What is Thomas Kuhn’s picture of scientific practice?
May 5, 2019
Introduction
Unlike many of his predecessors, Thomas Kuhn’s account of science given in The Structure of Scientific Revolutions is historically rooted rather than based on an a priori methodological or semantic criterion. He suggests that science takes place in cycles of conformity to a tradition, accumulation of unexplained anomalies, and revolutions, which yield new traditions. In his text, Kuhn outlines a number of seminal ideas, though many were felt by his contemporaries to be controversial. In assessing his overall picture, I begin by defining key terms that he uses throughout his text. Then, using these terms, I explain the historical advancement of science according to Kuhn. Finally, I defend Kuhn’s claim that scientific paradigms are incommensurable against the criticism that his view provides no reason to believe that science progresses over time.
Summarizing Kuhn’s Account
To begin, I define a few terms that are centrally characterize Kuhn’s account. A paradigm is a set of assumptions, methods, and standards of explanation that guide scientific practice. Only through paradigms can the notions of fact or anomaly be intelligible; fact is only fact in so far as it is deemed admissible truth or assimilated by some paradigmatic understanding of the world. Insofar as an accepted paradigm denies the possibility of an observation, the observation is plausibly erroneous and does not amount to discovered fact. Moreover, anomaly can only be identified by its nonconformity with what a paradigm implies. Normal science is scientific practice that attempts to assimilate observational data into a settled paradigm, an objective that Kuhn analogizes to solving a puzzle. In this endeavor, many scientific practitioners abide by particular rules, explicitly or implicitly, while others roughly and unsystematically reflect their various influences without recourse to rules. When anomaly amasses, scientists, to some extent, start to become worried about the credibility of their paradigm. When trying to account for these anomalies, they start to loosen rules in such a way that their experimentation starts to vaguely resemble pre-paradigmatic inquiry; observation becomes less directed, and hypotheses become more speculative and inventive. Eventually, chaos yields a new paradigm; some novel idea promises a markedly different approach to the problems of the old paradigm. If the initial deficiencies of the new paradigm can be shored up, it displaces the old paradigm and becomes the new mode of normal science in a community of practitioners. It is important to distinguish this account of scientific progress from the prevailing viewpoint before Kuhn published The Structure of Scientific Revolutions: many believed science to be characterized by linear accumulation of knowledge. On the other hand, basing his argument on historical case studies, Kuhn describes scientific progress in cyclical terms: paradigmatic practice yields observed anomalies that, in accumulation, yield crises and revolutions, which advance new paradigms.
Kuhn’s suggestion that scientists who practice according to different paradigms live in different worlds is one that is scarcely explained, but the claim unambiguously stands at odds with scientific realism—the view that the world is mind- and language-independent and that scientific theories describe or explain truly existent phenomena. The scientific realist must hold that all scientists, no matter their operative paradigms, live in the same world, and only their beliefs about the world differ. Peter Godfrey-Smith, too, considers this claim to be tenuous and weakly defended by Kuhn to the extent that he thinks it would have been best left unpublished or lost in a taxi cab. It is clear that this claim factors into Kuhn’s claim that scientific theories are incommensurable, which I shall discuss without delay.
The Incommensurability of Scientific Paradigms
When discussing how one paradigm gains more practitioners during a scientific revolution, Kuhn alleges that competing or successive theories are incommensurable—an appropriation of a term previously used in mathematics to describe dimensions that are related by a constant ratio. Kuhn’s use of incommensurable refers to the inability to compare two conceptually dissimilar frameworks using a uniform terminology. Ian Hacking points out three distinct notions of incommensurability that can be identified in Kuhn’s text. One is topic-incommensurability. While phenomena described by competing theories may be overlapping, they may overlap in such a way that one theory cannot be said to subsume the other. The example Hacking provides is the oxygen theory of burning and bleaching. This theory cannot be said to topically correspond to the prior phlogiston theory because there are clear differences in the phenomena accounted for by each; not all data explained by the phlogiston theory were initially dealt with by the oxygen theory.
Another type of incommensurability is what Hacking calls dissociation. Some theories employ a style of reasoning that is foreign to scientists of another paradigm. For instance, practitioners in Ayurvedic medicine believe that spiritual matters play a vital role in bodily health/functioning and, therefore, that spiritual interventions are just as much candidates for medical treatment as are pharmaceutical or biological interventions. The paradigms of Western medicine are dissociated from this conceptual possibility, and, therefore, Ayurvedic medicine’s postulation that the spirit is a causal force renders it incommensurable with Western medicine; each paradigm has different standards for a satisfactory causal account, and the reasoning employed in Ayurvedic medicine is unintelligible to Western researchers given the standards for admissible reasoning implicit in their paradigm.
As a third type of incommensurability, Hacking points to differences in the meaning of terms that signify unobservable entities or properties. For example, rationality to a neoclassical economist is perhaps different than rationality to a behavioral economist; a behavioral economist conceives of rationality as subject to fickle affective biases and uncertainty whereas a neoclassical economist speaks of rationality as something entirely different and at odds with a psychological account: rationality as consistency of choice, no matter the goal or process. Practitioners in these fields talk past each other when discussing empirical matters because the frameworks used to interpret the data are semantically inconsistent. A behavior identified as rational to the psychology-trained behavioral economist is likely to be different than that which is rational to the neoclassical economist. Therefore, an assessment of rationality by a neoclassical economist cannot be understood or evaluated using the paradigm of a behavioral economist.
One concern often articulated about incommensurability is that it leaves a deep chasm between scientists of different paradigms; the ideas generated by one paradigm are unassailable by another. Moreover, there exists no supra-paradigmatic standard of comparative appeal between two different paradigms. In principle, is it unclear why this implication poses a problem for the scientific enterprise. One might think that without some type of commensurability, it is impossible for science to be assessed as progressive if theories cannot even be apprehended through the lenses of preceding or competing paradigms. I argue forthwith that this conclusion is wrong for a few reasons.
First, while progress might not be assessible through the lens of any particular paradigm, prior or current, it can still be seen in terms of the quantity of problems that the accumulated body of theoretical knowledge has resolved. Through the process Kuhn has described of normal scientific inquiry, observation of anomaly, crisis, and revolution, new problems and areas of understanding have arisen, and a steadily increasing set of paradigmatic toolkits are available—in books, computers, and living scientific traditions—for modern and future scholars to recycle whenever needed. I offer recent attempts to explain gravity as an example of this alleged recyclability of old paradigms. When Newton put forth his theory of universal gravitation, he offered no causal account of gravity; his theory merely posed a mathematical relation with which gravitational force could be calculated. Some scientists of his time rejected his account on the grounds that it did not meet the explanatory standards of their paradigms. While the problem of explaining this mathematical relation is a newer one, one idea that many scientists have devoted attention to is a corpuscular theory of gravity—a novel application of an older paradigm of physics. Before gravity, some variant of corpuscularianism was used to explain light, chemical phenomena, and even planetary orbits. Now, a particle termed the graviton has been discussed as an explanation for the transmission of gravitational force through space. While other applications of corpuscularianism might have proven fruitless from a puzzle-solving standpoint, as a paradigmatic toolbox it has many potential applications, and its contribution to scientific progress is inherent in its past explanatory successes and potential future applications. The mere existence of many paradigms and their unique and overlapping applications constitutes progress, as does their accumulated successes in solving puzzles.
A second measure of progress is in the enduring communal satisfaction with solutions achieved. To the extent that certain areas of scientific knowledge have not been cast aside with paradigms, it seems that progress has been achieved. For instance, the widely accepted description of the water cycle as evaporation followed by condensation and precipitation has remain unchallenged for generations despite paradigmatic shifts in each of the fields in which the relevant phenomena are studied. While related areas of chemistry and physics may have undergone revolutions, it seems that the overall characterization of the water cycle have achieved enduring scientific consensus, even if many details about the water cycle remain to be understood. This example shows that there exist puzzles that scientists consider to be mostly solved (the overall picture is clear), even if small sections remain to be filled in. This accumulative clarity is a crucial manifestation of scientific progress that appears compatible with Kuhn’s account.
Solutional efficiency is the final type of progress that I wish to suggest. By solutional efficiency, I am referring to the strength and simplicity of scientific paradigms. The strength of a scientific paradigm is its analytical potency when applied to the problems in its domain. What motivates a scientist to choose on paradigm over another, according to Kuhn, is its usefulness in solving problems that the scientist is interested in, potentially at the cost of being helpful in solving other problems. This motive exerts a structural, selectionist pressure on paradigms that, over time, discards those that are least functional. Though the problems that scientists are interested in change over time, it seems empirically true that scientists seek to build out new paradigms such that their deficiencies in solving old problems are ameliorated. The process of selecting efficacious paradigms and patching up its holes often causes long-term gains in analytical strength, particularly for the problems that scientists are most interested in at any given time. It seems uncontroversial, for example, that heliocentric models of the solar system provides better insights and more accurate predictions about phenomena in the solar system than did Ptolemaic astronomy, no matter what initial holes existed in Copernicus’s account. In suggesting that scientific paradigms are progressively simple, I do not mean that any newer paradigm is less detailed than its antecedent. Rather, I mean that over time the overlapping networks of living paradigms increase in their collective simplicity because new paradigms add value and potency to old ones, and by virtue of their integrative synergies, less disparate paradigms are needed. Note that much more empirical analysis would be needed to support this generalization, but, for now, let the claim remain a curious suggestion. I offer an illustration to show what I mean. The scientific revolutions spurred by conservation of energy resulted in many paradigms being strengthened by the same theoretical assumption. For example, the ideal gas law, in all of its paradigmatic assumptions, assumes conservation of energy. So, too, do quantum theory, the special theory of relativity, and thermodynamics. Progress can be observed in the increasing theoretical common ground shared by different paradigms, as ideas are being selected that yield greater mileage in a multitude of scientific fields. For these reasons, I see no problem with Kuhn’s claim that scientific paradigms are incommensurable. Even if not identifiable a priori—before a scientific paradigm has been widely adopted and bolstered—, progress results in the long-term when scientists allow their interests and vexations to lead them in directions that seem promising for their puzzle-solving mission.
Conclusion
While authors such as Godfrey-Smith point out the inapplicability of Kuhn’s generalization to many fields such as genetics, Kuhn observes many neglected aspects of scientific progress. Even if his account is not exhaustive, it certainly captures the importance of paradigms and puzzle-solving as a scientific objective. For the aforementioned reasons, I do not consider incommensurability to be a problem. Many of its critics base their objections on the unfortunate implications of the claim on their philosophical commitments: science as progressive, scientific realism, and various others. However, incommensurability is aptly identified in a variety of cases, though perhaps not all cases, and it emphatically does not override claims of scientific progress. Rather, scientific progress ought to be reconceptualized in the manner that Kuhn suggests—as something non-linear but nonetheless identifiable in the increasing breadth of solvable problems and paradigmatic synergy.