Tearing Speech To Pieces - Smith
Tearing Speech To Pieces - Smith
Tearing Speech To Pieces - Smith
V
-DFRE6PLWK
Music, Sound, and the Moving Image, Volume 2, Issue 2, Autumn 2008,
pp. 183-206 (Article)
3XEOLVKHGE\/LYHUSRRO8QLYHUVLW\3UHVV
Access provided by School of the Art Institute of Chicago (27 Apr 2015 15:41 GMT)
MSMI 2:2 Autumn 08 183
JACOB SMITH
Thanks to Jennifer
Fleeger for organising The Voder and the Sonovox were voice technologies that had a signif-
the SCMS panel on icant impact on radio and film production during the 1940s. In this
which this research
was first presented, essay, I examine the cultural life of these mid-century talking
Alexander Russo, Jay machines, with a focus on the use of the Sonovox in radio advertising
Beck, Patrick Feaster,
Matthew Solomon and
and the Hollywood ‘women’s films’ Possessed (1947) and Letter to Three
the students and staff Wives (1949). An examination of these devices, the discourses that
at the College of Staten surrounded them, and their use in media texts can add to our under-
Island who attended
my talk, and Barbara standing of the cultural history of sound recording and vocal perform-
Klinger, a fantastic ance, the use of sound in film melodrama, and overlaps between radio
teacher who stimulated
my thinking on many and film sound production in the 1940s. Because of its ability to make
topics, melodrama recorded sounds take on the qualities of human speech, the Sonovox
being one.
found success in the relatively new area of radio spot advertising,
where it could make consumer products ‘speak’ to a largely female
audience. In Hollywood films, the Sonovox made objects speak the
unconscious thoughts and anxieties of female characters. Operated by
women, with the goal of speaking to a female audience, gender was
thus a determining factor in shaping the cultural life of voice tech-
nologies of the 1940s. Acknowledging the female operators behind
these voice technologies helps us to understand some of the factors
that shaped their use in media production, and also represents a small
gesture towards redressing the larger erasure of female labour in the
cultural industries.
Visitors entering the Bell Telephone exhibit at the 1939 World’s Fair in
New York encountered a striking demonstration of a new electronic
device. Once inside a large, streamlined building, fair-goers were greeted
by a face outlined in gold on the wall. Nearby, a man on a raised platform
invited questions from the crowd, and a young woman worked an
elaborate keyboard to produce an electronic simulation of human speech
(New York Times, 1939a: 20 – see figure 1). Audiences were fascinated by
the talking machine known as the Voder, as well as by the young female
184 2:2 Autumn 08 MSMI
Jacob Smith ♦ Tearing Speech to Pieces
devices, the discourses that surrounded them, and their use in media
texts can add to our understanding of the cultural history of sound
recording and vocal performance, the use of sound in film melodrama,
and the overlaps between radio and film sound production in the 1940s.
My analysis is informed by work on the social construction of sound
media technology. Rick Altman has outlined a ‘crisis historiography’ of
new media technologies that takes into account the multiple and
conflicting definitions of a new technology, and considers ‘unsuccessful
experiments and short-lived practices’ as well as those that become estab-
lished (2004: 22–3). The use of the Sonovox was a relatively ‘short-lived
practice’ in Hollywood film and American radio broadcasting, but one
that reveals a previously unexamined constellation of practices and
discourses in media production of the 1940s. Altman describes how new
media technologies tend to be recognised initially as several ‘different
phenomena, each overlapping with an already existing medium.’ Such
ambiguity leads to ‘conflicting definitions’ of the artifact, as various social
forces struggle to gain control over it. Eventually, new technologies take
on stable features as the result of culturally overdetermined factors (ibid:
1 Altman builds upon 19–21).1 Altman’s framework provides a structure for this essay, since
Trevor J. Pinch and after an examination of the multiple definitions and suggested applica-
Wiebe E. Bijker, who
write that the history of tions of mid-century talking machines, I will describe how several of their
technology has tended dominant uses in radio and film production were determined by cultural
to feature an ‘asymmetric
focus’ on successful discourses concerning gender, genre, and consumer culture.
innovations, leading Though I describe both the Voder and the Sonovox as ‘mid-century
scholars to ‘assume that
the success of an artifact talking machines’, the two devices represented different methods of
is an explanation of its speech simulation: the Voder simulated speech through the actions of an
subsequent develop-
ment’ (1989: 22).
operator’s hands and feet, whereas the Sonovox allowed recorded sounds
to take on the qualities of human speech by reproducing them through
the throat and mouth of a human ‘articulator’. The Sonovox’s ability to
shape the sounds made by inanimate objects into words allowed it to find
a niche in the relatively new area of radio spot advertising, where it made
consumer products ‘speak’ to what was thought to be a primarily female
audience. The history of the Sonovox thus reveals a little-known function
of sound technology in both radio advertising practice and the represen-
tation of the commodity. At the same time as these developments in broad-
casting, the Sonovox made objects speak the unconscious thoughts and
anxieties of female characters in Hollywood films such as Possessed and
Letter to Three Wives. Operated by women, with the goal of speaking to a
female audience, gender was thus a determining factor in shaping the
cultural life of voice technologies of the 1940s. The importance of women
in the production and reception of these machines was symptomatic of the
particular historical context of 1940s America, but the devices themselves
186 2:2 Autumn 08 MSMI
Jacob Smith ♦ Tearing Speech to Pieces
I
The Voder exhibit at the 1939 World’s Fair recalls earlier demonstrations
of talking machines such as Thomas Edison’s phonograph, Wolfgang von
Kempelen’s eighteenth-century speaking automaton, and Joseph Faber’s
nineteenth-century ‘Euphonia’. Faber’s device, which toured Europe and
America in the mid-1800s, worked by pumping air through a bellows
modelled on the lungs and throat that, through deft manipulation by the
operator, could thereby produce speech (figure 3). The machine used a
keyboard connected to an android
Figure 3 The ‘Euphonia’ figure variously dressed as a
woman or a ‘Turk’ (Feaster 2006:
59), simulating the form and
function of the human body, an
approach that James Lastra identi-
fies as one of the ‘master tropes
through which the nineteenth
century sought to come to grips
with the newness of technologi-
cally mediated sensory experi-
ence’ (2000: 21). Both Lastra and
Jonathan Sterne distinguish
between talking machines like Faber’s that were modelled on the lungs
and throat, and other devices modeled on the ear: a shift in design from
the simulation of ‘human sound production’ to that of the ‘human
perception’ of sound (ibid: 44). Leon Scott’s ‘Phonautograph’, Alexander
Graham Bell’s ‘Ear Phonautograph’ and Thomas Edison’s phonograph
were sound transducers that simulated what Sterne has called the
‘tympanic function’ of the ear (2003: 81).
Scholars have tended to end the story of talking machines with either
Edison’s phonograph or the cinematic soundtrack, as though they were
the goals towards which all of these devices led. The talking machines of
the 1940s reveal that research, innovation and public interest continued
in this area, albeit with variations on the models of bodily simulation
described above. One day in the late 1930s Gilbert H. Wright was shaving
with an electric razor when he noticed that ‘queer sounds came out of his
mouth’ when the razor passed over his Adam’s apple. He soon discovered
MSMI 2:2 Autumn 08 187
Jacob Smith ♦ Tearing Speech to Pieces
that by silently articulating words with his mouth and lips, the sound of
the razor was formed into speech. From this initial observation, Wright
developed a device that he dubbed the Sonovox, whereby a sound
recording was fed into two hand-held speakers that would be placed on
each side of the throat. Whatever sounds were on the recording were
transmitted to the larynx, so that they came out of the throat as if
produced there, and could then be shaped into speech by articulating the
desired words. Sounds could thus be made to speak, or as a 1939 Time
magazine article put it: ‘a grunting pig, relayed through the human
voice-box, can be made to observe: “It’s a wise pig who knows his own
fodder” ’ (1939: 45). In terms of the models of simulation outlined above,
the Sonovox is a hybrid: the phonograph record is based on sound trans-
duction that simulates the ear, while the human ‘enunciator’ becomes a
surrogate throat along the lines of Faber’s ‘Euphonia’.
At the same time that Wright’s Sonovox was the subject of national
press coverage, Homer Dudley, a technician at Bell Labs, was patenting
the Voder, and later a device called the Vocoder that manipulated speech
fed into a microphone. Dudley’s machines did not simulate the body, but,
as his patents stated, produced speech artificially, ‘by purely manual
operation independent of any of the parts of the human body which are
normally used in the production of vocal sound’ (Dudley 1938: 1). In
fact, the Voder and Vocoder are pioneering examples of what Remko
Scha has called gennematic speaking machines. He distinguishes
between two approaches to the technology of speech simulation. The first
is the ‘genetic method’, wherein ‘the hardware of the larynx and the oral
cavity [is] reconstructed in a stylised way’ in order to imitate the ‘physio-
logical processes that generate speech sounds in the human body’: a
category that corresponds to the simulation of human speech production
described above. By contrast, Scha’s gennematic method is ‘based on the
analysis of the speech sounds themselves, and reconstructs these sounds
without considering the way in which the human body produces them’.
A gennematic speaking machine does not imitate the mechanics of the
voice, but fakes its output: ‘The algorithm computes signals that evoke
the image of a physical process that never occurred.’ The ‘genetic’ eigh-
teenth-century automaton was ‘a mechanical body’, Scha writes, ‘a piece
of clockwork claiming the qualities of life.’ With twentieth-century
‘gennematic’ voice simulation, the ‘mechanics is abstract, the machine
dissolves into mathematics. The body has disappeared’ (1992: 2).
Edison’s phonograph was still a simulation of the human body: the
tympanic function of the ear. The Sonovox’s novelty was to combine the
phonograph with a surrogate throat. The Voder and Vocoder however,
had as their raw materials the electrons in a vacuum tube: the body, we
188 2:2 Autumn 08 MSMI
Jacob Smith ♦ Tearing Speech to Pieces
Though the Sonovox and the Bell Labs devices differed in their
design, they were meant to perform a similar function: to make objects
speak. In so doing, they blurred categories of speech and sound, animate
and inanimate, and amazed audiences at their early public demonstra-
tions. Following a 1940 display of the Vocoder given by Dudley at the
Franklin Institute in Philadelphia, the Hopewell Herald described how the
audience was enjoying what they thought was a pipe organ playing ‘The
Bells of St. Mary’s’, when they suddenly ‘sat bolt upright’: ‘Some dug at
their ears, certain their hearing was playing them false. Others sat in
puzzled wonder. For the pipe organ suddenly burst out singing the
words of the chorus.’ After making a musical instrument speak, Dudley
next animated the sounds of industry: ‘While the audience was still
gasping, a recording of a locomotive just chugging out of the station
obediently puffed out the words: “I’m – a – train. Watch – me – now.
Here – I – go. Here – I – go. Here – I – go!” An electric power hum
followed, chilling the audience with the high-pitched monotone “I am
power. I light your houses. I run your street cars. I work for you. I am
power!”’ (1940: 3). The rhetorical force of the demonstration had to do
with how the Vocoder could jumble categories of music, sound, and
speech. Such an ability could startle crowds, but what practical applica-
tions could be found for it in the early 1940s? Dudley and Wright had to
find marketable applications for their devices, and proposals put forward
at public demonstrations, in patent applications, and in the popular
press, illustrate the ways in which these talking machines were under-
stood. Though the body had disappeared from the technology of speech
simulation, we shall see that the body reappeared in the practices and
discourses that surrounded it.
II
George Basalla has described the importance of military imperatives in
the selection of technological innovation (1988: 158). Notably, several
military initiatives were prominent in the early career of Wright and
Dudley’s talking machines. Research on the Vocoder was partly driven by
its ability to send speech under ‘war-time conditions’ in a coded form that
would be undecipherable (O’Neill 1942: L6). The Albuquerque Journal
described in 1940 how the Sonovox was being used at a Naval Air School
to teach Morse code by articulating the signals into the letters they repre-
sented (1940: 20). Indeed, Wright was granted a patent in 1942 for using
the Sonovox as a means of teaching telegraphy (Wright 1942c). His
device was also used for more straightforward wartime propaganda.
190 2:2 Autumn 08 MSMI
Jacob Smith ♦ Tearing Speech to Pieces
maker did, in fact, take an interest in the Sonovox: Walt Disney licensed
Wright’s device and used it in both Dumbo (1941) and The Reluctant
Dragon (1941).
As we have seen, there were multiple, and sometimes conflicting defi-
nitions of these new speaking technologies: would they be special effects
for cartoons; tools for the military; devices for translation; supplements
for popular singers? In the end, two other suggested applications became
central to the implementation of the Sonovox in the media industries,
and reveal how discourses of gender shaped these voice technologies.
Consider a scene from Walt Disney’s The Reluctant Dragon (1941), in
which we are shown the Sonovox in action. This quirky film follows the
hapless Robert Benchley on a backstage tour of the Disney Studios.
Benchley stumbles into a sound recording session, and is shocked to hear
a voice emanating from a toy cow. ‘Don’t be afraid, I’m only the train
whistle’, says a young woman sitting nearby. ‘Well, they’ve certainly
improved the looks of them since I was a boy’, Benchley replies. The
woman goes on to explain how the Sonovox works, and demonstrates by
articulating the recorded sound of a train whistle.
Contemporary press accounts described how the Sonovox was
typically operated by female ‘enunciators’ or ‘articulators’: ‘usually a girl
who has had acting experience’ (Yoder 1943: 39). A similar configuration
of female operator and speaking machine was witnessed at the 1939
World’s Fair. This figure of the female enunciator can help us to under-
stand why and how talking machines came to be used as they did.
Indeed, notions of gender played an important role in shaping how the
Sonovox was used, not only in animated cartoons, but in radio adver-
tising and film melodramas, where it did not alter or improve an existing
voice, but made objects speak to what was assumed to be a female
audience.
III
Like many ‘talking machines’ that came before it, such as Faber’s
‘Euphonia’, the Voder was personified. Erika Brady has described a
similar tendency to personify Edison’s phonograph through such means
as framing demonstration recordings in the first-person (‘I am the
Edison Phonograph’). Brady places the phonograph in a larger history
of mechanical speaking automata, virtually all of which, she argues, were
‘given a human form, as though to re-embody the sound, even when that
form was not dictated by technical requirements’ (1999: 33–4). Dudley’s
invention was similarly personified, in part by being given the nickname
MSMI 2:2 Autumn 08 193
Jacob Smith ♦ Tearing Speech to Pieces
‘Pedro’, after the Emperor of Brazil, Dom Pedro. The Emperor was said
to have been shocked by the newly invented telephone at the 1876
Philadelphia centennial exposition, and in his amazement had said: ‘My
God! It talks’ (Nebraska State Journal 1939: 7). Though the Voder
apparatus was gendered male, Pedro’s operators were women. Press
coverage of the World’s Fair exhibit described the skills of ‘dextrous
operators’ like Mrs. Helen Harper, one of twenty-four telephone
operators chosen from more than 300 applicants to operate the Voder at
the World’s Fair. A photograph in the Christian Science Monitor showed a
group of female telephone operators waving to the camera, with the
caption, ‘these New York telephone girls are on their way to San
Francisco on a novel mission. They have been trained by the New York
Telephone Company to operate ‘Pedro the Voder’, the first device by
which speech can actually be created’ (1939: 4). As this caption indicates,
one factor in the gendering of talking machine operators lay in their
connection to the traditionally female occupation of the telephone
5 Venus Green describes operator.5
how young boys were Working the keys of the Voder was also reminiscent of an activity long
first hired as telephone
operators, but proved associated with women in the domestic sphere: playing the piano. One
to be unsuitable because article described the device in detail, and concluded with a musical
of ‘temperamental
defects’. To counter the analogy: ‘there are ten white keys which when pressed turn on vowel
‘negative effect on sounds. One other white key controls volume. Alongside the right hand
business’ from unruly
boy operators, telephone are three black keys which make the consonants, ‘k’, ‘p’, and ‘t’. Under
management hired the left wrist is a key that changes the sound from consonants to vowels.
women after the 1880s
in the belief that ‘girls,
The right foot presses a treadle for vocal inflections… the operator
socialized to defer on modifies the sounds by pressing more than one key at a time, like the
the basis of class, gender, player of an organ’ (Davies 1939:1). Given this description, it is not
and age, were best
qualified to give the kind surprising that the difficult and complex task of operating the Voder
of service Bell envi- took a year or more to learn, and was reserved for an elite team of profes-
sioned’ (2001: 55–7).
sional telephone workers. Nor is it surprising, given the gendered asso-
ciations with telephone switchboard workers and piano playing, that the
operators of the Voder and Sonovox would be women. Scha has argued
that the body disappeared with the kind of gennematic speaking
machines emerging in the 1940s; with the Voder, however, we see how an
explicitly gendered body could reappear in the discourses and practices
that surrounded the machines.
Though the operation of these devices became associated with women,
the gender of the Voder’s voice could be changed by flipping a switch: ‘A
knob turns to left for a male voice, right for women’ (Nebraska State
Journal, 1939: 7). Dudley described how the Voder could produce a
‘hybrid voice effect’ by changing the machine’s output ‘from that of a
man’s voice to that of a woman’s voice or vice versa’ (1941: 5). In fact,
194 2:2 Autumn 08 MSMI
Jacob Smith ♦ Tearing Speech to Pieces
The vocoder took normal speech and, slicing the pitch in half, made the
voice that of a toneless robot; then doubling the normal pitch made the
voice take on the exaggerated inflections of a ham actor’s speech.
Quadrupling the pitch range made the voice a sissy’s affected talk…
Popeye’s spinach-gobbling basso profundo was a mouse’s shrill squeak
compared to the vocoder’s rendition of Barnacle Bill the Sailor with the
pitch at low; and even a coloratura soprano canary would have burst its
vocal plumbing trying to match the 1,000 vibrations per second rendition
by the vocoder of ‘Happy Birthday to You’.
(Hopewell Herald, 1940: 3)
Reading this description, one might wonder whether this was a demon-
stration of sound technology or cultural stereotypes about the gendered
voice. Note the violence of the language in the phrase ‘slicing the pitch
in half ’: a motif that can be found in much of the press coverage of these
demonstrations. The Vocoder, one author noted, was a machine ‘that
tears speech to pieces and remakes it in new patterns’ (New York Times,
1939: 24). Another described how Dudley had trapped the human voice
in a ‘mysterious maze’ such that ‘it never had a chance’ (Williams 1940:
G8). The gennematic voice created by the Voder and Vocoder was felt to
do violence to the unity of the voice, and the disturbances of voice and
identity were often framed in terms of gender.
The use of female operators seems to have amplified the effect of the
carnivalesque gender reversals that the Voder, Vocoder, and Sonovox
could perform, as Benchley’s first reaction upon encountering the female
enunciator in The Reluctant Dragon confirms. The combination of the low,
gutteral train whistle and a conventionally attractive and demure female
operator became the source of surprise and fascination.6 This is not the 6 In this regard, see Kay
place for an exhaustive history of female vocal etiquette, but suffice it to Dickinson’s discussion
of the recent appropri-
say that since at least the turn of the twentieth century, American women ation of the Vocoder by
were encouraged to consider their voices as a potential problem, and female pop artists (2001).
urged to speak with a low tone, free from rasp or nasality.7 The sounds 7 See, for example, the
article ‘The Voices of
created by the Sonovox provided a surprising contrast to the female American Women’ in
bodies that operated them, as demonstrated in press copy such as this: the Washington Post, 9
‘The mouth from which these tough and angry tones emerged was the January 1905; and
Donnelly, 1932.
dainty kisser of Miss Adelaide Gerwig, a radio actress who had become
one of the first Sonovoxeuses. Miss Gerwig’s own voice is said to be soft
and pleasing’ (Yoder 1943). Janet Eberhardt, ‘one of only four Sonovox
“articulators” in the world’ was described as ‘a pretty, dainty, and diminu-
MSMI 2:2 Autumn 08 195
Jacob Smith ♦ Tearing Speech to Pieces
tive young lady’. When she used the Sonovox on the radio, however, she
opened ‘her lovely mouth’ and bellowed a ‘familiar foghorn wail’
(Johnson 1947).
That ‘familiar foghorn’ formed the sonic hook of a radio advertising
campaign for Lifebuoy soap. Indeed, the Sonovox found one of its most
lucrative implementations in the field of radio advertising. We have seen
how notions of gender shaped the operation of the Sonovox, and how
gender inversions made for a particularly stunning demonstration of the
voice. Gender was also key to the reception of the Sonovox effect, which
was used as a technique to address the largely female radio audience.
IV
Wright’s Sonovox made its broadcasting debut in September 1941, a time
when the radio industry was debating the issue of spot advertisements.
Alexander Russo has described the emergence of short, spot advertising
in the early 1930s, arguing that by the end of the decade, spot announce-
ments had overtaken more integrated programme advertising in terms
of ‘prevalence and sales volume’ (n.d.: 194). Kathy Newman writes that,
in comparison to the print advertising of the same period, radio spots
had less space for ‘the invocation of extra-product symbolism’ and so had
to be less metaphoric and more direct (2004: 29). Advertisers had
learned from market research that the repetition provided by musical
jingles was an important form of brand identification (ibid: 41–3). The
Sonovox was thought to provide some of the same sonic attributes and
benefits as the jingle, and Wright’s brother-in-law, James L. Free, who
worked in radio advertising, actively promoted the device to the industry
(Yoder 1943). The Sonovox’s distinction in the field of radio advertising,
one article stated, was that it could ‘make a vacuum cleaner talk’
(Newsweek, 1942: 60). In 1941, the Christian Science Monitor warned its
readers not to be surprised ‘if a tub of suds’ made a ‘fervent appeal at the
end of your favorite radio “soap opera”.’ With the Sonovox, the article
continued, any quality of the sponsor’s product could be made to
advertise itself (1941: 9).
On radio ads of this time, Wright’s invention sometimes made musical
instruments speak commercial messages: in a Shell ad, an organ said,
‘Stop at the sign of the Shell’; and a Novachord playing the tune of ‘Good
Night, Ladies’, was made to say, ‘V-E-L, my hands feel so soft and smooth
with Vel. Vel swell, ladies’ in a Colgate ad (Newsweek, 1942). The Sonovox
also made objects speak: pots and pans ‘sang that they just loved to be
washed’ in an advertisement for dish soap; a chugging locomotive
196 2:2 Autumn 08 MSMI
Jacob Smith ♦ Tearing Speech to Pieces
became the words, ‘Bromo Seltzer’; and a car horn was made to say
‘Better Buy Buick’ (ibid.; Yoder 1943). A 1942 radio spot in the Lifebuoy
soap campaign mentioned above begins with two long blasts of a
foghorn. We hear a male announcer state, ‘There’s nothing so dismal as
a foghorn’, after which the foghorn returns, this time made by the
Sonovox to say ‘B.O’. ‘Unless it’s somebody with B.O.,’ the narrator
continues, ‘that’s really dismal.’ A second narrator chimes in with, ‘Yes,
and the most dismal thing about B.O. is, nobody tells you about it. You
think it’s you they don’t like, when really it’s B.O.’ The trademark
foghorn, again made to say ‘B.O’. by the Sonovox, is reprised at the end
of the ad.
Compare the way in which the Sonovox effect was used in radio adver-
tisements to the simple, abstract techniques developed for outdoor
advertising in the 1930s. Catherine Gudis describes an ‘aesthetics of
speed’ that dominated billboard design at that time: as motoring speeds
increased, more ‘simplified, abstract, and streamlined means of repre-
senting and selling’ were sought, leading to an increased concern with
achieving simplicity in approach and speed of communication. In that
pursuit, outdoor advertisers used modernist forms and refined their use
of trademarks, slogans, and images that allowed for a ‘quick impression’
(2004: 66–8). Similarly, short radio spot ads had to make their sale
quickly and so were constrained by a sonic ‘aesthetics of speed’. In this
light, Sonovox ads might be understood as aural analogues to the stream-
lined logos and stylised lettering found on billboards of this era.
Sonovox spot ads spurred the imagination of writers and advertisers
who described a world of the not-so-distant future in which brand-name
goods would speak to consumers at every turn: one article gushed that a
day would come ‘when sound will talk and sing on every side – not
merely in the movies and on the air but all over the place, with bus horns
proclaiming the name of the bus company, delivery trucks calling out
their wares as they honk, and train whistles announcing the name of the
train. Once it would have been regarded as exotic if milk trucks could
moo. Thanks to Sonovox, they can now moo the name of the dairy’
(Yoder 1943). Such fantasies may not have materialised, but Sonovox
radio spots were certainly precursors to well-known animated television
ads of the 1950s that depicted square-dancing Lucky Strike cigarettes,
marching Rheingold beer bottles, and singing Muriel cigars. In his
famous description of the commodity fetish, Karl Marx offered the image
of a table that, as soon as it becomes a commodity, changes into a thing
that ‘stands on its head, and evolves out of its brain grotesque ideas, far
more wonderful than if it were to begin dancing of its own free will’ (2000
[1867]: 331). The Sonovox seemed to provide commodities the chance
MSMI 2:2 Autumn 08 197
Jacob Smith ♦ Tearing Speech to Pieces
not only to come to life, but to articulate some of those ‘grotesque ideas’
directly to consumers.
We should note that at this time, the majority of radio listeners and
8 See Hilmes 1997 and consumers of household goods were assumed to be women.8 The
1999; as well as
McCracken 2002: 186;
Sonovox was a new technology mobilised by advertisers to encourage
May 1988: 167; female audiences to go on what Gary Cross has called the postwar
Newman 2004; and consumer splurge. Michele Hilmes notes the paradox that, while women
Cohen 2003: 278.
were ‘the primary audience of broadcasting’s commercial address’, their
‘agency as active “speakers” and producers in the medium’ was ‘severely
and deliberately circumscribed’ (1997: 17). The female labour behind the
Sonovox both indexes women’s limited agency as radio speakers, and
points to a notable, though certainly limited, exception to that rule.
Female Sonovox enunciators and female radio listeners were at different
points in a circuit of communication that involved voice technology,
network broadcasting, and consumer culture. The film industry was
eager to create its own circuit of communication with female audiences in
the 1940s, and the Sonovox found a place in that process as well.
V
The 1930s and 40s were a high point in the production of the ‘woman’s
film’: a genre of Hollywood films that featured a female protagonist,
treated issues typically defined as ‘female’, and were directed toward a
female audience (Doane 1987: 3). Britta Sjogren argues that the film
industry’s desire to address the ‘subjective experience’ of female cinema-
goers in the form of the 1940s woman’s film was connected to ‘the World
War II home-front audience’s new spending power’: ‘Targeted to this
newly dominant, independent female market’, women’s films articulated
and responded to a ‘socio-historical context within which the “voice” and
the increased discursive visibility of women was a pressing issue’ (2006:
4). As we have seen, the Sonovox was a point of articulation between
women, voice technology and radio broadcasting during this era. Janet
Eberhardt – the Sonovox ‘articulator’ who worked on the Lifebuoy radio
ads – also worked in Hollywood, where she was said to have been a train,
a carousel, an automobile horn, an airplane, train wheels, a paddlewheel
on a riverboat, an anaesthetiser, the wind, and horses’ hooves (Johnson
1947: 4). One of Eberhardt’s uncredited appearances was in the Joan
Crawford star vehicle Possessed, an examination of which reveals a further
alignment between women, voice technology and Hollywood film.
Possessed tells the story of Louise (Joan Crawford), a woman with an
obsessive love for David Sutton (Van Heflin), but who marries her
198 2:2 Autumn 08 MSMI
Jacob Smith ♦ Tearing Speech to Pieces
employer, Mr. Graham (Raymond Massey), after his wife dies under her
care. Returning to the Grahams’ home, Louise – who blames herself for
Mrs. Graham’s death – hears a buzzer coming from the dead woman’s
bedroom. The Sonovox, presumably operated by Janet Eberhardt, makes
the buzzer speak her name: ‘Louise’. That sonic effect is paired with a
shaky hand-held camera, providing Louise’s point of view (figure 6).
the female voice seems to ‘speak’ directly to women ‘from’ a female subject.
A female character vocalises her thoughts or feelings to an audience of
(assumed) women, and a ‘hailing’ of the spectator is underscored in a way
that rarely, if ever, sustains itself on the image track.
(2006: 9)
The Sonovox took part in that hailing as a second order, subjective voice-
off. Following other film scholars such as Michel Chion, Sjogren empha-
sises the ‘mutability of the voice-off ’; the way it ‘slips free of the image,
glides in and out of its attachments to its apparent body’ (ibid). We might
say that the Sonovox both proliferated and grounded that mutability, by
allowing the voice to glide into more points in the image, and to anchor
the voice to objects.
Take, for example, the way in which the Sonovox was featured in a film
that is well known for its use of a female voice-off: Letter to Three Wives,
written and directed by Joseph L. Mankiewicz. The film’s narrative
concerns Addie Ross, a character that we never see, but whose voice we
hear. Ross leaves town, sending a letter to three friends saying she has run
away with one of their husbands. The letter prompts each woman to
reconsider her marriage, which the film presents through flashback
sequences that are each initiated by a sound bridge utilising the Sonovox.
Deborah Bishop (Jeanne Crain) hears Addie Ross’s taunting voice
emerging from the rhythmical sound of a boat engine; Lora Mae
Hollingsway (Linda Darnell) hears Addie’s voice speak through the sound
of a dripping water pipe (figures 8 and 9). As in Possessed, the Sonovox
200 2:2 Autumn 08 MSMI
Jacob Smith ♦ Tearing Speech to Pieces
Eddie!’ (figures 9 and 10). Despite the denigration of radio serials in the
film, this soundtrack embodied an important continuity with commercial
broadcasting: the Sonovox was part of the crass commercial clutter of
radio spot ads as well as being the signature sonic trope of the film.
VI
Bringing to light the relatively brief career of the Sonovox in radio and
film helps to situate the media texts that employed it within a larger
context of media practice: as Jonathan Sterne reminds us, media tech-
nologies must be understood as involving ‘a whole set of related institu-
tions, technologies, people and practices’ (2003: 182). The application of
Rick Altman’s ‘crisis historiography’ has helped us to better understand
the social construction of voice technologies of the 1940s, and appreciate
how their predominant functions in radio and film production were
shaped by cultural discourses of gender, genre and consumer culture.
The history of the Sonovox and the Voder reveals that research and
development in voice technologies continued after the emergence of
modern sound media such as the phonograph, radio and sound film.
Indeed, the 1940s were a time when military demands, incipient broad-
casting practices, and an industry focus on female audiences converged
to encourage the development of new voice technologies. When placed
in the larger history of talking machines, this was also a moment when
the overt simulation of the body was disappearing from the devices them-
selves, but returning in the gendered discourses and practices that
surrounded them. Acknowledging the female operators behind these
voice technologies helps us to understand some of the factors that shaped
their eventual use in media production, and also represents a small
gesture towards redressing the larger erasure of female labour in the
cultural industries at a time when, as Hilmes notes, women’s voices were
‘literally contained and controlled’ (1999: 20).
An AT&T US television advertisement from 2006 can provide a final
illustration of the erasure of female labour behind these mid-century
9 At the time of writing, talking machines.9 The ad begins as fluorescent lights illuminate a labo-
available for view at
http://www.youtube.co
ratory, accompanied by narration describing the history of AT&T’s
m/watch?v=IEp6ca9Pp research into voice technology, beginning with the Voder. That narration
ks is spoken by the Voder itself, but soon switches to Robbie the Robot from
the film Forbidden Planet (Fred M. Wilcox, 1956), then to Rosie the Maid
from The Jetsons (Hanna-Barbera, 1962–3), then to ‘WOPR’ from the film
War Games (John Badham, 1983), then to ‘K.I.T.T.’, the car from the NBC
television show Knight Rider (1982–6), and finally to the female voice of a
car satellite navigation system. The ad represents a strange mix of actual
electronic talking machines and human actors playing robots in film and
television: an apt illustration of the long-standing tendency to personify
talking machines. Most notable from the perspective of this essay
however, is the fact that the hands that we see operating the Voder’s keys
are those of an automaton. The female practitioners whose labour
204 2:2 Autumn 08 MSMI
Jacob Smith ♦ Tearing Speech to Pieces
♦ ♦ ♦ ♦ ♦
References
Albuquerque Journal (1940) ‘Science Speaks Up’, December 22, p. 20
Altman, Rick (1994) ‘Deep-Focus Sound: Citizen Kane and the Radio Aesthetic’,
in Quarterly Review of Film and Video, Vol. 15 (3), pp. 1–33
— (2004) Silent Film Sound, New York: Columbia University Press
Basalla, George (1988) The Evolution of Technology, Cambridge: Cambridge
University Press
Brady, Erika (1999) A Spiral Way, Jackson: University of Mississippi Press
Chion, Michel (1999) The Voice in Cinema, trans. Claudia Gorbman, New York:
Columbia University Press
Christian Science Monitor (1939) ‘Going to Golden Gate Fair’, February 4, p. 4
— (1941) ‘Talking Soapsuds on Air? Blame It on the Sonovox’, September 2, p. 9
Cohen, Lizabeth (2003) A Consumers’ Republic, New York: Knopf.
Davies, Lawrence E. (1939) ‘Machine That Talks and Sings Has Tryout;
Electrical Voder Will Speak at Fair Here’, New York Times, January 6, p.1
Dickinson, Kay (2001) ‘ “Believe”? Vocoders, digitized female identity and
camp’, in Popular Music, 20/3, pp. 333–47
Doane, Mary Ann (1987) The Desire to Desire, London: Macmillan
Donnelly, Antoinette (1932) ‘Woman’s Voice Reveals Charm or Lack of It’,
Chicago Daily Tribune, September 16, p. 21
Dudley, H. W. (1938) ‘System for the Artificial Production of Vocal or Other
Sounds’, U.S. Patent 2,121,142, filed April 7, 1937, granted June 21, 1938
— (1941) ‘System for the Artificial Production of Vocal or Other Sounds’, U.S.
Patent 2,243,089, filed May 13, 1939, granted May 27, 1941
— (1944) ‘System for the Artificial Production of Vocal or Other Sounds’, U.S.
Patent 2,339,465, filed July 10, 1942, granted January 18, 1944
Elsaesser, Thomas (1972) ‘Tales of Sound and Fury: Observations on the Family
Melodrama’, from Monogram, no. 4, pp. 2–15
MSMI 2:2 Autumn 08 205
Jacob Smith ♦ Tearing Speech to Pieces
Russo, Alexander (no date) On the Spot: The Spatial and Temporal Flow of Spot
Broadcasting, unpublished manuscript
Scha, Remko (1992) ‘Virtual Voices’, Mediamatic, vol. 7, number 1, Fall
Scheuer, Philip K. (1939), ‘Town Called Hollywood’, Los Angeles Times, July 2,
p. C3
— (1940) ‘Town Called Hollywood’, Los Angeles Times, December 15, p. B3
Sjogren, Britta (2006) Into the Vortex, Urbana and Chicago: University of Illinois
Press
Solomon, Matthew (1995) ‘Adapting ‘Radio’s Perfect Script: “Sorry, Wrong
Number” and Sorry, Wrong Number’, Quarterly Review of Film & Video,
Vol. 16 (1), pp. 23–40
Sterne, Jonathan (2003) The Audible Past, Durham: Duke University Press
Time Magazine (1939) ‘Sonovox’, 24 July, pp. 45–6
Tucker, George (1940) ‘Man About Manhattan’, Salamanca Republican-Press,
November 18, p. 4
Washington Post (1905) ‘The Voices of American Women’, January 9, p. 6
Williams, Greer 1940,‘The Graphic Laboratory of Popular Science’, Chicago
Tribune, July 21, p. G8
Wright, G. M. (1942a) ‘Means and Method for Producing Sound Effects’,
U.S. Patent 2,273,077, filed October 19, 1938, granted February 17, 1942
— (1942b) ‘Means and Method for Producing Sound Effects’, U.S. Patent
2,273,078, filed March 27, 1939, granted February 17, 1942
— (1942c) ‘Means for Teaching Telegraphy’, U.S. Patent 2,300,038,
filed September 9, 1940, granted October 27, 1942
Yoder, Robert M. (1943) ‘How To Talk Like a Cow From Boston’, Saturday
Evening Post, vol. 215, June 5, p. 11