Science, to borrow a phrase from Steven Shapin, is a social process that is “produced by people with bodies, situated in time, space, culture, and society, and struggling for credibility and authority.” This simple fact is difficult to remember in the face of intricate computer generated images and declarative statements in credible publications. Science may produce some of the most accurate and useful descriptions of the world but that does not make it an unmediated window onto reality.
Facebook’s latest published study, claiming that personal choice is more to blame for filter bubbles than their own algorithm, is a stark reminder that science is a deeply human enterprise. Not only does the study contain significant methodological problems, its conclusions run counter to their actual findings. Criticisms of the study and media accounts of the study have already been expertly executed by Zeynep Tufecki, Nathan Jurgenson, and Christian Sandvig and I won’t repeat them. Instead I’d like to do a quick review of what the social sciences know about the practice of science, how the institutions of science behave, and how they both intersect with social power, class, race, and gender. After reviewing the literature we might also be able to ask how the study of science could have improved Facebook’s research.
This sort of work has been done under a number of names, including social studies of science, science studies, science and technology studies, sociology of knowledge, and science, technology, and society. The banners that individual researchers march under is less important than the approaches and perspectives each take, so instead of concentrating on the changing names for this sort of work, I’ll instead focus on what these authors thought was aspect of science was most important to study.
The scientific method itself was born out of a debate between Thomas Hobbes (author of The Leviathan, best known for the “states exist so we don’t immediately kill one-another” hypothesis) and Robert Boyle (inventor of the air pump and widely considered founder of modern chemistry). The two argued vigorously over whether or not you could see something and declare it as fact (Boyle), or whether one had to understand underlying causes before contributing to natural philosophy (Hobbes). Whereas Boyle was willing to separate facts from causes ––birds die when you put them in a vacuum, exactly why was a mystery–– Hobbes saw this as sloppy philosophy. One had to build an argument from the ground up, starting with the cause (which may have been grounded in what would today be called “social” or “political” reasons) and ending with the observable phenomenon.
The division between Hobbes and Boyle (catalogued in Shapin and Schafer’s Leviathan and the Air Pump)represents in microcosm, the modern western worldview we have today: politics says what should be and science says what is. But science, whether it is making nuclear bombs or social media platforms, often works in the service of politics or becomes the center of political debates. You can’t neatly separate the two. It’s no coincidence, for example, that Newton’s calculus is particularly helpful at calculating cannon ball trajectories and statistical methods are particularly well-attuned to assisting a few people make definitive claims about lots of people.
Prominent French sociologists like Emile Durkheim and Max Weber have written about science; the former recognizing that just because something is socially derived does not make it not objective, and the latter noticing that science is a vocation and, like all vocations, demands loyalties to certain practices and people that are based on social considerations. By the 1930s social scientists were dedicating their entire careers to studying science. Robert Merton and Bernard Barber were some of the first sociologists of science. They saw the rise of authoritarianism in Europe as a threat to science and set out to show that science was inherently democratic and therefore science was good for democracy. Their work was concerned mainly with the practice of science and rarely made claims about the nature of facts and claims. They studied how scientific communities formed, how they rewarded desirable behavior, and the ways science was internally stratified by rank and prestige.
Starting in the 40s with Michael Polanyi and picking up in the 60s with the publication of Thomas Kuhn’s The Structure of Scientific Revolutions, science studies began to expand out and make observations about the content of scientific knowledge as well. Polanyi argued that underlying claims to knowledge are personal and collective convictions and science does itself a disservice to ignore such preconditions. If the sort of free debate that is necessary for a healthy scientific community is to occur, values must be stated plainly. To claim neutrality, Polanyi argued, was to hide your values. Kuhn makes a similar, but more systemic argument that all science happens within a paradigm. A paradigm is analogous to what sociologists call a s social world: a set of practices, widely held ideas, values, languages, and practices that mark a particular place and time.
According to Kuhn, most of the history of science can be described as punctuated equilibrium. Scientists live within a certain paradigm and do work based on that paradigm until something really big happens that threatens the dominant paradigm. The go-to example is the “Copernican Revolution” but not for the reason most people think. The story that is popularly told is that Copernicus “discovered” that the sun revolved around the Earth, not the other way around, but it is more accurate to say he re-discovered this fact. It was widely understood by the Ancient Greeks that the sun was the center of the solar system, but that “fact” fell out of commonly-accepted natural philosophy for over a thousand years.
Kuhn would say that this demonstrates that science is not a linear progression of ever-increasing understanding, rather it is a practice that generally works within incrementally changing social worlds until something big happens that causes a revolution into a radically different one. The important thing to note here is that the “something big” need not be a scientific discovery or breakthrough. It can be a war, a new tool coming to market, or (and Kuhn says this is usually the case) the death of a prominent member of a scientific community. If their “rivals” are able to take up positions as department chairs and journal editors, entire disciplines can change dramatically. Of course the ability to gain prestige in the field is based in no small part on the ability to do science but it is far from a pure meritocracy.
From the 60s to the 80s social scientists were largely preoccupied with describing exactly what contributed to success in science beyond the merit of work. Or, to put it more precisely, social scientists set out to understand how, what, and who was deemed meritorious and worthy of praise within scientific communities.
Early work in this field falls under the large banner of “social constructionist.” Radical social constructionists say that all claims to knowledge are power moves, not efforts towards truth or understanding. More moderate social constructionists only contend that scientific theories should at least be subjected to the same sociological analysis, whether they end up being “proven” true or false. This still means that a social constructionist analysis would never ascribe the success of a theory to its (to use a Colbertism) “truthiness.” Instead, the success or failure of a scientific program or theory comes from its ability to do useful work or are particularly suited to confirming the beliefs of powerful actors. At the center of the social constructionist approach are authors like David Bloor, Barry Barnes, and Donald A. MacKenzie. They and others are usually referred to as the Strong Programme of STS.
Shapin and Schafer describe the social constructionist approach as “playing the stranger.” They write in the introduction to their history of Hobbes and Boyle that they sought to “adopt a calculated and an informed suspension of our taken-for-granted perceptions of experimental practice and its products. By playing the stranger we hope to move away from self-evidence.” They try to understand why someone might disagree with the scientific method, especially at a point in history when it was far from clear that Boyle would win that particular controversy.
Social constructionist analyses tend to focus on the role of individual agency to effect change in scientific research programs but, as Daniel Kleinman has shown [paywall], the institutional structure of science can have a big influence on research practice as well. The standardization of lab equipment, for example, has a constraining influence on the variety of scientific research. Widely available equipment and chemicals “are created to suit a wide market of laboratories, not the local needs of individual labs.” Specialized research isn’t just a matter of modifying those widely available chemicals or tools, especially if they are covered under intellectual property laws. Standardized, proprietary equipment can make private companies indispensable to entire sub-disciplines. It can also mean the replicability of a study is directly tied to the business decisions of private firms
Aside from social constructionism, another large branch of science studies comes out of critical feminist studies. Feminists focus on the ways androcentric views of the world are embedded in scientific accounts of nature and scientific practice. Donna Haraway, Sandra Harding, bell hooks, Patricia Hill Collins, Karen Barad, Susan Traweek, Evelyn Fox Keller, Linda Layne, and many more authors have contributed extensive research in this field. Everything from the time to full professor (surprise, it takes much longer for women) to the tendency to ascribe the features of patriarchal white middle class family structures onto animal communities bend science toward male supremacy and away from other (and perhaps one could even say more empirically accurate) views of the world.
There are lots of examples here but Haraway’s concept of Teddy Bear Patriarchy is one of my favorites so I will use that as my example. * Haraway, in tracing a genealogy of primatology and natural history more generally in her book Primate Visions, notes that early naturalists and conservationists’ practices form the ideological bedrock for how present-day scientists go about cataloging and understanding the world. She points to the dioramas in the American Museum of Natural History in New York, some of which date back to Theodore Roosevelt’s efforts at nature conservancy, as the epitome of patriarchy’s counter-intuitive logic: taxidermy –the active hunting and killing of animals so that their skins may be presented in a museum– is somehow held up as a window onto life. She writes:
This is the effective truth of manhood, the state conferred on the visitor who successfully passes through the trial of the Museum. The body can be transcended. This is the lesson Simone de Beauvoir so painfully remembered in the Second Sex; mas is the sex which risks life and in so doing, achieves his existence. In the upside down world of Teddy Bear Patriarchy, it is the craft of killing that life is constructed, not in the accident of personal, material birth.
Teddy Bear patriarchy can hide in plain sight thanks to what Donna Haraway, in her essay Situated Knowledges calls the “god-trick.” The god trick is “seeing everything from nowhere.” It is the illusion of an all-seeing eye that doesn’t just passively view natural phenomena but also “fucks the world to make techno-monsters.”** Instead of a disembodied and universalistic approach to understanding nature, feminists like Haraway argue for a “Feminist objectivity” that situates knowledge in particular bodies and treats objects of study “as an actor and agent, not a screen or a ground or a resource.”
What might feminist objectivity look like in practice? How would the Facebook echo chamber report look if its authors had used feminist research practices? It is hard to tell because so much of Facebook is built upon a decidedly anti-feminist approach that capitalizes on the view from nowhere. But, if we were to start incorporating feminist epistemologies into Facebook research, a good place to start would be adopting what Sandra Harding and others have called “standpoint epistemology.” Standpoint epistemology is a way of reporting scientific findings without relying on the god-trick to assert legitimacy.
Facebook researchers using a standpoint epistemology would first recognize that their position of power in relation to users, not to mention their clear biases in showing that the algorithm is benevolent if not agnostic, has profound impacts on their results. If they still wanted to conduct the study they may mitigate their own biases by selecting users to review the data as well. They might pair their quantitative data with qualitative accounts and personal stories about avoiding or seeking out opposing viewpoints.
Most importantly though, a standpoint epistemology would recognize that different people avoid or seek out viewpoints for different reasons and not all echo chambers are created equal. A group of people intensely sharing news that confirms anti-choice legislation is not the same thing as a group of people sharing stories about the survival of trans people in the rural south. The latter acts as a safe space in a largely hostile world, whereas the former is means of distilling a hegemonic discourse.
Truly interesting questions arise when we think about how algorithms themselves may benefit from science studies in general and standpoint epistemologies in particular. Could algorithms help in the controlling of triggering content or might the formation and shaping of the algorithm become part of the daily practice of Facebook? It seems that having a single algorithm runs counter to the very basic precepts of standpoint epistemology in the first place. Perhaps the first step in a more feminist direction would be to acknowledge the obvious: that the Facebook algorithm is a product made for certain ends and to embrace that fact in future scientific reports.
*It is worth noting, reflexively, that Pierre Bourdieu (who founded what he called, reflexive sociology wherein the sociologist recognizes and announces his or her sociological positioning) identified summaries and textbooks as inherently political documents wherein the author defines a discipline by selecting its constitutive authors. That’s totally what I’m doing.
**Not completely sure what this means but see asterisk above. An interpretation of this quote might suggest that the view from nowhere gives scientists ideological cover when building or contributing to weapons development or other destructive technologies. Scientists can say that their work is value-neutral and it is the decisions of politicians that actually produce deadly results.
- Bauchspies, Wenda K, Jennifer Croissant, and Sal Restivo. 2005. Science, Technology, and Society: A Sociological Approach. 1st ed. Wiley-Blackwell.
- Bloor, D. 2007. “Ideals and Monisms: Recent Criticisms of the Strong Programme in the Sociology of Knowledge.” Studies In History and Philosophy of Science Part A 38 (1): 210–34. doi:10.1016/j.shpsa.2006.12.003.
- Haraway, D. 1988. “Situated Knowledges: The Science Question in Feminism and the Privilege of Partial Perspective.” Feminist Studies 14 (3): 575–99.
- Haraway, Donna J. 1990. Primate Visions: Gender, Race, and Nature in the World of Modern Science. Reprint. Routledge.
- Harding, Sandra, ed. 2003. The Feminist Standpoint Theory Reader: Intellectual and Political Controversies. 1 edition. New York: Routledge.
- jurgenson, nathan. 2015. “Facebook: Fair and Balanced.” Cyborgology. May 7. https://thesocietypages.org/cyborgology/2015/05/07/facebook-fair-and-balanced/.
- Kleinman, Daniel L. 1998. “Untangling Context : Understanding a University Laboratory in the Commercial World.” Science, Technology & Human Values 23 (3): 285–314. doi:10.1177/016224399802300302.
- Kuhn, Thomas S. 1996. The Structure of Scientific Revolutions. 3rd ed. University Of Chicago Press.
- Merton, Robert K. 1979. The Sociology of Science: Theoretical and Empirical Investigations. University Of Chicago Press. http://www.amazon.com/The-Sociology-Science-Theoretical-Investigations/dp/0226520927.
- Sandvig, Christian. 2015. “The Facebook ‘It’s Not Our Fault’ Study.” Social Media Collective. May 7. http://socialmediacollective.org/2015/05/07/the-facebook-its-not-our-fault-study/.
- Shapin, Steven, Simon Schaffer, and Thomas Hobbes. 1985. Leviathan and the Air-Pump: Hobbes, Boyle, and the Experimental Life : Including a Translation of Thomas Hobbes, Dialogus Physicus de Natura Aeris by Simon Schaffer. Princeton, N.J.: Princeton University Press.
- Tufekci, Zeynep. 2015. “How Facebook’s Algorithm Suppresses Content Diversity (Modestly) & How the Newsfeed Rules the Clicks.” The Message. May 7. https://medium.com/message/how-facebook-s-algorithm-suppresses-content-diversity-modestly-how-the-newsfeed-rules-the-clicks-b5f8a4bb7bab.