FACT/VALUE DICHOTOMY

Representatives of modern science and its social institutions have repeatedly claimed that science is value free, and this claim has contributed to marginalizing serious discussion of the relations among science, technology, and values. Lying behind this claim is the philosophical view that there is not just a distinction but a sharp separation, an unbridgeable gap or dichotomy, between fact and value. The supposed fact/value dichotomy arose at the beginning of the seventeenth century, accompanying the early works of modern science, underpinning an interpretation of their character and epistemic status and became part of the mainstream tradition of modern science (Proctor 1991). Prior to that, it was not a major issue in philosophical thinking about science.

Science and Technology as Value Free

The claim that science is value free is that science deals exclusively with facts and—at its core—admits of no proper place for ethical (and social) values. This is not to deny that important relations between science and values exist—for example, that scientific knowledge is a value (even a universal one), that the conduct of scientific research requires the commitment of scientists to certain virtues—such as honesty and courage to follow the evidence where it leads (Merton 1973), and that experimental activities are subject to ethical restraint. Rather, elaborating what it is to keep values out of the core of science, it is to affirm four theses: (1) Scientific knowledge is impartial: Ethical values should not be among the criteria for accepting or rejecting scientific theories and appraising scientific knowledge. (2) Ethical values have no fundamental role in the practices of gaining and appraising scientific knowledge, because the broad characteristics of scientific methodology should be responsive only to the interest of gaining understanding of phenomena. (3) Similarly, research priorities should not be shaped systematically by particular values. The point of both (2) and (3) is that scientific practices are autonomous. (4) Scientific theories are neutral: Value judgments are not among the logical implications of scientific theories (cognitive neutrality); and, on application (e.g., in technology), in principle these theories can evenhandedly inform interests fostered by a wide range of value outlooks (applied neutrality) (Lacey 1999). The theses of impartiality and applied neutrality have counterparts regarding the claim that technology is value free. This claim involves the theses: (1) The characteristic criterion of appraisal for technological objects is efficacy, the factual issue of whether they work or not. (2) Technology progressively makes it possible to effectively achieve more ends, but it does not privilege any particular ends; its products are available to be used to serve the interests of a wide range of value outlooks (Tiles and Oberdiek 1995).

Sources of the Fact/Value Dichotomy

Materialist metaphysics constitutes one source of the fact/value dichotomy. In the words of Alexandre Koyré (1957), one of the most authoritative historians of early modern science, it—by rationalizing the mathematical and experimental character of science—led to the "discarding by scientific thought of all considerations based upon value-concepts, such as perfection, harmony, meaning and aim, and finally the utter devalorization of being, the divorce of the world of value from the world of facts" (p. 4).

According to materialist metaphysics, the "world of facts" is identical to the "world as it really is in itself." This world consists of the totality of the underlying (normally unobservable) structure and its components, processes, interactions, and mathematically expressed laws, whose generative powers explain phenomena, in a way that dissociates them from any relation to human experience, social and ecological organization, or values—the totality of bare facts, purely material facts. On this view, because its aim is to gain understanding of the world, science will attend to grasping the bare facts. Thus, scientific theories should deploy only categories that are devoid of evaluative connotations or implications, such as the quantitative ones (force, mass, velocity, etc.) characteristically used in physical theories. No value judgments follow, for example, from Isaac Newton's law of gravitation, and it makes no sense to ask whether it is good or bad, or whether one ought to act in accordance with it or not. Newton's law expresses a bare fact; faithful to the way the world is, it makes an objective statement.

Representatives of modern science often argue that value judgments, by contrast, do not make true or false statements about objects of the world. Rather they serve as expressions of subjective preferences, desires, or utilities (perhaps grounded in emotions). In this way, the fact/value dichotomy is reinforced by the objective/subjective dichotomy. Science deals with facts; it is objective. Ethics deals with preferences; it is subjective. The efficacy of technological objects, attested to by confirmed scientific theories, stands on the side of facts. Legitimating their uses, however, involves ethical judgments, which cannot be derived from the bare facts that account for the technology's efficacy and the material possibilities that it makes available.

Epistemology is a further source of the fact/value and objective/subjective dichotomies. Scientific epistemologies identify facts—confirmed facts—with what is well supported by empirical data, and the results of established scientific theories. Those that inform technological practices are exemplary instances. Confirmed facts derive from intersubjectivity, that is, replicability and agreement, which cuts across value outlooks and cultural norms. Value judgments are not considered intersubjective. Whereas from the metaphysical source, objectivity derives from faithfully representing objects of the world in statements that express bare facts; from the [other] from the epistemological, it derives from the intersubjectivity of confirmed facts. In practice the two notions of fact tend to fuse together and, from both sources, value judgments appear to be subjective, unlike scientific results that are objective. Hilary Putnam (2002) reviews and criticizes much of the vast philosophical literature on the subjectivity of value judgments.

Finally, logic constitutes a third source of the fact/value dichotomy, and for many philosophers it is the principal one. David Hume, in A Treatise of Human Nature (1739–1740), is argued to have demonstrated an unbridgeable logical gap between fact and value, because factual statements cannot logically entail value judgments; ought is not logically entailed by is. The mark of a fact in Hume's argument is a linguistic one: the role of is and grammatically related verbs, and the absence of such terms as good and ought. Less discussed is the complementary thesis, defended by Francis Bacon in The New Organon (1620), with his famous injunction to avoid "sciences as one would," to avoid inferring is from ought, or good, or from what serves one's interests; for example, it may serve the interest of legitimating the use of a particular technology that it not occasion serious risks to human health, but that interest is irrelevant to determining what the facts are about the risks.

The Entanglement of Fact and Value

Many criticisms have been made of the fact/value dichotomy, including those of pragmatists and critical theorists. But they all come down to one basic argument, that rather than dichotomy there is some kind of entanglement (Putnam 2002) between facts and values. Some (not all) aspects of the entanglement, most of which were discussed by Dewey (1939), are identified below.

NO UNBRIDGEABLE GAP. Many significant factual statements are articulated in scientific theories (such as Newton's law of gravitation). Whether or not a theory is rationally accepted, and thus whether or not statements articulated in it represent confirmed facts, depends on the satisfaction of criteria that require that certain relations obtain between the theory and relevant observed facts. Exactly what these relations should be (inductive, abductive) remains disputed; nevertheless, it is clear that the theories are not logically entailed by the observed facts. the criteria that must be satisfied are those for evaluating the scientific knowledge and the understanding of phenomena represented in theories.

These criteria have been called cognitive values (McMullin 2000); they are a species of values in general, and include empirical adequacy, explanatory power, and consilience. Cognitive values are held to be distinct from ethical, social, and other kinds of values (Lacey 2004), although this is disputed (Longino 1990). Cognitive value judgments concern how adequately cognitive values are manifested in a theory in the light of available observed facts. Soundly accepting that a statement represents a confirmed fact amounts to making the cognitive value judgment that the cognitive values are manifested in the theory to a high enough degree. Far from there being an unbridgeable gap between fact and value, confirmed facts are partly constituted by cognitive value judgments.

FACTS AS PRESUPPOSITIONS AND SUPPORT FOR VALUES. Hume's argument by itself does not rule out that factual statements may provide support for value judgments; otherwise, it would also rule out that observed facts can provide evidential support for facts confirmed within scientific theories, for the fundamental hypotheses of scientific theories are not logically entailed by facts. Logical entailment need not be a particularly important relation in analyzing how facts may support other facts or other kinds of judgments. Consider, for example, the statement: "Recently enacted legislation is the principal cause of the current increase in hunger and child mortality rates." This is a factual statement, because it has the relevant linguistic marks, and empirical inquiry may confirm it to be true or false. At the same time, accepting that it is well confirmed would support holding the value judgment that the legislation should be changed, because, unless there are other factors to consider, it would make no sense to deny that the legislation should be changed, if it is accepted that the factual statement about the causes of hunger has been confirmed. Linked to this, the ethical value of the legislation presupposes that it does not have ethically undesirable causal consequences such as increased hunger (Lacey 2004; for a variant of this argument, see Bhaskar 1986).

SOME SENTENCES MAKE BOTH FACTUAL STATEMENTS AND VALUE JUDGMENTS. Declaring that legislation is the cause of hunger may be intended as the statement of a confirmed fact. Alternatively it may serve to express a value judgment, that is, ethical disapproval of the legislation. The logical and linguistic form of the declaration permits it to be used in either role, showing that there is an overlap of the predicates used in factual and ethical discourse. What have been called thick ethical terms, terms such as honest and unjust (also hunger and high child mortality)—in contrast to thin ethical terms, such as good and ought—may be used simultaneously to serve factual and evaluative ends (Putnam 2002).

Declaring that legislation causes hunger is simultaneously to describe it and normally to criticize it ethically. Using thick ethical terms in factual discourse is no barrier to arriving at results that are well confirmed in the light of the cognitive values and available empirical data; and when such results are obtained, the ethical appraisal is strengthened. Theories that contain such results are not cognitively neutral; they lend support to particular ethical appraisals. Of course, the ethical values of the investigators may explain why they engaged in the relevant research and used the thick ethical terms as their key descriptive categories. Ethical values may influence what facts a person comes to confirm; but they have nothing to do with their appraisal as facts.

SCIENTIFIC APPRAISAL MAY INEXTRICABLY INVOLVE EMPIRICAL CONSIDERATIONS AND VALUE JUDGMENTS. Empirical appraisal never provides certainty; in principle, even the best confirmed statements might be disconfirmed by further investigation. Thus, when a hypothesis is applied, the appraisal made is that it is sufficiently well confirmed by available empirical evidence so that, in considerations about the legitimacy of its application, it is not necessary to take into account that it might be disconfirmed by further investigation, and that, if it were, it might occasion negatively valued outcomes. In the light of this, the standards of confirmation that need to be satisfied depend upon how valuatively significant are these outcomes (Rudner 1953).

MODERN SCIENCE HAS FOSTERED THE VALUE OF EXPANDING HUMAN CAPACITIES TO EXERCISE CONTROL OVER NATURE. Because there are confirmed facts—that is, facts that reliably inform human action—that deploy thick ethical terms, not all confirmed facts are bare facts. This challenges the metaphysical view that the world "as it really is" is identical to the totality of bare facts; and, indeed, it is neither a bare fact, nor a confirmed fact, that the world "really" is that way. Scientists may make the choice to attend only to bare facts. Although this is not the only way to gain factual knowledge, it has generated an enormous amount of knowledge of inestimable social and technological importance. Moreover, because its categories are (by design) chosen to describe facts without the use of thick ethical terms, this knowledge has no ethical judgments at all among its implications. Approaching scientific research, attending only to bare facts, produces results that are cognitively neutral.

At the same time, the contribution of scientific knowledge to enhancing human capacities to exercise control over nature has been highly valued throughout the modern scientific tradition. It has been argued (Lacey 1999) that the approach to scientific research that attends principally to bare facts gained virtual hegemony because of its dialectical links with according high ethical value to enhancing human capacities for control, as well as the exercise of these capacities in ever more domains of life. Bare facts are especially pertinent for informing projects of technological control. Furthermore, sometimes the results of modern science (for example, the developments that have produced transgenic crops) have little application where competing values are held (such as the values of simultaneously gaining high productivity, ecological sustainability, protection of biodiversity, and empowerment of local producers [Altieri 2001]). Thus, while the results gained in this approach of modern science are cognitively neutral, they do not, on the whole, display applied neutrality. (For a variant of this point, see Kitcher 2001.) Humans have considerable knowledge of bare facts (in part) because the values about control are widely held in society and shape scientific institutions. It is not the nature of the world that leads humans to search out such facts but, contrary to the claim of autonomy, a choice highly conditioned by social and ethical values, one that Robert N. Proctor (1991) refers to as "political."

Assessment

Not all the components of the claim that science is value free can be sustained. While there are important results that are cognitively neutral, that is, results that do not logically entail value judgments, in general results do not fit applied neutrality; that is, they are not evenhandedly applicable for a wide variety of value outlooks. Moreover, because applied neutrality does not hold in general, the claim that technology is value free cannot be sustained. There is no objection, then, to engaging in research for the sake of obtaining results that could inform one's ethically favored projects. What confirmed facts are actually obtained reflect these values. That they are confirmed facts does not. The ideal of impartiality remains intact. Ethical values are not among the cognitive values, so that ethically laden commitments (ideological, religious, political, entrepreneurial) are irrelevant to appraising knowledge claims. Science does not need the strong separation of facts and values in order to protect the ideal of impartiality. It needs only a nuanced account of their entanglement.

HUGH LACEY

SEE ALSO Scientific Ethics.

BIBLIOGRAPHY

Altieri, Miguel. (2001). Genetic Engineering in Agriculture: The Myths, Environmental Risks, and Alternatives. Oakland, CA: Food First. Provides an example showing that developments of transgenic crops do not fit the thesis of applied neutrality.

Bhaskar, Roy. (1986). Scientific Realism and Human Emancipation. London: Verso. Qualified argument that factual statements may logically imply value judgments.

Dewey, John. (1939). Theory of Valuation, in Foundations of the Unity of Science: Toward an Encyclopedia of Unified Science, ed. Otto Neurath, Rudolf Carnap and Charles Morris. Chicago: University of Chicago Press. Primary source of arguments for the entanglement of fact and value.

Kitcher, Philip. (2001). Science, Truth, and Democracy. New York: Oxford University Press. Analysis of the interaction between the cognitive and value aspects of science.

Koyré, Alexandre. (1957). From the Closed World to the Infinite Universe. New York: Harper Torchbooks. Influential history of the changes that inaugurated modern science.

Lacey, Hugh. (1999). Is Science Value Free? Values and Scientific Understanding. London: Routledge. Comprehensive critique of the claim that science is value free.

Lacey, Hugh. (2004). "Is There a Significant Distinction between Cognitive and Social Values?" In Science, Values, and Objectivity, ed. Peter Machamer and Gereon Wolters. Pittsburgh, PA: Pittsburgh University Press.

Longino, Helen E. (1990). Science as Social Knowledge. Princeton, NJ: Princeton University Press. Important analysis of the social and value dimensions of science.

McMullin, Ernan. (2000). "Values in Science." In A Companion to the Philosophy of Science, ed. W. H. Newton-Smith. Oxford: Blackwell. Important source on cognitive values.

Merton, Robert, ed. (1973). The Sociology of Science. Chicago; University of Chicago Press. Contains the classic statement of the "scientific ethos," the virtues required for scientific research.

Proctor, Robert N. (1991). Value-Free Science? Purity and Power in Modern Knowledge. Cambridge, MA: Harvard University Press. A comprehensive critique of the thesis of neutrality of science.

Putnam, Hilary. (2002). The Collapse of the Fact/Value Dichotomy and Other Essays. Cambridge, MA: Harvard University Press. Provides many arguments, in the spirit of Dewey, against the fact/value dichotomy.

Rudner, Richard. (1953). "The Scientist qua Scientist Makes Value Judgments." Philosophy of Science 20: 1–6. Famous brief article that argues that value judgments cannot be kept out of science.

Tiles, Mary, and Oberdiek, Hans. (1995). Living in a Technological Culture. London: Routledge. Presents a detailed critique of the claim that technology is value free.