Total translation

Navajo song and the story of US modernism

In her deconstruction of Mary Austin’s ‘The American Rhythm,’ Leah Dilworth argu
In her deconstruction of Mary Austin’s ‘The American Rhythm,’ Leah Dilworth argues that modernists held American Indian culture to be fundamentally analogous to that of ancient Africa and China. Above: Austin's 'American Rhythm,' Eda Lou Walton's 'Dawn Boy,' and 'Navajo Songs,’ 1933 and 1940 field recordings from settlements in New Mexico and Arizona courtesy of Smithsonian Folkways Recordings.

American Indian[1] culture attracted many white poets to the Southwest in the early and mid-twentieth century. Educated in Anglo American traditions, but compelled by the modernist urge to develop new poetic forms, poets from Mary Austin to Jerome Rothenberg went to great lengths to represent what they were hearing and feeling. D. H. Lawrence wrote of his experience of Hopi dance, and he and Witter Bynner composed lyrics purportedly inspired by the Taos Pueblo. In American Rhythm, Austin “re-expressed” the music of several peoples, including the Paiute and Shoshone, and in Red Earth, Alice Corbin Henderson claimed to write “from the Indian,” naming the San Ildefonso and Tesuque pueblos. In the mid-twentieth century, Kenneth Rexroth listened to recordings by Frances Densmore and admired her renderings of songs from peoples of the upper Midwest, including the Teton Sioux and Chippewa. In working with American Indian music and oral expression, these Anglo modernists were taken back to basic questions about the nature of poetry as verbal art, and they were challenged to examine their assumptions about what constitutes the “primitive.”

Rather than surveying the poet translators and the wide range of Southwest American Indian cultures that interested them, this essay focuses on Navajo song and two fundamentally different poets in whose work the problems of translation are concentrated: Eda Lou Walton, a poet-anthropologist who was associated with many of the writers in the Greenwich Village and New Mexico of the 1920s and ’30s; and Jerome Rothenberg, whose study of recordings of Navajo singer Frank Mitchell was part of his efforts to define the field of ethnopoetics in the 1960s. I will examine Walton’s poetic rendering of her sources, and argue that her methods and poetry differ from the neo-Romantic work of her contemporaries. Walton’s writings demonstrate an unresolved tension between her determination to appreciate Navajo poetry as powerful in its own regard and her tendency to present it in relation to modernism. What Walton called her “re-creations” of Navajo song seek the authority of ethnography (in this case, Washington Matthews’s foundational studies), trade sound for image, and set aside the experience of rhythm in the interest of content.[2]

In contrast, Rothenberg’s “total translations” attempt to account for the entire soundscape of the source material.[3] This marks a departure from his predecessors, especially in the treatment of vocables, the nonlexical syllables in Navajo song.[4] Rothenberg’s performance poetics, as they manifest in his recordings of and writings about “The 17 Horse Songs of Frank Mitchell,” show poetry’s capacity to transpose something essential about music into a target language and poetic form. They do so while honoring difference and acknowledging the contingency of translational acts. In my estimation, Rothenberg’s work is an example of what Eve Kosofsky Sedgwick called “reparative reading,” a form of critical engagement that invites experimentation, acknowledges affect, and seeks to repair the damages of (in this case) white Anglo misrepresentation of American Indian expression. Rothenberg’s work should be understood as part of a line of such efforts to listen and respond across cultural boundaries — a line that includes Walton.[5]

Critics and anthologists have tended to treat Walton and Rothenberg as members of separate movements and periods. Rothenberg belongs to the sixties and seventies avant-garde and the translation debates of Dennis Tedlock, Dell Hymes, and others. Largely forgotten, Walton might be classed among the neo-Romantics and imagists of the first wave of modernism. Moreover, traditional American Indian song does not factor into accounts of twentieth-century US poetry even though, as we shall see, audio recordings and translations made it integral to Austin, Walton, Bynner, and others’ experiments in lyricism and imagism, as well as Rothenberg’s performance poetics. Studying Navajo song through its modern poet translators opens the possibility for a multicultural, cross-genre conceptualization of twentieth-century US poetry and poetics, one that expands upon accepted scholarly narratives and groupings. In their efforts to “carry across” without eliding difference, and to regard translation as a scholarly and creative act, Walton and Rothenberg are significant figures in this alternative story. 

Fascinating rhythm

In his history of translations of American Indian song, Arnold Krupat asserts that translators “cannot help but place themselves in relation to Western conceptions of art (literature) or of (social) science as they inevitably privilege either the Sameness of Native American verbal expression in forms aspiring to what is accessibly recognizable as literary, or its Difference, in forms committed to scientific authenticity and accuracy.”[6] Twentieth-century poet translators moved along this continuum of “identity” and “difference” with varying degrees of self-awareness. An intense shared interest in American Indian music, especially in the early part of the century, was an expression of high modernist primitivism. In her lucid deconstruction of Mary Austin’s The American Rhythm (1923), Leah Dilworth argues that modernists held American Indian culture to be fundamentally analogous to that of ancient Africa and China. These traditional cultures, modernists claimed, could provide models of social cohesion where modernity tended toward expansion and fragmentation. The daily life of the American Indian was said to integrate labor, spirituality, and art.[7] Without knowledge of the languages or, often, firsthand experience of the cultures, these white, Anglo poets often relied on late nineteenth-century ethnographies by Horatio Hale and Washington Matthews that included “literal” and “literary” translations in which American Indian song seemed to offer attractive alternatives to English prosody.[8] American Indian representations of the world were read as symbolic, their forms spare; poets such as Austin, Henderson, and Bynner considered those forms to be strikingly comparable to imagist aesthetics. As Austin in particular maintained, their poetics were native to American soil and thus a natural source of “authentic” American poetic voice. What Dilworth says of Austin and American Indian arts might be said generally of modernism’s perspective on so-called “primitive” culture: that “the main value of American Indian literature was not that it was a lasting voice unto itself, but that it could be used to revitalize a somnolent Euro-American tradition.”[9]

This troubling dynamic between modern poetry and traditional forms is, to an extent, true of Eda Lou Walton’s work with Navajo song, though Walton deserves a closer look than she has received by critics to date. A versatile scholar-poet and socialite, well known in the 1920s and ’30s, Walton rarely appears in critical assessments of the period. The most extensive writing about her in recent decades can be found in Steven G. Kellman’s biography of novelist Henry Roth, for it seems that Walton is now remembered principally as muse and lover of the author of Call It Sleep. Walton was an English professor at NYU, the only female instructor with a PhD in the college, and known among students as a champion of modernist writing. Roth met her through a friend, and it was in her apartment that Roth discovered books by Joyce, Eliot, and other avant-garde writers of the moment. Kellman reports that Walton associated with many of the artists and intellectuals in New York City — Margaret Mead, Edna Millay, Mark Van Doren, Louise Bogan — and that she had a knack for discovering and promoting young talent. Not unlike Millay, Walton’s “new woman” image could be as much a liability for her public reception as a benefit; according to Kellman, her sexual decisions often prejudiced opinions of her. Nonetheless, gatherings at her apartment in the West Village were an integral part of the literary and arts scene.[10]

Walton grew up in New Mexico and wrote a dissertation at UC Berkeley on Navajo culture. Some of this research was later published in the Journal of American Folklore; her article “Navajo Song Patterning” is an important reference point for understanding not only Walton’s “re-creations” but also the typical modernist assumption that American Indian song should be treated as poetry.[11] At least one reason Walton held this assumption seems clear. While Walton likely witnessed Navajo music and dance during her youth, her sources for the article and for her own Navajo-influenced poetry are all textual: the transcriptions, many in verse form, were made by Washington Matthews in his series of late nineteenth-century studies of Navajo music and legend.[12] I look comparatively at Matthews’s and Walton’s texts below; for now it is important to emphasize that the textuality of Matthews’s song transcriptions underwrites Walton’s treatment of Navajo song as poetry. David McAllester and Douglas Mitchell identify two broad categories of Navajo music: diné biyiin, which are casual and sacred nonceremonial songs, and hatáál sin, which are the hundreds of songs imbedded in Navajo ceremonials.[13] They list several general characteristics that they claim are unique to Navajo music: robust and nasal voice; downward-tending melody; introductory formulae that identify particular songs; a scale that emphasizes the tonic, third, and fifth; limited note values, predominantly eighth and quarter notes; and duple meter. Walton studies and translates various examples of both ceremonial and nonceremonial Navajo music into poetry, but many of the elements of sound are lost in those translations because of her emphasis on the formal structures of Navajo lyrics.

Walton’s analysis of Navajo song is, on one level, an effort to confer sophistication on a so-called primitive form by delineating its similarities to something in her estimation more advanced: the conventions of English verse. In The Invention of Native American Literature, Robert Dale Parker argues that even the choice to render oral literatures into textual form betrays an unexamined Eurocentrism. Parker shows that American Indian narratives were often translated as poetry both to promote an idealized affinity between the music and modernist free-verse poetics and to elevate an oral form to the status of Western lyricism.[14] The American Indian singer became, Parker implies, an exotic precursor to the modernist poet. Even so, some forms within the Navajo ceremonials are poetry “by most definitions,” as Sam Gill writes of prayers.[15] Gill shows that Navajo prayers develop from a set of constituent parts, an oral composition process similar to that identified by Parry and Lord in epic poetry. “Even in English translation,” he writes, “it is clear that rhythm and repetition are essential elements of style in Navajo prayer.”[16] What is lost in translation of this poetry to the page are the multisensory qualities of performance, a fact that would lead Jerome Rothenberg beyond print.

Walton assumes that all American Indian vocal art is poetry, and her analysis of the Matthews texts is at once a defense of “primitive poetry” as protomodern and a demonstration that it is less sophisticated than the modern. At the outset of her essay, she asserts that Navajo poetry is “no less formalized than that of advanced civilizations”; rather, our unfamiliarity with the unique formal patterns causes us to overlook them, and her task is to make them apparent: “The Navajo patterns are as important in poetic composition to the Navajo Indian’s mind as is any English verse form to an English poet.”[17] In insisting upon the complexity of the “primitive,” Walton’s strategy here is typical of early twentieth-century defenders of folk art, and like Matthews before her, she concentrates on such structural elements as repetition and parallelism to make Navajo song legible as poetry in English. Yet although she finds the poetry to be highly crafted, she also identifies its limits. Like “advanced” verse, its lines and stanzas evince incremental and antithetical parallelism, but Walton discerns little more of note, at least structurally, in the 150 songs she claims to know.[18] This, too, is typical of her moment. The primitive deserves intellectual attention and respect, it is argued, and yet Navajo art must be understood as a precursor.

This contradiction is central to the genealogy of US poetry that Austin, Walton, and other modernists wanted to claim, however unacknowledged. That genealogy traces a lineage from, in this case, the Navajo to modern American poetics and, at the same time, construes the two as radically contemporary. The primitive culture’s simplicity and cohesion (Walton ends her essay by noting that Navajo poetry and baskets follow the same simple patterns) must be recaptured in imitation and seen as one with the avant-garde modern, and yet primitive culture must remain remote, different from the present, unsullied by it. In Playing Indian, Philip Deloria explains that this strain of modernist primitivism is a feature of mimesis, that miming the American Indian allowed modernists to appropriate the primitive while keeping a respectable distance. “Recognizing that one is becoming Other,” he writes, “asserts a boundary line between one’s always-forming Self and an Other that is most certainly not the Self. On the other hand, as a way of becoming Other, mimesis seems to insist that Self and Other are, in fact, the same.”[19] As we shall see, Walton’s own “Indian” poetry enacts a similar mimesis. Even in her analytical writing about Navajo song, Walton casts the American Indian as her Self in an earlier form.[20]

When Walton and her contemporaries “play Indian” on the page, they carry out mimesis using the language of sensibility. Feeling Indian is essential to playing Indian. Rhythm in American Indian music is the engine that drives these modernists’ own feelings and experiences. We see this in the introductory note that the poet Witter Bynner provided for Walton’s Dawn Boy: Blackfoot and Navajo Songs. His note characterizes rhythm’s role in the physicality of American Indian song: “On one occasion, remaining several days [at a ceremony], I felt myself become an established piece of earth, vibrant to the rhythm of their feet and voices. On another occasion, I watched a fellow-poet shed tears to their rhythm and heaven shed prayed-for rain. Without understanding any words of the rhythm, I have been made sure of its intimate power and beauty.”[21] Bynner writes of contact, of being touched. First as the ground beneath the dancer’s feet, then as an empathic participant in his colleague’s tears, Bynner feels his way into Indian-ness. While Bynner’s presence at the ceremony may seem to authenticate his claims, the dancers (in this case, they were probably Taos Pueblo) remain Other. Unfamiliar with the languages, Bynner makes a case for the semantic weight of rhythm. He can be “sure” of his feelings because the rhythms communicate in what he assumes is a universal and intimate language, music.

Bynner appropriates Mary Austin’s argument, in her provocative treatise The American Rhythm, that rhythm is experienced before it is understood, because rhythm occurs in the body.[22] The rhythms of daily activity, dance, ritual, and poetry are “autonomic,” she asserts, and moreover all are tied to the land of one’s origin. By this logic, Austin is able to claim that when Europeans immigrated to the Americas, their “natural” rhythms necessarily changed. Because of the “intensity” of colonial experience, this evolution happened remarkably quickly: new earth, new quotidian demands, new push and pull, stress and unstress.[23] Austin’s essentializing anthropological discourse underpins her central claim: that it is unnatural for American poets to imitate European poetic forms, that instead they should feel their rhythmic connection to the American Indian and represent that in verse.

Austin and Bynner have a Romantic’s regard for poetry as a mean of expressing intense sensations “Felt in the blood, and felt along the heart,” as Wordsworth declared in “Tintern Abbey.” Walton had less faith that American Indian rhythms could be made legible to white poets and readers, and in her preface to Dawn Boy, she does not invoke firsthand, physical experience of the songs. Rather, her sense of Indian poetry comes from Matthews’s texts. If her personal memories of witnessing Indian music and dance did play a role in her thinking, she chooses to keep them in the deep background of her work. Instead of appealing to the intimate experience of rhythm, Walton develops a critique of translation: she contends that many elements of translated song militate against the song’s poetry. As poems in print, she argues, most translations are insufficient precisely because they fail to hew to the rhythms of the original, instead using “tiresome refrains, repetitive and parallelistic lines.”[24]

Walton has her source texts, the Washington Matthews translations, primarily in mind. In his ethnographies of Navajo ceremony, Matthews generally offers several translations of each song: a literal, lineated rendering, a prose explanation or elaboration, and sometimes a free poetic translation. Providing text and rich context, Matthews nonetheless stops short of attempting to represent or fully describe the sounds of the music; ethnomusicologist David McAllester has pointed out that Matthews “saw the futility of relying on words to describe the subtleties of pitch, timbre, rhythm, and the way melody moves.”[25] Where Matthews as ethnographer seeks to capture as much as possible, Walton as poet prefers to essentialize. Be that as it may, she does not believe she has deviated from her sources. Rather, she suggests that her knowledge of poetry in English and its readership has allowed her to detect and recreate the “Indian poetic conception” in a way superior to Matthews’s translations.[26] After enumerating her changes, she writes, “And yet, despite these various alterations, I have, I am sure, been closely true to the essence, the heart and spirit of the Indian poetic conception. I have presented these poems simply and directly, without artificiality of diction, letting the beauty of the idea or symbol stand clean-cut. This method harmonizes best with the Indian texts.”[27] There is an element of apologia in Walton’s “I am sure,” as if she were not unconscious of the audacity of her claim. She assumes not only that such a phenomenon as “Indian poetic conception” exists, but also that she has touched it, embodied it — been “closely true” to it. What Walton purports to have discovered is that this “Indian poetic conception” is proto-imagist, even if, as we shall see, her re-creations do not always come across as either. But her intent is clear. Shorn of those elements that make it primitive and jar the sensibilities of readers, she suggests, a Navajo song becomes “a contribution to English verse” in the 1920s, a modern poem. 

Walton’s re-creations

Walton’s “Indian” is a modernist to Austin and Bynner’s Romantic; yet all were trying to claim American Indian song for US poetry. Despite their differences over whether to represent the rhythm and repetition of their sources, Austin and Walton were similarly self-conscious about their own language and cultural deficits as would-be translators of American Indian song. In their essays and prefaces, they seek scholarly imprimatur for the translations they offer. Austin’s American Rhythm includes an anthology of songs from the Pueblo, Paiute, Shoshone, and others, and ends with explanatory footnotes about sources and methods. Austin admits the difficulty of translating the “poetic values” of the originals and writes that only with greater knowledge of the language might she be able to improve upon her reëxpressions.[28] While Walton believes that a Navajo song, properly re-created, exhibits the standalone qualities of an Amy Lowell or Ezra Pound imagist lyric, her re-creations are published as evidence of an “Indian poetic conception” that requires context and explication because, as she admits, the Indian-ness differs between tribes, in this case Blackfoot and Navajo.[29] Moreover, Walton presents her book of poetry as an extension of (and, at the level of lyric, an improvement on) Matthews’s ethnographic writings. The modernism of Walton’s Navajo is as much a problematic scholarly project as it is a poetic experiment.

There is, however, another side of Walton’s claims about the poetic deficiencies of Matthews’s translations of Navajo songs. If, as Walton argues, Matthews’s texts would not appear poetic to most readers of verse, that may be because the songs were not poems in their original language. Robert Dale Parker laments the fact that scholars and poets who treated American Indian legend and song as poetry unintentionally displaced actual Indian poetry.[30] Implicit in his critique is a distinction between performance and artifact; a song is the former, he implies, and a poem the latter. Parker’s point is that in the quest for oral, “primitive” poetries, modern writers reduced song to text and remade ancient music at the expense of active American Indian poets. To an extent, Walton is guilty of this. I want to suggest, even so, that like many poets, Walton took a broad view of poetry, one that encompassed song. Moreover, having accepted the authority of Matthews’s translations of the Navajo, Walton implies that the “poetic conception” is revealed by her mentor’s full record, which includes not only the translated lyric itself but also the paratextual materials that elucidate its contexts and purposes. If Walton’s term “re-creation” seems too bold an assertion of her mastery of the poetics and intent of Navajo song (she admits that her Blackfoot pieces are “largely my own poetizing”), it accurately represents her artistic license as well as the limits she openly admits. It suggests that she refines Matthews’s translations in order to bring forward the “poetic conception” that was always there. Her re-creations are not, of course, Navajo. They are attempts to make something new — variations on Navajo song informed by her interest in modernist poetics, twentieth-century social issues, and Navajo culture.[31]

I will turn now to several of Walton’s Navajo re-creations, their sources in Matthews, and audio recordings of Navajo performances from Walton’s period. I have not been able to listen to the Matthews recordings, so in order to bring the vocables into this discussion, I draw on recordings made in 1933 and 1940 by ethnomusicologist Laura Boulton, recently edited by David McAllester and Charlotte Frisbie and commercially available through Smithsonian Folkways.[32] In her search for the “conception,” Walton commonly deletes many song elements in favor of poetic compression; at the same time, she adds elements to a song in order to suggest its meaning. Comparison of these materials makes it clear that just as the rhythms of performance proved as difficult to represent as they were inspiring to experience, Navajo vocables have tested the poet translator, who is challenged to convey their power meaningfully in verse. Walton’s decisions and their limitations gesture toward a reconception of poetic translation rooted in the physical production and reception of sound. While some of Walton’s peers experimented with representing vocables in poetry, it would not be until Rothenberg’s work much later that Navajo song would be recast in terms of a poetics of sound.

Walton distributes her Navajo “re-creations” into three categories: Navajo Indian legends, which she has versified in the manner criticized by Parker; and gambling and ceremonial songs, which are mostly variations on texts from Matthews. One of the gambling songs, “Wild Cat Song,” offers insight into how Walton revised her source texts. In an 1889 issue of The American Anthropologist that includes several annotated translations, Washington Matthews describes how Navajo gambling songs accompanied a winter game, played by two teams, each with a row of four buried moccasins.[33] The first team hides a stone in one of its moccasins; the second tries to guess its whereabouts. The game is said to replicate a competition among diurnal and nocturnal animals in ancient times in which animals would gamble in order to do away with the divided day and lead to either eternal light or nighttime. Since the contest ended in a tie, legend suggests, the divided day remained. Matthews believed that the songs not only recalled the characters of the various animals in the original contest but also helped participants strategize. He records three wildcat songs, each in a phonetic transcription and an English translation, with glossary and explanatory notes. Walton’s poem combines two of the translations (numbers 17 and 18 in the American Anthropologist essay). For ease of comparison, I’ll provide the Matthews’s translations and Walton’s re-creation side by side:

Matthews’s Wild-Cat Song no. 17                                      Walton’s Wild Cat Song

Wild-cat’s foot is sore,                                                                     Look at Wild Cat
Wild-cat’s foot is sore.                                                                                   One night
                                                                                                              Coyote striped his thighs
Matthews’s Wild-Cat Song no. 18                                      With white.
He looks like a wild-cat,                                                                  Are his feet sore?
He looks like a wild-cat;                                                                  Are his feet sore?
The insides of his thighs are striped.[34]                                    Did he shut them in the door
                                                                                                              Of his Hogan?
                                                                                                              Why does he walk as if the ground
                                                                                                              Were wet?[35]

Walton elaborates the story in order to communicate some of what we would learn from Matthews’s introduction and notes. In the introduction, Matthews points out that Coyote was “cunning” and played for both sides in the diurnal versus nocturnal animals’ contest.[36] Matthews’s notes to his translations explain that these particular songs tease Wildcat for his unusual appearance and his “cautious, delicate tread.”[37] Walton integrates Coyote into the Wild-Cat song, offering a possible source of Wild-Cat’s “stripes” and casting him as a comical, pitiful figure. Elaboration of this kind is one of Walton’s methods. It indicates that for Walton, I suggest, locating the “poetic conception” of a song sometimes requires mining its contexts, which are in these cases footnotes and prefatory remarks. Walton poetizes informational discourse, such that her re-creation interprets the Navajo song for the reader, presumably to communicate meanings implicit in the original. Per her preface, her decision to add rhyme must be intended to appeal to English-speaking readers.

Walton’s elaborations are less typical than her efforts to pare down her source translations. In these cases, we are left with the volume’s more imagistic poems, whether or not they are accurately Navajo. Compare, for example, Matthews’s verse translation of the legend “Maid-Who-Becomes-a-Bear” and Walton’s version. Along with his extensive commentary, Matthews’s verse translation is intended to illustrate the role of “climax” and irony in Navajo song and to correct a misperception, common at the time, that Navajo song had no form. His examples of such devices as metaphor and end-rhyme build a case for Navajo poetic intentionality and complexity.[38] Aside from the poetic devices the poem supposedly illustrates, faith is the central concern of the Matthews version of “Maid.” This is clear from his footnotes and anecdotes about the character.[39] The maid’s faith helps her find the gods. The speaker of the poem, who seeks the maid’s intervention with the gods, must presumably follow her example. Hence the irony of the refrain “Somebody doubts it, so I have heard”: the speaker and perhaps his kin must suppress their doubts if they wish to find the gods.

Walton’s “re-creation” differs markedly in perspective, form, and content. Taken by the image of the maid on the mountain summits, Walton writes of a goddess-figure who “walks on the summits” and for whom great distances seem small and darkness is penetrable.[40] There is no “I” in her lyric, and there is no ironic reflection on the human tendency to doubt. Rather, the voice is hortatory, and if the purpose of the Navajo original (as represented by Matthews) is to acknowledge the intercessory powers of the Maid, Walton’s is rather to focus on the powerful female character herself. Consistent with her intent to appeal to readers of English verse, Walton deletes the simple refrain of the Matthews translation, “Sought the gods and found them,” and chooses instead a more stylized syntax, inverting adverbs and verbs, predicate and subject. While her sharp focus on an intense image might be described as proto-imagist, the syntax and diction of the poem seem designedly archaic. The result is a short, mystical lyric that pays homage to female power — a New Woman interpretation of the Matthews’s verse translation.

In Walton’s re-creation of “Maid-Who-Becomes-A-Bear,” a modern poet believes she has found in the Navajo a kindred spirit as well as a like-minded artist. One of the best examples of the latter in the volume is Walton’s re-creation of “Magpie Song.” Again for ease of comparison, I provide Matthews and Walton: 

Matthews’s Magpie Song no. 5                                               Walton’s Magpie Song

The magpie! The magpie! Here underneath                                   In the white of his wings
In the white of his wings are the footsteps of morning.               Are the footsteps of morning![42]
It dawns! It dawns![41]

In a note, Matthews describes the coloring of a magpie’s wing. In imagist fashion, Walton homes in on the white of the wing and light of sunrise and eliminates the song’s exhortations and repetition. But both ethnographer and poet choose not to represent the Navajo language vocables that, according to Matthews, are imitations of the magpie’s call and include the onomatopoetic name of the magpie.[43] This is Matthews’s practice throughout “Navajo Gambling Songs,” and Walton follows it. In contrast, in his important study of the Navajo Nightway, Night Chant, Matthews includes phonetic renderings of vocables in the translations as well as transcriptions, though Walton’s re-creations of those pieces do not. Matthews and Walton considered these non-lexical sounds to be meaningless, and on the page, neither seemed sure that the vocables could be represented poetically.

Nonetheless, insofar as a “poetic conception” inheres in the “Gambling Songs,” the vocables must surely be vital to it. That is the conclusion drawn by ethnomusicologist Charlotte Frisbie in her study of the subject.[44] Acknowledging that a long history of deleting vocables from the record limits her research, Frisbie argues that vocables served as sound symbols (i.e., sounds unique to deities; animal sounds; onomatopoetic sounds) and as poetic devices (e.g., in songs, sound combinations used to extend lines or connect meaningful portions of text). While their origins are obscure, some vocables may be archaic words, the meanings of which are no longer known, or may derive from foreign words that entered Navajo. Frisbie further identifies functions specific to vocables in ceremonial songs, where they provide links between verses and chorus, emphasize contrasts between words, and serve other purposes that are identifiably melodic and rhythmic. As Matthews also realized, these supposedly meaningless syllables do not occur randomly; on the contrary, their pronunciation and position in songs are often fixed by poetic form, as Frisbie notes: “While it is still impossible for me to predict which vocables will be used, when one studies Navajo texts in sung and spoken form, it is obvious that such choices are at least in part determined by poetic rhyme and meter.”[45]

Even without the benefit of knowing Navajo, it is possible to hear the poetic role of vocables.[46] In a recording made by Laura Boulton at the Chicago Century of Progress Exposition in 1933, a group of Navajo singers performs “Wild Cat Song.”[47] In this rendering, the rhythm and playful mood of the Wild Cat Song depends on repeated voicing of the vocables. Matthews transcribes these as “yooini yaani” and “yooini ya,” and they take up most of the Wellito performance. They drive the song’s insistent forward motion, and extensive melismas on their open vowels help develop the song’s meter. Repetition is surely essential to the “poetic conception” of this song in performance — purposeful repetition of sounds without denotative meaning.

While we might expect Walton to delete the vocables in her re-creations, Mary Austin does so as well, despite her argument for the centrality of rhythm. Only two of her re-expressions incorporate these sounds: “Young Men’s Song,” which is not attributed to a particular tribe, and “Commemorative Song and Dance of the Paiutes for a Victory Over the Mojave.” Of the two, “Young Men’s Song” is the most experimental. The few English words suggest that this is a chant before battle. It consists primarily of variations on the non-lexical words “ah,” “ahou,” “aou,” “ai,” and “hi.” Austin uses several techniques to suggest a rhythm for performance. All appear in the first two stanzas:

Ah — ahou!  Ahou — aou!
Hi!  Hi!   Hi!  Hi-ah-ee — ah!

Go we now,
       Go we now,
              Go … o … o now!
Ahou!  Ahou!  Ahou!  Ahou![48]

Varied line and sentence length pace our reading, as does the placement of the lines. Dashes and ellipses indicate differences in time value for each word or syllable; exclamations suggest emphasis and perhaps volume. Because Austin has been so intentional about representing rhythm, it’s especially surprising that she has not annotated this re-expression to fill out a scene of performance. We are told that the first line is a call to the war god, and a performance direction for a later line says, “Prolong by clapping hand on mouth.” Moreover, it’s impossible to determine whether the vocables here are drawn from an ethnographer or linguist’s transcriptions, or whether they are Austin’s free interpretation of sounds she heard.

While Austin’s representation of vocables and rhythm here offers information about sound that Walton’s re-creations lack, the poem draws our attention to its limits as printed verse and, consequently, underrepresents whatever “poetic conception” Austin may have noticed in her source material. As in sounding any printed poem, a reader must fill in the blanks, must determine tempo and rhythm through studying meter, line, and punctuation, all in relation to diction, mood, and tone. Variation in performance is to be expected, even when the poem in question follows conventions generally shared and understood by readers of poetry in English. The variations stand to be more extreme when a poet attempts to follow other conventions — and this in itself is a justification for the poet’s decision to act as ethnographer and educator and contextualize her re-expressions, re-creations, or otherwise denominated liberal translations. Austin generally does so, but the absence of references in “Young Men’s Song” makes its imitation of vocables troubling.

Austin’s faith in the prospects of representing American Indian rhythm is deeper than Walton’s, but they share a concern for remaining true to their sources — at least, true to what they perceive to be the poetry and spirit of their sources. That distinction is critical for understanding their projects, for whether “reëxpression” or “re-creation,” their poetry purports to be new and hew close to the original, to benefit from the poet’s close study of American Indian songs and her freedom from the constraints accepted by pure ethnography or literal translation. As we’ve seen, when Walton and Matthews eliminate the vocables, or when Austin haphazardly imitates them, they fundamentally alter the poetry of the Navajo original. And yet in doing so, their intent is, paradoxically, to translate the poetry meaningfully within the limits of English forms and printed text. In the aggregate of their successes and failures, Matthews and Walton, Bynner, and Austin point forward to a performance-based poetics that could compensate for some of those limits.

Rothenberg’s total translation

In an essay on translations of Zuni performances, Dennis Tedlock argued that the translated texts suffered not only from the translators’ misunderstandings of linguistic elements and generic differences, but also from their inability to account for the paralinguistic aspects of performed oral arts.[49] Jerome Rothenberg shared these sentiments, and “The 17 Horse Songs of Frank Mitchell” is his most innovative experiment in poetic translation to reckon with the problem of the paralinguistic aspect of performance.[50] Rothenberg’s Horse Songs are performance-translations based on audio recordings of the Navajo Blessingway singer Frank Mitchell, as well as texts transcribed and translated by ethnomusicologist David P. McAllester, who collected the recordings.[51] The Blessingway is one of the principal Navajo ceremonials, generally given precedence over the healing chantways. A Navajo myth of origin, its purpose is to invoke blessings and good luck on such occasions as the establishment of a hogan or the arrival of a girl’s first menstruation.[52] In an interview with McAllester, Frank Mitchell and his sister, Augusta, explained that while the Horse Songs were not part of the ceremony proper, they would be performed during a Blessingway upon request: for example, by a participant whose horses or other livestock had been suffering in some way, or by a medicine man seeking the revitalizing effects of song.[53] Mitchell’s versions of the songs tell the story of how Enemy Slayer, son of Changing Woman and the Sun, is summoned to his father’s house to fetch horses and bring them back to Changing Woman’s house on earth.[54] The sequence is cumulative, each verse changing slightly amid a series of burdens and vocables.[55]

Although Rothenberg consistently presents his work as true to the poetry of its source, translation-as-performance emphasizes the contingency of his acts of engagement with Navajo material. In his “total translations” of the Mitchell recordings, Rothenberg sets out to account for all the sound in the Navajo songs and begins to conceive of translation as a performance. A “total translation” is possible only in theory, a fact that Rothenberg clearly recognizes: “The English translation should match the character of the Indian original,” he writes; “take that as a goal & don’t worry about how literal you’re otherwise being.”[56] “Match the character” here is comparable to Walton’s finding the “Indian poetic conception” and Austin’s attempting to “re-express” the original, yet unlike his predecessors, Rothenberg’s move away from the “literal” returns us to sonorities that cannot be rendered by the letter. In this way, “total translation” is meant to point us to the necessity of full engagement with song, and to the ethical imperative of honoring even its intransigent elements. Rothenberg claims that full engagement with sound drove his efforts:

It was the possibility of working with all that sound, finding my own way into it in English, that attracted me now — that & a quality in Frank Mitchell’s voice I found irresistible. It was, I think, that the music was so clearly within the range of the language: it was song & it was poetry, & it seemed possible at least that the song issued from the poetry, was an extension of it or rose inevitably from the juncture of words & other vocal sounds.[57]

The general project engages him — “the possibility of working with all that sound” — but it is fundamentally the specifics of a voice that he finds “irresistible.” In performance, the song is embodied, and Rothenberg speaks of it in terms of affection — attraction, resistance; what I earlier called the language of sensibility. What is new here, relative to the other poets we’ve considered, is his recognition of the performer’s voice as part of the affective experience. His translation will need to account for what he hears in Frank Mitchell’s voice, specifically. The song and the performance are one, and the performed song’s relation to poetry in this passage is critical. Rothenberg seems to search for the right metaphor, but it becomes clear that for him song is poetry’s offspring: issue, “extension,” or child of the joining of “words and other vocal sounds.” The metaphor arises again when Rothenberg speaks of translation as “a means of delivery & of bringing to life.”[58] As in Walton’s “Indian poetic conception,” song is born of poetry. For Rothenberg, truly hearing the poetry necessitates listening studiously to Mitchell’s performance, and as “total translations,” his “Horse Songs” are the issue. They are extensions of his engagement with Mitchell’s recorded voice, but they are also distinct works.

I have referred to Rothenberg’s concern for the ethics between translator, performer, and performed text. Because he deems it essential to honor all the sounds of the Navajo song, language is pushed toward pure vocal sound, and the poem as artifact depends on the audio recording. In his case, an ethical stance — a sense of responsibility for the “total” experience of the performed text — fosters artistic innovation. Likewise, Rothenberg writes of the importance of entering into the experience of the performer without merely playing Indian: “I translate … as a way of reporting what I’ve sensed or seen of an other’s situation: true as far as possible to ‘my’ image of the life & thought of the source.”[59] Rothenberg’s essay was published in 1969, and in keeping with his poststructuralist moment, he assumes that identity is fluid and perception contingent. “Translate” in his discourse is closer to Derrida than to Benjamin,[60] and in the negative capability he exercises with Frank Mitchell’s performances, he never presumes to have entered “an other’s situation” absolutely. Be that as it may, the purpose of translation is, in his estimation, the dissolution of cultural boundaries.

Some of Rothenberg’s contemporaries were not persuaded. In a review of American Indian verse translations, William Bevis condemned Shaking the Pumpkin and the theory behind it.[61] Shaking the Pumpkin: Traditional Poetry of the Indian North Americans was Rothenberg’s 1972 anthology of American Indian works, translated by Rothenberg and others. In his review, Bevis maintains that it’s possible to delight in Rothenberg’s poems and to object in the strongest terms to calling them “translations.” Bevis argues that Rothenberg ignored critical differences of genre and distorted imagery and language.[62] In the final analysis, a literal translation with extensive notes is much preferred.[63] Similarly, in “Faking the Pumpkin,” William Clements objects to Rothenberg’s presentation of the materials as translations, except in cases where Rothenberg worked alongside someone who spoke the language in question.[64] Rothenberg’s contemporary reviewers rarely mentioned the audio performances, although Clements would later write favorably of them.[65] Moreover, short of accepting Rothenberg’s poststructuralist orientation, we can see that Rothenberg followed a procedure that respected the demands of working outside one’s language and culture. In “On the Translation of Native American Song,” Arnold Krupat maintained that the “conditions for any approximation to translational accuracy” would be “philological control,” “ethnographic control,” and literary competence.[66] Through collaboration and emersion, Rothenberg attempted to gain the first two, having obvious command of the last.[67]

Rothenberg later professes a faith in empathy — a faith that underwrites his total translations. In his 1972 “Pre-Face” to Shaking the Pumpkin, he claims:

By its very nature, translation asserts or at least implies a concept of psychic & biological unity, weird as such an assertion may seem in a time of growing dis-integration. Each poem, being made present & translated, flies in the face of divisive ideology. The question for the translator is not whether but how far we can translate one another. Like the poet who is his brother, he attempts to restore what has been torn apart. Any arrogance on his part would not only lead to paternalism or ‘colonialism’ (LeRoi Jones’s term for it from a few years back), it would deny the very order of translation. Only if he allows himself to be directed by the other will a common way emerge, true to both positions.[68]

Following this logic, we could imagine that Mitchell’s performance “directs” Rothenberg, that an experience Rothenberg described as the most “physical” of his many translation efforts is at once an intimate and a political act.[69] It is intimate insofar as the translator would embody his source in an attempt to know the other’s experience as fully as possible. By that same token it is political, and the goal of the knowledge is restorative.

Rothenberg here anticipates Eve Kosofsky Sedgwick’s concept of reparative reading. Speculating at the turn of the millennium about the future of queer studies, Sedgwick argues that in its identification with “paranoid” practices — in the prestige that has become attached to self-consciousness and suspicion in the act of reading critically — theory has failed to acknowledge, let alone develop, critical modes of affective engagement. As she puts it, the “desire of a reparative impulse … is additive and accretive.”[70] Borrowing her terms, I would argue that Rothenberg attempts to strike a balance between “paranoid hermeneutics” and the “reparative impulse” — between, that is to say, his suspicion that any act of studying, embodying, and translating an other is necessarily colonialist, and his politically motivated interest in using poetry to repair differences between historically divided peoples.[71] Invoking LeRoi Jones (later Amiri Baraka), Rothenberg describes a self-conscious repositioning of the poet-translator as student to the American Indian mentor. This metaphor might seem disingenuous, insofar as we’ve seen similar dynamics invoked in Matthews, Austin, and Walton, all of whom regarded the American Indian as primitive. In his early work, Rothenberg attempted to recuperate the word “primitive” with the assertion that “primitive is complex.” Given the high value that twentieth-century avant-gardes have placed on complexity, that gesture risks clothing the American Indian in modernist habiliments. But I reiterate that the “Horse Songs” differ from other efforts in Rothenberg’s emphasis on the individual relationship and the contingent performance. The “Horse Songs” are reparative precisely because they are the result of engagements at once affective and intellectual and because Rothenberg’s grand gestures are self-consciously ephemeral. They are the manifestation of the hopefulness and irony implied by the “total” in “total translation.”

Rothenberg’s treatment of Navajo vocables distills his difference from the other poets I have discussed and provides a concrete example of a reparative hermeneutics. Recognizing that vocables were integral to the soundscape of Navajo song, Rothenberg was dissatisfied even with the few English translations that included them because the vocables, left in the original language, seemed merely disruptive or extraneous to the translated poem. In a letter to David McAllester from July 23,1968, Rothenberg writes that listening to Mitchell on tape and singing along with him using McAllester’s translations led him back to the “question of how to make the English sound correspond to the sound in Navaho”; it was this collaboration with McAllester and the recorded Mitchell that yielded his solution.[72] In McAllester’s translation of the tenth Horse Song, Navajo vocables follow English words and always appear at the end of the poetic line. Here are the burden and first verse:

      To the Woman, my son, ‘ehye-la,
      To the Woman, my son, ‘ehye-la, nana, yeye ‘e,

      It is I, I who am that one, ‘ohye-la,”
      To the Woman, my son, ‘ehye-la, nana, yeye, e[73]

It’s a solution that diminishes the sound poetry of Mitchell’s performance, even as it accurately acknowledges the untranslatable syllables and closely follows the syntax of the original. Rothenberg replaces the vocables with English words that resonate sonically with the rest of the translated line, irrespective of meaning but following the example set by the Navajo vocables. For example, in the tenth Horse Song, the words “one,” “none,” and “gone” become part of the repeated first line “Go to her my son”: “Go to her my son & one & go to her my son & one & one & none & gone.”[74] Rothenberg laments the fact that his “vocables” are actual words, but I suggest that their meaning dissolves in the sound stream of the performance. The transcription happens to reinforce the fact that the semantics of the poetic line are secondary to the sound; the line makes little sense without being sounded. The nasal n in “son,” for instance, blends into “and one,” which themselves are enunciated as a single, two-syllable sound, roughly “anwan,” such that word and vocable might be transcribed “sonanwaan.” The emphatic gutturals in “go” and “gone” mark the rhythm. The performance, that is to say, reminds us that spoken and sung language is not segmented phonologically, at least not in the manner represented on the page. The sounds do not merely supplement the lyric’s meaning. In the “Horse Song,” we experience language as performed sound, not just carrier of poetic content.

Furthermore, Rothenberg’s English “vocables” imitate the sonic patterns of the original Navajo as Mitchell performed it. As Rothenberg points out in his letter to McAllester, Mitchell’s performance of the tenth Horse Song shows the sonic connection between words and vocable, and as we saw earlier, Charlotte Frisbie’s study of Navajo song identified and elaborated upon that connection. The vocables flow from and into the words, acting as connectors; they extend lines rhythmically, and they sometimes appear within words. The final voiced vowel of ‘Esdza shiye — the Navajo that McAllester translates “The Woman, my son” — melds with the initial vowel of the vocable ‘e hye-la. In fact, throughout the performance, assonance is connective; in this verse, the long “a” sound dominates. Later, the vocables will shift slightly along with the prevalent vowel sound of the words, and they will be added to or inserted between words, as Frisbie’s research indicates is common in Navajo song. For instance, in the first verse, ‘e hye-la becomes ‘o hye-la, and Mitchell interweaves vocables with the words that McAllester translates as “Boy-brought-up-within-the-Dawn, / It is I, I who am that one.” Mitchell sings in a steady, duple meter, and with few exceptions the note-range is within five whole steps of the tonic. Rothenberg imitates Mitchell’s voicing of the gutturals and nasals in the Navajo and his emphatic “unh” at the end of the song’s burdens. These punctuate Mitchell’s performance and are essential to its rhythm, and they become so in Rothenberg’s translation, making all the more critical the experience of his translation as sound poem. Rothenberg sings his lines, and the English vocables blend into each other and the surrounding words.

Rothenberg’s sound poems demonstrate his argument that Navajo song is poetry, when “poetry” is expressed in sound and is not always, or even fundamentally, oriented toward discrete meanings. The performed poem — we might call it the sound text — urges response but resists conventional textual explication of the kind that Walton’s imagist renderings, for instance, invite. But like the work of his predecessors, Rothenberg’s performances are embedded in a larger discourse that is part of their meanings and effects. We do not listen only to the Horse Song itself when we engage with his total translation. On the contrary, the performances and “Pre-Faces” should be understood as a unified project. The performances are situated in Rothenberg’s research and explanation of his process. All of this discourse, from essay to audio-recorded performance, is an act of reparative reading — in this case, a reimagining of Navajo song that attempts to restore, poetically, what we had traditionally misread or misheard. McAllester makes this point beautifully in a July 30, 1968, letter to Rothenberg. He has just read and listened to Rothenberg’s first translation of the tenth Horse Song: 

When I first saw the poem I had something of a Life Shock, like seeing a baby born or something. That’s a good sign. You produced something strong. At first I felt it was a very long way from the Navaho, but as it sank in, slowly, it felt closer and closer in basic intent. Many thanks for the tape. I thought the chanted version the better by far. That took it out of the usual context of our spoken modalities so that extraneous ideas almost inevitable in spoken communication were simply dropped.[75]

Again poetic translation is described in the language of conception and birth, and McAllester acknowledges having moved from a shock at a misreading to a sense of gratitude for a form that approximates Mitchell’s “basic intent.” Note McAllester’s “closer and closer”: like Rothenberg, he does not imagine that the translation will ever hit the mark precisely. That, in fact, may not be either Rothenberg’s or McAllester’s “basic intent” in trying to carry song over into poetry, or one language over into another.

By acknowledging and exploiting the contingent nature of performed art, Rothenberg creates a poetic response to American Indian song that is radically different from the earlier generation we’ve considered. As Michael Castro shows in his study of poetic representations of American Indian culture, Rothenberg drew from Stein, Williams, and Olson and partnered with Jackson Mac Low, David Antin, and others who sought to reconceive poetry as a language event, as opposed to a textual artifact. Castro writes that Rothenberg was “interested in translating events as rich sources of usable forms that can extend our culture’s sense of the poem into new, participatory dimensions.”[76] Insofar as the Horse Songs were ceremonial, it follows that Rothenberg’s “total translation” would include performance. Sound recordings of his live performances of the Horse Songs are traces of poetry-in-translation as “participatory,” including his interaction with the audience and their response.[77]

I agree that Rothenberg is more successful than his predecessors in crafting a distinct work that simultaneously bears the imprint of the Navajo, and his uses of reparative reading depend upon a greater degree of self-consciousness about embodying an other than poets of the 1920s and ’30s seemed capable of. At the same time, I would argue that for all their differences, his “The 17 Horse Songs of Frank Mitchell” are a logical outcome of the less dramatic experiments in Walton’s imagist “re-creations” and Austin’s “reëxpressions.” Like Walton’s work in particular, they are the result of sustained study of Navajo song, and they purport to find in it a sensibility and a poetic that are radically contemporary. Rothenberg’s differences should not distract us from these underlying continuities. All these poets, within the horizons of understanding that constituted their historical moment, imagined US poetry to be inclusive of American Indian song and understood their own responses — re-creation, re-expression, total translation — to be of a second order in relation to the Indian poetry itself. The crucial point was to engage American Indian song as source and influence and to recognize the singer as equal in the construction of American poetic voice.


1. Whether to use “American Indian,” “Native American,” or another broad term has been debated since the 1960s. A helpful summary of that debate is provided by the North Carolina Humanities Council.

2. Eda Lou Walton, Dawn Boy: Blackfoot and Navajo Songs (New York: E. P. Dutton and Co., 1926; Whitefish, MT: Kessinger Publishing, 2010), i. Citations are to the Kessinger edition.

3. Jerome Rothenberg, Pre-Faces and Other Writings (New York: New Directions, 1981), 78.

4. Describing vocables as nonlexical is admittedly problematic. A discussion of the problems follows. For the sake of consistency with my sources, I will use the term “Navajo” (rather than Diné) to refer to the language and people.

5. Eve Kosofsky Sedgwick, Touching Feeling: Affect, Pedagogy, Performativity (Durham, NC: Duke University Press, 2003), 123–51.

6. Arnold Krupat, “On the Translation of Native American Song and Story: A Theorized History,” in On the Translation of Native American Literatures, ed. Brian Swann (Washington, DC: Smithsonian Institution Press, 1992), 4.

7. Leah Dilworth, Imagining Indians in the Southwest: Persistent Voices of a Primitive Past (Washington, DC: Smithsonian Institution, 1996), 192.

8. Krupat, “On the Translation of Native American Song and Story,” 6–7.

9. Dilworth, Imagining Indians in the Southwest, 187.

10. Steven G. Kellman, Redemption: The Life of Henry Roth (New York: W. W. Norton, 2005). See especially chapters 3 and 4.

11. Eda Lou Walton, “Navajo Song Patterning,” The Journal of American Folklore 43, no. 167 (January–March 1930): 105–18.

12. Songs from Matthews’s translations figure in two twentieth-century novels: Oliver La Farge’s Laughing Boy and N. Scott Momaday’s House Made of Dawn. La Farge’s protagonist sings prayers from the Night Chant. The songs are woven into the texture of Momaday’s novel, sometimes as part of a character’s consciousness.

13. David P. McAllester and Douglas F. Mitchell, “Navajo Music,” in Handbook of North American Indians, vol. 10 of Southwest, ed. Simon Ortiz (Washington, DC: Smithsonian Institution, 1983), 605–6.

14. Robert Dale Parker, The Invention of Native American Literature (Ithaca, NY: Cornell University Press, 2003), 85.

15. Sam D. Gill, Sacred Words: A Study of Navajo Religion and Prayer (Westwood, CT: Greenwood Press, 1981), 181.

16. Ibid., 183.

17. Walton, “Navajo Song Patterning,” 105.

18. Ibid., 118.

19. Philip J. Deloria, Playing Indian (New Haven, CN: Yale University Press, 1998), 120.

20. On the shift from Romantic era “noble savage” to modernist perceptions of American Indian culture, see Deloria, Playing Indian, chapters 2 through 4.

21. Walton, Dawn Boy, xxi–xxii.

22. Mary Austin, The American Rhythm: Studies and Reëxpressions of Amerindian Songs (New York: Houghton Mifflin, 1930; New York: Cooper Square, 1970), 5. Citations are to the Cooper Square edition.

23. Ibid., 9–14.

24. Walton, Dawn Boy, vii–viii.

25. David McAllester, “Washington Matthews and Navajo Music,” in Washington Matthews: Studies of Navajo Culture, 1880–1894, ed. Katherine Spencer Halpern and Susan Brown McGreevy (Albuquerque, NM: University of New Mexico Press, 1997), 48. McAllester notes that in his Navaho Legends, Matthews gave the task of musical analysis to John C. Fillmore, a music historian who annotated several of the songs. Following Matthews’s lead, Fillmore sets out to demonstrate the complexity of “primitive expression” by seeing it as a precursor to sophisticated forms. He transposes Matthews’s audio recordings of Navajo performances into Western tonal keys and scales, and this helps him corroborate a general claim about folk melody, that “harmonic perception is the formative principle in folk melody.” See Washington Matthews, Navaho Legends (Boston: Houghton Mifflin, 1897), 257. Rhythm is undoubtedly the origin of music, Matthews acknowledges, in what seems to be commonplace at the time. But his survey of world music, now including Navajo song, indicates that harmony naturally and invariably follows.

26. Walton does not seem to have known of Matthews’s poetry. Paul G. Zolbrod argues that Matthews’s knowledge of poetry complemented his ethnographic work, an opinion that McAllester shares. Zolbrod analyzes two of Matthews’s poems — “The Eyes of Judah” and “The Pagan Martyrs” — and compares them to Matthews’s well-known translation of a prayer from the Nightway beginning “House made of the dawn.” It’s unlikely that Walton would have liked Matthews’s verse. In form and religious imagery, “The Eyes of Judah” follows from the work of poets such as Heber, Hemans, and Whittier. But Zolbrod’s analysis of “House made of the dawn” — that in it “ethnography and poetic skill merge fully” — is convincing (107). He demonstrates the ways that Matthews’s knowledge of the English ballad supported his recognition of similar modes of repetition in Navajo song. See Paul G. Zolbrod, “Washington Matthews Writes a Ballad,” in Washington Matthews: Studies of Navajo Culture, 1880–1894, ed. Katherine Spencer Halpern and Susan Brown McGreevy (Albuquerque, NM: University of New Mexico Press, 1997).

27. Walton, Dawn Boy, viii–ix.

28. Austin, The American Rhythm, 174–75.

29. Walton, Dawn Boy, ix.

30. Parker, The Invention of Native American Literature, 4.

31. Remaking translations of American Indian songs as poems was a common practice in the 1920s and ’30s, as Austin’s reëxpressions suggest and Michael Castro’s study amply demonstrates. Along with Austin, we could consider Alice Corbin, coeditor with Harriet Monroe of Poetry magazine in the 1920s. She recasts several American Indian songs in her collection Red Earth (1920), recently reprinted by the Museum of New Mexico Press. In her introduction to that edition, Lois Rudnick points out that Corbin “recognized and attempted to emulate many of the characteristics that ethnomusicologists attribute to Indian ceremonial song” (30). Austin and Walton are notable for their blended presentation of their poetry and research.

32. Navajo Songs: Recorded by Laura Boulton in 1933 and 1940, ed. David P. McAllester and Charlotte Frisbie (Smithsonian Folkways, 1992). Matthews’s wax cylinder recordings are stored at the Library of Congress and the Indiana University Archives of Traditional Music. Matthews’s notebook lists of the cylinder contents are reprinted in Halpern and McGreevy, Washington Matthews.

33. Washington Matthews, “Navajo Gambling Songs,” American Anthropologist 2 (1889): 1–19.

34. Ibid., 17.

35. Walton, Dawn Boy, 49.

36. Matthews, “Navajo Gambling Songs,” 4.

37. Ibid., 17.

38. Matthews, Navaho Legends, 22–29.

39. Matthews discusses the character in relation to the Navajo Mountain Way ceremony, described by Matthews in 1887 and by Leland Wyman in 1975. Matthews gives the Navajo for Maid-Who-Becomes-a-Bear as Tsiké Sas Natlehi, and reports that the name appears to refer to several legendary maidens whose aid is sought by the sick (Navaho Legends, 228).In The Mountainway of the Navajo (Tucson, AZ: University of Arizona Press, 1975), Wyman reports that the legend of the Changing Bear Maiden belongs more intimately to the Evilway ceremony (104–5). Citing work of Father Berard Haile, he writes that Changing Bear Maiden’s influence was associated with insanity (55).

40. Walton, Dawn Boy, 67.

41. Matthews, “Navajo Gambling Songs,” 9.

42. Walton, Dawn Boy, 46.

43. Matthews, “Navajo Gambling Songs,” 9.

44. Charlotte J. Frisbie, “Vocables in Navajo Ceremonial Music,” Ethnomusicology 24, no. 3 (September 1980): 347–92.

45. Ibid., 381.

46. I do not speak Navajo; Walton’s knowledge of the language is unclear. For consistency I will use Matthews’s phonetic renderings of the vocables, supplemented (where applicable) by Frisbie and McAllester’s notes to the Boulton recordings, and McAllester’s notes to the Mitchell recordings.

47. Matthews, “Wild-Cat Song no. 17.” The liner notes identify the performers as Frank and Pablo Huerito. Ethnomusicologist Aaron Fox has researched these recordings extensively and, through contacts with the families, corrected the attribution. The performers are Pablo Wellito and four of his sons (email correspondence with the author).

48. Austin, The American Rhythm, 94.

49. Dennis Tedlock, “On the Translation of Style in Oral Narrative,” Journal of American Folklore 84, no. 331 (January–March 1971): 114–33.

51. For access to recordings of Frank Mitchell and to the David McAllester Papers, the author gratefully acknowledges the World Music Archives and Special Collections and Archives, Wesleyan University, Middletown, CT.

52. Leland Wyman, Blessingway (Tucson, AZ: University of Arizona Press, 1970), 3–8, 37.

53. David P. McAllester, interview with Frank and Augusta Mitchell, July 16–20, 1965, audiorecording, World Music Archives, Wesleyan University, Middletown, CT. In addition to the audio-recorded interview, see Frank Mitchell, Navajo Blessingway Singer: The Autobiography of Frank Mitchell, 1881–1967, ed. Charlotte J. Frisbie and David P. McAllester (Tucson, AZ: University of Arizona Press, 1978). In the autobiography, Mitchell provides an account of how he learned the Blessingway. The autobiography includes his telling of some of the myths in the Blessingway and his description of its rituals. Leland Wyman’s study of Blessingway also includes a Mitchell version of the ceremonial.

54. Wyman’s study of Blessingway includes a different account of the origin of horses. See version 3, pp. 621ff. in Blessingway.

55. McAllester’s extensive notes to the Mitchell recordings include synopses of the songs, transcriptions in Navajo of most, line-by-line translations, and editorial footnotes from comments made by the Mitchells. The translations range in date from 1965 to April 18, 1970.

56. Rothenberg, Pre-Faces, 78.

57. Ibid., 86.

58. Ibid., 91.

59. Ibid., 78.

60. Rothenberg’s performances and theory are inseparable, and the “total translations” operate in a manner that Derrida describes as basic to translation — that is, a kind of symbiosis between the two languages in question and the meta-language of theory and the translating act. In “What Is a ‘Relevant’ Translation?,” moreover, Derrida takes up the word “relevant” as a “vocable,” a group of sounds with divergent meanings across the languages in which he works. In contrast, the importance of a hierarchy of genres and expressiveness in Benjamin seems at odds with Rothenberg’s project. Advocating for literalness in translation, Benjamin insists that the higher the literary quality of a work, the more translatable it will be because of its proximity to the “true language” that is the essence of all discourse. Jacques Derrida, “What Is a ‘Relevant’ Translation?,” trans. Lawrence Venuti, Critical Inquiry 27 (Winter 2001): 174–200. Walter Benjamin, “The Task of the Translator,” in Selected Writing, Vol. 1: 1913–1926, ed. Marcus Bullock and Michael W. Jennings (Cambridge, MA: Harvard University Press, 1996), 253–63.

61. William Bevis, “American Indian Verse Translations,” College English 35, no. 6 (March 1974): 693–703.

62. Ibid., 694–95.

63. Ibid., 701.

64. William M. Clements, “Faking the Pumpkin: On Jerome Rothenberg’s Literary Offenses,” Western American Literature 16, no. 3 (November 1981): 193–204.

65. See William M. Clements, Native American Verbal Art: Texts and Contexts (Tucson, AZ: University of Arizona Press, 1996). However, Clements develops a comparison of anthology and museum that further problematizes Rothenberg’s many collections of world “poetries.”

66. Krupat, “On the Translation of Native American Song and Story,” 23–24.

67. Bevis and Clements also object to Rothenberg’s characterization of all the American Indian oral forms as “poetry.” However, in Reading the Voice: Native American Oral Poetry on the Page (Salt Lake City, UT: University of Utah Press, 1995), Zolbrod conceives of “poetry” on the same broad terms as Rothenberg — “the art form,” as he puts it, “whose primary medium is language, whether written or spoken (or sung) …” (7).

68. Rothenberg, Pre-Faces, 93.

69. Ibid., 90.

70. Sedgwick, Touching Feeling, 149.

71. Ibid., 143.

72. Jerome Rothenberg to David McAllester, July 23, 1968, David McAllester Papers, Special Collections and Archives, Wesleyan University, Middletown, CT.

73. David McAllester, translation of Horse Song no. 10, David McAllester Papers, Special Collections and Archives, Wesleyan University, Middletown, CT.

74. Rothenberg, Pre-Faces, 88.

75. David McAllester to Jerome Rothenberg, July 30, 1968, David McAllester Papers, Special Collections and Archives, Wesleyan University, Middletown, CT.

76. Michael Castro, Interpreting the Indian: Twentieth-Century Poets and the Native American (Norman, OK: University of Oklahoma Press, 1991), 135.

77. Note, for instance, the SUNY Buffalo performance, March 3, 1993, introduced by Dennis Tedlock and archived on PennSound.