In Praise of Partiality in Science

We all grow up with an image of science as a pillar of truth and nothing but truth. This ideal is so deeply embedded in us, that the very idea that scientists should take responsibility for the normative aspects of their work is anathema. Of all the things I have written here on Sweet Talk, my series on this subject provoked the most ferocious responses by far.

But science itself is far more than just truth. Elizabeth Anderson thoroughly dismantles the notion that it is. Our very ability to discern the whole truth, according to her, depends heavily on what we would call normative values, rather than value-neutral considerations. The whole truth is not a representation of “every fact about the phenomenon being studied.” If it were, it would “end up burying the significant truths in a mass of irrelevant and trivial detail.”

Theoretical inquiry does not just seek any random truth. It seeks answers to questions. What counts as a significant truth is any truth that bears on the answer to the question being posed. The whole truth consists of all the truths that bear on the answer, or, more feasibly, it consists of a representative enough sample of such truths that the addition of the rest would not make the answer turn out differently.

Anderson’s whole truth can only be determined by honing in on what is significant, an inherently value-laden concept. And that significance is determined by the questions we ask, which are based on our interests.

She argues that we have three broad areas from which we can criticize a theory on the basis of values.

The theory, although it asserts nothing but truths, may be trivial, insignificant, or beside the point: it doesn’t address the contextual interests motivating the question. Or, although it asserts nothing but truths, it may be biased: it offers an incomplete account, one that pays disproportionate attention to those pieces of significant evidence that incline toward one answer, ignoring significant facts that support rival answers. When the question which the theory seeks to answer has moral or political import, the charge of bias can only be made relative to an assessment of the moral and political relevance of the evidence the theory cites. Such assessments of course depend upon moral and political value judgments. Finally, the theory may be objectionable for trying to answer a question that has illegitimate normative presuppositions.

Triviality, partiality, and illegitimacy. The first is obvious: “so what?” is not a value-neutral question, but it should be a key consideration in deciding whether to pursue some line of thought. Among its many faults, the publish or perish model arguably encourages the production of too much triviality.

Partiality, as Anderson discusses it here, means being selective in a way that is deliberately misleading. I am not satisfied with this precise framing, but the basic notion is also fairly obvious: very small differences in the facts we present can paint radically different pictures. Anderson is quite right that missing some significant pieces of the puzzle is a valid criticism of a theory. Whether incompleteness or partiality per se are the problem is a matter we will return to.

The last source of criticism, illegitimacy, is the hardest for most people to accept. This touches on what I have called dangerous questions, or the range of applications for particular ideological frameworks that lead to atrocities. As an example, she looks at the Nation of Islam’s book The Secret Relationship Between Blacks and Jews.

The question that The Secret Relationship purports to answer is thus not adequately specified by such seemingly value-neutral questions as “What was the role of the Jews in the Atlantic slave system?” or even “How did Jewish roles in the slave system compare with the roles of other ethnic groups?” For these do not specify which roles and which comparisons are of interest. The question that The Secret Relationship implicitly purports to answer is rather “Do Jews deserve special moral opprobrium or blame for their roles in the Atlantic slave system or bear special moral responsibility for that system’s operations?”


Thus, if it is a moral mistake to pass judgments of collective guilt or merit on whole ethnic groups, then there is no justification even to make “Jewish” a significant classification in historical studies of the slave trade that are aimed at addressing questions of responsibility. What justification could there be for singling out Jews as a comparison class in such studies, rather than, say, the class of people who have drooping eyelids?

Asking how much we ought to blame the Jews for slavery was an illegitimate question from the start, borne of pernicious values.

It seems simply correct to me that we ought to criticize scientific theories for being trivial or illegitimate in the sense described. But I do not think partiality itself is a flaw. Everything in human experience is partial. Our sense of the whole is always only projected, and frequently revised. Partiality in the sense of incompleteness is also impossible to do away with; we can only argue that the picture we paint from our interpretation of the partial evidence we have is superior to someone else’s interpretation of the same or different, but still equally partial evidence. Moreover, to value anything is to be partial towards it; so we cannot embrace value-laden theory and do away with partiality. Making impartiality one of the chief values of science is a mistake on Anderson’s part. 

Justice already does enough in her framework without recourse to impartiality:

Not value-neutrality, but justice, offers the proper model of objectivity in science. Justice includes the demand to do justice to the subjects of study as well as the demand to do justice to other inquirers: to respect them as equals, to respond to their arguments, evidence, and criticisms, to tolerate the diversity of views needed to secure the objectivity of science as a social practice.

Do justice to the subjects of study, even if we increasingly find a conclusion we don’t wish for more persuasive than one that we do. Anderson rightly cries foul at the notion that our values warp science into wish fulfillment:

[The] claim that value-laden inquiry leads to wishful thinking makes sense only if value judgments express nothing more than idle wishes or desires: propositions one would like to be true, quite independently of whether they or any other propositions are likely or possible. No serious contemporary theorist accepts such a crude account of value judgments. Even those who believe that value judgments express something more like emotional states than beliefs argue that emotional states can he warranted or not, depending on the facts. So warranted value judgments, too, must he attentive to the facts. [The] assumption that value-laden inquiry leads to dogmatism makes sense only if value judgments are essentially matters of blind, overbearing assertion, not subject to critical scrutiny or revision in light of arguments and evidence. Again, no serious moral theorist accepts this primitive emotivist view any more. [The] assumption that value-laden inquiry will be dishonest comes from the thought that morally value-laden inquiry can only be inquiry designed to reach a foregone conclusion, hence inquiry that will neglect, cover up, or misrepresent evidence tending to show that the conclusion is false. Yet, this supposes that honesty is not itself an important moral value that should guide inquiry. Finally, [the] charge that politically value-laden inquiry will invite totalitarianism supposes that political values are essentially totalitarian. But feminist empiricists, including Longino, are virtually all democrats and aim to extend principles of democracy to scientific practice, notably in insisting on tolerance of diverse value-laden research programs and on the equality of inquirers. [The] alarm seems based on the nihilistic view that there is no such thing as moral inquiry at all, only arbitrary moral commitment.

The fear of people who oppose a vision of science as driven by values is that this would make scientists dishonest, or self-deceptive, or invite injustice. But honesty, integrity, and justice are values!

Anderson sees this quite clearly, which is why I find it odd that she would resort to the language of the objective, impartial, and unbiased. Science is persuasion, it is the art of using your judgment and making the best case that you can and attempting to build a consensus. Scientists ask questions they think are worth asking and attempt to find answers worthy of the questions. It is a value-laden and political enterprise, and each individual scientist’s contribution is utterly partial.

And there is nothing wrong with that—so long as they strive to be honest and do justice to their subject matter, and are willing to enter the arena to be criticized by their peers and offer them good faith criticism in turn.

Related Posts:

7 thoughts on “In Praise of Partiality in Science

    1. To be rude and answer a question with a question: what would that look like?

      I know what honestly looks like. If you see something that works against the conclusion you want, you include it anyway. Or if you’ve become persuaded of a conclusion you didn’t want, you present it anyway, and make the strongest case you can.

      I’m not sure what objectivity looks like, or impartiality. I see everything from my perspective, which is incomplete. I can make sure to do my experiments in a way that other people could easily attempt to replicate, but that’s a matter of doing justice to your peers or having integrity or just observing the proper standards of evidence.

      I no longer know what being objective or impartial would mean.

      1. Paul Crider

        I think objectivity or impartiality are still useful concepts. They’re related to honesty but I’m not sure they’re reducible to it. Think of Smith’s Impartial Spectator. The IS thinks from the perspective of an idealized person without the attachments an actual individual has. This is imperfect, as what we feed into the idealization always has some residual, but it’s not an incoherent concept either.

      2. Ok, I admit it can be a useful thought exercise to say “if I were someone who didn’t have a stake in this, would I believe what I believe?”

        If that is what is meant by impartial, I think that’s right and good. I’m just wary of some of the baggage that comes with the term.

  1. Your distillation of Anderson’s thesis reads like a description–a good one, at that–of what scientists presently do, acknowledged or not. (And I suspect most scientists would accept this broad-based value-driven description of science.)

    However, it’s unclear to me how articulating the framework might compel any sort of change in Science. At present, funding agencies direct money towards questions that bear on social goods. Scientists pursue research they find interesting and important. Hopefully these purposes harmonize, but, when they don’t, scientists honour their obligations to provide for their families, studying that which attracts funding (deemed important by someone somewhere). Researchers develop dry, heavy prose, and controversial results filter through years of analysis and debate before coming to light. Etc. etc.

    Professors and graduate students and national labs employees and the like generally realize their work is value-laden. (With a few exceptions, typically found near the furthest reaches of theoretical disciplines.) And here we tell scientists: Your work is and ought to be value-driven. And then what? It already is, no? Which practices are primarily at issue?

    Any change Anderson thinks reasonable is, at some level, arguable within already-existing social conventions of science. Pleas to the NIH to consider some disease, to a journal editor or readers of a scientific magazine to publish null results, or whatever. Social sciences are already value-laden to the extreme–if anything, ideals like truth and the dispassionate eye are probably most needed at the margins. Anderson’s descriptive contribution is nice, but her language in the excerpts you provide suggest a normative program. And I’m left wondering how she cashes it out.

    What is the relationship between Anderson’s positive account and normative project?

    (If I were a good reader, I’d go through the 1995 paper. But I’m currently off campus, so that will have to wait until tomorrow.)

    1. What I find fascinating about Anderson and her allies is that they’re part of a wave of feminist theorists who saw that in order to ground feminist social criticism, they needed to flesh out the implicit epistemology and social ontology. (Here’s a paper on the latter from another feminist

      You can see Anderson’s paper as a response to those who claim feminist critiques of social science especially are “politicizing” and the “positive” (that is, value-neutral) considerations should take priority. So her epistemology is one step removed from her actual normative project, which involves specific feminist critiques.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.