capitalized the n in “never” to make the “ottoman” clause a sentence of its own. Presumably a misspelled “serglio” was more acceptable to Gansevoort than Macheath’s prostitutes. Melville finalized his brother’s censoring of the one and not the other by inscribing his cancellation in ink over his brother’s in pencil. In due course, the “ottoman” line and the subsequent reference to the mad, wife-killing Sardanapalus were also censored, by whom we cannot be certain, before Typee reached print. Perhaps Herman thought better of it as he prepared his fair copy; perhaps John Murray’s copy editor, Henry Milton, excised the line before sending the text to a printer; perhaps Gansevoort in London made the cut while he read page proofs.

A second instance of penciled sexual censorship occurs with Melville’s “college of Vestals” joke following Kory-Kory’s fire-lighting scene (RS12ms179). As discussed in Melville Unfolding (chapter 8), the joke is that there are no virgins on the island to supply such a “college.” Initially, Melville had winkingly masked his punch line by vociferously refraining from discussing the implied “slanderous aspersion” on the native girls; it is something, he says, that “I courtiously beg I may be excused from more distinctly pointing out.” A penciled dot in the left margin calls attention to the phrasing, and Melville tinkered with it at a later point changing “I may be excused” to “to be excused.” Still later in the British edition, the entire sentence was drastically revised to obscure the joke even more.

The sexual nature of the pencil revision sites involving MacHeath and the vestal joke reveal the consequences of Melville’s development beyond the audience within himself to actual audiences represented by his brother’s sensibility. In recollecting the sensual scenes that in his later revisions became compensations for his despondent guilt over Toby, Melville had expanded his sexual self-awareness about as far as his own sense of decorum would allow. But with their mother and family in mind (and the reading public as well), Gansevoort insisted—his looping lines seem particularly intense—that his brother rein in his sexual self-exposure. The partial censorings indicate that Gansevoort was no prude; he could take oriental sensuality but not John Gay. Herman’s straightforward cancellation in ink indicates full and unambiguous agreement. Thus in a stroke of the quill, Melville was acknowledging his brother’s collaboration and through him negotiating his sexuality with a palpable, representative audience.

Equally important here is the amount of energy a set of men—two American brothers and two British editors—were willing to expend, seriatim and probably not in direct consultation but nevertheless collaboratively, in order to preserve a joke (no matter how obscured) about Polynesian women for Victorian readers, male and female. At any stage of the process, an editor (including Melville himself) could have deleted the punch line altogether, but Melville and his revisers retained the joke through incremental editorial mutilations. The repeated gendered revising amounts to a kind of male fetishizing of the joke. More than a winking acknowledgment of the comic impossibility of virginity, the strategy of preservation through obscuration expresses a deeper cultural anxiety over sexual promiscuity, and perhaps the nineteenth-century system of brothels and prostitution.


Nine revision sites involve a marginal pencil mark connected to penciled revision instructions, but no inked revision. And yet, these pencil revisions do appear in the first British edition of Typee. For example, Melville first described a native crowd as “a group of gossiping idlers gathered about the tavern door of a village” (RS10ms3e3). Later, Gansevoort penciled the numbers 1, 2, and 3 over “door,” “village,” and “tavern,’ respectively, to indicate a reordering of the phrase to “door of a village tavern,” but no finalized ink revisions were made in manuscript. The first edition version of the phrase, however, appears with its properly revised sequencing. In another instance, Gansevoort inserted the word “like” in pencil between the words “cathedral gloom” to give the phrase “cathedral like gloom” (RS10ms43). Although no change is made in ink, this pencil revision does appear in print.

Most likely, Melville assumed that these more specific pencil markings could stand as a revision to be adopted, without his inking over them, when the fair copy was prepared; and his confidence that this would happen argues for the relative lateness of the pencil revisions in the composition process, perhaps as late as Stage VI, after he had consulted his source books, made significant expansions to his text, and was about to prepare his fair copy.

Pencil revision sites of this kind dealing with Polynesian language support this late dating. Gansevoort placed a penciled dot in the margin near Melville’s rendering of the Polynesian word “kikino” with the alternative “keekeeno” placed (uncharacteristically) below the baseline (RS11ms135). Gansevoort’s intended focus in this line is not entirely clear. If Herman had actually inscribed two alternative spellings at this site and left them hanging for a later decision, then Gansevoort’s penciled dot would simply call Melville’s attention to his need to make a choice. But on the same line, Melville also has Kory-Kory saying “Tipi kikino” instead of what the context requires, “Happa kikino,” and his point may be that Melville needed to correct the erroneous designation. In fact, the manuscript shows that Melville inscribed “Hapa” over “Tipi.” But if Melville had already made that correction before submitting the manuscript to Gansevoort and if he had not yet supplied an alternative to “kikino,” a third revision scenario at this site would be that Gansevoort’s dot refers to Melville’s need to clarify the sound of the Polynesian word “kikino.” If so, Melville would have at that time then inscribed the alternative “keekeeno” below “kikino” and left his choice between the two for another day. Only in the first British edition do we find his decision to follow “keekeeno.”

These and other possibilities are elaborated in the revision narrative for this site, but for our purposes here, these odd dots and scribblings in pencil and ink record a moment of both decision and indecision over what to do about the Polynesian language he had been using. Most urgently he needed to regularize the orthography. But what was not entirely clear to him was how to do that and for whose ears. Should he render the Polynesian in a format found in books he had consulted, or render it in a style that would approach the sounds familiar to an English speaker?

One revision site indicates Gansevoort’s active participation in solving the problem. Originally, in the scene in which Toby’s limp body is brought into the village, Herman has the women announce, “‘Awa! Awa! Toby mucke moe!’—(Alas, Alas, Toby is killed!)” (RS11ms63). Gansevoort placed a dot to the left of the line and underscored “mucke moe,” perhaps to call attention to how the phrase might be pronounced. In response, Melville squeezed an additional “e” at the end of each word to render “muckee moee,” a less ambiguous spelling and a transliteration of the Polynesian more consistent with the double e’s in “keekeeno.” Also, in this manuscript line, h’s have been inserted in “Awa! Awa!” to give “Awha! Awha!” Here, Melville uses the English wh orthography for the aspirated hw vocable to familiarize and westernize the Polynesian sound. This revision probably occurred to Melville as he was altering “mucke moe.” The decisiveness of the transliteration of “muckee moee” alongside the indecision of “kikino / keekeeno” only enhances our sense of Melville’s indeterminacy, at this particular compositional moment, over the language he would translate for readers.

Gansevoort’s apparent tinkerings with “keekeeno” and “muckee moee” indicate that he may have collaborated with Herman on the problem of Polynesian sounds. Before the discovery of the Typee manuscript, readers and scholars had doubted the authenticity of Melville’s engagement with the Marquesan language because his Polynesian transcriptions in print did not seem to conform to what later visitors to Nuku Hiva thought they were hearing. But the Typee manuscript reveals that Melville understood more than we had been previously allowed.

Although Marquesan grammars and word lists appeared as early as 1842, Melville claims in Omoo not to have seen them and says that his word usage “has been mostly governed by the bare recollection of sounds” (NN Omoo, xiv). And the numerous variant spellings of Polynesian words (like Taipi and Nuku Hiva in particular) suggest that Melville was indeed experimenting with spelling as he wrote to render sounds that could match his “bare recollection.” Even so, the word “mostly” suggests, as Harrison Hayford believes, that Melville also borrowed spellings from his Polynesian sources. This view applies to Typee as well, although it is clear from the author’s preface to Typee (written well after the manuscript stage of composition) that certain aesthetic and rhetorical considerations also “governed” his orthography. Here, Melville relates that he has adopted spellings “which might be supposed most easily to convey their sound to a stranger.” He notes that in other books about the South Seas (presumably sources like William Ellis, Stewart, and David Potter) , “the most beautiful combinations of vocal sounds have been altogether lost to the ear of the reader by an over-attention to the ordinary rules of spelling” (NN Typee, xiv), and goes on to imply that he will counter that tradition by departing from overly precise phonetic transliterations to help the reader hear the beauty of Marquesan more directly, through regular English orthography. Presumably, we will hear better (and more accurately) if the words we read more closely resemble familiar English spelling conventions. The manuscript bears out this strategy: Melville’s original “Tipii” (which we might mispronounce as “tippy” or “tie pie”) becomes “Typee,” and we now read the romanticized “Fayaway” and more familiar “Happar” rather than the odd but more precisely spelled “Faaua” and “Happaa” found in manuscript.

Max Radiguet, who published his account of Du Petit-Thouars’s Polynesian exploits some years after Typee, found Melville’s title to be “bizarrement orthographié.” And Robert Louis Stevenson, a late admirer of Melville in most respects, seemed to have missed Melville’s point in the preface altogether when he remarked that Melville “had no ear for language whatever: his Hapar tribe should be Hapaa.” Modern scholars, such as Anderson, translator Jacqueline Foulque, Harrison Hayford, and Walter Blair, have defended Melville’s spelling from various angles, all of which are treated in Hayford and Blair’s thorough note on the matter in his and Blair’s 1969 Hendricks House edition of Omoo.13 Hayford contributes the idea that Melville may have also altered spellings to conform to his own Northeast American dialect in which the syllable ar would be pronounced “ah”; hence, his “Happar” would actually be pronounced “Happah,” and therefore more closely resemble the Polynesian sound he was aiming to reproduce.

This is compelling logic, although it is clear that in this particular transmutation of the word, Melville does not come any closer to the actuality of Polynesian but rather is moving away from the original, double ah ending of Happaa. To be sure, the doubling and tripling of vowels in Oceanic languages is, especially to modern Western “strangers,” the least familiar aspect of Polynesian pronunciation, as is evident in the word Faaa (meaning airport and pronounced “Fah-ah-ah”), which is almost comically exhausting for English speakers to reproduce. Melville’s ar substitution in Happar (presumably pronounced “Happah” in his accent) would have certainly eased the burden for many readers from having to pronounce “Happah-ah,” but, again, at the expense of what a Polynesian would consider was a beautiful vocal sound in its own right. Melville’s word “Tior” for the place name “Taioa” (pronounced “Ti-o-ah”) is another case in point, for the or ending here again conflates two repeated syllables into one . Further complicating Hayford’s thesis is that Melville’s eastern pronunciation of or (unlike his ar for “ah”) would not be heard as a more Polynesian “ah” sound in his dialect.

To a certain limited degree, then, Radiguet and Stevenson were right: Melville’s Polynesian, in print, is off the mark, even if we take his eastern accent into account. But Melville’s manuscript renderings of Polynesian spelling indicate (contra Stevenson) that he heard Polynesian well and initially transcribed the sounds with an informed (although not always consistent) sense of the conventional transliteration of Polynesian printed in Melville’s day. But as he was getting ready to make his fair copy, he liberalized his literal renderings of Polynesian to familiarize the oddly beautiful sounds for Western ears. His transliterations were most decidedly a conscious translation.

Initially, Melville did use the “ordinary rules of spelling” in rendering his Polynesian, at least as much of those rules as were evident in what he would have absorbed through reading, and he did not seem immediately concerned with modifying his own experimental spellings to suit his or anyone else’s ear. Throughout the manuscript we find Melville always opting for what might be called the more typical-looking Polynesian spellings. Contrary to Stevenson’s specific example, Melville uses “Happaa” consistently. Initially, he uses “Maheyo” instead of “Marheyo” and “Kori-Kori” instead of “Kory-Kory.”14 He spells Typee without the anglicizing y, writing instead the word “Tipi” or “Tipii.” Chances are he picked up his final spelling of Typee from Porter, a source he did not have at hand until late in the process. In addition, Melville spells Nuku Hiva in so many different ways that it seems likely that he had not, in the initial stages of composition, seen it written in any source book but was indeed operating under his own “recollection” of the sound. He gives us “Nuiheiva,” “Nuheva,” “Nuuheva,” and what finally ends up in print, “Nukuheva.”

Gansevoort’s pencilings have led us to surmise that Melville did not consider standardizing his spellings until late in the compositional process, just before Melville prepared his fair copy. And the appearance of Melville’s use of Porter’s spelling of Typee, not in manuscript but in print only, not only confirms the idea that Melville composed his personal narrative first without source books but also leads us to believe that his decision to Westernize his spellings came quite late in the process and only after he did in fact resort to certain source books. Additional evidence involving Fayaway and Marheyo corroborate this idea.

Originally, Melville consistently named Tommo’s lover using the thoroughly Polynesian spelling of “Faaua” (pronounced Fah-ah-oo-ah).15 The word “Fayaway” appears only once in manuscript (RS12ms56), the product of a curious revision. Here, letters have been squeezed within and around “Faaua” to create “Fayaway.” Since this careful over-inscription occurs on the same leaf on which Melville straight-pinned the insertion slip that dramatizes Fayaway’s gaze (discussed in Melville Unfolding, chapter 9), we might conclude that the shift to the Westernized spelling was part of Melville’s late-stage sentimentalization of the young girl. That is, once Melville had decided to elaborate upon his lover’s sympathetic glance→gaze, he also considered romanticizing her name: Faaua→Fayaway.16 Thus, in creating out of the guttural Polynesian an English word romantically evoking (let’s say) “departing faith or fairies” (fay away), Melville was quite clearly giving his Faaua a wistful Western presence designed to familiarize her gaze, just as Melville had Tommo succumb to those eyes.

A final bit of altered Polynesian spelling indicates a still later dating of Melville’s Westernizing translations. On the inside of the cover folder Melville made to contain his manuscript, the writer tinkered with spellings for Fayaway’s eccentric father, “Maheyo.” Arranged in a column, the trial names are “Marheyoo,” “Marheeyo,” and “Marheyo.” Presumably, Melville was experimenting with sound variations in the old man’s name to evoke certain comic associations: the come hither of “Hey you” in Marheyoo, the braying of “Hee yo” in Marheeyo, and the more genial “Heigh O” of Marheyo. Since “Maheyo” appears throughout the manuscript and since the revised “Marheyo” appears only once on the cover but then consistently throughout the print version, we can deduce that Melville composed this list either just before or during his fair-copy stage.

Gansevoort’s precise participation in Melville’s revisions of his Polynesian cannot be fully determined. At the very least his pencilings suggest a need for his brother to standardize spelling, and to the degree that the double e’s in “keekeeno” and “muckee moee” create in Western ears a more familiar sound, we can argue that Gansevoort was also urging his brother to consider the needs of his audience. Gansevoort may not have been the one to suggest the “Fayaway” revision but his penciled suggestions elsewhere may very well have sustained or even initiated a discourse that led to the change. With some help from his brother, Melville was beginning to resolve his quandary over his “interpretation” of the Polynesian speech that gave Tommo rheumatic pains and the headache. He was beginning to liberalize the literal —not to demean Polynesian language, but to make it more accessible.

Overall, the twenty-eight penciled revision sites offer an intriguing view of the kinds of issues that concerned Melville at a late moment in the creative process when he had reason to feel that he had reached a certain degree of completion. They are striking material evidence of Melville’s translation version of Typee, for they show the writer responding to the behest of a third reader representing readers to come by revising to achieve the more fluid expression of an idea, the more restrained exposure of sexuality, or a more liberal rather than literal treatment of language.

Previous: Writing Typee


1. Review of Typee in New Englander [New Haven] 4 (July 1846): 449; rpt. in Higgins and Parker, Contemporary Reviews, p. 51.

2. In completing The Confidence-Man, Melville compiled a list of chapter titles, and we might be tempted to use this document to suggest that Melville compiled a similar list of subheadings for Typee. But The Confidence-Man was the last prose work Melville would prepare for publication in his lifetime, and he had most certainly developed procedures at that point in his career which he would not necessarily have followed in preparing Typee .

3. These added materials are largely comic and demeaning in their insistence upon the infantile nature of the natives in both warfare and dress. One can speculate, then, that these additions were made before the revisions we find in manuscript that reduce this kind of comic stereotyping in Kory-Kory (see Melville Unfolding, chap. 11).

4. As noted in Melville Unfolding (chap. 5, n. 4), Charles Anderson contends that Melville drew his description of tappa-making from William Ellis’s Polynesian Researches, but there are virtually no textual similarities between Melville and Ellis in the passages Anderson cites; see Melville in the South Seas (New York: Columbia University Press, 1939), p. 749. In fact, tappa-making was something anyone in Polynesia was likely to witness.

5. While the insertion slip itself does not survive, the text written on that document, and intended for insertion at a place marked on a surviving manuscript page, does exist in print at the designated place.

6. Anderson, assuming that Melville consulted sources before writing his narrative, suggests that Melville’s reading of David Porter’s Journal may have influenced his writing of the skirmish in print chapter 17 (Melville in the South Seas, p. 135). But the passages Anderson cites are a general depiction of warfare with no verbal echoes of Porter in Melville’s description; moreover, Porter does not mention any musketry. In my view, Melville had not yet read Porter when he wrote up the skirmish, and whether or not he witnessed any actual skirmish during his island stay, the muskets seem to be derived from his personal experience. As discussed in Melville Unfolding (chap. 15), Melville did add specific appropriations from Porter’s Journal to chapter 17 after he had finally read that notorious text in Stage VI.

7. Henry Hughes, “Fish, Sex and Cannibalism: Appetites for Conversion in Melville’s Typee,” Leviathan: A Journal of Melville Studies 6, no. 2 (October 2004): 3–16.

8. Leon Howard argues that Melville sent these additional materials at Murray’s request (NN Typee, pp. 279–80), but according to Hershel Parker, shipping schedules would not have allowed Melville enough time to receive the request and comply, and it is more likely that Melville himself initiated the addition of these three chapters. See “Evidence for ‘Late Insertions’ in Melville’s Works,” Studies in the Novel 7 (Fall 1975): 409–13.

9. From the outset, we need to recognize that this backward-projecting, arithmetical procedure, although rooted in material facts, cannot be conclusive. The deductions assume that Melville did, in fact, use booklets from the beginning and that he used them in a consistent way, mixing into them from the beginning the same writing phases he applied to the one surviving booklet. Moreover, the assumption is that the ratio of manuscript production to print yield is a constant throughout the unrecovered manuscript, and that the one surviving booklet with its particular yield can be applied to the entire production. And even if these assumptions could be proven true, there is always the possibility that a factor utterly unaccounted for entered the process. Any line of deduction is open to counterclaims, so that these calculations bear a remarkable resemblance to a house of cards. Nevertheless, we have enough data to speculate responsibly, if not definitively, upon various combinations of material Melville might have added. The 296 print-line limit cannot prove the greater likelihood of one revision scenario over another but only offers one manuscript-based way to fine-tune such speculation.

10. Parker, Herman Melville, p. 373.

11. New York Public Library, Gansevoort-Lansing Collection Box 308/2 and 4. Most of Augusta’s essays appear on quarto-size composition booklets, shorter but similar in paper stock to Melville’s manuscript booklet. A typed note accompanying one booklet (which is closer in size to Melville’s Typee manuscript booklet) erroneously identifies the material as Allan Melville’s composition corrected by Peter Gansevoort. However, a second note by Melville scholar Henry Murray gives the proper attribution of Gansevoort correcting Augusta, which is evident from Augusta’s own attributions on the document itself.

12. David Ketterer was the first to discuss the passage in print; see his “Censorship and Symbolism in Typee Revisited: The New Manuscript Evidence,” Melville Society Extracts, no. 69 (February 1987): 6–8.

13. Max Radiguet, Les derniers sauvages: la vie et des mœurs aux Iles Marquises, 1842–1859 (Papeete: Editions du Pacifique, 1981), p. vi. For Stevenson’s pronouncement in an 1890 letter to E. L. Burlingame, see Jay Leyda, The Melville Log, (New York: Harcourt, Brace, 1951), p. 822. Anderson brings Radiguet and Stevenson together in a note on Melville’s spelling (p. 451, n. 24) as do Hayford and Blair in their extensive note on orthography (pp. 343–47); Herman Melville, Omoo: A Narrative of Adventures in the South Seas, ed. Harrison Hayford and Walter Blair (New York: Hendricks House, 1969).

14. Parker (Herman Melville, p. 361) asserts that Melville first considered “Kori” as the name for Kory-Kory, and while the name “KoKori” (obviously an uncorrected misspelling made in haste) exists on one manuscript leaf, I have found no evidence to support his assertion. Elizabeth Renker’s reading of one instance of “Kori Kori” as “Kiri Kiri” seems unlikely; Strike through the Mask: Herman Melville and the Scene of Writing (Baltimore: Johns Hopkins University Press, 1996), p. 9.

15. In transcribing the only manuscript leaf available to them before the 1983 discovery of the larger manuscript fragment, the NN editors misread “Faaua” as “Faawai” because the original word had been obscured by inserted letters and over-inscribed markings so as to transform “Faaua” to “Fayaway.” The misreading “Faawai” was a good guess under the circumstances, but it does not appear elsewhere in the manuscript fragment, which consistently gives “Faaua.” In this spelling, faaua does not appear in any Marquesan word books I have inspected and its meaning is not clear. It may be a combination of faa (Pandanus) and au (leaf), i.e., pandanus leaf, but this is just another guess. The Marquesan dialect, however, is linked to Tongan, and John Martin’s Tongan “Grammar and Vocabulary” defines “Feáooagi” as “mistress lover or sweetheart” (see William Mariner, An Account of the Natives of the Tonga Islands of the South Pacific Ocean with an Original Grammar and Vocabulary of their Language. Compiled and arranged by Dr. John Martin. Edinburgh, 1827, p. lviii).

16. This singular conversion to Fayaway is the only evidence of orthographic tinkering on this particular manuscript page. That is, other Marquesan names (Kori Kori, Maheyo) have not been Westernized. Interestingly, the loose leaf on which the Fayaway revision occurs has a provenance independent of the rest of the manuscript. This single leaf had been part of the Paltsits donation to the New York Public Library decades before the 1983 discovery of the larger Typee manuscript with its other loose leaves and booklet. It is not known how or why the Paltsits leaf was separated from the other loose leaves. One remote possibility is that after the publication of Typee, Melville (or someone else) gave the single manuscript leaf away as a souvenir and at that time altered “Faaua” to “Fayaway” to provide the more familiar spelling of the name made famous through its publication.