Listening in Black and White: Race and Voice in the 1920s

Despite the perceived innovation presented by the blackness of their vocal style, the Boswell Sisters grew up in an era in which the impersonation of blacks by whites was a commonplace of popular entertainment. We may presume that the Boswells' blackness resided in those features of their singing that were substantially different from other white popular singers: a southern accent making “correct” and pure vowel sounds more malleable; using the lower, speaking register of the female voice as the general tessitura; an expressive flexibility of tone enhanced by the closeness of the speaking with the singing voice; and vocal phrasing borrowed from instrumental technique. To understand what made their “blackness” something other than blackvoice, we must be clear how blackvoice, described by Gage Averill as “the stereotyped representations of black dialect, vocal mannerisms, and musical style that was the product of the minstrel stage,” was constituted for the Boswells' audience.¹⁵

Blackface minstrelsy, although in sharp decline by the 1920s, had been the predominant form in American musical theater for nearly a century. Eric Lott comments on its decline provocatively, arguing that although minstrelsy was originally necessary to introduce black and white cultures to each other, by the 1920s that work was largely done, but instead of dying out, it began to “transmogrify,” or metamorphose. He further suggests that blackface “provided a channel for the black cultural ‘contamination’ of the dominant culture,” quoting John Szwed: “The fact that, say, a Mick Jagger can today perform in the same tradition without blackface simply marks the detachment of culture from race and the almost full absorption of a black tradition into white culture.”¹⁶

Minstrelsy is often described as an appropriation of black cultural forms by white entertainers. Nonetheless, minstrelsy's practice (whether enacted by white or black actors) relied upon establishing and retaining racial markers, and in doing so defined blackness, its cultural forms and its stereotypes on its own terms through what W. T. Lhamon calls the blackface lore, a unique repertoire of performative signifiers.¹⁷

Lott gives preeminence to the visual marker of skin color and implies that without black makeup, a white artist's performance may be racially identified with his or her “biological” culture. Pamela Brown Lavitt has argued, however, that at least for white female performers during the first years of the twentieth century, the perceived blackness of a performance resided much more in a physical embodiment—a combination of “discrete physical, comic, aural, linguistic, and sexual cues”—than it did in makeup alone.¹⁸ Moreover, by the end of the 1920s, entertainment was delivered through the aural media as much as or even more than through the visual. Whereas the appearance of traditional minstrelsy was instrumental in defining its lore as representative of blackness, in its transmogrification the sound of minstrelsy provided a more significant arena for the defining, contesting, and crossing of racial boundaries. According to Houston A. Baker Jr., “The sound emanating from the [minstrel] mask reverberates through a white American discursive universe as the sound of the Negro.”¹⁹ Minstrelsy's legacy cannot be just looked for but must also be listened for: in the way the last vestiges of traditional minstrelsy were captured in the cinema of synchronized sound, radio, and phonography and in the way the qualities of these new formats challenged and were challenged by its practice.

In the first three decades of the century, the kaleidoscope of short acts that made up black and white vaudeville, variety, and minstrel shows—comic, sentimental, devotional, spoken, musical, sung, danced—were integrated into the new media and the lore developed in each according to both their possibilities and their limitations. Common to all of these developments was the preservation of the garbled vocality or the “bizarre phonic legacies” of minstrelsy's lore, the “mispronunciations and misidentifications” of a pidgin dialect and a faux regional accent that approximated southern origin.²⁰

Even on its own, the southern drawl, performed with a relaxed larynx and a predominance of lazily articulated consonants and elongated vowel sounds, sometimes functioned as marker for blackness, a fact noted—and resented—by contemporary African American commentators.²¹

In nonmusical situations, blackface comedians clearly had their own coded vocality that was intended to sound or connote “black,” but it is difficult to know just how far earlier blackface singers went in modifying their vocal delivery in order to produce blackness in the voice. The few male recorded singing artists still known for and regularly performing in blackface during the early 1920s—especially Al Jolson and Eddie Cantor—did not sing in the same way blackface comedians spoke.²²

Nevertheless, Stephen Banfield points out that blackface makeup could have a direct, physical effect on voice production, in that “if you are in blackface and your mouth is a white painted ring, the expressive mobility of that ring is crucial.”²³ This need for facial movement inevitably alters diction, providing an opportunity for the expressive or playful manipulation of both vowel diphthongs and consonants, as happens in southern dialects. Moreover, Banfield has made a case for a particular version of female stage vocality, the belt, with its feature of a forced tone at the top of the range, being derived from the practice of “coon-shouters”: specifically, white female vaudeville artists who did not always wear blackface as a matter of course.²⁴ Using the example of Stella Mayhew, recorded in 1910, whose spoken “virtuosic ethnic cameo” (clearly representative of the “southern black mama”) is sandwiched between parlando song verses, he notes her “crystal-clear diction that is still largely Anglicized (with hard 't's) but not entirely so, for her … coarse contralto quality … privileges a southern diphthong mouthing of ‘plan’ and ‘man.”’²⁵

Banfield implies that Mayhew's singing diction was different from that of her speaking voice; this is consistent with observations of the one female blackface star of the twenties who left a body of recorded work to examine. Rosetta Duncan's singing was largely in the context of one character: Topsy, derived from the feisty urchin in Harriet Beecher Stowe's Uncle Tom's Cabin. In speech, Duncan's Topsy combines a gruff, babyish voice (as one might imagine a hoarse Shirley Temple) with the standard minstrel dialect, but in song she uses instead a well-moderated mezzo delivery, like Mayhew highlighting only key parlando phrases with Topsy's voice.²⁶

Her unblackened sister, Vivian, on the other hand, both speaks her lines and squeaks through musical numbers with an insubstantial and forward, facially placed soprano tone, a sound associated with the twenties' flapper. The differences in vocal production between the two sisters are linked to their visual stereotypes: if Rosetta employs (at least in speaking) a blackface vocality, then we can hear Vivian's light nasal tones as indicative of whiteface, and whiteness.

As is shown by the examples of Mayhew and Duncan, a lower tessitura, especially in women's voices, seems to have been used as a signifier of blackness. Linguistically, of course, adjectival phrases such as “lowdown” or “gully-low” were applied to the blues and to the playing or singing that voiced them, but there also seems to have been an abiding fascination for the black female contralto.²⁷

The Duncan Sisters clearly exploited this interest by emphasizing the range difference between Topsy and Eva (the difference disappears when they are singing as themselves). There is also a sense that a white woman singing in a lower range was so unnatural as to invoke a double mask. Rosetta Duncan occasionally cross-dressed Topsy as part of her act. Sophie Tucker's alto delivery was famously regarded as a more convincing cross-gendered performance than that of a celebrated female impersonator.²⁸ Lee Morse, a major musical and vaudeville star throughout the 1920s, was billed as singing “Character Songs, Darky Ditties”; yet an early review in Variety from November 1922 mentions that her act included “the impressions of a male impersonator,” and a review from the following year states: “She accomplishes a low register and a volume which equals in tone a male bass singer, yet her voice has a feminine quality, a richness and sweetness which no male voice could produce.”²⁹ Although she eventually dropped the impressions, once recording she was still billed as “Miss Lee Morse,” presumably to remove any ambiguity about her sex.

The traditional combination of comedy and music in blackface performance allowed some further variations on how blackness might be invoked by musical comedians; principal among these appears to have been the use of scat, a long-standing performance practice commonplace among the vocal quartets featured in black minstrelsy and vaudeville.³⁰

In a 1924 recording of their signature tune “Rememb'ring,” the Duncan Sisters inserted a light, untexted vocal break over the bridge between the first and second choruses, imitating the polyphonic soloing styles of early jazz bands.³¹ The technique was taken to its vaudevillian limits by the comedian Cliff Edwards, or Ukulele Ike, the first white singer to use scat (developed during his time in blackface) “as the cornerstone of his style.”³² Another blackface comedian of the 1920s, Emmett Miller, retained a heavily inflected Georgian accent throughout both the spoken patter and the songs of his comic turns.³³ But he was also a vocal novelty artist, and his freewheeling blues yodeling simultaneously represented qualities of the minstrelsy “coon”: inebriation, rebellion, lack of discipline—and jazz innovation. Both white and a genuine southerner, Miller is unusual among recorded blackface vaudevillians, and although his performances are maverick by the standards of the late 1920s, it is possible that they reflect an earlier or less prestigious (i.e., working-class) performance tradition.

It could be that there were other unrecorded blackface singers who, like Miller, took a more physical approach to black vocality. However, at the other end of the class/cultural spectrum, recorded performances by Jolson, Cantor, and Duncan (and the more upmarket “coon-shouters,” such as Tucker) imply a less comprehensive practice, instead relying heavily on surface stylistic elements or vocal embellishments that were intended to connote blackness. The ideal of the sentimental, emotional Negro extended well into the twentieth century, becoming a defining element in white reception of blues singers and blues singing. Instructions in How to Play and Sing the Blues Like the Phonograph and Stage Artists recommended: “If one can temporarily play the role of the oppressed or the depressed, injecting into his or her rendition a spirit of hopeful prayer, the effect will be more natural and successful…. Without the necessary moan, croon or slur, no blues number is properly sung.”³⁴

Sobbing and sighing, yodeling, moaning and wailing, glissandi, and portamenti all formed part of the blackface artist's portfolio of vocal lore that hypervocalized emotion.

With the release by OKeh Records in 1920 of the first specifically marketed race record, Mamie Smith's “Crazy Blues,” new and wider audiences became aware of and experienced black artists speaking and singing for themselves. Connoisseurship of the “real” African American experience now developed rapidly in the North, finding its apogee in New York's salon culture and resulting in white patrons such as Carl Van Vechten nurturing and endorsing black artists.³⁵

But even lower down the social and financial scale, audiences—no longer just spectators—could listen to black music in the comfort of their own homes via the phonograph or, eventually, the radio. “Authentic” black culture (inasmuch as it was created, if not controlled, by black artists) began to be available to all, but at a distance that did not require the potentially repugnant, or even illegal, mutual presence or physical interaction that might take place at a theater, music hall, or speakeasy. Although Mamie Smith's recordings purportedly were intended initially to exploit the African American market for phonograph records, within months black music was being marketed directly to white audiences. OKeh's 1921 advertisement to distributors for a series of recordings by the Norfolk Jazz Quartette is explicit: “The initial release of ‘JELLY ROLL BLUES’ totalled a sales four times greater than any popular hit in that bulletin…. You may be interested to know that it isn't the colored race which is responsible for this jump in record sales. The big demand comes from the white people.”³⁶ OKeh's 1923 advice to distributors, pointing out the “growing tendency on the part of white people to hear their favorite ‘blues’ sung or played by famous colored ‘blues’ artists,” suggests that authenticity may have influenced the record buyer's preference between white and black artists' recordings of the same song.³⁷ It is generally understood that the white audience for race records, however influential, was relatively small, consisting primarily of disaffected youths in search of an alternative culture, intellectuals, and perhaps most importantly, musicians.³⁸ Nonetheless, OKeh's early promotional material gives the impression that a huge white market for its “real” black products existed from the outset, as the traditional audience for minstrelsy embraced the new media. Even if this was more hype than fact, white consumers may have experienced records made by black artists without actually taking them home from the store. As the race catalogues grew, so musicians' aspirations and audiences' expectations changed; and their perceptions of blackness, and of what black music and voices sounded like, had to incorporate new approaches to rhythm, intonation, voice placement, and vowel modulation.

It may not have been easy to express what constituted the changing criteria for racialized vocality, but it is certain that judgments of a performance's success or failure would have at least partially been based on them. For instance, the Phonograph Monthly Review of August 1930 criticized two sides by the African American singer Ethel Waters, saying she “apes white styles in My Kind of Man and loses much of her own individuality, but You Brought a New Kind of Love, with some magnificently tortured exhortations, is more striking.”³⁹

There are indeed significant stylistic differences between the two sides. Even though it contains sixteen bars of wordless doodling and a significant amount of swooping and ornamentation (and indeed emotive hypervocalization), the “white” performance comes across as mannered: with precise diction; alternation between a babyish, flapper voice and a plummy operatic tone; a wide, pervasive vibrato; a rhythmically rigid declamation, and a very centered attitude to pitch. The “black” side, on the other hand, while hinting at “white” qualities in the verse, gradually metamorphoses over the two choruses: the enunciation becomes much closer to speech; the tone and vowels become much more flexible and expressive; the roughness of Waters's developing vocal damage is much more apparent; thirds and fifths become much more blues-inflected; the rhythm relaxes; the vibrato is used ornamentally rather than as an imperative; and the tune is subjected to a breathy stop-time variant, eventually almost disappearing in improvisational invention. For the reviewer, the more valued, authentic, and black performance was the one characterized by jazz elements, but his preference is expressed as approval for the gestures of natural, emotive singing (“exhortations”) that involve the degradation of musical material (“magnificently tortured”). Moreover, the review is judged solely on vocal rendition, and the opinion regarding her performance's racial propriety formed only on an auditory experience.

Before the phonograph industry, and more specifically its marketing strategy, was inflected by racial categorization, it was enough for white musicians to adopt black forms and for white blackface singers in vaudeville to employ simple gestural representations of blackness rather than to try to sound truly black. Indeed, by 1920 it may have been crucial to some comedians' delivery that they didn't attempt authenticity in at least part of their act, so that the obvious counterfeit became part of the humor. But with the advent of the race record, black musical artists set their own standards of “hot,” “jazzed,” and “black.” As white musicians' abilities to rely on the visual masks of blackface disappeared through the agency of recorded sound, a need to cultivate a more convincing aural black mask arose to retain white artists' popularity in a market hungry for black cultural forms; to ensure that white concerns could retain control over an increasingly lucrative and socially powerful industry; and to continue to tap in to, and even develop, the audience's fascination for cross-racial performance.⁴⁰

If white musicians could learn again to sound and play black, they would be a better commercial proposition for club owners, booking agents, record companies, and retailers. Sophie Tucker, realizing she could no longer rely solely on her material and gesture for authenticity, engaged Ethel Waters for singing lessons to make her sound more black.⁴¹ And although erstwhile “coon-shouter” Marion Harris continued to bill herself as “Queen of the Blues,” her accompaniment was almost exclusively provided by the sweetest of white orchestras, and by the end of the 1920s she had moved to Europe. But younger white singers such as Annette Hanshaw and Bing Crosby, whose recording careers began after the advent of race records, could choose which elements of the white and black popular styles, and which of minstrelsy, and which of jazz they wished to nurture in their own, a process that happened gradually as part of their artistic development. Hanshaw and Crosby both began recording in 1926, already individually distinctive in their tonal and rhythmic approaches but still recognizably emulating their avowed white models (Harris and Jolson, respectively). By the middle of the 1930s, however, they sounded quite different from their previous selves; more mature artists, certainly, but also ones who had thoroughly studied and assimilated the new markers of blackness. Close contemporaries of Hanshaw and Crosby—slightly younger, though a year sooner to record for the first time—the Boswell Sisters entered a music business in which white musicians were busily learning the new language of black music and in which white singers were coming to terms with irreversible changes in, or at least additions to, the popular style.

Not All Black and White: The Boswell Sisters in the 1920s

The Boswell Sisters' sound was so convincingly black to their audience that misunderstandings constantly arose during the first year of their stardom. A letter addressed “To Whom It May Concern” was written soon after they began appearing on national radio: “I should like to know if the Boswell Sisters are white or colored. I am asking you to settle a long argument to the above question. I'm thanking you in advance, Herbert A. Bailey, Roxbury, Mass. October 17, 1931.”⁴²

Vet Boswell also remembered a publicity caricature depicting the sisters as black women that was issued in Paris prior to their 1933 tour, presumably drawn on the basis of the artist hearing their blackness rather than seeing it.⁴³ Such misperceptions were not altogether without foundation; New Orleanian creoles of color formed a significant proportion of the original jazz community.⁴⁴ Furthermore, fair-skinned women far outnumbered dark-skinned in black chorus lines, and—in a neat reversal of white minstrelsy—theatrical “passing” was an established variety turn by fair-skinned black vaudevillians such as the Whitman Sisters and the Washington Sisters.

The Boswells (shown in Figure 1), with their dark eyes and hair, were not appreciably “whiter-looking” than Alice Whitman or Fredi Washington; if visual assessment could not establish their racial origin and aural assessment suggested they must be black, Bailey's perplexity was at least understandable. And in a society where Jim Crow existed everywhere, in both southern and northern states, racial origin mattered. For all the enthusiasm with which black cultural forms were embraced by whites during the Harlem Renaissance, by the 1930s social anxiety over blacks passing as white was at fever pitch; and passing could not have been an issue were the boundaries between black and white not still deeply inscribed. As George-Graves notes, theatrical passing, be it black-to-white with its “hint at miscegenation,” or white-to-black courtesy of burnt cork, was “intriguing and titillating,” but only because it “allowed the forbidden to be worked out” as part of a risk-free masquerade.⁴⁵

But the Boswells, to all intents and purposes, were not masquerading; they spoke in the accents of their upbringing and sang in the musical language of their hometown, perhaps so making their vocal ambiguity initially startling—and worrisome.

Underpinning the anxiety of passing and the threat of miscegenation was the legally enshrined doctrine of “hypo-descent,” by which even a drop of African American blood qualified an individual as black.⁴⁶

Literary, theatrical, and cinematic examinations of the socio-legal paradox created by miscegenation were manifest around the turn of the decade—the novels Imitation of Life and Show Boat and their stage and film adaptations, for instance—presenting stories of fair-skinned African Americans passing in white society.⁴⁷ may have had both fortunate and unfortunate resonances for the Boswells' new radio audience, inasmuch as the two main musical characters, Magnolia and Julie, represent the sisters' two possible biological identities. Julie, the mixed-race singer who is unmasked, initially betrays her origins through her knowledge of a “colored” song (although her manner of delivery is strongly coded “white”—reliably performed by the white singer Helen Morgan).⁴⁸

Although both Martha and Connie were born in Kansas City, the Boswell Sisters grew up in New Orleans. Brought up in a middle-class home, all three girls were given a privileged musical education: Martha played the piano, Connie the cello, and Vet the violin. Yet Martha learned to play stride piano from her father, who “had a way of playing with that left hand that was something else,” and the sisters' jazz abilities and sensibilities were fostered in what could be seen as a slightly bohemian salon culture that grew up around their Camp Street home.⁴⁹

There the sisters mixed with other young white musicians interested in jazz, and for these gatherings, Connie eschewed the cello for the trombone and the saxophone; Vet picked up the guitar and the banjo. As much as their playing, singing was part of the sisters' daily experience, not only in the home but as a crucial element of the city's musical life. The sisters reportedly learned gospel and blues from the domestic help, and they would have heard quartet harmony singing on a regular basis. Recreational harmony was endemic in early twentieth-century New Orleans, cutting across white, black, and creole communities; it figured importantly in the Boswell house, too, for the adults (parents, aunt and uncle) regularly sang together as a quartet.⁵⁰ The Boswells undoubtedly absorbed every sound the city had to offer, and the legacy of their early years can clearly be heard in the scat breaks that characterize the sisters' recordings of the 1930s.⁵¹ But the girls also claimed to have experienced black music firsthand at the black theaters that opened to whites on weekend nights. Although smaller venues existed, there was only one formal circuit theater exclusively for blacks in New Orleans: the Lyric Theater, operated by the black vaudeville circuit TOBA (the Theater Owners' Booking Association), which allowed white audiences into balcony seats on Saturdays.⁵² The Whitman Sisters were headliners on the TOBA tour throughout the 1920s. No evidence exists to prove that the Boswells ever saw their show, but if they did they might have been impressed by the affinities they could draw between themselves and the classically trained, self-accompanying, harmony-singing Whitmans. Connie was known to have admired Mamie Smith, who was another TOBA regular. Ethel Waters also headlined on TOBA; in 1922, she opened at the New Orleans Lyric to rapturous acclaim.⁵³

In March 1925, the Victor Recording Company brought their mobile recording equipment to New Orleans seeking out new talent and the girls were offered the chance to record. Between 22 and 25 March they recorded five songs, becoming the first artists to record for Victor in the city. Two of their self-penned songs, “Cryin' Blues” and “Nights When I Am Lonely,” made it onto a single disc.⁵⁴

The songs, while wholly unrepresentative of the sisters' later output, reveal the early influences and elements that would eventually be subsumed and transformed in their mature style. Although quite different, not surprisingly they share common roots in both New Orleans popular culture and contemporaneous vaudeville styles. “Cryin' Blues,” primarily a solo for the seventeen-year-old Connie, effectively bundles up elements from “coon-shouting,” vaudeville blues, and recreational quartet singing.⁵⁵ “Nights When I Am Lonely” is more of an ensemble piece, also relying on close-harmony conventions but heavily influenced, too, by the sound of other popular sisters acts such as the Brox Sisters and the Duncan Sisters, both recording for Victor at that time.

The structure of “Cryin' Blues” follows the established pattern of instrumental introduction/sung verse/sung chorus/instrumental chorus/sung chorus common to many vaudeville blues recordings. Connie's delivery closely emulates the classic blues singers, with the occasional guttural growl, sliding, sobbing and blue-note inflections to the thirds and fifths. Just as the song's lyrics (“I ain't no Queen of Sheba, I don't dress fine, but I'm the kind of mama that's home any time,” and the rhyming of go with mo') are filled with intimations of the singer's supposed race, Connie's performance is doubtlessly intended to invoke blackness, if not actually to sound black. Yet four other features of the recording—Connie's belting projection, her accent, the excessive yodeling hiccup on the word cryin', and the treatment of the instrumental chorus—suggest that “Cryin' Blues” is as much about minstrelsy's legacies as it is about the blues. The song is placed at the top of Connie's chest register, and the resultant belt brings to mind Banfield's comments regarding the optimum pitch for “coon-shouting.” The style is also implicit in her bizarrely modified pseudo-southern accent; not only are her vowels inconsistent with all of her later recordings (and all records of her adult speaking voice), they are inconsistent within the song alone, suggesting that she was deliberately, if not entirely successfully, adopting a specific performative style.

Connie's yodel/hiccup is a theatrical hypervocalization of emotion, reminiscent of Emmett Miller and, more particularly, Lee Morse. Morse did not work in blackface, but she was described as a singer of “quaint Southern melodies,” and when recording her studio band was always listed as The Bluegrass Boys. The southern element presumably was implied only by her yodeling and the subject matter of her songs, since her accompanists were all northern jazzmen playing hot jazz, not bluegrass.⁵⁶

The “instrumental” chorus on “Cryin' Blues” is not provided by the piano accompaniment alone but features the only vocal ensemble work on the recording. Connie and Martha imitate a banjo and a muted trumpet through the full sixteen bars of the chorus—eight bars in the original feel, eight double-timed—scatting in a polyphonic duet. The Duncan Sisters were already using this type of effect in vaudeville, and Victor had released the Duncans' “Rememb'ring” only the year before, but the Boswells' instrumental imitations are harder-edged than the Duncans', showing an even closer affinity with the “authentically” black practice of imitating instrumental jazz as a feature of quartet singing.⁵⁷

The second of the Victor songs, “Nights When I Am Lonely” (shown in Example 1) also has musical roots in recreational harmonizing. The verse is a note-on-note duet, but the chorus is sung by all three sisters.⁵⁸

The structure of the complete arrangement is broadly the same as “Cryin' Blues,” but instead of an “instrumental” chorus, the sisters sing a chorus in what would later become known as “Boswellese,” a private gibberish play language they had spoken together since childhood.⁵⁹ On both sides, then, the sisters use the traditional space for improvisation in an instrumental jazz arrangement for a vocal novelty. Certain techniques that are familiar in barbershop lore also make an appearance, such as vocal slides filling in large melodic intervals (mm. 28–29), vocal lines crossed for obvious effect (mm. 23–24) and the use of passing-note quasi-heterophony decorating the melody (m. 32). The pinched, helium-nasal vocal delivery owes a huge debt to white flapper recording stars such as the Broxes and Helen Kane; perhaps here, as with Rosetta and Vivian Duncan's different vocal qualities, the nasal ingenue sound is meant to convey whiteness (in contrast to the belting of the “coon-shouter”). As a debut offering, a disc advertising the girls' abilities to sing both black and white might have been considered useful for promoting them to the public.

After the Victor session, the Boswells did not record again until 1930. There is no reason to suppose that the hiatus in their recording career had anything to do with a lack of interest in the initial disc; it is much more likely that personal circumstances dictated the subsequent course of events, limiting their professional activities between 1926 and 1927.⁶⁰

However, in the following two years the trajectory of their revitalized career, together with the technological innovations that were sweeping the entertainment media, brought new and different influences to bear on their developing style. Having made the decision to leave New Orleans and transcend local fame, the sisters had a brief flirtation with minor vaudeville circuits in the Midwest and California in 1928 and early 1929. They settled in Los Angeles and began working in both the infant “talkie” industry and on the radio, doing short sustaining programs in between networked and scheduled sponsored shows.⁶¹ Hollywood life brought them into contact with established and rising stars that could have influenced them musically, such as Cliff Edwards who, like the Boswells, provided off-screen musical tracks for the MGM musical They Learned About Women in 1930. The sisters also began a long-standing friendship with Bing Crosby, who was in Hollywood filming Paul Whiteman's The King of Jazz at Paramount and singing at the Cocoanut Grove with the Rhythm Boys.⁶² It seems reasonable to assume that they would have been aware of, if not actually met, the Brox Sisters, who also appeared in The King of Jazz, and the Duncan Sisters, who were making their feature It's a Great Life at MGM.

Although Victor had introduced electrical recording in 1925, the field studio in New Orleans where the Boswells first recorded used acoustic apparatus. Film and radio work in Los Angeles gave the sisters daily experience with the latest sound technology, including electrical recording and condenser microphones, and soon they found ways to tailor their singing style to exploit the new equipment. They no longer had to shout into an acoustic horn, so perhaps the most dramatic change in their style is the way they began to sing softly. This gave them more freedom of expression; they could subtly articulate emotion through tone and breath rather than hysterically hiccupping to the gallery. The greater intimacy of the microphone allowed the sisters' natural vocal placement to develop; and though they could sing nasally in the conventional popular white-girl-singer sound, they began to use the tone only for a certain effect, the rest of the time letting their voices resonate more freely in their throats or their chests. Furthermore, the sisters created a more natural support for the voice; rather than sustaining tone quality for its own sake as a trained singer would do, they let the ebb and flow of their natural speech cadences inflect the vocal line. Moreover, singing with a greater degree of dynamic and tonal range allowed their New Orleans accent to be heard more clearly; the open, modulated vowel sounds and relaxed consonants enhancing the “instrumental” attitude of their phrasing, whether it was in block harmony or solo passages, in lyrics as well as scat.⁶³

From this time, too, the sisters found they could exploit the lower parts of their ranges (in effect, the speaking range), and Martha and Connie began almost to compete with each other for the lowest notes in the harmony. Connie later explained how their drop in range occurred:

Frequently, one sister held the melody at the bottom of the triad with two voices above (rather than the more conventional placement of the melody in the middle voice), as in the two sides recorded with Jackie Taylor's Hollywood-based orchestra in July 1930: “We're On the Highway to Heaven” and “That's What I Like About You.”⁶⁵ The tessituras of both melodies rest below middle C and descend to E-flat, well below the accepted or acceptable range of the classical female mezzo-soprano or even contralto. This suggests that the orchestra's charts for both songs were originally devised for use with a male singer (or singers), and the Boswells simply inserted their arrangement where a male vocal would have been.

The difference in the Boswells' voices between the Victor recordings and the Hollywood sessions is stark. Although maturity and experience had also given them a much tighter blend, three factors of vocal production—tonal variety (leading to greater expressiveness), tessitura, and regional accent—provided the technical basis for the Boswells' new sound, the sound that would be identified by their audiences as “authentically” black. All three qualities had been exploited by minstrelsy's blackface artists; what set the Boswells' sound apart from that tradition was the lack of travesty or hypervocalization, or the necessity for another contextual marker (blackface makeup, black-identified material) to confirm the racial categorization. Furthermore, the Boswells sang with voices that sounded naturally produced; the quality of tone, ornamentation, vibrato, and enunciation was not overly manufactured, and the accent was genuine. For one later reviewer, authenticity was the key to their initial and their lasting success:

The radio audience in California could only compare what they heard from the Boswells with known models, and up against the most popular white female voices of the period, the Boswell Sisters were, quite literally, something else. Connie and Martha's ultra-low tessitura and the sisters' informal and natural diction contrast sharply with those of the established white sweethearts of stage and screen, such as Ruth Etting, Gertrude Lawrence, and the soon-to-be archetypal Hollywood soprano Jeanette MacDonald. Vaughn de Leath, the “First Lady of Radio” who was one of the pioneers of microphone singing, may have perfected singing softly, but her delivery sounds derived from European parlor song, and she hypervocalized emotion as part of a more melodramatic musical presentation, often interpolating spoken recitations into her performances. Helen Kane may have sung as she spoke—without refined elocution—but hers was a character voice, a squeaky, nasal travesty of baby-doll speech. Lee Morse, of the masculine range and “bluegrass” yodeling, for all her skillful collaboration with jazzmen was also still a character voice, a master of minstrelsy's legacy of hypervocalization. Annette Hanshaw, who of all the established white female pop singers of the 1920s came closest to the Boswells' timbral freedom and tessitura, even manages to modulate her vowel sounds somewhat (in her 1929 “Big City Blues,” for instance, she sings “Ah wonder whah-y” for “I wonder why”), but there is no consistency to her accent, which reverts to her native New York dialect frequently and without warning.⁶⁷

It is perhaps the matter of an authentic southern accent that sealed the Boswells' vocal racial ambiguity, for outside the South, in an age when transcontinental travel was still fairly unusual and most broadcast or recording concerns were concentrated in the North and West, audiences were as yet unaccustomed to hearing white southern voices. What they thought they knew of black southern voices had, of course, been filtered through minstrelsy, but performers such as Bessie Smith and Louis Armstrong had begun to disseminate the accents of southern African Americans. What passed for a white southern accent in Hollywood in the early days of talkies would have been (and still is) unrecognizable to actual southerners; even a year after the Boswells gained international fame, Stuart Erwin's Texas millionaire, as portrayed in Paramount's The Big Broadcast (a major, big-budget production in which the Boswells also starred), speaks his lines in an almost incomprehensible garble, not remotely identifiable as any southern regional accent. “Pseudo-southern” was recognized as a ruse, especially when it was inconsistent, but regardless of its inauthenticity it still operated strongly as a signifier, normally with disadvantageous (class- and education/intellect-related) connotations. As such, white southern actress-singers, such as Ginger Rogers and Tallulah Bankhead, worked hard to eliminate their native accents, at least in performance. One might have expected the Boswells to do the same, but they did not. Together with the lower tessitura and the natural expressivity, the Boswells' genuine accent provided the complete racially identified vocal package.

Gary Giddins's analysis of Bing Crosby's developing style offers an appealing comparison. He states, “To Bing's generation, southern was a synonym for black,” and he makes a particular note of the way southern speech cadences “chimed well with [Crosby's] bottom notes and vocal affectations.”⁶⁸

For singers like Crosby, striving to assimilate the gestures of blackness both musically and vocally, the Boswells must have been a revelation of how to be white but to sound black. Giddins nails the three qualities—intimate expressivity, tessitura, accent—as crucial to Crosby's sound, but he fails to acknowledge the early impact of the Boswells. Universally, Crosby's development as a jazz singer is seen solely as part of a reciprocal influence with Louis Armstrong, unfolding over the second half of the 1920s. However, if one compares the facial tone, the hard r's and the tight vowels of Crosby's 1928 performance of “There Ain't No Sweet Man That's Worth the Salt of My Tears” (a song laden with lyrics identified with black female blues singers, such as the appeal to “broken-hearted sisters, aggravatin' misters,” so crying out for a black vocal treatment) with his relaxed, throaty, diphthong-rich pronunciation of “Life Is Just a Bowl of Cherries” (“l-ah-ife is jus' ta bah-owl of chuh-erriz”), which he sings with Connie and her sisters in 1931, the Boswells are also heavily implicated in his transformation.⁶⁹

Along with the blackness of their vocal quality, the sisters' musical style incorporated many elements that could be associated with blackness, blues, and hot jazz. “Boswellese” gibberish, although primarily an effective way of making a straight tune more rhythmically sophisticated, was also alienating and foreign to the listener, redolent of both scat and the “jive talk” being developed by the alternative black jazz community.⁷⁰

Their rhythm was loose and swinging, their drawls elongating the downbeat, and melodic embellishment and scat-singing, both solo and in harmony, were frequent features of their arrangements. Their keen sense of intonation allowed them to exploit blue-note inflections as effectively in harmony as in solo passages, both in parallel movement or as a way of exoticizing the blander harmonies of their Tin Pan Alley material.⁷¹ They were also increasingly irreverent in their treatments of songs, breaking down lyrics, melodies, harmonies, and whole song structures in service to their musical imagination.

The Boswells gave their first coast-to-coast broadcast in July 1930, and almost immediately they became a hot property. Record companies renewed their interest, and in October the sisters cut four sides for OKeh; by November, they were negotiating to visit NBC headquarters in New York. Nevertheless, the incongruity of the Boswells' true racial origin on the one hand and their output on the other may have been disturbing, even more than confusing, to certain sectors of the public. This is suggested by a letter of complaint, written to a radio station in San Francisco by an Oakland woman: “But please, please, if you are going to keep those Boswell Sisters tell them to change their stuff and quit that squawking and harmonize a tune. All my friends say the same thing. They call them the savage chanters and tune them out.”⁷²

Clearly the listener was responding and objecting to what she perceived to be a conscious invocation of blackness, but one that was not safely mediated by the traditional blackface conventions. If a notion of cultural geneticism defined to which race certain musics, styles and gestures belonged, then a notion of “cultural hypo-descent” could have helped establish the blackness of a performance through the incorporation of single or multiple stylistic elements. The consternation of some members of the Boswells' listening audience was consistent with a general hardening of attitudes towards people of mixed race during this period, especially those in the public eye. It seems that policing the boundaries between black and white, either by requiring light-skinned performers like Cab Calloway to wear dark makeup (as in The Big Broadcast) or by asking the radio station to tell the Boswells to “quit that squawking and harmonize a tune,” was important to some in upholding the validity of hypo-descent, culturally or biologically, while it still remained enshrined in the laws of many states.⁷³ The entertainment industry's careful attitude to racial difference might be seen as symptomatic of wider concerns. In the summer of 1930, America was coming to terms with the gradual recognition that economic upturn was not just around the corner, with a concomitant effect of heightening racial tensions as interracial competition for jobs and welfare relief became all too real. For entertainers, the attraction and assimilation of blackness was more or less a one-way street, but there are hints that the business was looking for ways of naturalizing blackness for whites: for instance, before Amos and Andy appeared in their first motion picture in October 1930, NBC arranged for their white portrayers Correll and Gosden to appear on national radio—for the first time—as themselves, speaking without their blackface accents.⁷⁴ Publicity material and press coverage issued early in the Boswells' national career appeared to play on their vocal ambiguity (particularly for northern audiences), with headlines like “Girls Want It Known That They Are Creoles” and “The Boswell Sisters Are Three-Quarters French,” but the racial titillation promised by such phrases was invariably dispelled in the body of the articles, with ample reassurance of the girls' racial origin and class status.⁷⁵

The need for such reassurance was, of course, heightened by the mode of broadcast and the performers' invisibility. As the Boswells' first medium for performance, radio accentuated the ambiguity of their sound by forcing the listener to speculate on its origins. In his analysis of black musical performance on film, Arthur Knight suggests that “radio's divorce of sound from image added a modern, technological anxiety to America's visually grounded but invisibly ‘blood’-based system of racism…. Insofar as radio was blind, it was also color blind, but in the context of a society that was extremely color attuned, radio's blindness made masking at once unavoidable and uncertain.”⁷⁶

Knight states further that “what music looks like relates crucially to how it sounds and what it can mean…. In the United States especially, a key determinant of the look of any music has long been the race—the ‘color’—of the musicians who play it.”⁷⁷ The language the 1930 PMR reviewer used to criticize Ethel Waters's performance of “My Kind of Man”—“aping white styles” (my italics)—supports the validity of Knight's statement, particularly for black performers during the historical period under review. However, the anxiety created by the Boswells' sound suggests that a reversal of his hypothesis is also true, that what music sounded like related crucially to how it looked, particularly in the imagination of the hearer. The need to visualize a category, moreover one that existed only within a binary system, meant that when dealing with radio and recorded sound, the processes of racial etiquette were reversed. Instead of approaching visual ambiguity by listening out for vocal clues—accent, vocabulary—that might aid in racial categorization, the listeners had to look for visual confirmation and reassurance through record company promotional material or radio fan magazines that their assessment was correct.

For audiences like the letter writer from Oakland and her friends, used to listening in black and white, it was almost impossible to tell where whiteness ended and blackness began in the Boswells' collective voice. Their sophisticated arrangements suggested white jazz, recalling and surpassing Paul Whiteman's Rhythm Boys, who “hotted up” the traditional sound of male harmony by interspersing their Tin Pan Alley material with short scat breaks, but the blackness implied by the sisters' voices, their rhythm, and the influences they chose to bring to their material mitigated their whiteness.Yet it is clear that the sisters appreciated the difference between sounding white and sounding black, and they were not afraid to have a little fun playing one off against the other. Their arrangement of the Tin Pan Alley standard “Don't Tell Him (What's Happened to Me),” which they recorded while still in California, has a kind of split personality, rocking back and forth between white and black styles from the opening bars.⁷⁸

The first eight bars of the chorus map out black styles: first the sisters sing a fast chordal recitation in a banjo-strum rhythm, then Connie breaks into a bluesy solo for the title line (“but don't tell him what happened to me”), accompanied by Martha's barrelhouse slow drag piano. The second eight bars, the lyrics of which mirror the first, are sung in closely corresponding white styles: the sisters sing a straight, slow, block harmonization of the melody in barbershop precision, then Connie repeats the title line imitating the torch-singer style of Helen Morgan, complete with a refined, arpeggiated accompaniment from Martha. The rest of the arrangement invokes stomping four-in-a-bar blues, hot jazz scatting and the original sweet ballad almost in equal measure. Enunciation and intonation (“tell him” or “tell 'um”; major thirds or blue notes), the use of vibrato and tonal variation change with the racial implications of the style at any given time. So it seems clear that the Boswells were as conscious of performing whiteness as they were of the implications of blackness in their work. And Herbert Bailey, the listener from Roxbury, Massachusetts, would have had every right to be deeply confused—not least because he would have been hard put to choose which of these styles was more “natural” to the girls: the one that had been racialized to connote a biological and ethnic type other than their own, or the one that required an acculturization of the voice that took it away from its “natural” production.

Among the other sides the Boswells released on OKeh was one that critiqued and almost redefined an existing recording that had already achieved “definitive” status in the new jazz canon. One of the most enduring myths of jazz history is the “creation” of scat singing by Louis Armstrong, caught on wax in the recording of “Heebie Jeebies,” made by Armstrong and his Hot Five in February 1926.⁷⁹

In reality, scat was nothing new, but Armstrong's record was the hit side that started both a dance craze and the craze for scat. Before “Heebie Jeebies,” it could be that (like a pseudo-southern accent) scat had begun to signify “blackface” rather than “black,” but its adoption by an African American jazzman reclaimed it as a legitimately “black” practice. Armstrong's performance on the recording has also been credited with establishing new standards for the jazz vocalist. In Will Friedwald's words, it “transliterate[d] patterns Armstrong had conceived for instrumental music very directly into vocal terms, starting with lyrics, then modulating into scat phrases, and returning to the words at the conclusion.”⁸⁰ Certainly, hot jazz vocals would increasingly follow this pattern thereafter; however, it is probably misleading to attribute this innovation to Armstrong's genius rather than to his powerful agency. The lyric chorus/scat chorus/lyric chorus structure, or its more concise form lyric chorus/scat to middle eight/lyric finish, must have already been widespread common practice.⁸¹ It would appear that Armstrong was drawing on his roots in quartet singing for the application of scat within the structure of an otherwise texted vocal performance; but whatever its origin, it was Armstrong's performative skill that brought “Heebie Jeebies” and its anarchic babble of scat to an international audience.

The sheet music for “(The Original) Heebie Jeebies” was published after Armstrong's recording was issued, but a white artist, Paul Ash, is pictured on the front and given credit for “featuring” the tune.⁸²

The printed verse and chorus correspond to the head arrangement played by the Hot Five rather than Armstrong's vocal performance. The published lyrics also differ considerably from Armstrong's; he sings something like the words given for the second chorus but it is impossible to tell which version is derived from which. Nevertheless, the sheet music's debt to Armstrong's recording is manifested in its incorporation of a (not wholly accurate) piano transcription of his scat chorus, with an added gloss, “Note: For correct interpretation of “SKAT” CHORUS HEAR OKEH RECORD No. 8300.”⁸³ Other recordings of the tune were issued quickly by both black and white artists. Ethel Waters's was the first to appear in July 1926.⁸⁴ Although all show the influence of the Hot Five recording with cornet solos aplenty, most of the performances—apart from Waters's—refer less to Armstrong's vocal than to the published lyrics and melodies. Waters's version is in a way surprising for its lack of scat, especially as she sings the first chorus almost note for note (but not word for word) from Armstrong's, but her performance is revealing in other ways. Like Armstrong, Waters avoids the melody almost completely, but instead of operating as a variation, her single-note declamation of the verse material is superimposed on a full piano rendition of tune; the effect is not unlike a pedal note in Dixieland polyphony. Equally instrumentally inspired, Waters swaps licks with her cornettist in the final chorus, fragmenting the words to the point that they become rhythmic accessories to the phrasing rather than lyrics. The whole performance has the feel of a truly integrated dance band rather than a singer with accompanists; it is also noticeably faster than any of the other 1926 versions, matched only by the breakneck speed of the Boswells' side.

The Boswell Sisters recorded their first version of “Heebie Jeebies” in October 1930, shortly before they were invited to New York to join NBC's national network. The tune was something of a talisman for the sisters: they recorded it twice for issue on 78 disc (in 1930 on OKeh and in 1931 on Brunswick with the Dorsey Brothers), and they also used it as an opening number for every new broadcast, largely retaining the arrangement intact.⁸⁵

It is unlikely that the 1930 recording itself represents a “new” arrangement, for the sisters always claimed that they knew the tune from their New Orleans childhood.⁸⁶ Melodically, the Boswells' arrangement (shown in Example 2) is completely different from most of the previous recordings; for instance, it never uses the sheet music's verse material. Nevertheless it does contain elements identical or very similar to Armstrong's performance: many of the lyrics (especially where his differ from the sheet music); the substitution of the chorus melody with single- or two-note declamation; the opening of the scat chorus; the use of the interjected patois phrases “suite cause evant” and “yoz ma'am.” The vocal fills in the stop-time coda are also the closest to Armstrong's.

These similarities may arise from a common performance tradition in New Orleans, or they could be conscious modeling on the part of the Boswells; most likely, they are a combination of the two. But regardless of where the Boswells originally learned the tune, their treatment of “Heebie Jeebies” is every bit as radical as the schizophrenic “Don't Tell Him (What's Happened to Me”); if anything, it is more explicit about its many stylistic origins. From quartet singing come the sliding horn imitations of the opening and from instrumental brass band “jazz,” the material the voices imitate (Ex. 2, mm. 1–16). Connie's solo verse invokes the female blues tradition, slow and low-down (mm. 34–49).⁸⁷

The scat chorus begins (m. 52) by referring to Armstrong's vocal—or a common model—but then the sisters continue towards a stop-time fill from a “doo-wack-a-doo” phrase (mm. 53–59), the sort of systematized scat that had come to signify the white youth culture of flappers and sheiks, and which characterized the sound of Whiteman's Rhythm Boys. Finally, in a direct invocation of minstrelsy comes the exaggerated hokum of the coda, the dialect enunciation of “Watcha doin' wid duh jeebies” and the banjo-imitation tag at the close (mm. 71–81).⁸⁸ From Armstrong's emerging standards of jazz singing comes a modified diction, sometimes closing a word down with a hum, “Don't feel blue-mm”; sometimes, like both Armstrong and Waters, using hard consonant sounds rhythmically to articulate an über-syncopation on the backbeat, “COME on down JUST a little BIT of joy.” These vocal features are clearly conscious stylistic gestures because they occur in three-part harmony, which means the sisters would have had to rehearse them.

The recordings made for OKeh in California show the more or less raw material the Boswells brought to New York—they have no real instrumental involvement apart from Martha's piano and, judging by the number of rejected waxes, quality control seems to have been done at the pressing stage rather than in the studio.⁸⁹

The lightning pace at which the stylistic disruption occurs in these recordings may also have contributed to their racial ambiguity. Apart from labeling them the first white vocalists to sound black, the one consistent comment made by jazz musicologists about the Boswells is that their arrangements were extraordinary.⁹⁰ Years into her retirement, Connie said, “At that time I'm sure that to the average ear we must have sounded like little green people from outer space.”⁹¹ For black and white performers alike, the traditions of minstrelsy provided a medium for the bizarre or the marginal, and for satire.⁹² The most outlandish of the Boswells' musical propositions—turning 3/4 to 4/4, major to minor, fox-trots into blues, or ragtime into rumba—could be accommodated by minstrelsy's language of inversion and subversion. At their most mercurial, the Boswells' arrangements also can be seen to correspond to the black blackface technique of “indefinite talk,” which the historian of black humor Mel Watkins considers a “type of comedy that was rarely seen outside black honky-tonks, variety shows, theaters and the occasional race movie.”⁹³ Indefinite talk is Arthur Knight compares the appeal of indefinite talk to the appeal of jazz, noting that its affect and meaning depend on “association, timbre, beat, rhythm and volume,” and that it creates its humor through “discrepancies” of syntax and process.⁹⁵ Certainly the Boswells' arrangements of “Don't Tell Him (What's Happened to Me)” and “Heebie Jeebies,” as well as many others not discussed in this essay, are models of disjunction and discrepancy, manipulating those musical attributes that Knight lists as essential. In the case of these two early sides, the interrupted dialogue is between blackness and whiteness, between masculinity and femininity, between voice and instrument. What it speaks of is the undefined middle term. As Knight points out, “There is no clear ‘it’ to get in indefinite talk, but nonetheless, some audiences—those who collaborate in the process of filling in the discrepancies—will ‘get it’ and some will not.”⁹⁶

White Is the New Black: The Boswell Sisters in the 1930s

Once they arrived in New York, the Boswell Sisters began to record for Brunswick Records, coming under the control of Jack Kapp, the record mogul who eventually broke with Brunswick to form the international label Decca. In a way, they seem to be exactly the sort of act that Kapp would have rejected out of hand; he is best remembered, and almost universally vilified, for his constant exhortations to singers to “stick to the melody.” His stylistic policing has been blamed for dulling the jazz sensibilities of Bing Crosby, Mildred Bailey, Lee Wiley, and later in the 1930s, Armstrong and Ella Fitzgerald (among many others). Connie later said that Kapp was “a wonderful guy … a terrific commercial man,” but it is clear she regretted many of his commercially oriented impositions.⁹⁷

He was certainly a man with a mission, a figurehead in the popular music industry's quest for “the lowest common denominator in its material: that music which will please as many and offend as few potential customers as possible.”⁹⁸ By 1930, Kapp had begun to expand his empire by blurring the audible color line, hitting paydirt by extracting the most irresistible elements of jazz and placing them in the context of mainstream popular music, so making jazz truly commercial. After ten years' exposure to race records from the major labels, white audiences had fully graduated from the coarse travesty of minstrelsy; the record-buying public felt it could discern between black and white, between authentic and inauthentic sounds, and it also had a sense of appropriateness of material and style. Only once such an environment was established could the Boswells have made not only a feature but also a virtue of manipulating the audience's expectations and judgments of how white women should sound.

In summer 1930, Kapp was promoted to head of all Brunswick recording laboratories; his ascendancy coincided with dramatic change in the main catalogue.⁹⁹

Whereas in 1929, literally dozens of tenors, sopranos, and stars of vaudeville and the musical stage were recorded by Brunswick, by the autumn of 1930, only four singers (all male) were still regularly issuing discs. Then in March 1931, within two weeks of each other the Boswell Sisters and Bing Crosby cut their first sides for Brunswick, heralding Kapp's new elite stable. Until well into 1932, Kapp concentrated his effort on just a handful of vocal artists, all rising or established stars on the radio: along with the Boswells (the sisters and Connie alone) and Crosby, Arthur Tracy (The Street Singer), Mildred Bailey, and the Mills Brothers also joined the roster, with Nick Lucas and bandleader Cab Calloway the only regulars remaining from 1929. It is difficult to ascertain how much the prevailing economic conditions may have influenced Kapp's strategy, especially early on, but the stock market crash of October 1929 and the ensuing economic meltdown must have had an overwhelming effect. It was Kapp's job to keep Brunswick afloat, and his complete rethink must have involved deciding which acts could continue to sell records, and which could not; additionally, those records would feature only the repertoire Kapp felt the American public would buy.

In the Boswells, Kapp had found his marketing Holy Grail. They were white singers that sang white material (i.e., Tin Pan Alley rather than blues) but sounded black, who could ride the crest of the popularity of black music while simultaneously drawing in more racially conservative white audiences. But in order to do so, a public image needed to be cultivated for them that both dispelled ambiguity and nonetheless permitted them to retain ownership of black signifiers. From the Boswell's second Brunswick recording, “Roll On, Mississippi, Roll On,” backed by what would become their signature tune, “Shout, Sister, Shout,” much of their material authorized for release by Kapp capitalized on their southernness: “Got the South in My Soul,” “Lawd, You Made the Night Too Long,” “An Evening in Caroline,” “Swanee Mammy,” “Down on the Delta,” and “Louisiana Hayride” are just some examples.¹⁰⁰

Publicity material echoed the theme, emphasizing the girls' birthright to their racialized sound by invoking fantasies of the Old South: “‘Hot’ rhythms were a natural outgrowth of their listening to the negro plantation singers around New Orleans.”¹⁰¹ But even sides that ostensibly had nothing to do with the South were likely to feature Connie's low-down, preaching gospel style, or to somehow incorporate Dixie references: for instance, the arrangement of “Put That Sun Back in the Sky,” an unremarkable but upbeat ditty on the rhythm changes, interpolates a slow march spiritual version of the middle eight which then suddenly breaks into a fast swing chant, superimposing the song's lyrics onto the chords and rhythmic patterns of “When the Saints Go Marchin' In.”¹⁰²

The Boswells left behind few visual records of their performances, mostly as a star turn in larger features or publicity shorts. In one way or another, the southern theme was carried through most of these cameos, their rise in status as performers echoed in the social elevation of their image. In the Boswells' first film, a short titled Close Farm-ony shot in 1931, the sisters are characterized as rural farm workers or hillbillies, an established white southern stereotype, dressed in overalls and seated on hay bales (even the piano is disguised in hay).¹⁰³

Later in 1932, they appeared in the first of Vitaphone's Rambling Round Radio Row series that provided “backstage” glimpses of the nation's favorite radio stars. Although they wear modest dark daydresses suitable for middle-class working women, again their southernness is underscored through the musical material (a song called “Lou'siana Waddle”) and through their spoken dialogue, which clearly showcases their heavy accents. In the Boswells' first major film appearance for Paramount, the feature-length star vehicle The Big Broadcast, the sisters' image is gentrified further, as shown in Figure 2 File Name: s1752196307070083s.sgm See this text:; dressed in white frilled chiffon gowns with silk corsages on their shoulders, they are introduced as “those three little girls from New Orleans.”¹⁰⁴ This southern-identified look, however moderate, nonetheless contrasts sharply with that of the other female performers in the film—including Kate Smith, Gracie Allen, and Boswell wannabes Major, Sharp and Minor—who wear either the fitted, sophisticated wardrobe of the urban socialite or sensible shirtwaist dresses.

Shortly after the release of The Big Broadcast, the sisters were the subjects of a short by Dave Fleischer, When It's Sleepy Time Down South, based on the song of the same name—and yet another tune that was strongly associated with Louis Armstrong.¹⁰⁵

Typically, the song's lyrics are infused with nostalgic images of “darkies” and mammies, and the impact of seeing and hearing white women singing these words is profound. When sung by Armstrong (and reinforced by his introductory patter), they conjure an image of the South—that of cotton fields, riverboats, and banjos—with which he, as an African American, would be identified and hence objectified by his audience. When the Boswells sing, they appear to share the listener's subjective position as narrators outside the scene, at least until they long “to be in my mammy's arms”—at this point, the Boswells become southern ladies, nurtured by a (slave or servant) “mammy.” The film presents the sisters in a garden “frame,” the piano stool replaced by a baroque-style white stone bench, and scenery overhung with lush vegetation.¹⁰⁶ The plantation backdrop of the minstrel show is directly referenced, but with the belles center stage; privileged white women take the place of contented slaves. Although they wear white satin pajama suits, not hoop skirts, they clearly are intended to invoke an excess of southernness; and again, the arrangement adds to the emphasis (as if any more were needed), as Connie interpolates another of her low-down verses, this time to the reworked tune of “Dixie.” The film makes manifest the way in which southernness becomes a way of understanding and accepting the Boswells' blackness; by working within an “economy of visibility,” blackness is not so much erased as rendered invisible by a white masquerade.¹⁰⁷

Given both his interest in cashing in on musical passing and his well-documented involvement in his artists' musical development, it seems logical that Kapp was strongly influential in promoting the Boswells' southernness, yet the strategy must also have met with their approval. Southernness remained the foundation of the sisters' catalogue throughout their recording and broadcasting career, even after Kapp left Brunswick to form Decca in late summer 1934.¹⁰⁸

The material used to connote southernness by the Boswells shared themes with, and in some cases was virtually identical to, that used by both blackface and “coon-shouters” to connote blackness. It would seem inevitable that at least traces of established vaudevillian codes could be detected in the sisters' work, but certainly the embodied practice of the “coon-shouters,” which was particularly bound to the depiction of specific sexualized characters, is not present (with one exception, to be discussed below). However, on rare occasions more general references to minstrelsy's comedic traditions can be heard. In some of their earliest recordings and radio transcriptions, Connie and Martha duet like a vaudeville double act—a short section in “Wha'dja Do to Me?” has Martha introducing a line, with Connie providing the sidekick's comments.¹⁰⁹ A single session in 1932 also produced two sides on which the sisters interact with the musicians in a more vaudevillian fashion. Both songs, perhaps unsurprisingly, have racial connotations. On “Minnie the Moocher's Wedding Day,” a sequel to Cab Calloway's “Hi-De-Ho Song,” Connie invokes Calloway's style by creating a call-and-response “Yeah, man!” with the band. “Louisiana Hayride” features horse-drawn carriage sound effects and an internal skit involving a roll call of musicians (to which the sisters reply with a unison “Yowzah!”) by an unnamed interlocutor, invoking an important feature of the traditional minstrel olio.¹¹⁰ Visually, the sisters' performances did not—and could not—incorporate the more physical aspects of vaudeville: Connie had been paralyzed by polio at the age of three and had very limited movement, so the sisters were static on stage, grouped around Martha's piano, most often with Connie seated next to her and Vet standing behind.¹¹¹ Nonetheless, vague performative references to mugging—shaking heads, wide eyes, open-mouthed smiles—form part of their expressive palette. Film footage from 1932 shows how Martha, in particular, uses these facial gestures to characterize her singing while her hands are busy playing the piano.

In 1933 and 1934, the Boswells appeared in two larger-scale Hollywood musicals, Transatlantic Merry-Go-Round and Moulin Rouge; in both films, the sisters' numbers are framed as musical variety turns within the plot. The agency of Hollywood in shaping the performances and in the way it translated the Boswells' music to the screen is strongly felt, particularly in Moulin Rouge: without a visual tradition to draw upon for the portrayal of white jazz singers, the designers fall back on minstrelsy and vaudeville, but not without some discomfort. In the film, the Boswells sing a suggestive lyric, “(Coffee in the Morning and) Kisses in the Night,” as part of a lavish production number. The staging incorporates marked visual clues that reinforce the racialized musical traits in their performance and directly invoke the mammy, but it also simultaneously places the Boswells in opposition to those images.¹¹²

The song is first sung by Constance Bennett and Russ Columbo, she voicing the coy line “coffee in the morning,” he the more explicit “kisses in the night.”¹¹³

A curtain is then pulled back to reveal the Boswells, who are arranged on and around Martha's piano (shown in Figure 3 top); they are framed by a false proscenium advertising the “Melodee Coffee Shoppee.” The sisters are engulfed in cooks' uniforms, with gleaming aprons, broad white collars and huge, billowing chefs' hats (shown in Figure 3 bottom). Already low (to accommodate the speaking range of the non-singer Bennett, who sings in the same range as Columbo), the key drops by a further semitone, and the tempo slows. In harmony, the sisters slur and drawl their way through the lyric in its entirety: plosive sounds are minimized and smaller words swallowed; the title line becomes “cawffe' 'n'm-mawnin' ‘n’ gizzuhz 'n-n'nahht.” In an original coda, Connie almost entirely uncharacteristically assumes the sexualized persona of the “coon-shouter,” at once yielding and predatory, complete with a double entendre in the delivery: “Oh sweetheart I really want you [pause] to take me in your arms and hold me tight, now honey I'm warnin', you'll get coffee in the mornin' and kisses in the night.”¹¹⁴

The vaudeville theme is continued in the final section of the musical production, in which white adult chorus girls dressed in polka dot petticoated outfits tap dance like so many “picks” outside a schoolhouse set, a large family of daughters for Columbo and Bennett—as the song says, the inevitable “consequence” of all that coffee and all those kisses.¹¹⁵

Although the song's lyrics respect the Hays Code's pronouncements on marriage and sex in principle (especially when shared between a couple), when sung by women only, the song becomes risqué, and its purpose in underlining the film's will-he-or-will-he-not-commit-adultery plot is revealed. But nice white girls with impeccable southern manners do not sing about sex—at least not on film—therefore the sisters' performance is propped with visual allusions to a black stereotype.¹¹⁶

Their costumes identify them with the southern cook-servant, a role already signified through the Aunt Jemima trademark; the reference is explicitly reinforced by the huge caricature of a pancake-tossing chef on the backdrop. Yet the glittering opulence of their sequined dresses and the substitution of the (male) chef's hats for the more customary bandana place the sisters in a poorly defined but nonetheless privileged space in relation to the backdrop caricature; they are not common or garden-variety hash-slingers or domestics, they are virtuosi, and they are white—or at least not-black. This curiously ambivalent version of minstrelsy's masquerades—“mammies” who are clearly not mammies, “picks” who are clearly not picks—possibly unintentionally echoes the main plotline, which involves Bennett impersonating her own sister in order to return to the stage, and subsequently ensnaring her own husband, who has not seen through her disguise. But it also constructs the Boswells' whiteness literally against a backdrop of blackness, and though it confirms the blackness of their music, it does not contest their ownership of both styles.¹¹⁷

Black and White and Reb All Over: The Boswells as Belles

Ruth Frankenberg recognizes that whiteness can be “the unmarked marker,” an empty signifier “more about the power to include and exclude groups and individuals than about the actual practices of those who are to be let in or kept out.”¹¹⁸

Before broadcasting and phonography in the 1920s allowed black voices to be heard, at least in ways they had not been heard before, the construction of blackness by whites in American entertainment was less about absolute authenticity than it was about proclaiming, “Whiteness is not this.” The Boswells initially disrupted notions of racial propriety in music because they did not eschew their own “natural” use of somewhat arbitrarily ascribed practices that marked the practitioner as excluded from whiteness. Why they chose not to is a matter of conjecture, but clearly it was their choice if they were conscious enough of racialized musicking to be able to contrast white and black styles from the start. Furthermore, their stylistic consistency throughout their career suggests they had few serious qualms about their perceived blackness. Had the Boswells been particularly sensitive to misapprehensions regarding their skin color, it was well within their musical gift to modify their performance style to convey a whiter sound from the outset. And it hardly needs to be stressed that as the racial markers for sonic non-whiteness were determined within a predominantly northern matrix, a whiter sound would also be a more northern, non-ethnically marked sound. Northern, like white in American entertainment, was an empty signifier—the absence of accent, the absence of attitude. Northern was normal, and therefore an invisible standard against which difference could be measured.

But unlike Mezz Mezzrow, for instance, the Boswells should not automatically be taken for “exceptional whites”: those whose assumed blackness is an expression of a desire to be black.¹¹⁹

Until they reached New York, their whiteness, although implied by their acceptance by radio sponsors, was not visible; and without being seen positively to reject whiteness, they could not be seen to be “exceptional.” Furthermore, it is possible that whereas the Boswells clearly recognized the racial implications of their music for their audiences, the dualism they were operating was to them as much one of North and South (or South and not-South) as it was black and white. Qualities they had as birthright, or that could be perceived as such—the naturalness of their feminine tonal expressivity, their accent—were not necessarily qualities they would themselves have recognized as “black.” On the other hand, the acculturated, apparently accentless voice that the sisters use in “Don't Tell Him What Happened to Me” would, to them, have been a performance of “not-southern.” By the same token, the sweet jazz that is and was so commonly characterized as white was almost exclusively a northern (or not-southern) phenomenon; if the sweet styles in “Don't Tell Him” indicated whiteness, they also indicated northernness. So what of the hot styles the sisters chose to contrast with the sweet—did these styles first mean “southern” to them, before they meant “black”? Dale Cockrell acknowledges that his youthful practice of black musical forms and repertoire was not a conscious theft or appropriation; to him, “It felt more like sharing, or loving.” He sees the “amalgamation” of white and black popular musics as a specifically southern phenomenon, arguing that recognizing its regional genesis is more significant than demarcating its racialized components.¹²⁰

Many scholars have commented on how minstrelsy offered a nostalgic Neverland of the agrarian South that assuaged the anxieties of the industrialized North, but the racial longing that minstrelsy and racialized vaudeville satisfied was no longer of the age.¹²¹

In the desperate first years of the Depression, which saw such a catastrophic failure of white northern masculinity, a southern masquerade could still furnish comfort for those struggling to find a psychological foothold while the nation was in free fall. But as the economic tragedy of the early 1930s unfolded, the ambiguity that titillated audiences of the twenties threatened rather than enthralled; at a time when whites were replacing blacks in jobs previous thought beneath them (for instance, garbage collection and domestic service), a strong white identity was needed.¹²² Of the stereotypes on offer, the southern belle had many more potentially positive and useful qualities than Simon Legree, the hillbilly, or the oil-rich millionaire, and she was one with which both northerners and southerners were happy to identify. The belle had been a mainstay of the imagined South since Reconstruction, especially through the cultural representations, theatrical and literary, that refigured the conflict in both North and South. In her own region, the belle had long “served as the linchpin of nineteenth-century revisionist versions of the Old South, in which the Lost Cause ideology of southern nationalism conveniently fused the figure of the southern lady onto a celebration of the rebirth of a ‘nation’ defeated.”¹²³ Northern minstrelsy may once have used blackness to signify southernness, and vice versa, but the (white) South did not see itself as black, it saw itself as female.¹²⁴ Tough and resilient yet quintessentially ladylike, the belle epitomized an adherence to strict behavioral codes, and therefore order in society; in the South, even the most gracious etiquette “functions as social control,” and social control was vital as the nation's economic infrastructure crumbled.¹²⁵ Furthermore, the belle offered the entertainment industry of early 1930s America the perfect vehicle for an optimism that nonetheless did not deny the seriousness of the nation's situation, and a strong hyperfemininity that would still defer when the time came (no matter how wrong the northern white males got it).¹²⁶ Like the southern Democrats who helped the northern Democrat Roosevelt to victory in 1932, hers would always ultimately be a supporting role.

Although not the first white southern musicians to make it in the North (the Original Dixieland Jazz Band preceded them by over a decade), as singers the Boswells gave an authentic voice to southern whiteness that contrasted—in class and in culture—with the other widely disseminated white southern stereotype, the hillbilly. Their media personae were early manifestations of what was soon to be an inundation of belles in American popular culture, epitomized by the character Scarlett O'Hara in Margaret Mitchell's 1936 novel, Gone With The Wind. There were good belles and bad belles, repentant belles that learned to be good after being bad, smart belles, and occasionally dumb belles (although these characters tended to be class-inflected to be closer to hillbillies). What these women shared, from Daisy Mae Scragg in the comic strip Li'l Abner (1934) to Bette Davis in Jezebel (1938), was a fierce sense of determination, a refusal to accept failure without a fight, and a willingness to risk all in order to attain their goals. In films, aspirational southern women working their way up in the entertainment business became a popular cliché, providing roles for bona fide southern actresses including Texans Ann Miller (Stage Door, 1937; Radio City Revels, 1938) and Mary Martin (Rhythm on the River, 1940; Kiss the Boys Goodbye, 1941, which also starred Connie Boswell in a speaking/singing role). That a southern accent had become a desirable asset in a singer, even for northern girls and despite its potential for racial marking, is reflected in the dialogue of the Oscar-winning 1937 film The Awful Truth. Cary Grant's character asks his date, a second-rate entertainer calling herself Dixie Belle Lee, “How long have you been talking like Amos and Andy, huh?” and she responds, “Oh, for quite some time. I got wise to the fact that it helps me in my work. So as long as I love my work, y'all'll have to pardon my southern accent.”

As southern white women, the Boswells would have been well inculcated into their societal role, and it was one based on an overt bargain: white privilege and limited access to the public sphere accorded in return for “mounting the pedestal on which southern femininity was popularly situated.”¹²⁷

Southern white privilege gave them ownership of all things black, and when constructed as southern ladies, their use of signifiers of blackness—without masquerading as black—became explicable and acceptable. A 1944 Down Beat retrospective emphasizes that the Boswell household had “colored women” in the household who “jived,” “crooned,” and “rocked”; and because of this, “No wonder [they] … sang the way they did! It was only natural for them to sing blues on the pop tunes, and spirituals on the blues.” Yet in the following paragraph the writer claims, “They sang like nobody else,” even while relating the story of how Connie learned the blues from “an unknown boss of a riverboat gang.”¹²⁸ It is useful to compare the quoted reviews of the Boswells and Ethel Waters in this respect: the Boswells are admired for their “authenticity,” while Waters can only “ape” white styles. Nevertheless, the sisters rarely ventured far into blues territory for their material, though their white southernness allowed them access to black signifiers that could be applied to “unmarked” repertoire. To perform unseen in such a heavily marked genre could have been regarded as a step too far.¹²⁹ The sexually charged traditions of vaudeville blues would have been entirely inappropriate to the southern lady. Accordingly (and without denying the presence of double-entendre in many Tin Pan Alley songs), only one side of the seventy-odd cut by the Boswells in the 1930s, “Coffee in the Morning,” has overtly salacious connotations.

The ultrafemininity of the belle is also constructed as the binary opposite of the manly man. If the Depression can be read as a crisis in northern white masculinity, it is also reflected in the crisis in white masculine identities that occurred in American entertainment at the turn of the 1930s. While as a trope the flapper constituted a blurring of gender attributes, both visual and behavioral, from one direction (flattened chest, short hair, sexually predatory, substance using, eager to move into the employment market), her counterpart the sheik blurred in the other (slim-hipped, sensual, emotional, also substance using, financially dependent on women). The urban “pansy craze” of the 1920s and the concurrent “rise of the crooners” made the bandstand a prominent arena for the contesting and fashioning of gendered identities.¹³⁰

In her article on 1930s male singers, Allison McCracken constructs a strong narrative tracing the development of Bing Crosby's masculinized crooning and a correspondingly manly visual persona as a response to public anxieties created by Rudy Vallee's more effeminate delivery. The Boswells-as-belles contributed to Crosby's masculinization by providing him with a feminine counterpart, frequently appearing on his radio shows—a role that Connie eventually performed on her own after the sisters disbanded.¹³¹ It is worth noting that the Mills Brothers performed a similar role with respect to Crosby's racialization as the model (white) American male; the two groups (Sisters and Brothers) alternated in the same supporting slot on Crosby's Woodbury Hour on CBS during 1934.¹³²

For the belle, “access to the public sphere” obtained through the deployment of hyperfemininity could include involvement in more masculine pursuits such as business.¹³³

The Boswells were not the first or only women for whom the performance of femininity worked to open doors (and to keep them open); as Sherrie Tucker notes, “Women's jazz history contains many examples of how women have combined or overshadowed their musical expertise with skilled performances of gender in order to make themselves palatable to the public.”¹³⁴ The Boswells, however, differed from many women in jazz, and particularly those whom Tucker has studied, in that they gave instructions to male musicians with whom they worked, supplying them with arrangements and musical direction rather than accepting direction (as singers) from a male bandleader or working within a strictly female environment.¹³⁵ Their performance of femininity needed not just to reassure their audiences, but also to ease relations with a musical fraternity that could at times be quite hostile to women.¹³⁶

Tara McPherson's analysis shows how the belle can be deployed in both overt and covert representations of race. Overtly, she is positioned in the frame with a black counterpart, as with Gone With the Wind's Scarlett and Mammy; covertly, she is presented without reference to black bodies, but since her very existence is predicated on the social and economic conditions of her upbringing, she can operate on her own as a paradigm for whiteness that simultaneously presents and then takes ownership of specific (opposing) criteria for blackness. For example, the privileged status accorded by her whiteness to Magnolia, the white heroine of Show Boat, is enacted twice in the plot: first, she assumes Julie's star position in the showboat troupe when Julie's passing is exposed; second, Julie gives up her job as a nightclub chanteuse to Magnolia, so that Magnolia can provide for her daughter (and so competently fulfill her role as mother). In the same way that Magnolia absorbs the role of the black woman, effectively rendering invisible the blackness of her position and her performance, the Boswells absorb the blackness of “When It's Sleepy Time Down South,” recasting it through the subjectivity of southern white womanhood.

But perhaps the most startling aspect of the Boswells' southernness is how it intersects with their emulation of black male jazz musicians, particularly Louis Armstrong. Not only do the Boswells specifically critique his performances in their own, as in “Heebie Jeebies” and “When It's Sleepy Time Down South,” but throughout their, and later her own, recording career Connie also frequently performed “trumpet” solos in their arrangements (produced vocally à la Harry Mills), mirroring Armstrong's dual role as vocal and instrumental soloist. Certainly it could be argued that, as the black male popular singer was still a liminal figure in American entertainment, the Boswells' references to Armstrong helped fix his identity in the public's imagination. Yet as southern white women, their apparently easy and persistent deployment/manipulation of the same elements that made Armstrong into such a potent force in early jazz, and indeed marked him for who he was (for example, his accent; his irreverence; his “bilingualism” as vocalist and instrumentalist; his eclectic repertoire; the naturalized understanding of the polyphonic textures of New Orleans jazz), possibly also worked to destabilize the black jazzman's image—and performance—of masculine virility that, in turn, so threatened the already beleaguered white masculinity of the Depression years.¹³⁷

The year after the stock market crash was a turning point for American entertainment. Throughout the decade of the 1920s, new technologies had made a new form of blackness—black vocality—increasingly available and recognizable to the public, for whom a frisson of impropriety, racial or otherwise, could satisfy an appetite for titillation. “Passing” performers may have risked offending some audiences with their transgressions, but they also knew they delighted others. Yet the frisson turned to a shiver in the colder social climate of the Great Depression, and the balance of audience desire tipped from titillation to reassurance, so that such anxieties, while still introduced, were increasingly to be assuaged (or eventually avoided, thanks to the Hays Code). The Boswells' swift rise to stardom coincided almost exactly with the nation's equally swift plummet into economic catastrophe, and one could argue that, bizarrely, the timing could not have been more fortuitous for the sisters. The overtly affected racial and gendered crossings that were so popular during the 1920s—Valentino's mysterious, made-up sheik; the loose, rhythmic shuffle of Clara Bow's flapper—were also symbolic of moneyed, northern, urban decadence, and as the Depression loomed they lost their cultural currency as rapidly as the Federal Reserve Bank lost its own. Although the Boswells also disrupted racial and gender etiquette, the mask of the southern belle allowed them to restore equilibrium with gracious decorum; they were seen as natural and authentic, their cultural miscegenation effectively rendered invisible by their southern charms. As good southern Democrats, moreover, they were ideal, New Deal sweethearts: fresh models of aspiration for a wholly different economic and social climate.¹³⁸

When the Boswells Sisters disbanded as a performing unit in early 1936, the reason given was that Vet and Martha, both recently wed, were leaving show business to concentrate on marriage and family; Vet, although it was not revealed at the time, was also pregnant. Connie, too, had recently married, to their manager, but after the group's demise she concentrated on developing her solo career. Her sisters' sudden departure from the entertainment stage serves to underline the nature of the bargain they had struck with their audience: their performance of a specific kind of white femininity legitimized their access to both public life and blackness (and hence to a musical career and to jazz), but to remain respectable eventually they would have to fulfill their side of the exchange.¹³⁹

With the ultimate assumption of respectability came silence and then obscurity; having willingly ceased performing, Martha and Vet withdrew from the limelight as creative artists and were refashioned instead as wives and mothers.¹⁴⁰ Connie remained musically active for many years to come, primarily on the radio, headlining the March of Dimes annual gala and even hosting her own show on ABC in the 1940s. She also enjoyed significant recording success, on her own and with both Bing and Bob Crosby, and her live cabaret performances continued into the 1960s. She continued to voice southernness and to play the belle, although she also used the role's masquerading function in a new way, effectively to render invisible another marker of difference, her disability.¹⁴¹ The longevity of her career perhaps ensured that familiarity eventually erased any notion of “difference” in her vocal persona. Yet through her influence, and even more through the influence of those who imitated her—Crosby, Fitzgerald—those salient qualities that had once marked her voice as “black”—her loose, drawling diction; a more speechlike, naturally emotive delivery; and a lower, mellower tone and tessitura—became fundamental to the twentieth-century American songbook style.¹⁴²