Author Archives: Jingyi Wu

Knowledge, Human Interests, and Objectivity in Feminist Epistemology

In this paper, Anderson aims to show that

  • The seeming competition between normative values and evidential values in science is unwarranted.
  • Simply justifying a theory based on evidential truth is not enough; we need a justice model that could bring in the “whole truth.”
  • We should integrate some specific feminist normative value—ontological heterogeneity and complexity of relationship—to all stages of science including discovery, justification, application, and critique of science.
  • Such interaction between normative and evidential value would result in a new phase of science where we apply dual justification for theory choice—significance and truth—instead of a single justification based on truth, as Haack argues. This new phase of science with dual justification, Anderson argues, offers a proper model of objectivity.

A. Longino’s argument for taking values into consideration in science:


Contextual Values:

  • “Political, moral and other values taken from the social context in which science is practiced.” (28)
  • External to science

Epistemic or Cognitive Values:

  • Values that are internal to science, and do not have explicit political or moral content. (29)
  • e.g. “accuracy, consistency, fruitfulness, breadth of scope,and simplicity” (29)


  • A situation in which “hypotheses are logically underdetermined by the data cited in their support” (28). So when two sets of hypotheses are equal in logically explaining the datas, our theory choice is underdetermined.
  • Underdetermination happens, such as the dispute between Einsteinian and Newtonian science, and between classical and marginal utility theory in economics (29).

Longino’s argument:

P1 If underdetermination happens in scientific theory choosing, then we resort to background assumptions to justify a certain hypothesis in science. (28)

P2 Underdetermination happens in scientific theory choosing. (28)

C1 We resort to background assumption to justify a certain hypothesis in science.

P3 Values are embedded in background assumptions.

P4 If we resort to background assumptions to justify a certain hypothesis in science, and values are embedded in background assumptions, then values play a role in science. (29)

C Values play a role in science.

“When the data run out, values legitimately step in to take up the “slack” between observation and theory” (29).

Some scientists are willing to accept Longino’s argument, but they only permit epistemic or cognitive values and refuse to consider contextual ones as important to science (29).

Longino’s response to that: the binary between epistemic and contextual interests break down when we focus attention on the grounds for supporting our theories, “we see that epistemic, metaphysical and practical interests” all contribute into the play (30). She especially focuses on practical interests, because she thinks that we all have practical interests in predicting and controlling phenomena (30).

My question about this: Longino goes on to distinguish the contemporary quantitative vs. the Aristotelian qualitative approach to characteristics of objects of study, and claims that the former represents a practical interest in predicting and controlling phenomena, and the latter represents a practical interest in self-understanding and successful communication. For me, Longino creates this practical interest language and puts both approaches in terms of this language, and then claims that every approach has a practical interest. I’m to some extent sympathetic to her, but language creating game still looks like cheating to me, since this argument is very similar to people’s reasons about how everything is political—once we create a political language, then everything is political, even apolitical becomes a political stance.

Now Longino can say that practical interests, which are highly connected with political and moral values, play a role in science. She then defines two specific feminist theoretical virtues that are representative of the contextual values we need to take into consideration:

  • Ontological Heterogeneity: “a preference for ‘splitting’ over ‘lumping’—for emphasizing the qualitative diversity and individuality of subjects of study and the distinctions among properties commonly classified together” (30)
  • Complexity of Relationship: “a preference for dynamic, interactive causal models that emphasize multiple causes of phenomena over single-factor linear or reductionist models” (31)

Longino further argues that if anything, taking political and moral values into consideration would only strengthen the conception of objectivity in science because our new values make inquirers accountable to others’ observations and criticisms while maintaining the empiricist adequacy standard of the theory. (32)

My question about this: as we advocate for the need to take heterogeneity and complexity into consideration, will these values become the new epistemic (constitutive) values in science over time? In that case, will they acquire a position of power that would supposedly obscure other values?

B. Haack’s Critique of Politically Value-Laden Science

Basically, Haack believes that values and evidences are competitive to each other. “Either theory choice is guided by the facts, by observation and evidence, or it is guided by moral values and social influences, construed as wishes, desires or social-political demands” (33). Therefore, she thinks that politically value-laden science is dangerous and dishonest because it lets value totalize theory choice.

Haack’s argument:

  1. Significant truth is the sole aim of theoretical inquiry.
  2. Whether a theory is justified depends only on features indicative of its truth, not its significance.
  3. One shows that a theory is (most probably) true by showing that it is (best) supported by the evidence.
  4. A theoretical proposition is supported by the evidence only if there is some valid inference from the evidence (in conjunction with background information) to it.
  5. Value judgements take the form “P ought to be the case.”
  6. There is no valid inference from “P ought to be the case” to “P is the case” (or any other factual truths).
  7. There is no valid inference fro value judgements to factual truths (5,6).
  8. Value judgements can provide no evidential support for theories (4,7).
  9. Value judgements can play no role in indicating the truth of theories (3,8).
  10. Value judgements can play no role in justifying theories (1,2,9).

C. Anderson’s Defense of a Longinoian Value-Laden Theory against Haack’s Critique

Anderson starts by recognizing that before her, critiques of Haack’s argument have been taking issues with premise (6), claiming that Haack’s argument “covertly relies on a background metaphysical assumption that the universe is not governed by teleological laws” and that “ought” might be interpreted as “can” which implies capabilities (34).

However, Anderson believes that a stronger criticism of Haack’s argument could be achieved by poking holes at premises (1) and (2). She claims that the justification of a theory depends on both its truth and significance. Moreover, she argues that not all values are totalitarian values that conflict with truth-seeking. Feminists only try to incorporate important moral and political values that would not interfere with evidential truth and would function as a second axis of theory justification. Haack’s worry about dishonesty would not happen because dishonesty is simply not an important moral value.

My question would be: so then do we get to choose what moral and political values are important? So who do “we” represent? Will these values be universal? Anderson clearly has heterogeneity and complexity in mind because they offer the most potentials (as we’ll see in a second), but I’m also wary of the tremendous authority that “we” claim to have. Thoughts?

D. Anderson on How Values Play a Role in Justification of Science

Anderson starts by noting that many scientists are willing to concede that values play a role in discovery and application of science, but refuse to admit that values influence the justification of science. Anderson then shows that values are important in justification of science for two reasons: first, a true theory might well be biased, so we need a justice model to ensure impartiality; second, contextual values in scientific inquiry guide theoretical classification, so we justify a theory by choosing a theoretical classification that corresponds with our contextual values.

I. Justice and Justification

Anderson distinguishes between telling some truth and telling the “whole” truth, and argues that a theory that tells some truth satisfies the truth criteria but can still easily be biased or distorted because it is not telling the “whole” truth. We need to incorporate the value of justice into standards of theory choice to ensure impartiality. (Example of the Jewish Slave Owner) (37-38)

Anderson further distinguishes between impartiality with value-neutrality. She claims that value-neutrality is not desired because it might present every small fact about the phenomenon in question and “bury the significant truths in a mass of irrelevant and trivial details” (39). On the other hand, impartiality makes sure that we present the evidence according to the goals and context of our theoretical inquiry. The whole truth under impartial lens “consists of a representative enough sample of such truths that the addition of the rest would not make the answer turn out differently” (40).

Question: This is sort of vague, how do we make sure that the truth in front of us is indeed the whole truth? Maybe we need to be fallibilistic?

II. Contextual Values Guide Theoretical Classification

Anderson argues that we classify our objects of scientific study differently according to our contextual values in our theoretical inquiry, in turn, we choose the theory that uses the specific theoretical classification related to our contextual values.

She gives an example about medicine: We classify organisms based on whether they cause disease in human or not. This classification is tremendously connected to our interest in human health.

She gives another example about gender test: The Terman-Miles M-F test only allows two options “masculine” and “feminine,” which reveals a gender binary value at the background. In contrast, the BSRI test does that more open choices: besides male and female, it has androgyny and gender undifferentiated as choices. The classification in BSRI test reveals a value for heterogeneity.

These examples show that the specific contextual values that we adhere to influence drastically our theoretical classification. If we subscribe to feminist virtues, every step in science will look differently.

Small question for this section: at some point she talks about pure vs. applied science, and blurs the line between them. I’m wondering if pure science is really illusory though. As a math major who hangs out with pure math, I am kind of reluctant to think that all math inquiries are interest-laden. It’s just fun!

E. What The Two Feminist Theoretical Virtues Can Do in Science?

Anderson recognizes that the two virtues: Ontological Heterogeneity and Complexity in Relationship are clearly feminist but are not exclusively feminist because other disciplines share these virtues also. These virtues would drastically change how we approach science from classification to method, and allow us to criticize science in a powerful way.

First of all, these virtues provide us concrete standards of good science, to name a few:

1. “[H]eterogeneity and complexity represent a desirability of human flexibility, autonomy, and creativity…[B]ecause these are valuable potentialities, it is important that our conceptual schemes be able to represent us as having them, if indeed we do” (50). In other words, it is important that we leave room for potentialities in our conceptual scheme (for example, classification) in case that empirical facts fill the gap.

2. The virtues of heterogeneity and complexity require a more qualitative research method that values open-ended, face-to-face interviews rather than a restrictive and distanced one. (51)

Secondly, the above-mentioned standards of good science allow us to criticize science in a more powerful way. If a certain scientific theory does not match the above standards, we can push for a more significant revision.

F. Conclusion

In short, Anderson argues for taking value into consideration in science, and specifically provides two feminist virtues as candidates of important moral and political value that science needs to incorporate. She envisions a revision of scientific justification which requires dual standards—significance and truth.