Values in science? Diversity, scientific objectivity and criticism

One of the liveliest discussion topics in the philosophy of science in the last three decades is whether non-epistemic (or non-cognitive) values can have a legitimate role in science –that is, not whether they do intervene but whether they should intervene. It is much less controversial that non-epistemic values influence how scientist choose their topics, the social/institutional organization of scientific activity, or how scientific findings are employed in practical applications or policy-making. By a legitimite or illegitimate intervention people usually mean the context of theory appraisal (including data collection and analysis strategies) –the scientific evaluation of the empirical support of scientific claims and their acceptance/rejection.

This philosophical discussion is directly related to several issues that the practicing scientist would also be immediately familiar, such as discussions of diversity/ inclusivity and representation in science, or research on sensitive topics like race.

There has been a significant number of different arguments against the value-free ideal of science (that is, free of non-epistemic values) and in favor of the intervention of non-epistemic values in science. One of the ‘hottest’ takes on this is called the ‘gap’ argument in Elliott’s (2022) Cambridge Element Values in Science. In this post I will try to sketch some preliminary points in defense of the value-free ideal against the gap argument, and do so from the perspective of Popper’s critical rationalism.

The gap argument is chiefly due to Longino (1990). It goes from the problem of underdetermination of hypotheses by data to the claim that science should not be value-free. The problem of underdetermination refers to the impossibility of conclusively refuting or verifying hypotheses based on empirical evidence alone, because we need to make certain auxiliary or background assumptions that link the data to the hypotheses (like how to best measure a concept, how to decide what is causally relevant and irrelevant in an experimental setting etc). Thus, there is an interpretive Spielraum in deciding what to take as evidence for or against a claim, depending on how we choose our auxiliary or background assumptions. And its in this choice where we have a potential (inevitable for Longino) value-encroachment from the non-epistemic realm.

Longino’s suggestion is to abandon the value-free ideal completely and embrace the value-ladenness of the innermost activities of science. This is not an argument against the objectivity of science though. Quoting from Elliott, “Longino argues that the best strategy for maintaining the objectivity and trustworthiness of science is not to try to eliminate values from science but rather to create a social context in which background assumptions and the values associated with them are subjected to critical scrutiny.”

The first claim by Longino is that we should not criticize scientific claims for being value-laden, because this is inevitable.

Firstly, value-freedom in science is not necessarily a descriptive claim. A descriptive claim would say that non-epistemic values never or rarely intervene in scientists’ reasoning. Such a claim should be studied empirically, and I would not be surprised if it turns out to be false currently or in some period in the past. But the question itself is not very interesting from a philosophical perspective, it is probably more interesting to a sociologist or historian. Value-freedom is a normative claim, and people who endorse the value-free ideal can easily accept that non-epistemic values do encroach into science. The catch is that they view such encroachments as irrational influences or biases, which are by default detrimental to scientific objectivity. On the basis of the value-free ideal, we acquire a standing to criticize scientists when they make value-laden choices in the research process or in their appraisal of other’s claims. Under the value-laden model, we would lose this standing.

Secondly, value-free science does not mean value-free individuals. It might as well be the case that it is absolutely impossible for individuals to be value-free. So, we do not have to assume or require that scientists show a skeptical attitude or impersonal detachment (cf. the Mertonian norm “disinterestedness”). On the contrary, if we must accept that individuals cannot be value-free, the only way to scientific objectivity through critical social interactions between scientists can be conceived under the value-free ideal. Critical discussion reveals and problematizes the encroachment of values into scientific reasoning.

A prominent example is Popper’s critical rationalism, which purports that all values introduce bias into epistemic judgments but scientific objectivity arises out of a social process of criticism that involves revealing such biases from diverse standpoints. Value-diversity means that even if scientists are blind to own biases (highly probable), they will not be so to the biases of others. Thus, there will always be others in the scientific community who are not blind to the same biases, if we have viewpoint diversity in science and no systematic obstacles in the way of criticism. If diversity decreases, or some views are shielded from criticism, the likelihood that the biases of certain worldviews will be accepted as part of scientific knowledge through scientific consensus increases, and scientific objectivity suffers.

What may be described as scientific objectivity is based solely upon that critical tradition which, despite all kinds of resistance, so often makes it possible to criticize a dominant dogma. In other words, the objectivity of science is not a matter for the individual scientist but rather the social result of mutual criticism, of the friendly-hostile division of labour among scientists, of their co-operation and also of their competition.

Popper (1994)

The value-free ideal is the only framework in which we can argue for diversity in science on epistemic grounds. When we take the framework of value-freedom away, diversity can only be valuable on other, non-scientific criteria (e.g., social representation), and we do not have an argument anymore for what makes science a special institution. Also, if diversity is not defended by virtue of its epistemic value, as an error-probing social mechanism or on some other epistemic ground, the implementation of diversity promoting practices would be servile to popular political leanings among scientists at the time.

The second claim by Longino is that scientific objectivity depends (or should depend) on critical scrutiny. This is already a meaningful claim in under the value-free ideal, but this is misleading, because Longino  changes the accustomed meanings of both scientific objectivity and critical scrutiny, where neither is grounded in epistemic values (alone). By critical scrutiny Longino does not (and cannot) mean an assessment of background assumptions based on epistemic values, which directly undermines the value-ladenness argument and restores value-freedom. Instead, she does (and must) mean an assessment based on non-epistemic values.

Longino’s second claim actually means:

Scientific claims should be criticized based on non-epistemic values.

Since according to Longino we should not criticize scientific claims for being value-laden (which is inevitable), we can criticize them only for being laden with values we do not endorse. Who would do this criticizing, then? We can either appeal to majority or hegemonic values (social, political, ethical) as some sort of a common sense, or say that everyone (or group) who subscribes to divergent values should voice criticism.

Appealing to majority values has some obvious problems:

Values that are hegemonic in a society change over time.

If we take the values that shape the Zeitgeist as our benchmark in appraising scientific claims, our appraisal of scientific claims becomes essentially time and fashion dependent. This implies that with no change to a given theory or set of data, their evidential relationship would change just as a matter of social change.

Minority values would be disregarded.

This is not a consequence for science itself but for the society in large. Science’s value free-ideal means that science is an impartial arbiter of facts that different social groups can all appeal to in resolving inter-group disputes. If the scientific “facts” are necessarily value laden, being laden with majority or hegemonic values makes science likely an institution of oppression from the viewpoint of minorities. What is worse, minorities lose one of their strongest tools (i.e., scientific facts) to counter oppressive discourses and practices.

The distinction between is and ought would be blurred.

The value-free ideal means that science does not make any normative suggestions. It (ideally) just tells what is the case, what would be the case if we did X, or how best to achieve Y. It does not tell us what should be the case, whether we should do X, or aim for Y. These latter are for the society to decide. Under the value-laden model, we would lose this important distinction.

Appealing to divergent values has other problems:

Scientific objectivity would become parasitic on social consensus.

Decisions to accept or reject scientific claims would require a much broader agreement on the legitimacy of a certain set of values in relation to those claims. Scientific claims may often be highly controversial for a long time until a scientific consensus takes shape on their fate. But when non-epistemic values also take a legitimate seat at the table, the process of reaching scientific consensus would be essentially social/political, and such a consensus is notoriously difficult to reach (without autocratic interventions). If consensus is sought only within groups that comprise scientists with similar views, that would mean (by implication) accepting the legitimacy of “alternative facts” for different political groups.

A further, more problematic implication is that the very distinction between objectivity and consensus becomes vacuous. How can critical scrutiny lead to epistemic objectivity if it does not rely on epistemic values?

The problem of “voice”.

Political representation is already a thorny issue and many viewpoints constantly take the short end of the stick. Following a value-laden view of science, one strategy to facilitate scientific objectivity (in Longino’s sense) could be to diversify the representation of non-epistemic values in the scientific community in a way that potentially goes way beyond the diversity of the broader society. This may sound like a reasonable approach at first sight, but comes with the caveat that most decisions to hire, promote, reward or fund scientists would have to become political, rather than based on their scientific competence or achievements. To maintain a (non-epistemic) viewpoint diversity in science, we would have to actively engineer scientific communities in light of non-epistemic concerns. This would entrust science a social mission which it does not currently have; namely, creating a ‘better’ society within science to solve the problems of the broader society.

At this point we face a still bigger problem: a better society under whose light, for whose problems?

What counts as value diversity, and which values should be included or excluded, is highly context dependent. The values represented (or could be represented) in a scientific community in the US would be very different from those in a scientific community in eastern Europe, southeast Asia or Africa. Will we then operate with different senses of objectivity and social mechanisms for critical scrutiny depending on the context? If so, we can end up with multiple, mutually conflicting scientific truths, and science would possibly lose any claim to universality and authoritativeness. Longino does not argue for such a radical epistemic pluralism, but she opens the way to this implication by not grounding the link between objectivity and critical scrutiny in epistemic values.

To sum it all up, the value-free ideal neither means that scientists are socio-politically, morally, or aesthetically disinterested, nor that they can be. The ideal of value-freedom in science means that what enters the canon as scientific knowledge should be free-from the influence of non-epistemic values. One of the best arguments proposed so far for how science achieves (or can achieve) value-freedom, Popper’s critical rationalism, underlines the crucial role of criticism – Value-freedom in science is a social achievement that is conditional on viewpoint diversity and openness to criticism as an institutionally endorsed virtue. If we lose sight of this ideal, any sanctioning of the encroachment of values into science would lead to undesired consequences for everyone, including the critics of the value-free ideal. Last but not least, neither diversity nor criticism in science can preserve their meaning and significance if we abandon the value-free ideal in science.


Elliott, K. C. (2022). Values in Science. Elements in the Philosophy of Science.

Longino, H. E. (1990). Science as social knowledge: Values and objectivity in scientific inquiry. Princeton: Princeton University Press.

Popper, K. R. [1962](1994). “The logic of the social sciences.” In In Search of a Better World: Lectures and Essays from Thirty Years, 64–81. Routledge.