Seeing like a Rover: How Robots, Teams, and Images Craft Knowledge of Mars by Janet Vertesi (review)

Janet Vertesi’s Seeing Like a Rover provides a detailed and engaging ethnography of scientific image production within NASA’s Mars Exploration Rover mission. Drawing on considerable fieldwork, including participant observation, the book provides thick description of mission scientists’ and engineers’ practices of image production, consumption, and dissemination during a multiyear period when Vertesi was permitted to observe many aspects of mission planning and implementation.

Analytically, the book focuses on the technologies and sociotechnical practices of image capture in the Rover mission, as well as on the sociological aspects of image manipulation and control. Vertesi starts with a reliance on the concept of “seeing as” to illustrate how visual information from Mars can be interpreted in different, specific ways. She then builds a new concept of “drawing as” that explains how scientists and engineers and technicians make decisions in advance of Rover action about where to aim their cameras and what types of images they want to produce.

The structure of the book takes readers on a detailed and enjoyable guided tour of mission activities. Chapter 1 presents an ethnography of daily mission meetings, which involve a ritualized method of deciding what the Rover will do and what images it will capture. Particular attention is paid to the variety of decisions made regarding the manipulation of image direction, macro settings, and wavelength filters. Chapter 2 reviews the mission’s ritualized practices for image calibration, noting how the team members’ need for images to be rendered objective and trustworthy also constrains the ways in which visual data for Mars can subsequently be “drawn.” Chapter 3 zeroes in on the delicate work of image processing, in which scientific questions are investigated and scientific claims are made through very specific forms of “drawing Mars as” a variety of potential hypotheses or interpretations. I was especially struck by the assertion in this chapter’s conclusion that “drawing as makes epistemology look like ontology” (p. 103).

Chapter 4 follows this thread to another analytical dimension, in which team members draw on or “annotate” images (especially maps) that have already been “drawn as” a specific view of Mars. Chapter 5 shows how image work is used to form a collective sense of team that supports a fairly flat social hierarchy that is generally effective in surmounting disciplinary differences. Chapter 6 explains how team members are socialized to “see like a rover” by imagining their own robotic embodiment when conceptualizing various aspects of image production. Chapter 7 discusses the ways that team members form a common vision for acceptable image manipulation that allows them to agree on what is objective and true and avoid concerns about image tampering. Chapter 8 goes even further beyond the idea of “drawing as” to point out the political importance of images to the Mars Rover missions and to discuss how this political importance becomes a constraint in image production and in the social productions it allows within the team.

Vertesi’s concept of “drawing as” is an important contribution to STS scholarship and is applicable not only to collaborative team-based scientific endeavors like the Mars Rover missions, but also to other more individualistic forms of science as well. By showing so clearly that decisions about image production are in fact decisions about science itself, and then analytically linking image production to socialized norms for image processing, Vertesi shows how mission team members limit the possibilities for how images can subsequently be seen and interpreted. These practices of image production thus “draw” Mars in specific ways so that future image-based interpretations of its nature are actually limited to specific views even before the imagery is ever created.

Overall, this is a compelling and important book that will be highly appealing to the academic community of STS scholars and will be especially useful in graduate courses. The early chapters show how visual data are drawn in specific ways to enable or produce...