Abstract
Harry Collins has written three books on the sociology of gravitational wave detection. Changing Order: Replication and Induction in Scientific Practice (Sage, 1985) was a concise (187 pages), witty, provocative, ground-breaking ethnography of knowledge being made and unmade. Gravity’s Shadow (Chicago, 2004), as the Pynchonesque title implies, was longer (870 pages) and murkier. An epic romance of gravity wave physics, that book narrated the field’s evolution from kludgy, erratic, home-built “bar” detectors in the 1960s to today’s miles-long interferometric detectors costing hundreds of millions of dollars and built by international teams of hundreds of people. Collins’ portrait of a community he has known for forty years was remarkable, but the book’s best observations (especially those about how scientific communities simultaneously judge evidence and acquire resources) rested on a daunting wealth of detail.
Gravity’s Ghost, then, is a return to form. It is concise, easy to grasp, and the sociological implications are kept at the fore. The story unfolds over about two years, though the previous forty years shape the plot at every turn. Collins sketches that deep background where necessary, though he also relies heavily on notes to “see Gravity’s Shadow, chapter X.”
If Gravity’s Shadow was an epic, Gravity’s Ghost is a whodunit—or, as Collins puts it, a “whatwosit.” The mystery concerns a blip in the data recorded on the autumnal equinox in 2007—the “Equinox Event.” Was the Equinox Event a bona fide detection of the gravitational signature of an actual astronomic event (such as a supernova or the “inspiral” of two colliding neutron stars)? Or was it the kind of combination of wishful thinking and electronic and mechanical noise that has gained the gravitational wave detection community a suspect reputation in the past?
Or was it a third option: a deliberate test of that community’s ability to sort real events from noise? The need for such a test arises from a methodological Catch-22. Because of their field’s reputation for error, gravitational wave detection physicists perceive a need for extreme caution in declaring that a blip is an “event” rather than “noise.” But if they are too cautious, then their instrument will, effectively, be insensitive to all but monumentally unlikely events, and taxpayers and politicians may feel they have been sold a bill of goods. Thus, the (largely European and American) collaboration responsible for the three largest interferometric detectors assigned two of its members to “inject” anywhere between zero and three false events into the data stream at random times during the first year of full operation. The Equinox Event might or might not be one of those fakes. In theory, that possibility should create an incentive against overcautiously declaring it “noise.” In practice, as Collins shows, things are more complicated.
Collins uses the Equinox Event to explore some standard themes from the sociology of scientific knowledge: the inability of rules to fully specify their own application, the inability of tests to fully simulate the conditions that the device being tested might face, the “experimenter’s regress” of using incomplete (perhaps incorrect) knowledge of phenomena to calibrate instruments that will be used to characterize those same phenomena, and the value-ladenness of statistical analysis.
It is only in a final “Envoi” that Collins stakes out a new, normative position. There, he argues that the gravitational wave detection community’s relentless self-questioning makes it a model for other sciences and for any society dependent on technical expertise. Collins marvels at gravity wave researchers’ experimentation with organizational structures and incentive schemes in order to make their knowledge more robust—and sees in that social experimentation a way to rescue science and society from science-skeptical market and religious fundamentalists.
Collins also casts his admiration for scientific social experimentation as a rebuke to the field of Science and Technology Studies (STS) that he helped create in the 1970s and 80s. Back then, some philosophers and scientists believed that Collins was the corrosive science skeptic. As the acknowledgments to Gravity’s Ghost show, however, former adversaries now embrace Collins’ overt enthusiasm for science (and disaffection for much of STS). Collins frames this turnabout as a “Third Wave” of STS—the First Wave being pre-Kuhnian apologia for (and rational reconstruction of) science, and the Second Wave the constructivist turn typified by Changing Order. The Third Wave swings the pendulum back: “Wave Three makes it explicit that in spite of the logic of Wave Two … science is still the best thing we have where knowledge about the natural world is concerned. Here, the processes of science are unapologetically spoken of as the most valuable models for the making of technological knowledge” (p. 5).
I am sympathetic with Collins’ admiration for the gravitational wave detection community’s social experimentation and his desire to harness constructivist STS to a pro-science agenda. The Third Wave framework, however, turns so far back to something like Merton’s norms that it undermines some of Collins’ own Second Wave achievements. Changing Order was eye-opening because it paid attention to the things scientists themselves took to be important in real time, rather than to the things philosophers and Mertonian sociologists retrospectively said were “really” important. Scientists argue with each other. They make seemingly extrascientific judgments about each other. They have trouble replicating each other’s—and even their own—experiments, even ones they know are supposed to work. Many spend much of their time teaching, cultivating funders, traveling, and writing and reading grants, textbooks, and even science fiction. Many of them would like to profit monetarily from their research; even those who do not must participate in various markets in order to build their experiments.
Gravity’s Ghost is strongest in showing how scientists fuse these multiple identities and loyalties into a complicated, inherently social (yet still admirably reliable) logic for establishing truth. Yet in the Envoi, Collins argues that some scientific identities—e.g., scientists as entrepreneurs, celebrities, or New Atheist gadflies—are dispensable and even inimical to proper science. It is an argument that only tenuously connects to the rest of the book. Let’s hope Collins eventually explains this notion more fully, with the empirical basis that it deserves (and that we would expect from his earlier work).