“Above all, rational knowledge does not pretend to disengagement: to be from everywhere and so nowhere, to be free from interpretation, from being represented, to be fully self-contained or fully formalizable. Rational knowledge is a process of ongoing critical interpretation among “fields” of interpreters and decoders. Rational knowledge is power-sensitive conversation. Decoding and transcoding plus translation and criticism; all are necessary. So science becomes the paradigmatic model, not of closure, but of that which is contestable and contested. Science becomes the myth, not of what escapes human agency and responsibility in a realm above the fray, but, rather, of accountability and responsibility for translations and solidarities linking the cacophonous visions and visionary voices that characterize the knowledges of the subjugated.” – Donna Haraway, “Situated Knowledges: The Science Question in Feminism and the Privilege of the Partial Perspective”, 1988
We are reading Donna Haraway’s Situated Knowledges and Cyborg Manifesto for our department’s “Classics” reading group. An odd institution at Berkeley’s School of Information, the group formed years ago to be the place where doctoral students could gather together to read the things they had a sneaking suspicion they should have read, but never were assigned in a class. Since we bridge between many disciplines, there is a lot of ground to cover. Often our readings are from the Science and Technology Studies (STS) tradition.
I love Haraway’s writing. It’s fun. I also think she is mostly right about things. This is not what I expected going into reading her. Her position is that for feminists, rational objective knowledge has to be found in the interpretation and reinterpretation of partial perspectives, not a “god trick” that is assumed to know everything. This “god trick” she associates with phallogocentric white male patriarchal science. This is in 1988.
In 1981, Habermas published his Theory of Communicative Action in German. This work incorporates some of the feminist critiques of his earlier work on the formation of the bourgeois public sphere. Habermas reaches more or less the same conclusion as Haraway: there is no trancendent subject or god’s point of view to ground science; rather, science must be grounded in the interaction of perspectives through communicative action aimed at consensus.
Despite their similarities, there are some significant differences between these points of view. Importantly, Haraway’s feminist science has no white men in it. It’s not clear if it has any Asian, Indian, Black, or Latino men in it either, though she frequently mentions race as an important dimension of subjugation. It’s an appropriation and erasure of non-white masculinity. Does it include working class white men? Or men with disabilities of any kind? Apparently not. Since I’m a man and many of my scientist friends are men (of various races), I find this objectionable.
Then there is Haraway’s belief that such a conversation must always be cacaphonous and frenetic. Essentially, she does not believe that the scientific conversation can or should reach consensus or agreement. She is focusing on the critical process. Reading Habermas, on the other hand, you get the sense that he believes that if everyone would just calm down and stop bickering, we would at last have scientific peace.
Perhaps the difference here comes from the presumed orientation or purpose of interpretation. For Habermas, is it mutual understanding. For Haraway, it is criticism and contest. The “we” must never completely subsume individual partiality for Haraway.
Advocates of a cyborg feminist science or successor science or science of situated knowledges might argue for it on the grounds that it improves diversity. Specifically, it provides a way for women to participate in science.
In UC Berkeley’s D-Lab, where I work, we also have an interest in diversity in science, especially computational social science. In a recent training organized by the committee for Equity, Inclusion, and Diversity, we met together and did exercises where we discussed our unconscious biases.
According to Wikipedia, “Bias is an inclination of temperament or outlook to present or hold a partial perspective, often accompanied by a refusal to even consider the possible merits of alternative points of view. People may be biased toward or against an individual, a race, a religion, a social class, or a political party. Biased means one-sided, lacking a neutral viewpoint, not having an open mind.” To understand ones bias is to understand ones partial perspective. The problem with bias in a diverse setting is that it leads to communication breakdown and exclusion.
A related idea is the idea of a statistical bias, which is when a statistic is systematically out of sync from the population of interest. In computational social science, we have to look out for statistical biases because we aim for our social scientific results to be objective.
Another related idea is cognitive bias, a psychological phenomenon more general than the first kind of bias. These biases are deviation from rationality in psychological thought. This Nobel Prize winning psychological research has found systematic ways in which all people make mental shortcuts that skew their judgments. I’m not familiar with the research on how these cognitive biases interact with social psychology, but one would imagine that the answer is significantly so.
Haraway’s situated knowledges are biased, partial knowledges. She is upholding the rationality of these knowledges in opposition to the “god trick,” “view from nowhere,” which she also thinks is the position of phallogocentric subjugating science. Somehow, for Haraway, men have no perspective, not even a partial one. Yet, from this non-perspective, men carry out their project of domination.
As much as I like Haraway’s energetic style and creativity, as a man I have a difficult time accepting her view of science as situated knowledges because it seems to very deliberately exclude my position.
So instead I am going along with what we learned in Equity, Inclusion, and Diversity training, which is to try to understand better my own biases so that I can do my best to correct them.
This experience is something that anybody who has worked collaboratively on source code will recognize as familiar. When working on software with a team of people, everybody has different ideas about how things should be organized and implemented. Some people may have quirky ideas, some people may be flat out wrong. The discussion that happens on, for example, an open source issue tracker is a discussion about reaching consensus on a course of action. Like in statistics or the pursuit of psychological rationality, this activity is one of finding an agreement that reduces the bias of the outcome.
In machine learning and statistics, one of the ways you can get an unbiased estimator is by combining many biased ones together and weighting their outcomes. One name for this is bagging, short for ‘bootstrap aggregating’. The idea of an unbiased democratic outcome of combined partial perspectives is familiar to people who work in data science or computational social science because it is foundational to their work. It is considered foundational because in the “exact sciences”–which is how Haraway refers to mathematics–there is a robust consensus on the mathematics backing the technique, as well as a robust empirical conclusion of the technique’s successful performance. This robust consensus has happened through translation and criticism of many, many scientist’s partial perspectives.
It is frustrating that this kind of robust, consensually arrived at agreement is still sometimes rejected as radically contingent or historical by those from the Science and Technology Studies (STS) tradition who find their epistemic roots in Haraway. It’s especially frustrating, to me, because I would like to see more diversity–especially more women–in computational social science, or data science more generally. Haraway seems to believe that women in science are unable to overcome their own bias (partiality), or at least encourages them to not try. That seems like a non-starter for women in STEM, because I don’t know how you would ever learn statistics or programming without orienting yourself towards unbiased agreement with others.
So I have to conclude that teaching people Haraway as an epistemology is really bad for science, because it’s bad for diversity in science. That’s a little sad because obviously Haraway had the best of intentions and she is a really interesting writer. It’s also sad because a lot of STS people who base their work off of Haraway really think they are supporting diversity in science. I’ve argued: Nope. Maybe they should be reading Habermas instead.