Event Title

Using transient underdetermination to undermine value-free science

Presenter Information

David Willmes

Start Date

26-6-2010 1:00 PM

End Date

26-6-2010 2:30 PM

Description

This presentation is part of the Feminism and Empiricism (Quinean Themes) track.

Much ink has been spilled over the problem of underdetermination of theories by evidence. Especially in the science and values debate, in which philosophers and sociologists of science have engaged since the 1970s, the underdetermination thesis has been prominently used to argue against the conception of a value-free science. A reason for the ongoing popularity of the argument from underdetermination can be seen in its targeting the epistemic core of scientific research, the “context of justification”, that is, the evaluation and choice of theories. The argument proceeds as follows:

(1) If science is value-free, then values (at least) play no role in theory choice.

(2) Theories are underdetermined by evidence, because any given body of data can be represented by different, incompatible theories. (Underdetermination thesis)

(3) If theories are underdetermined, there is a gap between theory and data, and in case of theory choice, this gap is filled by employing values.

(4) It follows, by (2) and (3), that values play a role in theory choice.

(5) Thus, science is not value-free.

In philosophy of science, the underdetermination thesis (on which the argument for the influence of values on the justificatory process of science rests) is contentious; there have been many attempts to attack the conception of underdetermination of theories by evidence. For instance, Laudan and Leplin (1991) argue that, although incompatible theories can represent the same data, they are not equally well supported on evidential grounds. Moreover, they contend that the notion of empirical equivalence is defeasible, since the observable consequence s of a theory change in the course of scientific development. Further arguments against underdetermination try to show, epistemologically and by virtue of historical case studies, that what at first sight seem to be two conceptually different, empirically equivalent theories are in fact only variant formulations of one theory, that is, on the structural level both theories are indeed consistent (see, e.g., Norton 2008). Of course, these attacks on the underdetermination thesis add some grist to the mill for those philosophers of science that are arguing against the influence of values in the context of justification.

However, the arguments against underdetermination are itself contestable.Besides this known critique, I want to highlight a common feature of many attacks on underdetermination, namely, that they are dealing with an in principle notion of underdetermination, which states that all theories at any time are underdetermined by the data. It is this conception of underdetermination that has raised most of the attention from philosophers of science. Yet, in concentrating on one specific argument against underdetermination launched by Laudan and Leplin (1991), I will show that the notion of transient underdetermination suffices to make the role of values in theory choice apparent. Transient underdetermination maintains that incompatible theories are equally supported by the present evidence. I think transient underdetermination is part and parcel of scientific decision making and it should thus earn more attention from philosophers of science.

The argument on which I want to concentrate is the defeasibility of underdetermination. As has already been stated above, Laudan and Leplin want to cast doubt on the notion of empirical equivalence by claiming that the observable consequences of a theory must be relativized to a specific state of science. The advice for scientists who are confronted with a case of two empirically equivalent, though rival, theories would simply be to postpone their judgment – hoping that new scientific findings provide new grounds for theory choice.

I will provide two arguments in order to counter this critique. The first one is historical. Taking a glance at the history of science helps us to see that scientists who are confronted with alternative theories did not wait until more evidence is available. As the dispute between Ptolemaic geocentrism, Copernican heliocentrism and Tychonic geoheliocentrism in late 16th century exemplifies, the scientific community eventually accepts only one account if it is confronted with empirically equivalent alternatives (Carrier 2009, pp. 15-16). Since this decision making of scientists in the face of empirically equivalent rival theories is not a unique phenomenon, transient underdetermination is, of course, important; it helps us to detect which values are employed in the justificatory process of science.

Yet from the fact that scientists do make a choice one cannot infer that this is something they should do. This leads us to my second argument. The question remains how we should deal with the advice that scientists who are confronted with empirically equivalent rival theories should better postpone their judgment? This brings us to the following question: How long should they postpone their judgment? Until enough empirical evidence is collected, would possibly be an adequate answer. But, then again, the questioning continues: When is enough empirical evidence collected? Here, a convincing response is not easy. It gets even more difficult if we keep in mind that those who propose to wait until enough empirical evidence is at hand adhere to the notion of a value-free science – in the sense that empirical adequacy suffices to determine theory choice. Thus, the answer that enough empirical evidence would be gathered, if all scientists agree, is not available for them, since this means employing something like a consensus criterion. Maybe answering the question “When is enough empirical evidence collected?” is indeed based on social agreement, but I don’t want to push that point. What seems to be clear is that any noncircular answer must invoke other criteria than empirical adequacy. These other criteria can be characterized as values.

We should hence bid farewell to the idea that empirical evidence unequivocally dictates scientists’ choice of theories. Seen in this light, transient underdetermination is a benefit, because it provides the epistemological tool to make the role of values in science apparent.

This document is currently not available here.

Share

COinS
 
Jun 26th, 1:00 PM Jun 26th, 2:30 PM

Using transient underdetermination to undermine value-free science

This presentation is part of the Feminism and Empiricism (Quinean Themes) track.

Much ink has been spilled over the problem of underdetermination of theories by evidence. Especially in the science and values debate, in which philosophers and sociologists of science have engaged since the 1970s, the underdetermination thesis has been prominently used to argue against the conception of a value-free science. A reason for the ongoing popularity of the argument from underdetermination can be seen in its targeting the epistemic core of scientific research, the “context of justification”, that is, the evaluation and choice of theories. The argument proceeds as follows:

(1) If science is value-free, then values (at least) play no role in theory choice.

(2) Theories are underdetermined by evidence, because any given body of data can be represented by different, incompatible theories. (Underdetermination thesis)

(3) If theories are underdetermined, there is a gap between theory and data, and in case of theory choice, this gap is filled by employing values.

(4) It follows, by (2) and (3), that values play a role in theory choice.

(5) Thus, science is not value-free.

In philosophy of science, the underdetermination thesis (on which the argument for the influence of values on the justificatory process of science rests) is contentious; there have been many attempts to attack the conception of underdetermination of theories by evidence. For instance, Laudan and Leplin (1991) argue that, although incompatible theories can represent the same data, they are not equally well supported on evidential grounds. Moreover, they contend that the notion of empirical equivalence is defeasible, since the observable consequence s of a theory change in the course of scientific development. Further arguments against underdetermination try to show, epistemologically and by virtue of historical case studies, that what at first sight seem to be two conceptually different, empirically equivalent theories are in fact only variant formulations of one theory, that is, on the structural level both theories are indeed consistent (see, e.g., Norton 2008). Of course, these attacks on the underdetermination thesis add some grist to the mill for those philosophers of science that are arguing against the influence of values in the context of justification.

However, the arguments against underdetermination are itself contestable.Besides this known critique, I want to highlight a common feature of many attacks on underdetermination, namely, that they are dealing with an in principle notion of underdetermination, which states that all theories at any time are underdetermined by the data. It is this conception of underdetermination that has raised most of the attention from philosophers of science. Yet, in concentrating on one specific argument against underdetermination launched by Laudan and Leplin (1991), I will show that the notion of transient underdetermination suffices to make the role of values in theory choice apparent. Transient underdetermination maintains that incompatible theories are equally supported by the present evidence. I think transient underdetermination is part and parcel of scientific decision making and it should thus earn more attention from philosophers of science.

The argument on which I want to concentrate is the defeasibility of underdetermination. As has already been stated above, Laudan and Leplin want to cast doubt on the notion of empirical equivalence by claiming that the observable consequences of a theory must be relativized to a specific state of science. The advice for scientists who are confronted with a case of two empirically equivalent, though rival, theories would simply be to postpone their judgment – hoping that new scientific findings provide new grounds for theory choice.

I will provide two arguments in order to counter this critique. The first one is historical. Taking a glance at the history of science helps us to see that scientists who are confronted with alternative theories did not wait until more evidence is available. As the dispute between Ptolemaic geocentrism, Copernican heliocentrism and Tychonic geoheliocentrism in late 16th century exemplifies, the scientific community eventually accepts only one account if it is confronted with empirically equivalent alternatives (Carrier 2009, pp. 15-16). Since this decision making of scientists in the face of empirically equivalent rival theories is not a unique phenomenon, transient underdetermination is, of course, important; it helps us to detect which values are employed in the justificatory process of science.

Yet from the fact that scientists do make a choice one cannot infer that this is something they should do. This leads us to my second argument. The question remains how we should deal with the advice that scientists who are confronted with empirically equivalent rival theories should better postpone their judgment? This brings us to the following question: How long should they postpone their judgment? Until enough empirical evidence is collected, would possibly be an adequate answer. But, then again, the questioning continues: When is enough empirical evidence collected? Here, a convincing response is not easy. It gets even more difficult if we keep in mind that those who propose to wait until enough empirical evidence is at hand adhere to the notion of a value-free science – in the sense that empirical adequacy suffices to determine theory choice. Thus, the answer that enough empirical evidence would be gathered, if all scientists agree, is not available for them, since this means employing something like a consensus criterion. Maybe answering the question “When is enough empirical evidence collected?” is indeed based on social agreement, but I don’t want to push that point. What seems to be clear is that any noncircular answer must invoke other criteria than empirical adequacy. These other criteria can be characterized as values.

We should hence bid farewell to the idea that empirical evidence unequivocally dictates scientists’ choice of theories. Seen in this light, transient underdetermination is a benefit, because it provides the epistemological tool to make the role of values in science apparent.