The Pattern Recognizers: Surveillance, Security, and the Making of Identity in 20th Century America

Abstract

This dissertation considers the historical relationship between science and surveillance in the 20th century in the United States. It centers its study on the emergence and development of pattern recognition as a scientific methodology and as a surveillance technique for recognizing identity. Across seven chapters, the dissertation shows how social scientific and technical research transformed patterns into a scientific object primed for classification. The dissertation begins with research in cultural anthropology in the early 20th century. It examines how, at the nexus of methodological, professional, and political concerns, anthropologists developed the notion that culture contains patterns. Throughout World War II, cultural anthropologists brought notions of patterning and their recognition into U.S. government agencies. Following the war, in response to existing research in anthropology and psychology, cyberneticists developed pattern perception mechanisms. While engaging with this work, in the 1950s scientists developed computers that could recognize patterns with formalized rules. Pattern recognition researchers soon advocated to U.S. intelligence agencies that their techniques could solve problems of surveillance. By the end of the 1960s, pattern recognition research had changed how intelligence agencies could conduct surveillance. Machine-classifiable patterns had become a ubiquitous object of scientific knowledge and object of surveillance. The dissertation foregrounds the often-mundane ways in which scientists have shaped the practice of surveillance, as well as how problems of surveillance have shaped the practice of science. More generally, the dissertation bridges historical epistemology with surveillance studies to excavate the historical conditions of possibility for surveillance practices today.