1. SEMANTIC AND METHODOLOGICAL INCOMMENSURABILITY

Kuhn's primary critical target in The Structure of Scientific Revolutions is the common view that scientific inquiry is a linear, cumulative process. He argues that normal scientific inquiry, which necessarily takes place against the background of a shared paradigm that includes standards, methods, and presuppositions, is punctuated by periods of crisis which ultimately result in the abandonment of the prevailing theoretical paradigm in favour of a new set of standards, methods, and presuppositions. While progress can be made in solving problems within a paradigm, Kuhn insists that the revolutionary move from one paradigm to another cannot be made on rational grounds, though this is inevitably the way it seems retrospectively. Kuhn's main argument for this position stems from his view that rival paradigms are incommensurable. Kuhn scholars have distinguished two types of incommensurability that are of particular importance to his account of theory-choice: semantic and methodological.Footnote ¹ In this section, we will outline both of these doctrines and clarify how Kuhn uses them to support his controversial account of theory-choice.

Kuhn's doctrine of semantic incommensurability is a consequence of his semantic holism, according to which, the meanings of the terms in a scientist's vocabulary are determined by the theoretical principles in which they appear. Consequently, when a scientist abandons one set of theoretical principles in favour of another, she thereby changes the meanings of the terms of her theory. Kuhn claims, for example, that the terms ‘space’, ‘time’, ‘mass’, and ‘energy’ mean something entirely different within Newtonian physics than they do in Einsteinian physics. It is on these grounds that he insists that a new paradigm cannot recover the empirical content of its predecessors. Newton's Laws cannot be derived from relativistic dynamics for the simple reason that the laws mean something quite different in relativistic dynamics than they do in Newtonian dynamics; their syntactic form has not changed, but their content has. Furthermore, because there is no theory-independent means of determining the meanings of these concepts, there can be no non-circular grounds on which to argue that the fundamental terms of physics should be defined in the Einsteinian way rather than the Newtonian way. To put the point another way: the acceptance of a paradigm entails the adoption of a conceptual framework that makes theoretically informative observations possible. Consequently, the adoption of a conceptual framework is a precondition of scientific inquiry within a paradigm, and for this reason, is not subject to such an inquiry. If two scientists subscribe to different paradigms, then they work within distinct conceptual frameworks that cannot be evaluated without taking one or the other framework for granted.

Another one of Kuhn's key insights is that scientific theories can differ not only in their results, but in the methods and standards that they use to generate and evaluate results:

New paradigms bring with them new methods that can be used to solve the problems that led their predecessors into crisis. Newton's powerful method of using the relative accelerations of celestial bodies as sources of information about the forces acting on those bodies proved invaluable for solving the problem of the System of the World. However, new paradigms also bring with them new problems: Newton's law of universal gravitation drew many critics because it seemed to violate the Mechanist's principle of action by contact. Indeed, Newton himself searched unsuccessfully for a mechanistic explanation of gravity. For some physicists, this conflict was reason enough to reject the Newtonian program, while for others the program's massive predictive power more than offset its metaphysical difficulties. In hindsight, we think of the Mechanists as being foolishly obstinate and the Newtonians as being enlightened, but Kuhn says that this is the result of the triumph of Newtonian physics, not the other way around. He insists that there can be no rational way to evaluate incommensurable methods and standards. A paradigm's methods and standards make it possible to evaluate theoretical claims, and therefore, they are not themselves subject to a non-circular evaluation within a paradigm. As in the case of conceptual change, there is no objective way of determining the merits of incompatible methods and standards, for such an evaluation must take place within a paradigm that functions by recognizing some methods and standards while eschewing others.

Theoretical paradigms consist, at least in part, of conceptual and epistemic frameworks that cannot be rationally defended because any such defence must be given within the very paradigm whose frameworks are at issue. On the other hand, they are immune from rational attacks, since those attacks must be formulated within a paradigm that takes other conceptual and epistemic frameworks for granted. Consequently, the move from one paradigm to another in times of crisis cannot be the result of rational deliberation:

And like political revolutions, Kuhn says that scientific revolutions are accomplished by means of subjective persuasion, after rational deliberation has failed.Footnote ²

As we have seen, this account of theory-choice is motivated by the doctrines of semantic and methodological incommensurability which seem to entail very specific doctrines of conceptual and epistemic relativism. However, some philosophers have noticed that these doctrines can be generalized in a way that threatens the rationality of science itself.Footnote ³ It is to this threat that we will now turn.

2. GENERALIZING KUHN

One of the necessary features of scientific inquiry, from Kuhn's perspective, is that it is practiced by a community of inquirers that subscribe to a common paradigm. However, this condition is not sufficient to demarcate scientific inquiry from other modes of inquiry, such as philosophy, theology, or pseudo-scientific disciplines, such as astrology and psychoanalysis. This being the case, why should we think of scientific forms of knowledge as being epistemically privileged? Are there not other modes of inquiry whose theoretical infrastructures are semantically and methodologically incommensurable with scientific paradigms? If there are, then Kuhn's arguments would seem to entail the conclusion that no rational argument could establish the epistemic authority of science.

In his discussion of incommensurability, Rorty makes precisely this point by drawing attention to the disagreement between Galileo and Cardinal Bellarmine about the motion of the earth:

Presumably, these arguments will begin by highlighting the fact that Galileo and Bellarmine subscribe to semantically and methodologically incommensurable paradigms.Footnote ⁴ Galileo was one of the first thinkers to have an inertial conception of motion, while Bellarmine recognizes motion only in the Aristotelian sense. Furthermore, Galileo was a skilled mathematician and experimenter who gave quantifiable empirical results the last word on metaphysical questions, while Bellarmine allots this role to Scriptural revelation. Once again, we think of Galileo as having been faultlessly rational and Bellarmine as having been rigidly dogmatic, but on the Kuhnian line of thought, this is because Galileo's scientific paradigm eventually won out. There can be no rational impetus for siding with Galileo in this debate because there is not enough common ground on which to adjudicate this disagreement. Any argument that Galileo was right and Bellarmine was wrong will presuppose a paradigm that more closely resembles Galileo's than Bellarmine's, making it viciously circular.

On this view, science is one possible mode of discourse that takes place against the backdrop of conceptual and epistemic frameworks whose rational justification necessarily involves presupposing those very frameworks. For this reason, science can be deemed neither more nor less rational than modes of inquiry that take place within alternative frameworks that are incommensurable with those in which science is practiced.

The remainder of this paper is dedicated to outlining a line of response to the relativist arguments from semantic and methodological incommensurability. The strategy will not be to argue from within the conceptual and epistemic frameworks of science, but to establish that these frameworks capture presuppositions that are common and indispensable to scientists and non-scientists alike. This being the case, anyone who rejects the basic components of these frameworks is simultaneously undermining the paradigm within which they themselves investigate the world.Footnote ⁵

3. CONCEPTUAL ANALYSIS

Despite his criticism of what he takes to be the Positivist account of theory change, Kuhn's argument from semantic incommensurability is nearly indistinguishable from the Logical Positivists’ doctrine of conventionalism. Like Kuhn, the Positivists claim that primitive theoretical terms receive an empirical interpretation from the fundamental principles of a theory.Footnote ⁶ For example, following Poincaré, they argue that in Newtonian mechanics, the terms ‘force’ and ‘mass’ are empirically defined by the laws of motion. And since dynamical laws, such as the law of universal gravitation, cannot be tested until their constituent terms are properly defined, the laws of motion cannot themselves be subjected to empirical tests (Poincaré Reference Poincaré and Greenstreet1902: Ch. VI; Schlick Reference Schlick and Blumberg1925: Ch. 11). Instead, they are adopted because they provide a convenient framework in which dynamical laws can be formulated and tested. Consequently, the Positivists see the move from Newtonian to Einsteinian physics not as an empirical development, but as a conceptual revolution motivated by purely pragmatic considerations (Carnap Reference Carnap and Gardner1966: 153).Footnote ⁷ The move from an Aristotelian physics to an inertial physics would presumably be understood in the same way, in which case, Galileo's argument against Bellarmine rests strictly on considerations of expedience.

This conventionalist view rests on the assumption that the terms of a scientific theory must be defined by means of a stipulation that governs their use. Yet, as Robert DiSalle points out, definitions are often motivated by a conceptual analysis which reveals a significant feature of the definiendum that is implicit in our established empirical judgements and practices. This sort of analysis figures prominently in Galileo's Dialogue Concerning the Two Chief World Systems, in which Salviato seeks to convince Simplicio that an inertial conception of motion is implicit in his empirical judgements, despite his explicit commitment to Aristotelian physics. As DiSalle explains:

Thus, Galileo is not urging Aristotelians to adopt an entirely new concept on the grounds of convenience; instead, his dialectical argument is drawing attention to the fact that his definition of inertial motion captures an existing, though unrecognized, commitment to treat all instances of uniform motion in the same way. Galileo is not urging Bellarmine to see the world in a completely different way; he is showing him something about the way that he already interprets motion. Not only is this not a viciously circular argument, it is an argument that illuminates the respect in which revolutionary science can result in an improved set of fundamental principles.Footnote ⁸

DiSalle (Reference DiSalle2002a , Reference DiSalle2006) makes a persuasive case that these kinds of dialectical arguments are found throughout the development of space–time theory, especially during periods of revolutionary science.Footnote ⁹ Conceptual analysis can also be seen to play an important role in mathematical theories, as is evident in Frege's logicist reconstruction of arithmetic. Frege subjects empiricist and Kantian conceptions of number to devastating criticisms which show that they fail to recover the generality and objectivity of statements involving number (Frege Reference Frege and Austin1884: Parts I and II). In these arguments, Frege is highlighting facts that his philosophical opponents already know: statements of number apply not only to objects presented in experience or spatio-temporal intuition, but to the creatures of the understanding as well, and indeed to all things thinkable. Frege's motivation for providing definitions of arithmetical concepts in strictly logical terms is that, from his perspective, this is the only way to secure the characteristics that everyone recognizes as belonging to statements of number.Footnote ¹⁰ Furthermore, conceptual analysis is used outside of the exact sciences to generate important results. For example, Nisbett and Wilson (Reference Nisbett and Wilson1977) recommend revising the traditional conception of introspective awareness, whose criterion of application is a subject's ability to produce an accurate report of their cognitive processes. They survey an extensive range of experimental results in support of the view that subjects report on their cognitive processes by consulting a priori causal theories to explain their behaviour rather than inward observation. This being the case, observers can provide reports of a subject's cognitive processes that are often as accurate as the report of the subject themselves. Yet, Nisbett and Wilson suppose, along with all other psychologists, that introspection yields essentially private information about one's mental life i.e. information that the subject has privileged access to that others can know only indirectly. Therefore, they claim that introspection actually takes place when a subject is able to provide a report of their cognitive processes that ‘exceeds in accuracy that obtained from observers provided with a general description of the stimulus and response in question’ (Nisbett and Wilson Reference Nisbett and Wilson1977: 251). Given the empirical evidence, this criterion seems to do a better job of capturing what we all take introspection to be.Footnote ¹¹

These conceptual analyses function as components of non-circular arguments for revolutionary new theories whose force ought to be felt by anyone who uses the concept(s) in question; this will include not only proponents of the new paradigm, but as I have argued, proponents of the old paradigm as well. Thus, while scientists may not be able to argue that their conceptual framework truly cuts nature at its joints without presupposing the framework at issue, they can show that it does a better job of capturing the presuppositions that govern our use of theoretical concepts than its competitors. It is this insight that allows us to block the move from semantic incommensurability to conceptual relativism. As we shall see in the next section, a similar move can be made to block the move from methodological incommensurability to epistemic relativism.

4. METHODOLOGICAL ANALYSIS

An important part of Galileo's case for the Copernican model involved the astonishing observations he made using a telescope of his own creation. To make these findings convincing to his many opponents, Galileo knew that he would have to justify his use of this new instrumentation. How could he be sure that the speckles of light that were visible through his telescope were in fact the heavenly bodies that he took them to be? To settle this question, Galileo used his telescope to make terrestrial observations that could be checked with the naked eye. He could see details of buildings and ships at great distances that could be confirmed upon closer inspection. Kitcher keenly observes that this Galilean strategy of justifying a new epistemic method proceeded ‘… by showing that the telescope would deliver conclusions that could be verified using methods that his contemporaries, including his critics, would accept’ (Kitcher Reference Kitcher2001: 173).

However, this strategy has limitations, as Galileo was to discover. While he and his critics accepted observation as a legitimate epistemic method, they differed on whether or not it is a basic source of information about the world. Galileo, like all true scientists, thought that it is, but his theological critics, such as Cardinal Bellarmine, did not. For Bellarmine, Scripture is the ultimate authority on all questions; the methods of empirical science may expand on and confirm Biblical doctrine, but they cannot falsify it. Indeed, he may insist that the only reason we have to trust our rational faculties is that they have been instilled in us by a benevolent Creator. Here we have a case of methodological incommensurability. Galileo and his colleagues treat observation, memory, and induction as basic epistemic methods, while Bellarmine and his colleagues recognize only Scriptural revelation as a basic method. Galileo could have argued for his scientific methods by appealing to their track record of empirical success, but this would be a blatantly circular argument. Bellarmine could just as easily argue for the authority of the Bible by highlighting those passages which confirm its divine source.

Many naturalists have now resigned themselves to the conclusion that they cannot provide a non-circular argument for their scientific methods of inquiry. Penelope Maddy, for example, says that the naturalist ‘… sets aside this objection with the now-familiar observation that her methods are the best she knows, justified by considerations so-and-so, but many of us aspire to a more conclusive refutation, a case whose force [our opponent] must feel’ (Maddy Reference Maddy2007: 35).Footnote ¹² The purpose of this final section is to outline just such an argument.

To formulate a more conclusive argument, we must recognize that epistemic methods depend on one another not only for the justification, but for their successful application. Galileo's use of the telescope to yield information about the cosmos depends for its justification on (unaided) observation and induction. Galileo's strategy was to put forward an inductive argument for the power of his telescope to deliver information about far-away phenomena by using it to make predictions that could be checked by means of independent observations. However, Galileo's use of the telescope to yield information about the cosmos also depends on observation and induction insofar as he must be able to apprehend the images that the telescope produces and interpret them by appealing to his past experiences; without observation and induction, he cannot so much as generate predictions that can be empirically tested. Thus, Galileo's new method of gathering information about the cosmos depends on observation and induction in the strong sense that they can be used in its justification, and in the weak sense that if the deliverances of observation and induction are unjustified, then so too is Galileo's new method. There are two corresponding senses in which an epistemic method can be basic with respect to an epistemic framework: it is strongly basic if it does not depend on another method for its justification, and weakly basic if it does not depend on another method for its successful application.Footnote ¹³ So, for example, within most scientific frameworks, set theory and induction are foundational, and thus strongly basic, but not weakly basic because they presuppose logic and observation respectively, which are both strongly and weakly basic. With these distinctions in place, we are now in a position to offer a non-circular argument against Bellarmine's epistemic framework.

Galileo's epistemic framework is incommensurable with Bellarmine's because observation and induction are strongly basic in his framework, while only Scriptural revelation is strongly basic in Bellarmine's. Consequently, it would seem as though there is no common ground on which their disagreement about the System of the World can take place. Moreover, there is no argument for the truth-conduciveness of either framework that does not take some subset of these methods for granted. But Galileo does not need to show Bellarmine that his empirical methods are more reliable than Scriptural revelation in order to establish the superiority of his epistemic framework; he need only show Bellarmine that Scriptural revelation weakly depends on these empirical methods. Bellarmine must be able to trust his eyes to accurately apprehend Biblical passages, recall the passages that he has read, and suppose that they remain the same over time.Footnote ¹⁴ If he cannot trust observation, memory, and induction, then he cannot trust Scriptural revelation as a source of information about the world.

I believe that this point can be generalized to yield a non-circular argument for the epistemic frameworks in which scientists investigate the world. Science treats as strongly basic those methods on which the application of epistemic practices – both within and outside of science – essentially depends, including: logical reasoning, observation, memory, inductive reasoning, and abductive reasoning. Whatever one regards as a source of information about the world, be it a religious text, animal entrails, or a crystal ball, one must have epistemic access to the source itself, which requires the use of the aforementioned methods. There is, then, an asymmetrical relation of weak dependence between scientific and anti-scientific methods: anti-scientific methods depend on scientific methods for their successful application, but scientific methods do not weakly depend on anti-scientific methods.Footnote ¹⁵ This is both a distinguishing feature and a principal virtue of scientific inquiry; it is certainly not grounds on which to justify the claim that the epistemic frameworks shared by scientists are truth-conducive, but it does make possible a non-circular vindication of those frameworks, for it follows from this insight that if any epistemic methods are truth-conducive, then the scientist's strongly basic methods are truth-conducive. As we saw in the case of Galileo, it is these methods that scientists use to justify the introduction of non-basic methods of inquiry, such as telescopic observations. Thus, by following this Galilean strategy, scientists give themselves the best possible chance of tracking the truth.

This is a heretofore unrecognized vindication of scientific frameworks that should be distinguished from Reichenbach's somewhat similar pragmatic vindication of induction (Reichenbach Reference Reichenbach1938). Reichenbach argues that if any method tracks natural regularities, induction will be just as reliable because it can be used to track the method's predictive success.

Luper generalizes this argument in favour of a broadly scientific epistemic framework that he calls the ‘ordinary epistemic standard’:

The epistemic frameworks of science are vindicated, according to this argument, because if anti-scientific methods are successful, they can be justified by means of scientific methods. The argument in this section is making a different point, namely that epistemic methods that are strongly basic with respect to scientific frameworks are required to apply anti-scientific methods. In a world ruled by magic, one must correctly see or hear and recollect a spell in order to use it, and suppose that certain strings of words retain a constant magical influence. The result of this argument is the same as Reichenbach's and Luper's: if any epistemic method is successful, the scientist's strongly basic methods will be successful. However, this argument exploits the relation of weak dependence between anti-scientific and scientific methods, rather than strong dependence.

In the previous section, it was argued that conceptual frameworks are justified in science by means of conceptual analyses that seek to reveal the presuppositions that govern our use of theoretical terms. Such analyses provide a principled basis on which to choose between semantically incommensurable frameworks. In this section, it has been argued that a non-circular evaluation of incommensurable epistemic frameworks will involve a methodological analysis that seeks to reveal the presuppositions that govern our epistemic judgements. The result of our analysis has been a vindication of the frameworks in which scientists conduct their inquiries.