Beth Semel

Grant Type

Dissertation Fieldwork Grant

Institutional Affiliation

Massachusetts Inst. of Technology

Grant number

Gr. 9251

Approve Date

April 8, 2016

Project Title

Semel, Beth M., Massachusetts Inst. of Technology, Cambridge, MA - To aid research on 'Speech, Signal, Symptom: Psychiatric Diagnosis and the Making of Algorithmic Listening in the United States,' supervised by Dr. Graham M. Jones

Preliminary abstract: While traditional techniques of psychiatric diagnosis in North America pivot on clinicians’ capacity to interpret the content of patients’ speech, this dissertation follows mental health research teams in the academic, commercial, and military arenas that have enlisted the work of computer engineers to develop alternate means of deciphering the biomedical significance of behavioral symptoms. These teams of psychiatrists and signal analysts–computer engineers trained to parse, extract, digitize and process complex signals–are all working to produce technology that they hope will identify connections between inner, psychological states and paralinguistic features of speech (pitch, intonation, prosody, etc.), bypassing the content of speech altogether. I investigate these multidisciplinary research projects with attention to researchers’ talk about language and mind and to how the software is designed and tested. Why do researchers insist that their technologies will be agnostic to culture, language, and gender differences, and how are these assumptions objectified in the algorithms they build? How do they envision this diagnostic technology in terms of its capacity to reconfigure the relationship between the speaking and listening subjects in the diagnostic encounter? Moreover, how does the flow of techniques, technologies, and technologists across the domains in which the researchers work pave the way for the permeation of the models of listening, speaking, and self that signal analysts enact into other arenas of listening? Against the backdrop of increased ambivalence toward technologies of surveillance in the U.S. after 9/11, this dissertation considers how signal analysis itself may reinforce linkages between mental health, national security, and commerce.