Distraction Free Reading

“Doing Being a Latina,” or Performing Identities Through a Computer Voice

The ability to communicate is often taken-for-granted and imperceptible, despite being vital to everyday life. It defines our social performances as family members, professionals, and neighbors. Moreover, institutions as well as identities need to be “talked into being” (Heritage and Clayman, 2010). Although in many mundane situations we get by with meaningful bodily gestures (Goodwin, 1986) such as nodding, shaking the head, waving, and pointing, other interactions require us to use complex language processing skills and muscular control over the vocal organs and hands.

People with what are known as “language and communication disorders”—whether developmental or acquired—perhaps more than anyone understand the intricate mechanics of human embodied speech. Articulating sounds, controlling prosody, taking one’s turn at the right time, as well as deploying various pragmatic skills—such as using metaphors and ironies—are jointly involved in the work of communication (Wilkinson, 2019). When one or several of these functions are impaired, a person may need some form of augmentative and alternative communication (ACC) to develop the habit of “not only listening to the world but participating in it” (e.g. such as a computer-mediated voice). This is especially important for people who don’t use sign language in everyday life (unlike Deaf communicators) but, because of their impaired hand and arm mobility, are only able to use a limited number of gestures, or sometimes only eye gaze.

Unlike many other human-machine hybrids analyzed in STS and disability studies (Haraway, 1991; Star 1991; Kafer 2013; Hamraie and Fritsch 2019), the use of synthetic voice by non-speaking individuals has been largely underrepresented. Several recent works have shed some light on practicalities of using AAC in the classroom setting (Hillary, 2019), the socio-technical networks supporting the life and work of the scientist Stephen Hawking (Mialet, 2012), as well issues of accommodating the tone of voice in the design of communication devices (Pullin, 2013). However, a nuanced analysis of both enabling and constraining aspects of various speech-generating tools, such as specialized apps for laptops, tablets, and smartphones, as well as dedicated Text-To-Speech (TTS) devices, has not yet been fulfilled.

Against this background, the book by Meryl Alper (2017) is a rare and precious exception, since it is the first and only qualitative research into the worlds of families whose children use a computer or a smartphone and a communication app (Proloquo2Go) to interact with others. Most parents studied by Meryl were middle-class Californians, with college-degree education, making more than $100,000. Almost half of the families were non-white, including Latino/a/x (3), mixed-race (3), Asian (2), and Black (1) families. What this blog post explores is some ways of enacting an identity (e.g. a Latina girl)—either through speech synthesized by a TTS program or via pre-recorded sequences of one’s “voice donor.”

giving voice = giving identity?

The book Giving Voice: Mobile Communication, Disability, and Inequality is a precious contribution to the growing field located at the intersection of STS and disability studies (see also Laura Mauldin, Mara Mills). The work by Meryl Alper is even more unique since it describes the population of people often ignored in the STS literature, i.e. people who speak through a TTS program, rather than oral speech (“augmented speakers”). Through qualitative interviews with parents, observations of at-home training with two speech-language pathologists, conversations with children’s school reps, Alper maps out a rather complex picture of how families’ use of speech-generating software is incorporated into their larger lives. Focusing on one particular program, Proloquo2Go for the iPad, produced by the Dutch company AssistiveWare, Alper documents the “social lives” of these technologies and explores issues of access, especially among less privileged families. The book does a great job describing the various concerns of parents of children with speech disabilities: from how to make a child and their friends distinguish between an iPad for communication and an iPad for fun; to issues of access, especially among less privileged families. As Alper describes, the procurement of an iPad or Proloquo2go could happen through direct purchase, charitable donations, temporary loans, or school provision, but is oftentimes a matter of negotiation and contestation, especially among the less economically privileged families (the current price of Proloquo2Go is $249.99, the price for an iPad varies from $300 to $1,000).

In what follows, I want to discuss some examples from Chapter 2 that clarify some relations between voice and identity. Many characteristics of a synthetic voice, such as age, gender, ethnic and expressive characteristics, are defined by the ACC developers—tech companies who inscribe their vision of the world (Akrich 1994) in the speech-generating program. Proloquo2Go, for example, suggests a range of 24 voices that speak “US English” (as of June 2016), of which 16 are male voices and 9 female ones. Alper argues that, in the program, male voices are endowed with more expressive choices in terms of the level of intimacy and affect, such as “Up Close,” “From Afar,” “Sad,” and “Happy,” while female ones lack such expressive possibilities and are all named simply “Female Adult.”

A list of 24 voice options arranged in a table of 3 columns and 8 lines. Each cell contains a name for the voice option and a description that specifies whether it's an adult or child voice, as well as the gender.

Screengrab from AssistiveWare’s Proloquo2Go website. Source: Alper 2017, p. 39.

Analysis of gender in the affordances of Proloquo2Go detects dissimilar opportunities available for male and female voices. Expressive features offered to adult male voices (like the voice of a hip-hop male adult, “Saul”; or a cartoonish voice of Star Wars’ Yoda, “Little Creature”) are much more well-thought and personalized than female voices which lack individuality. Imagine how gratifying for female users it would be if they could pick a voice that sounds like Emma Fitzgerald, or Rihanna, or any other of their favorite artists.

Alper’s research shows how a range of default options for voices at Proloquo2Go accommodates some identities (e.g. a Texan man, an old man, a sad man), and restricts the performance of others (e.g., a female teenager, a happy woman, or a Latinа girl). [1]

In the world of ACC, some of the most challenging have been issues of the tone of voice and its expressiveness. To a large extent, this is because “tone of voice, in all its subtlety, is not an easy quality to talk about—even within the discipline of phonetics” (Pullin 2013, 24). Alper talks about such issues faced by the parents of children using a synthetic voice. In this line, the father of 10-year old Beatriz, thought that the Proloquo2Go sometimes “sounds too automated.” The previous app that this family used to communicate with Beatriz was TapToTalk. In this program, Beatriz’s older sister, Ariana, was able to pre-record her speech, so that Beatriz could use it in different situations by tapping a corresponding button on the app. Ariana, hence, served as a “voice donor” to her beloved sister, and all the family seemed to be happy about this decision. “Because it was her sister’s voice, and it was according to her age, and it was one the tone of voice that we wanted,” explained Beatriz’s father (Alper, 2017, 54). This way, Beatriz could sound like her sister, a Latina girl of the same age, and also talk about things with much richer intonation, afforded by Ariana’s pre-recorded natural speech. Through this and similar examples, Alper shows how families enact their ideas about kinship, culture, and age in their choice of a synthetic voice. In selecting a particular voice for their non-speaking children, parents try to tailor the expressive options of available ACC devices to their notions of what it means to sound “like a child,” “like a human,” “like a family/culture member.” The construction of children’s identity through voice, therefore, is shaped both by their parents and specific affordances of ACC devices.

***

There are endless ways to communicate with others—through oral speech, body language, and gaze, as well as communication apps, text messages, chats, and sign languages—and all of them place high expectations on one’s behavior. As Erving Goffman points out, “any deviation… on any occasion when the rule is supposed to apply can give the impression that the actor is delinquent with respect to the whole class of event” (1971, 97). This way, Goffman connects one’s voice or communication style with one’s identity. Our social self depends on daily rituals to maintain one’s identity, “doing being” an x, y, or z (Sacks, 1984): communicating like a child, like a parent, like a school teacher, like a radio host, etc. These ways of making one’s identity through specific ways of talking are what makes Alper’s observations so interesting and promising in terms of further research. The enabling and constraining functions of TTS technologies and their affordances, such as a range of voices available for a particular user to pick, hopefully, will have an effect of a chain reaction, and the world will see other good texts on the social worlds of families using TTS technologies, as well as detailed analyses of organization of augmented communication.


Endnotes

[1] The newer version of Proloquo2Go, which is currently featured on the AssistiveWare website, includes 40 voices and has made some progress towards including more diverse female voices, such as an Indian female (“Deepa”), a Scottish (“Rhona”), and an old British (“Queen Elizabeth”) female options (as of December 2019).

References

Akrich, M. (1992). The De-Scription of Technical Objects. In W. Bijker and J. Law (eds.) Shaping Technology, Building Society: Studies in Sociotechnical Change. Cambridge: MIT Press, pp. 205-224.

Alper, M. (2017) Giving Voice: Mobile Communication, Disability, and Inequality. Cambridge, MA: MIT Press.

Goodwin, Ch. (1986). Gestures As a Resource For the Organization of Mutual Orientation. In Semiotica, Vol. 62, No. 1/2, pp. 29-49.

Hamraie, A., Fritsch, K. (2019) Crip Technoscience Manifesto, Catalyst: Feminism, Theory, and Technoscience. In Catalyst, Vol. 5, No. 1. URL: https://catalystjournal.org/index.php/catalyst/article/view/29607

Haraway, D. (1991). A Cyborg Manifesto: Science, Technology, and Socialist-Feminism in the Late Twentieth Century. In Simians, Cyborgs, and Women: The Reinvention of Nature. Routledge.

Heritage, J., & Clayman, S. (2010). Talk in Action: Interactions, Identities, and Institutions. Malden, MA: Wiley-Blackwell.

Kafer, A. (2013). The Cyborg and the Crip: Critical Encounters. In Feminist, Queer, Crip, Bloomington: Indiana University Press, pp. 103-128.

Mialet, H. (2012) Hawking Incorporated: Stephen Hawking and the Anthropology of the Knowing Subject. Chicago: The University of Chicago Press.

Pullin, G. (2013). Seventeen Ways to Say Yes, Exploring Tone of Voice in Augmentative Communication and Designing New Interactions with Speech Synthesis. Doctoral Dissertation, University of Dundee.

Sacks, H. (1984). On Doing “Being Ordinary.” In Structures of Social Action: Studies in Conversation Analysis, J. Maxwell Atkinson and John Heritage (eds.), Cambridge: Cambridge University Press, pp. 413–429

Star, S. L. (1991). Power, Technology, and the Phenomenology of  Conventions. On Being Allergic to Onions. In A Sociology of Monsters: Essays on Power, Technology and Domination, J. Law (ed.), Routledge, p. 26–56.

Wilkinson, R. (2019). Atypical Interaction: Conversation Analysis and Communicative Impairments. In Research on Language and Social Interaction, 52(3), 281–299.

Leave a Reply

Your email address will not be published. Required fields are marked *