There are few philosophical distinctions as central to 20th century thought as the distinction between fact and value. Indeed the distinction has proven its utility for enlightenment and emancipation by providing a powerful tool for exposing ideological distortion and political manipulation. When individuals or groups confuse fact with value or value with fact; when they embrace values on the basis of distortion of fact, or embrace facts on the basis of distortions of value; and are unconscious of the irrational mechanisms at work, the fact/value distinction may be deployed as a powerful resource of critique and enlightenment. Certainly this logic played a key role in the process through which many came to question the value of racial segregation or gender hierarchy, on the basis of an empirical/scientific critique of the pseudo-science(s) and factual distortions underlying beliefs in essential difference or inferiority.

Yet, in the last decades of the 20th century, it has become fashionable to claim that facts and values are both socially constructed and depend, for their credibility, on the subjective interests or needs of specific groups - which is what facts and values embody or express, rather than any independent world of nature or morality.

In this essay, I hope to clarify these issues by a rigorous focus on post-Kuhnian insights and debates concerning the relativity of scientific knowledge, and their impact on rationality, objectivity, and realism. From my standpoint, the focus on scientific knowledge will take us to the nub of the question of whether or not science is, or can be, "value-free". For the claim that science, as we know it, is not and cannot be value-free, while gaining an almost theological status for many scholars today, really obscures distinct dimensions of science and the roles different sorts of values may or not may play in each of these dimensions.

We can and should recognize that human beings' value-commitments and interests shape and inform the practices of science(s) in many ways. But many or perhaps all these modes of influence may not imply the value-ladeness of scientific knowledge itself. In particular, we can acknowledge that such commitments typically shape scientists' motives for practicing science, the particular questions and problems they tackle, the concrete concepts, methods, and theories they embrace, and the uses to which scientific knowledge are put, including who has power over its uses, for whose benefit, and at whose expense. Beyond that, we can also recognize that value-commitments and interests inform the direction of scientific funding or patronage more generally, the distribution of credit or recognition for scientific work, the division of labor in science, the institutions which structure scientific work, the education or training of scientists, and the class, racial, ethnic, gender, or religious composition of scientific groups or communities.

If and when we recognize all these ways that groups' interests and values typically shape the practices of science, the whole game seems to be over for many scholars: value-free science is an illusion, end of story! Yet none of these lines or types of normative influence, singly or taken together, necessarily implies that scientific knowledge is itself defined by, or relative to, this or that group's interests or value-commitments. All of the above value dimensions of science aside, the question of whether or not a group succeeds in producing scientific knowledge may be simply and primarily, a matter of whether its theories accord in the right ways with the relevant domains of empirical evidence. Do they succeed in explaining, predicting, or unifying already well-known patterns of phenomena with accuracy? If they do so succeed, it may be this empirical success alone which provides the criterion of scientific knowledge and reality. Value commitments can shape the knowledge we desire, the concepts, methods, or hypothesis at our disposal, our motives, how the 'we' is constituted, etc. - but maybe none of these determine whether or not it is scientific knowledge that we have achieved, when it is indeed achieved.

There are undoubtedly, many scientists, philosopher, and lay people who accept some version of this picture of 'pure' scientific knowledge. For them, when science succeeds, its empirical success alone provides the yardstick of knowledge, and the true 'mirror of nature'. To this extent, in its moments of triumph, science is value-free. Thus, the bottom-line for the defense, or rejection of the value-freedom of science concerns the not-yet-dead, age-old question of the nature or criterion of (scientific) knowledge. Is knowledge itself value-free?

I want to defend the view that scientific knowledge is itself essentially value-laden, but that the value-commitments or interests at stake are, in the final analysis, cognitive or epistemic values, I have three sort of things in mind: (1) normative commitments to (or interest in) the value of certain kinds or patterns of phenomena, and not others, as what the theories of a science must or should explain, predict, etc., in order to constitute knowledge of the domain of that science (e.g., a physics of motion); (2) normative commitments to (or interests in) the value of certain kinds or types of theory and not others, e.g., deterministic or indeterministic, possessing this or that mathematical structure, utilizing this or that ontology, as what theories must or should be like to constitute scientific knowledge of the domain (or in the discipline); and (3) normative commitments to the value of certain kinds of inference, reasoning, or proof and not others (deduction, empiricist induction, the method of hypothesis, the consilience of induction) as how scientific theories must or should be established in order to constitute a knowledge of phenomena. These three sorts of value-commitments are bound to the historical development of science (or natural philosophy) as a tradition in which knowledge-makers necessarily share such value-commitments up to a point, contest them up to a point, and transform them in great and small scientific revolutions. Scientific knowledge thus requires a community of practitioners who implicitly share a certain normative consensus - commitments to the value of certain (but not other) types of phenomena, standards of reasoning, proof or argument, and virtues of theory - as 'the' values or goals which must or should be jointly actualized by their practice if genuine knowledge is to be attained. Clearly, such normative consensus, while necessary for the attainment of scientific knowledge is not sufficient. Success or failure in the achievement of community values and thus scientific knowledge is obviously contingent on many factors and forces at play: among others (1) the way the world is - as realists argue; (2) how effectively scientific groups are able to renegotiate their common values, standards, and goals when they conflict and fail to be conjointly realizable in their practice of inquiry - as social constructivists argue; and (3) the ability of new groups to break with established traditions of inquiry and empower more or less novel practices which devalue at least some of the old sorts of phenomena, theories, and standards of proof, and establish the value of others - as the mark of genuine knowledge.

If I am right, socially shared and enforced value commitments are essential to scientific knowledge. Moreover, fundamental scientific controversies, and large changes in the way science is practiced typically involve conflicts over, or transformations in, the value commitments which dominate genuine knowledge production. The best way to defend my view is to consider some of the objections made against it.

First of all, postmodern scholars of the politics of knowledge may object that my notion of epistemic or cognitive values depends on a false separation between the epistemic and the wider social or political interests that define the cultural contexts of all knowledge-production. On my view, epistemic and cognitive values are social and define the cultural dimension and politics of specific practices or communities of knowledge-production. Furthermore, I readily admit that practitioners have reasons or motivations for their commitments to epistemic values which are rooted in wider interests, goals, or values of the society, epoch, and social/gender/racial situation they occupy. Nonetheless, on my view, it is the epistemic value-commitments I have characterized above that define scientific knowledge and the normative agreements which enable knowledge-making practices to exist. Put differently, it is only when such wider social or cultural interests get expressed and embodied by special groups of knowledge-makers in shared, epistemic, inquiry-defining, value commitments, of the three sorts I discussed above, that scientific knowledge becomes possible. Beyond that, the epistemic values which inform knowledge-making practices may persist or undergo modification somewhat independently of those specific wider social and political interests that initially motivated the credibility of these epistemic values. So while the practices of science are always shaped by a wider context of socio-political interests and values, the value-ladeness of knowledge cannot be reduced to these wider interests.

This has important implications for the effective critique of knowledge-making practices - whether by lay actors or scientific/technical professionals. Effective critique requires a challenge to the way established practitioners define the boundaries of the knowledge they produce - the epistemic values or interests embodied in their practice. Those who advance critique, may justify their allegiance to rival epistemic values by appeal to wider ethical or political concerns; but such concerns need to be articulated at the level of epistemic values, and transformations in the sorts of phenomena, methodologies, or theories, taken as normative for knowledge-making. I will further clarify my argument here below in relation to issues of rationality, objectivity, and relativism.

A second set of objections to my view are inspired by philosophers of science who deny that scientific knowledge, in the final analysis, is relative to local, contingent, and historically variable epistemic values and standards, as I maintain. I will consider three such objections to the value-relativity of scientific knowledge (or, for short, the value-relativity thesis), as follows:

(1.) Against the value-relativity thesis, there are fairly neutral, external, and universal epistemic values and standards in science which provide the criteria of scientific knowledge independent of local context.

(2.) Against the value-relativity thesis, we can embrace a normative naturalism which marshals empirical evidence for evaluating the reliability or effectiveness of local values and methods in attaining the aims of science.

(3.) The value-relativity thesis should be rejected because it is incompatible with the evident rationality, objectivity, progress, and empirical success of science.

For the sake of argument, let me concede the existence of universal epistemic values in all scientific inquiry - empirical success, predictive accuracy, breadth of explanatory scope and unification, simplicity, or problem-solving effectiveness - to take some well-known candidates. At this level of abstraction, such values could plausibly be taken as motivating at least a great deal of scientific inquiry. Nonetheless, such values can only function as criteria of scientific knowledge when they are given flesh, and articulated in terms of the more local values and standards to which communities of scientific practitioners are actually committed. Consider the epistemic value of empirical success, or the ability of theories to "save the phenomena" - perhaps the value or standard with the strongest claim to scientific universality.

Clearly, this goal or norm cannot function as a criterion of knowledge, until questions like the following are answered: What sorts of phenomena are most important to 'save', and which can be neglected? What kind of theory is valuable or useless to 'save' the phenomena? What type of reasoning or proof is valuable, invaluable, or of little value, if the phenomena are to be saved by a theory or empirical law? In the history of science, groups and communities of practitioners of inquiry answer such questions in radically different ways - relativizing the possibility of scientific knowledge to the epistemic values to which such communities are actually committed.

What follows are some historical examples of such disagreements over epistemic values:

My conclusion is that the existence of ‘universal’ epistemic values such as empirical success, or simplicity, or problem-solving effectiveness does not by itself refute my thesis of the value-relativity of scientific knowledge. Of course, my critic may reply that I have missed her whole point - which is to articulate or flesh out a universal value or standard of empirical success (explanation, predictive accuracy, etc.) that is independent of history. Only then can we gain a genuinely universal and normative criterion of knowledge, one based on a philosophical analysis/theory of the very nature of science, not a description of what this or that scientific group actually takes knowledge to be.

Of course this is the traditional method of a priori epistemology which has been effectively challenged by the historicist, sociological, and naturalist turns of our present period. Purely philosophical criteria of scientific explanation, confirmation, verification, etc., always elicit a host of contemporary or historical counter-examples; that is, examples of scientific explanation, reasoning, etc., which we value as such, but fail to fit the a priori philosophical model in question. For my part, I do not draw the conclusion that philosophy cannot say anything important and distinctive concerning scientific inquiry as a whole. Indeed, I take my claim concerning the value-ladeness, or value-relativity, of scientific knowledge, to represent a distinctively philosophical conception of scientific inquiry, through one informed by history and the social sciences.

Among most philosophers today, the rejection of traditional a priori epistemology leads not to my brand of value-relativity, but to science-based, or naturalistic epistemology. Here, the project is to utilize the methods of science, in order to characterize scientific knowledge. Normative naturalists can recognize the value-ladeness of scientific practices while resisting the conclusion that knowledge itself is value-relative. They propose that we evaluate the value-laden practices in science empirically, as more or less effective means to the ultimate aim(s) of science. Scientific knowledge can be neutrally or scientifically characterized as those locally valued methods, standards, aims, theories, etc., which prove in fact to be more effective and reliable than rivals in achieving the aims of science.

Normative naturalism comes in different versions, depending on how the aims of science are characterized. On one view, (the multiplicity view), scientific groups even within the same discipline(s) and area(s) of inquiry, may embrace different ultimate aims - for example, prediction, rather than explanation, or probabilistic explanation, rather than nomological explanation. As I read it, the multiplicity version of normative naturalism concedes most of my value-relativity thesis. If it evaluates the efficacy of local values and methods relative to aims, and allows aims to vary from one scientific group to another (in history, or contemporary science), then, on its own view, scientific knowledge will be relative to the larger epistemic values/aims to which some but not all scientific groups are committed. This conforms to my view of value-relativity.

On a second version of normative relativism, there is but one unitary aim of all science - for example, to discover the truth about nature. Suitably refined, this view challenge my value-relativity thesis. We should refine the view to disambiguate realist from instrumentalist goals. If we fix the aim of science as the attainment of true theories, or maximally reliable theories, then the naturalist can characterize knowledge as whatever local epistemic values, methods, theories, etc., prove in fact to be the most effective means to this aim.

The normative naturalist's language of unitary aim, efficiency, and empirical evidence, promises a value-neutral science of scientific knowledge. In my view, this language is deceptive. It masks the fact that in science, the attainment of theoretical truth (or reliability) is no more a value-neutral unitary aim than empirical success, or saving the phenomena, as I argued above. Suppose in the naturalists' fashion, we set out to determine which of the value-laden practices of scientific groups is in fact most effective in producing true (or reliable) theories of nature or empirical phenomena. What values are embodied in the theories which a group takes to be true? For example, should true theories be explanatory or predictive, or simple and unifying, or deterministic, in each case as some but not all scientific communities or groups have insisted, or denied? Do we value and count as true any theory which succeeds in accounting for or implying already well-known kinds of phenomena, as Mill and other scientists were committed to. Or, do we restrict the value of theories and thus those that we will count as true, to theories which succeed in predicting previously unknown, surprising phenomena, different in kind from those they were designed to explain, as Whewell and others were committed to? And, which sorts of phenomena define the most central, valuable, or significant core of the domain, those which it is most important for a true (or reliable) theory to be true to or of? The normative naturalist seeks to circumvent all of these issues - either obscuring the questions concerning epistemic values they raise, or smuggling in the epistemic values of current scientific communities and practices.

Put differently, normative naturalism requires an epistemology - some account of the standards of reliability or confirmation which the naturalistic scientist of scientific knowledge can use to determine which methods, goals, and theories are the most effective and thus warranted as scientific knowledge. But short of producing and justifying some such universal standard (who has done that?), the naturalists' criterion raises all of the same issues of value-relativity that 'empirical success' raises. On the other hand, once a scientific community or tradition has implicitly committed itself to the value of predicting and explaining certain sorts of phenomena and the value of certain standards of reasoning and theory (or models) for attaining this value, then judgments of reliability, empirical success, accuracy, etc. can be made, and scientific realism can be secured.

Should my value-relativity thesis be rejected because it is incompatible with the rationality, objectivity, progress, and/or empirical success of science, as, for example, anti-Kuhn critics have claimed? The value-relativity, or value-ladenness of scientific knowledge is only incompatible with certain philosophical conceptions of the rationality, objectivity, progress, and empirical success of science. The thesis of value-relativity forces us to rethink, not abandon, these norms or concepts. Indeed this value-laden conception of scientific knowledge opens the way onto a more fruitful model for a critical theory of scientific argument.

But first some preliminary observations:

(a) Obviously, the thesis of value-relativity does not imply an ‘anything goes’ position or one which discounts ‘the claims’ of nature. The work of scientific communities is driven by their attempt to modify their value-laden practices in order to accommodate natural phenomena which so far resist the forms of description, explanation, prediction or control they respectively value.

(b) The thesis of value-relativity allows plenty of room for objective and rational criticism and decision-making. Scientific groups are committed to various epistemic values and aims, as well as the devices, methods, models, techniques, technologies, computations, etc., which embody their values. Obviously, these commitments constrain inquiry and enable scientific practitioners to question or modify some of their commitments (values) in order to more effectively realize others <<Give examples of Newtonian inductivists arguments with Lesage, Hartley on ‘method of hypothesis>>

Scientific rationality, controversy, criticism, and objectivity involves evaluation in which practitioners appeal to some epistemic values or the results attained with them in order to criticize or modify other results or epistemic values.

(c) The thesis of value-relativity does not imply the Kuhnian picture of scientific revolution in which one whole set of epistemic values is replaced wholesale by another. The thesis is compatible with continuities as well as breaks in the epistemic value-commitments of scientific groups within historical contexts and across them. Thus, even though epistemic value commitments are changed in the development of science, it is possible to identify some powerful elements of continuity or overlap. As a result, current or later theories, methods, values, etc. may represent progress over earlier or outdated ones in the straightforward sense that they realize some epistemic values or goals prominent in the past tradition of science as well as its current practice, which were unrealized in past science; this is a value-relative ‘progress’ because past scientific groups may have embraced and more or less realized other epistemic values which have simply been abandoned by later scientific practitioners.

The thesis of value-relativity abolishes unilinear, cumulative, one-dimensional and algorithmic conceptions of scientific objectivity, rationality, and progress. It generates the basis for an alternative conception of these norms. I want to conclude by indicating how this thesis opens the way onto a critical theory of scientific argumentation that can accommodate ‘the politics of scientific knowledge.’

On the thesis I’ve developed, many fundamental arguments over scientific knowledge, among experts and lay people, are normative conflicts concerning epistemic values. On the surface, they often appear as arguments concerning what is and is not known; what can or cannot be known through this or that method, approach, or theory; who is or is not in a position to know or speak authoritatively about some subject matter. On my analysis, conflict over the facts of the matter, or scientific knowledge, often embody rival epistemic value commitments. Extreme relativism or irrationalism only threaten if we assume that epistemic value commitments generally are beyond the scope of reason. What forms of reasoning can be exploited by a critical theory of scientific argument grounded in the thesis of the value-relativity of knowledge? I conclude with my answer:

(1) First, such a theory may be used to expose the rival epistemic value-commitments at stake in scientific controversies that seem to participants, on the surface, to be exclusively about matters of fact or matters of objective scientific knowledge. Such reasoning can reveal that the roots of controversy raise questions not being confronted by the parties to such controversies.

(2) Secondly, a critical theory may seek out the larger political, economic, or cultural interests or values at stake in groups’ rival epistemic value-commitments. Once again, such reasoning may yield the insights that there are good political or social reasons, or that there are bad political or social reasons for embracing particular epistemic values which define the boundaries of science/knowledge in a particular way.

(3) Thirdly, a critical theory of scientific argument confronts the root question of whose political or social interests/values are getting embodied in a scientific practice, and how to determine whether or not they provide good or bad reasons for the epistemic value-commitments embedded in that practice. Once again, such reasoning may reveal that the epistemic value-commitments are rooted in group-interests at the expense of wider, more credible, human values or political ideals. Such reasoning could certainly provide good reasons for revising the epistemic value-commitments embodied in that scientific practice, and generating new boundaries of knowledge in service to the human values or political values at stake in a given domain of scientific inquiry.

The thesis of the value-ladennes of scientific knowledge, as I construe it, opens the way onto a richer terrain of normative reasoning concerning the knowledge-producing practices. It inspires a critical theory of scientific argument which seeks to clarify the epistemic values and larger social interests or political values at stake in knowledge-making practices and the controversies over them. Because commitments to epistemic values and larger social interests or values are amenable to reasoning and the logic of justification, the thesis of the value-relativity of scientific knowledge promises a more, not less, rational practice of inquiry.