Perspectives on Kuhn: Contemporary Approaches to the Philosophy of Thomas Kuhn 3031163702, 9783031163708

This book presents essays and commentaries that continue on Thomas Kuhn’s work from where he left off at the time of his

204 54 2MB

English Pages 194 [195] Year 2023

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Perspectives on Kuhn: Contemporary Approaches to the Philosophy of Thomas Kuhn
 3031163702, 9783031163708

Table of contents :
Preface
Contents
Chapter 1: Kuhn, Coherentism and Perception
1.1 Introduction
1.2 The Coherentist Reading of Kuhn
1.3 A Shortcoming of the Coherentist Reading
1.4 The Input Objection to Coherentism
1.5 A Kuhnian Response to the Input Objection
1.6 Kuhn, the Cartesian Paradigm and the Myth of the Given
1.7 Conclusion
References
Chapter 2: Sankey on Kuhn and Epistemological Coherentism: A Commentary
2.1 Introduction
2.2 Kuhn and Lewis’s “the Given”
2.3 Kuhn and Wittgenstein’s on Certainty
2.3.1 Kuhn’s New Questions
2.3.2 Kinds of Belief
2.3.3 Modes of Change
2.3.4 Kuhn’s Attitude Towards Epistemology
2.4 Conclusion
References
Chapter 3: A Defense of Structure in Structure of Scientific Revolutions
3.1 Introduction
3.2 Is Structure a Contribution to the History of Science?
3.3 Structure and Structures in Philosophy of Science
3.4 What Structure Is Not
References
Chapter 4: A Vindication of Structure in Structure of Scientific Revolutions: A Comment to K. Brad Wray
4.1 Introduction
4.2 Wray vs. Daston
4.3 Kuhnian Historiography
4.4 Structures as Historical Regularities
4.5 Conclusion
References
Chapter 5: Kuhn’s Reconstruction of Structure: The Theoretical Background
5.1 Introduction
5.2 Ethnography and Reconstruction
5.3 What Constitutes a Lexicon?
5.4 A Recurring Outline
5.5 Reconstructing Our History
5.6 Lexicons, Relativism and Truth
5.7 Conclusion
References
Archival Sources
Published Texts and Secondary Literature
Chapter 6: A Role for Cognitive Agents from a Kuhnian Point of View: A Comment to Juan Vicente Mayoral
6.1 Introduction
6.2 The Relevance of Individuals in Kuhn
6.3 A Compatibility Account: Rational and Interpretative Individual
6.4 Learning a Lexicon: Variability and Their Limits
6.5 Reconsidering the Activity of an Interpretive Individual: The Context of Discovery
References
Chapter 7: Incommensurability and Metaincommensurability. Kind Change, World Change and Indirect Refutation
7.1 Introduction
7.2 What Causes Incommensurability?
7.3 What Are the Consequences of Incommensurability?
7.4 How Does Feyerabend’s Conception of Incommensurability Differ from Kuhn’s?
7.5 What Does Incommensurability Imply About Theory Comparison?
7.6 What Do Kuhn’s Views on Incommensurability Imply About Theory Comparison?
7.7 What Do Feyerabend’s Views About Incommensurability Imply About Theory Comparison?
7.8 What Do Howard Sankey’s Views on Incommensurability Imply About Theory Comparison?
7.9 What Does Incommensurability Imply About Theory Comparison? A Response to Sankey
7.10 What Does Incommensurability Imply About Scientific Realism?
7.11 What Does Incommensurability Imply About Scientific Progress?
7.12 What Is the Justification of Incommensurability?
7.13 What Is Metaincommensurability?
7.14 What Causes Metaincommensurability?
7.15 What Are the Consequences of Metaincommensurability?
7.16 What Is the Justification of Metaincommensurability?
7.17 What Can We Conclude About Incommensurability and Metaincommensurability?
References
Chapter 8: The Landscape of a Metaphysical Battlefield: A Comment on Eric Oberheim
8.1 Introduction
8.2 The Exegetic Dimension
8.3 The Phenomenological Dimension
8.4 The Metaphysical Dimension
8.5 Conclusion
References
Chapter 9: The Plausibility of Thomas Kuhn’s Metaphysics
9.1 Introduction
9.2 The Problem: Scientists and the World
9.3 World Change and Shifts of Vision
9.4 A Philosophical Interlude
9.5 Shifts of Vision, Again
9.6 The Historian’s Viewpoint
9.7 “Change of World View” or “Change of World”?
9.8 Conclusion
References
Chapter 10: Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment to Paul Hoyningen-Huene
10.1 Introduction
10.2 First Concern: What If the World Doesn’t Change?
10.3 Once Upon a Time, a World
10.4 Navigating Two Worlds
10.5 Don’t Forget to Look Back
10.6 Second Concern: Same Moon, Different moon
10.7 Layers in the Meaning, Layers in the World
10.8 Give Me the Lexicon and I’ll Give You the Laws
10.9 Ways of Lexicon Making
10.10 Conclusion
References
Index

Citation preview

The Western Ontario Series in Philosophy of Science 84

Leandro Giri Pablo Melogno Hernán Miguel   Editors

Perspectives on Kuhn Contemporary Approaches to the Philosophy of Thomas Kuhn

The Western Ontario Series in Philosophy of Science A Series of Books in Philosophy of Science, Methodology, Epistemology, Logic, History of Science, and Related Fields Volume 84

Series Editors Robert Di Salle, University of Western Ontario, London, Canada Stathis Psillios, University of Athens, Athens, Greece Assistant Editors David Devidi, Philosophy of Mathematics, University of Waterloo, Waterloo, Canada Wayne Myrvold, Foundations of Physics, University of Western Ontario, Ontario, Canada Editorial Board Members John L. Bell, University of Western Ontario, London, Canada Yemina Ben-Menahem, Hebrew University of Jerusalem, Jerusalem, Israel Jeffrey Bub, University of Maryland, College Park, USA Peter Clark, St. Andrews University, St Andrews, UK Jack Copeland, University of Canterbury, Christchurch, New Zealand Janet Folina, Macalester College, Saint Paul, USA Michael Friedman, Stanford University, Stanford, USA Christopher A. Fuchs, University of Massachusetts, Boston, USA Michael Hallett, McGIll University, Montreal, Canada William Harper, University of Western Ontario, London, Canada Clifford A. Hooker, University of Newcastle, Callaghan, Australia Jürgen Mittelstrass, Universität Konstanz, Konstanz, Germany Thomas Uebel, University of Manchester, Manchester, UK

A series of books in Philosophy of Logic, Mathematics and Natural Science, History of Science, History of Philosophy of Science, Epistemology, Methodology, and Philosophy of Cognitive Science. The Western Ontario Series in Philosophy of Science is devoted to studies in philosophy of science broadly considered, including work in the philosophy and the foundations of the particular sciences and the history of the conceptual development of science. The Series includes monographs and collections of papers in these fields.

Leandro Giri • Pablo Melogno • Hernán Miguel Editors

Perspectives on Kuhn Contemporary Approaches to the Philosophy of Thomas Kuhn

Editors Leandro Giri Institute of Philosophical Investigations, Argentinian Society for Philosophical Analysis National University of Tres de Febrero – CONICET Buenos Aires, Argentina

Pablo Melogno University of the Republic Montevideo, Uruguay

Hernán Miguel Argentinian Society of Philosophical Analysis University of Buenos Aires Buenos Aires, Argentina

ISSN 1566-659X     ISSN 2215-1974 (electronic) The Western Ontario Series in Philosophy of Science ISBN 978-3-031-16370-8    ISBN 978-3-031-16371-5 (eBook) https://doi.org/10.1007/978-3-031-16371-5 This work was supported by School of Information and Communication (FIC) at the University of the Republic (UdelaR) National Agency for Research and Innovation (ANII) Sectorial Commission for Scientific Research (CSIC) National Agency for the Promotion of Science and Technology (ANPCyT) Institute of Philosophical Investigations-Argentinian Society for Philosophical Analysis © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

100 years after Thomas Kuhn’s birth, his work continues to be topical and of interest to an important community of researchers interested in the philosophy and history of science. This text is tangible proof of the perennial relevance of the author’s thought. There are many of us who still debate our interpretations and seek to untie the knots of Kuhn’s ideas, and here we take advantage of the excellent historical opportunity provided by the author’s centenary to offer the public a state of the art of the most current discussions in the community of researchers on the subject. In 1962, Thomas Kuhn published his celebrated The Structure of Scientific Revolutions, where he coined notions such as “paradigm,” “incommensurability,” and “normal science.” These notions—encompassed by an unconventional and newfangled view of science—would change the course of philosophy of science in the second half of the twentieth century. Describing this with one of Kuhn’s most influential concepts, we could say that a new paradigm took over most of the academic work done in professional philosophy of science since then. Moreover, Kuhn’s strong influence contributed to consolidating and stimulating novel research areas outside traditional philosophy of science. For instance, the rise in science accounts based on sociology and anthropology corresponds—at least partially—to the gradual predominance of the Kuhnian program both in philosophy of science and social studies of science. We could open a debate to determine if this is a faithful legacy of the image of science advocated by Kuhn throughout his work. It is a fierce debate, mainly because of the very complexity and evolution of Kuhn’s position. The tension between his notion of incommensurability and his ideas of scientific revolution and scientific progress is useful for framing several issues that still require and deserve a response. In the same vein, the debate on Kuhn’s work should consider the stages of his intellectual evolution, mainly as he devoted much of his later work to refining and clarifying many of his most controversial assertions. Maybe he did so because he was aware of how complex his position was and the misunderstandings that could be created. v

vi

Preface

Notions such as paradigm, normal science, scientific revolution, and incommensurability have become so popular that they are now part of our more intuitive picture of science and pervade our more specialized conceptual approaches. On the other hand, as criticism and hostile exegesis have gained ground and conceptual density, finding a critical balance in the current state of Kuhnian thought is essential. In this context, we must also consider the projection of several and diverse research programs aiming to understand science and inspired by Kuhn’s ideas. Therefore, it is relevant to have a forum where such an influential picture of science can be discussed in its roots, structure, and consequences. This was the aim of the II Colloquium in the Philosophy and History of Science Río de la Plata 2018 “Kuhnian Studies: Past, Present and Future” held at the Argentinian Society for Philosophical Analysis (SADAF) in Buenos Aires, Argentina, and at the University of the Republic in Montevideo, Uruguay. The event brought together four prominent worldwide specialists in Kuhn’s work for the first time: Paul Hoyningen-Huene, Howard Sankey, Eric Oberheim, and Brad Wray. They were joined by Spanish-­ speaking specialists such as Juan Vicente Mayoral from Spain; Hernán Miguel, Leandro Giri, and Pío García from Argentina; and Jorge Rasner and Pablo Melogno from Uruguay. The Colloquium mainly reflected on Kuhn’s influence from the second half of the twentieth century to this day, both within and outside philosophy of science. In this way, Kuhnian philosophy engaged in dialogue with humanities, social sciences, and history, as well as with experimental sciences: the original realm of Kuhnian thought. This event coincided with the Spanish edition of Scientific Development and Lexical Change—a collection of Kuhn’s unpublished papers—published by Pablo Melogno and Hernán Miguel and translated by Leandro Giri. These writings are the basis of the Thalheimer Lectures delivered by Kuhn at Johns Hopkins University in 1984. They include essential ideas to understand Kuhn’s mature thought. A considerable effort went into bringing these scholars together in Buenos Aires and Montevideo. We had long and intense debates between the guests and the audience. Several topics related to Kuhnian thought were unpacked for critical examination; the aim was to go beyond the mere exegesis and interpretation of the primary source. The event also focused on various philosophical research programs inspired by Kuhn, programs that have evolved in recent years and still pose challenges and hints that pave the way to a more precise picture of science and scientists’ work. Below we present the results of the debates. These texts are a valuable contribution to contemporary philosophy of science, since they show that Kuhn’s thought is still a pillar for issues of fundamental interest. “Kuhn, Coherentism and Perception” is the title of Howard Sankey’s contribution. Sankey explores the compatibility of Kuhn’s epistemological thinking and a coherentist approach, as Jouni-Matti Kuukanen did previously. Sankey takes Kuukanen’s exploration of Kuhn’s allegedly coherentist thinking as interesting but argues that such characterization is not fully addressed. Although Sankey

Preface

vii

understands the lack of evidence about Kuhn’s coherentism, he provides clues as to how Kuhn would deal with one of the most critical issues in coherentism: the “input objection.” Sankey states that Kuhn was not a coherentist tout court but could have been one without resigning any relevant part of his epistemological theory. In “Sankey on Kuhn and Epistemological Coherentism,” Juan Vicente Mayoral’s commentary to Sankey, Mayoral agrees with Sankey and Kuukanen regarding the existence of some coherentist elements in Kuhn’s epistemology. However, a more plausible hypothesis is that Kuhn deliberately avoided being identified with coherentism or foundationalism. This hypothesis is based on Kuhn’s admitted inspiration from Ludwig Wittgenstein’s epistemology (at least as presented in his classical work On Certainty). Mayoral shows that the way in which Wittgenstein and Kuhn view the change of belief (changing the community’s viewpoint by applying value-based criteria) is incompatible with coherentism as it is not based on proving a new core of belief through doxastic justification (or nondoxastic justification as empiricist foundationalism would require). In “A Defense of Structure in Structure of Scientific Revolutions,” K. Brad Wray examines Lorraine Daston’s interpretation of the current value of Kuhn’s work in the history of science. She points out that the notion of structure in The Structure of Scientific Revolutions is essential because of how popular the term was in social sciences when Kuhn wrote the book. Daston further states that the ideas of a historical structure and long-term historical regularities have been abandoned, both in history and social sciences. Once Kuhn’s notion of structure is framed in this zeitgeist, it can be rejected and becomes useless for the history of science. Conversely, Wray declares that Structure is not primarily a book on the history of science, so the notion of structure is not a tool for historical narrative. Furthermore, Wray aims to show that Kuhn’s use of “structure” is intensely philosophical rather than historical, making his position immune to Daston’s criticism. This is how Kuhn’s notion of structure retains its philosophical legitimacy and conceptual interest. In “A Vindication of the Structure in Structure of Scientific Revolutions,” his commentary to Wray’s text, Pablo Melogno also aims to vindicate the notion of structure in Kuhn’s work through a discussion with objectives analogous to Wray’s, but with differences in method and scope. Melogno examines some implicit assumptions in Wray’s discussion and tries to discuss the Kuhnian notion of structure in the historiographical realm, and not only within philosophy. In “Reconstructing Structure: the Theoretical Background,” Juan Vicente Mayoral reconstructs and explains (mainly from Kuhn’s viewpoint in The Presence of Past Science, his 1987 Shearman Lectures) Kuhn’s lexicon model, which enhances (and does not leave behind) his position in Structure about the nature of the dynamics of scientific change as an alternative to positivistic, relativistic, or pragmatic models. The connections between The Quest for Physical Theory (Kuhn’s 1951 Lowell Lectures), Structure, and later Kuhn works show continuity in his project (although presented with evolving language). This continuity was forged during Kuhn’s transition from physics to history and philosophy of science in the

viii

Preface

late 1940s and he never cast it aside. On the contrary, it only became more sophisticated as Kuhn achieved maturity. Pío García presents “A Role for Cognitive Agents from a Kuhnian Point of View” as a comment to Mayoral’s work. He complements Mayoral’s view by indicating that it could be fruitful to include an analysis from the individual scientist’s viewpoint as a unit in Kuhnian research programs. One of the most influential dimensions in Kuhn’s philosophy is a shift from the individual researcher’s viewpoint—typical of positivistic and Popperian philosophies—to scientific disciplinary communities. However, García argues that Kuhnian philosophy has room for a “compatibility account,” where important topics like the role of individual learning, the shift from the “rational individual” to the “interpretative subject” (who decides based on community shared values but weighing them independently), and the nature of individual processes in scientific innovation, among other topics, can be followed to enrich and fill some gaps in Kuhn’s program. Eric Oberheim’s “Incommensurability and Metaincommensurability: Kind Change, World Change and Indirect Refutation” includes a detailed and in-depth analysis of the historical roots and fluctuations of the notion of incommensurability in both Kuhn and Paul Feyerabend. He argues that discovering incommensurability led both of them (based on Einstein’s philosophical writings) to defend a “Kant-on-­ Wheels” metaphysical position, where ontology is partly defined by the prevailing theories. This kind of metaphysical position is incompatible with scientific realism, which led Howard Sankey, one of its most tenacious defenders, to attack the very notion of incommensurability as Kuhn and Feyerabend presented it. Oberheim states that Sankey—along with many other scientific realists—is mistaken in his general strategy when arguing against incommensurability. It is impossible to understand both Kuhn and Feyerabend’s frameworks from the viewpoint of a scientific realist. This is because incommensurability occurs at the meta-theory level (where notions like “truth,” “progress,” and others describe the structure and dynamics of scientific theories). This type of meta-level incommensurability is called “meta-incommensurability” and precludes fruitful discussions among people with different metaphysical positions unless they become bilingual. According to Oberheim, Sankey’s sin is not being aware of meta-­ incommensurability in sustaining the demise of the incommensurability notion from the scientific realist point of view. Finally, Oberheim also argues that meta-level incommensurability is an indirect refutation of scientific realism, as it is a metaphysical position that cannot formulate the very phenomenon of incommensurability coherently. Leandro Giri, in “The Landscape of a Metaphysical Battle,” his reply to Oberheim, analyzes his work in the context of a long controversy between Oberheim and Hoyningen-Huene and Sankey on the interpretation of the notion of incommensurability and the metaphysical consequences of each interpretation. Giri argues that there is no relevant difference between how they hermeneutically interpret Kuhn and Feyerabend’s versions of incommensurability but rather in what they do with that interpretation. Giri acknowledges the incommensurability and

Preface

ix

meta-incommensurability between the contenders. He also supports Oberheim in that endorsing scientific realism is insufficient to show that Kuhn and Feyerabend were wrong about incommensurability. However, Giri defends the plausibility of Sankey’s weakened version of incommensurability, which enables the use of incommensurability to assess the viability of a metaphysical position. Every viable metaphysical framework should be able to explain incommensurability in its terms, a test that both scientific realism and Kant-on-wheels realism easily pass. According to Giri, this shows that a metaphysical plurality is possible. Paul Hoyningen-Huene, in “The Plausibility of Thomas Kuhn’s Metaphysics,” addresses the issue related to taking the term “world change” instead of “change in the worldview.” He reminds us of the comparison between the Gestalt change and the revolutionary change to underline the differences, as pointed out by Kuhn. He also focuses on the fact that scientists have no unconceptualized framework to see the world or talk about it. The position to take—as if there were a change in the worldview without a world change—is based on the view that epistemology and metaphysics are strictly separated. But the “principle of economy” proposed by Kuhn is critical for analyzing the way scientists can refer to the existing things in the world. Hoyningen-Huene also considers the point of view of historians of science and analyzes how plausible the Kuhnian way of talking about world change is. His analysis includes the reasons that scientists have to take something as real, given how they see and detect things or because there is a “considerable number of empirical phenomena that can best be understood as causal effects of existing unobservable objects.” Thus, world change is the result of scientists’ ontological commitment when talking about the world, and this scenario is not completely captured by the term “change in the worldview.” In “Seeing, Talking and Behaving... Ways of Inhabiting the World,” Hernán Miguel presents two concerns regarding the issues addressed by Paul Hoyningen-­ Huene. First, he focuses on how scientists engage in their research according to the current paradigm. This is how their behavior, and even that of non-scientific people, must be considered to understand that they are in a world and behave in that world, apart from seeing that world and talking about it. In this way, behavior analysis helps support use of the term and understanding the change as a “world change” instead of a “change in a worldview.” Second, Miguel analyzes the preservation of terms after a revolution, although those terms are related to nature in different ways through different paradigms. In this regard, he analyzes the problem of the features associated with the definition and identification of objects as corresponding to a node in the lexical network. Scientists must make decisions when selecting these features and this task is part of the process to obtain a new lexical network used to understand and describe the world. The points made on these two topics help cast light on conversations about world change. We would like to thank the institutions which made all this possible: in Uruguay, the School of Information and Communication (FIC) at the University of the Republic (UdelaR), the National Agency for Research and Innovation (ANII), and the Sectorial Commission for Scientific Research (CSIC). In Argentina, the

x

Preface

Colloquium was supported by the National Agency for the Promotion of Science and Technology (ANPCyT) and the Institute of Philosophical Investigations-­ Argentinian Society for Philosophical Analysis (IIF-SADAF, National Council for Scientific and Technological Investigations, CONICET). Buenos Aires, Argentina

Leandro Giri

Montevideo, Uruguay

Pablo Melogno

Buenos Aires, Argentina May 2022

Hernán Miguel

Contents

1

 Kuhn, Coherentism and Perception ������������������������������������������������������    1 Howard Sankey

2

Sankey on Kuhn and Epistemological Coherentism: A Commentary ����������������������������������������������������������������������������������������   15 Juan V. Mayoral

3

 Defense of Structure in Structure of Scientific Revolutions��������������   25 A K. Brad Wray

4

A Vindication of Structure in Structure of Scientific Revolutions: A Comment to K. Brad Wray ����������������������������������������������������������������   41 Pablo Melogno

5

Kuhn’s Reconstruction of Structure: The Theoretical Background������������������������������������������������������������������   53 Juan V. Mayoral

6

A Role for Cognitive Agents from a Kuhnian Point of View: A Comment to Juan Vicente Mayoral����������������������������������������������������   83 Pío García

7

Incommensurability and Metaincommensurability. Kind Change, World Change and Indirect Refutation������������������������   93 Eric Oberheim

8

The Landscape of a Metaphysical Battlefield: A Comment on Eric Oberheim��������������������������������������������������������������  127 Leandro Giri

xi

xii

Contents

9

 The Plausibility of Thomas Kuhn’s Metaphysics����������������������������������  139 Paul Hoyningen-Huene

10 Seeing,  Talking and Behaving… Ways of Inhabiting the World: A Comment to Paul Hoyningen-Huene������������������������������  155 Hernán Miguel Index������������������������������������������������������������������������������������������������������������������  183

Chapter 1

Kuhn, Coherentism and Perception Howard Sankey

1.1 Introduction In the latter half of the twentieth century, foundationalist approaches to epistemology and philosophy of science were widely rejected in favour of holist and coherentist approaches. Kuhn was not oblivious to this trend. Indeed, Kuhn may be regarded as a contributor to this anti-foundationalist tendency. In this paper, I wish to consider the extent to which Kuhn’s epistemological thinking was coherentist in nature. This is a task that has already begun in the work of Jouni-Matti Kuukkanen. However, I wish to go beyond Kuukkanen by raising a number of questions that he does not address. In addition, I wish to focus specifically on epistemological questions rather than on issues relating to the theory of truth to which Kuukkanen devotes considerable attention. The structure of the paper is as follows. First, I present Kuukkanen’s coherentist interpretation of Kuhn. I regard the interpretation as a promising one that is worthy of further exploration. However, in the next section, I argue that the interpretation has a crucial shortcoming. It fails to show that Kuhn rejects non-doxastic justification in favour of coherence relations among beliefs. In the fourth section, I present a standard objection to coherentism that any adequate formulation of the doctrine must confront, the “input objection”. I then frame the objection in terms of Kuhn’s model. In the fifth section, I argue that Kuhn’s account of science is able to satisfactorily deal with the input objection. In the sixth section, I explore the relationship between Kuhn’s treatment of perception and the anti-foundationalist tendency that contributed to the rise of coherentism. Kuhn was critical of the idea of the given. But I will argue that he did not reject the given in a sense that entails commitment to coherentism. Finally, I briefly summarize the outcomes of the discussion. H. Sankey (*) Melbourne University, Parkville, VIC, Australia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 L. Giri et al. (eds.), Perspectives on Kuhn, The Western Ontario Series in Philosophy of Science 84, https://doi.org/10.1007/978-3-031-16371-5_1

1

2

H. Sankey

1.2 The Coherentist Reading of Kuhn My point of departure is a paper by Jouni-Matti Kuukkanen which makes the case for a coherentist reading of Kuhn (Kuukkanen 2007). Kuukkanen argues that Kuhn mistakenly rejected the correspondence theory of truth because he wrongly understood it in epistemic terms.1 He suggests that a coherentist reading of Kuhn may enable Kuhn’s views to be reconciled with a convergent realist view of scientific progress.2 But my primary interest here is in neither of these points. What interests me is the case that Kuukkanen makes that Kuhn’s view may be understood in terms of a coherentist account of epistemic justification. Kuukkanen does not claim that Kuhn explicitly endorsed a coherentist epistemology. Rather, the coherentist reading of Kuhn is “an extension of his philosophy” which Kuukkanen believes “does not distort his thinking” (2007, 559). Kuukkanen points to several factors which set the stage for his coherentist reading of Kuhn (2007, 557–9). Kuhn’s adoption of a historical perspective on the nature of science formed part of his rejection of foundationalist forms of empiricism such as logical positivism.3 The historical perspective led Kuhn to see scientists as working with historically situated sets of beliefs that are in place at various stages in the history of science. Such historically situated sets of beliefs form the epistemic background against which specific beliefs and theories are evaluated. The focus of the evaluation of belief is not justification of an individual belief but change of belief or modification of the set of beliefs. As Kuukkanen explains: “The focus is not on the evaluation of beliefs as such, because the whole inherited system of belief is more or less just taken for granted, as if presumptively justified, making it pointless or even impossible to ask justification of individual beliefs.” (2007, 558). On what basis are beliefs to be evaluated? Kuukkanen takes Kuhn to deny that beliefs may be directly compared with reality to determine correspondence (2007, 557–8). So, truth cannot serve as a criterion of appraisal in the evaluation of belief. Instead, beliefs are evaluated with respect to their compliance with epistemic criteria or values other than truth.

 Kuukkanen goes too far in agreeing with Kuhn when he writes that Kuhn “correctly maintained that it is impossible to evaluate correspondence between beliefs and reality” (2007, 556). We make evaluations of correspondence in an entirely routine way. For example, I may check to determine whether my belief that I am now typing on a computer keyboard corresponds with reality, and in fact it does so correspond. The interesting question is not whether we do make such evaluations, but whether we may do so with certainty or in a way that is neutral with respect to competing theories or viewpoints. 2  This aspect of Kuukkanen’s view has been the topic of an earlier exchange (ŠeŠelja and Straßer 2009; Kuukkanen 2009). 3  Kuukkanen names logical positivism and empiricism as the main examples of “epistemological foundationalism” against which Kuhn reacted (2007, 557–8). It is important to note, however, that the contrast between foundationalism and positivism, on the one hand, and coherentism, on the other, requires qualification. Some positivists, e.g. Otto Neurath, flirted with a coherentist theory of truth (see Hempel 1935 for discussion). 1

1  Kuhn, Coherentism and Perception

3

Apart from these stage-setting factors, Kuukkanen specifically bases his interpretation on two key aspects of Kuhn’s view. The first aspect is an epistemic conservatism which Kuukkanen finds in Kuhn’s account of belief change. The second is the role played in Kuhn’s account by the idea that science is essentially a problem-­ solving activity.4 The conservative aspect of Kuhn’s view relates to the way in which an existing body of beliefs is taken by scientists to be presumptively justified. Rather than wholesale change of belief, scientists seek to improve the belief system in a piecemeal manner.5 Where a system has a high degree of coherence, there is no incentive to alter it in fundamental respects. Kuhn’s focus on change of belief rather than justification of individual beliefs also has a coherentist rationale. As Kuukkanen explains: “If our beliefs are holistically justified, that is, if the beliefs, in a set, mutually justify each other, then we have to understand the whole set as given and try to improve it, rather than to try to find a justification for individual beliefs on a one-by-­ one basis or (even less) to start the construction of a new system from scratch.” (2007, 559). Scientists work with a mutually supporting set of beliefs that are “already in place” in a historical situation. In such a context, questions of the justification of belief relate to the rationality of altering beliefs in a way that may be justified with respect to the background system of belief. It is the existence of a historically given set of background beliefs, and the need to justify belief change with respect to that set of background beliefs, that gives rise to the conservative element in Kuhn’s view.6 As for the role played by problem-solving in science, Kuukkanen sees this as closely connected with the criteria of theory-choice. In work subsequent to Structure, Kuhn sought to clarify his view of the rationality of theory-choice by characterizing a set of non-algorithmic values that inform scientists’ decision-making in the context of theory-choice (e.g. Kuhn 1977). The values include accuracy, consistency, breadth, simplicity and fruitfulness. Kuukkanen understands the values employed in theory-choice as factors which contribute to judgements of the problem-solving capacity of theories. They are the standards that scientists use to determine the  In Structure, Kuhn distinguishes between two main kinds of problems. On the one hand, there are the puzzles which are the focus of ordinary research activity in normal science (e.g. 2012, 36). On the other hand, there are the anomalies which arise in normal science, which resist solution and may ultimately give rise to crisis and revolution (e.g. 2012, 82–3). 5  This is not to deny, of course, that there may be scientific revolutions. To put it in terms of Kuhn’s model of scientific change, there is a tendency on the part of normal scientists to resist fundamental alteration of paradigm and to make instead piecemeal adjustments of the existing paradigm. Revolutionary displacement of paradigms only occurs once a candidate paradigm capable of resolving crisis-inducing anomalies has been developed. 6  Kuhn’s views are often associated with those of Paul Feyerabend. But the conservatism of Kuhn’s approach may constitute a point of difference between their views. In apparent recognition of the conservative element in Kuhn’s view, Feyerabend attributes to Kuhn a “principle of tenacity”, which he contrasts with his own “principle of proliferation” (Feyerabend 1970, 203–5). 4

4

H. Sankey

adequacy of a proposed solution to a problem. When a scientist employs the values in the context of theory-appraisal the scientist is conducting an evaluation of the problem-­solving effectiveness of the theory. Kuukkanen sees a close fit between the problem-solving conception of science and coherentist epistemology (2007, 561). The existence of anomalous phenomena that are unable to be explained on the basis of a set of beliefs will reduce the coherence of the set of beliefs. Solving a problem may increase the coherence of the set of beliefs. The value of consistency promotes coherence directly since coherence requires consistency. The values of simplicity and breadth promote coherence because more powerful explanations reduce the number of sub-systems within a system of beliefs, thereby increasing overall coherence of the system. The value of accuracy serves to reduce the number of anomalies facing a system since accuracy indicates agreement between theory and observation. Fruitfulness may conduce to future increase of problem-solving capacity, leading to an increase in overall coherence. Thus, Kuukkanen suggests that “all criteria are linked either directly or indirectly via problem-solving to coherence, which makes Kuhn’s philosophy consistently coherentist.” (2007, 560).

1.3 A Shortcoming of the Coherentist Reading Kuukkanen makes a suggestive case for a coherentist reading of Kuhn. While I agree that there are coherentist elements in Kuhn’s view, I do not find the case fully compelling. The reason is that there is an essential ingredient of coherentist epistemology which is lacking from Kuukkanen’s account. Without this ingredient, Kuhn’s position fails in a strict sense to be coherentist. The key issue that separates coherentism from foundationalism in epistemology relates to the possibility of a non-doxastic source for epistemic justification.7 For the foundationalist, justified basic beliefs receive non-inferential justification from a non-doxastic source, such as sense experience. Unlike basic beliefs, non-basic beliefs receive inferential justification from a doxastic source, namely from other beliefs. By contrast with foundationalism, the coherentist denies the possibility of a non-doxastic source for epistemic justification. In effect, there are no basic beliefs for the coherentist. All justification is inferential justification. Beliefs receive

 The need to avoid a non-doxastic source of justification within a coherentist epistemology is illustrated by Laurence BonJour’s treatment of observational beliefs. For BonJour, the justification of an observational belief is not based on the experience that produces the belief. Instead, the justification derives from beliefs about the circumstances in which the belief is produced (1985, 118). In her criticism of BonJour’s coherentism, Susan Haack focuses on the role played by experience in the production of observational beliefs. Her foundherentism departs from coherentism precisely in granting a justificatory role to the experience that prompts such belief (e.g. 1995, 60). 7

1  Kuhn, Coherentism and Perception

5

justification by way of inferential relations from other beliefs within the belief system. For the coherentist, no beliefs receive justification from a non-doxastic source.8 This is the crux of the issue between coherentism and foundationalism. Kuukkanen fails to show that Kuhn rejects the possibility of a non-doxastic source for epistemic justification for basic beliefs. For this reason, he fails to show that Kuhn is in a strict sense a coherentist. Still, Kuukkanen is on safe ground when he describes his coherentist reading as an “extension” of Kuhn which “does not distort his thinking”. There are coherentist elements in Kuhn’s thought. Indeed, Kuukkanen’s case for the coherentist reading is sufficiently strong to warrant further exploration of the coherentist tendency in Kuhn’s thought. In the next section, I will commence this task by bringing Kuhn’s view into contact with a major objection to coherentism.

1.4 The Input Objection to Coherentism Though not entirely compelling, the coherentist reading of Kuhn warrants further investigation. To initiate this task, I will first present a central problem which faces coherentism. The question that I wish to consider is whether Kuhn’s position has the capacity to deal with this problem. There are a number of problems that an adequate coherentist epistemology must address. I focus here on a problem that is of particular relevance to the epistemological aspects of Kuhn’s position with which I am currently concerned. Following Laurence BonJour, I refer to the problem as the “input objection”.9 It is also known as the “isolation objection”. According to a coherentist account of epistemic justification, a belief is justified by a relation of coherence which obtains between the belief and surrounding beliefs within a belief-system. The problem is that a system of beliefs might be internally coherent even though the system of beliefs fails to have any contact with the world. There could be a set of beliefs which is internally coherent even though there is no

 This may require slight qualification. For the coherentist, a belief is justified by coherence with other beliefs within a surrounding belief-system. Given that the source of justification is coherence with other beliefs, there is a doxastic source of such justification. But it is important to bear in mind that the justification is understood to be inferential. Inferential relations (e.g. deductive or inductive relations) are non-doxastic relations. This appears to suggest that the source of justification contains both doxastic and non-doxastic elements. One way for the coherentist to avoid this implication is to require that the inferential relationships be believed to obtain by the epistemic subject. 9  Apart from the input objection, BonJour considers two others (1985, 107–10). One is the alternative coherent systems objection: there may be mutually incompatible but internally coherent systems of belief; hence, there is no reason to hold one belief rather than its opposite, since both the belief and its opposite will be justified within some system of belief. The second objection is the problem of truth: the internal coherence of a system of beliefs provides no reason to think that a belief is true, where truth is understood in a correspondence sense. 8

6

H. Sankey

input from the external world which contributes to that set of beliefs. Given that the set of beliefs is internally coherent, beliefs within the system of beliefs would be justified by way of their coherence with other beliefs in the belief-system. Yet, surely, a crucial ingredient is lacking with respect to the justification of beliefs within the coherent system. What is lacking is input from the world outside the system of belief. To respond to the input objection, the coherentist must show that the external world has an impact on the system of beliefs.10 The question that I wish to pursue in relation to Kuhn is whether his account of science can meet the input objection. Kuukkanen has correctly identified a coherentist tendency in Kuhn. I wish to go a step further than this. I wish to ascertain, not just whether there are coherentist elements in Kuhn’s work, but whether his account of science satisfies a minimal condition of adequacy for a coherentist epistemology. If Kuhn is unable to deal effectively with the input objection, his account will fail to have what it takes for an adequate coherentist epistemology. To set the question up, let us first frame the input objection in Kuhnian terms. According to Kuhn, in normal science scientists devote themselves to solving puzzles on the basis of a shared paradigm. On occasion, anomalies which resist solution within the paradigm give rise to a crisis. This may result in revolutionary overthrow of the paradigm, and the resumption of normal science under a new paradigm. On a coherentist reading of Kuhn, scientists maintain an internally coherent set of beliefs throughout this process. In normal science, a coherent set of beliefs is maintained because the puzzles that arise during normal scientific research are all solved using the resources of the paradigm. In the context of a revolution, scientists who adopt the new paradigm seek to restore the coherence which was undermined by the anomalies that confronted the previous paradigm. The input objection may now be presented in terms of Kuhn’s model. On the coherentist reading, scientists maintain a coherent set of beliefs throughout normal science and revolutionary transition between paradigms. Is there any scope in Kuhn’s model for input from the world into the internally coherent beliefsystems of scientists? Or are the systems of belief cut off from reality altogether? If there is no input from the external world, then Kuhn’s view falls prey to the input objection.

 I assume, of course, that the coherentist is not an outright idealist for whom the world and belief-­ system are one and the same thing. It is precisely because BonJour wishes to combine a coherentist account of epistemic justification with realist commitment to a mind-independent reality that the input problem arises for his coherentist theory (e.g. 1985, 108). On the assumption that Kuhn is not an idealist for whom the world collapses into belief, the problem arises as well for a Kuhnian coherentist. 10

1  Kuhn, Coherentism and Perception

7

1.5 A Kuhnian Response to the Input Objection The answer that I wish to propose is that Kuhn is indeed able to deal with the input objection in terms of his model. My answer will turn on two considerations. The first relates to the role of problem-solving in Kuhn’s model of scientific change. The second relates to the causal role played by the world in the production of scientists’ perceptual experience. Starting with the issue of problem-solving, we have seen that Kuukkanen takes there to be a close connection between the problem-solving account and coherentism. But, apart from that, the idea that science is a problem-solving enterprise provides a basis for a Kuhnian response to the input objection. This may be seen, in the first place, from the problems to which normal scientific research is dedicated. Normal scientific puzzles involve the precise measurement of significant facts (e.g. physical magnitudes, stellar positions), improvement of fit between prediction and observed fact (e.g. detection of the neutrino), and articulation of the paradigm (e.g. work on the value of the gravitational constant after Newton) (Kuhn 2012, 25–8). All such normal scientific puzzles involve empirical information that stems from the world investigated by scientists, with the result that substantial scope exists for empirical input from the world into the belief-systems of scientists. The point is even more striking in the case of anomalies. For Kuhn, an anomaly is an empirical phenomenon that fails to fit with the existing paradigm, and whose existence may prove impossible to reconcile with the paradigm. As such, an anomaly introduces incoherence into the belief-system of scientists, since the anomaly is incapable of being explained within the prevailing belief-system. Given this, the anomaly constitutes an imposition of the world on the paradigm. Recognition that such an unexpected phenomenon exists is the result of input into the belief-system from the world outside that system of belief. The problem-solving apparatus of Kuhn’s model requires that a role be played by the external world in the production of problems. This indicates that Kuhn’s model has the capacity to deal with the input objection, since problems are injected into the belief-system from the world outside of the belief-system. This point may be reinforced by considerations about the causal role played by the world in the perceptual experience of scientists. For this point, I will draw upon Paul Hoyningen-Huene’s discussion of this aspect of Kuhn. In his neo-Kantian interpretation of Kuhn, Hoyningen-Huene introduces a distinction between the phenomenal world which varies with paradigm and the world-­ in-­itself which remains stable throughout change of paradigm (1993, 33–5).11 On Hoyningen-Huene’s analysis, Kuhn implicitly allows the world-in-itself to play a causal role in the production of scientists’ perceptual experience (1993, 34).

 The contrast between the neo-Kantian view and my own scientific realism is, of course, the focus of significant disagreement between Hoyningen-Huene and myself. However, for present purposes, I set this issue to one side. As far as the input objection is concerned, both the realist and the neo-Kantian may grant a role to the world as a source of input into the system of belief. 11

8

H. Sankey

As evidence for this, Hoyningen-Huene cites several passages in Structure in which Kuhn credits the world-in-itself with partly determining a scientist’s experience (e.g. Kuhn 2012, 112). In work published in the early 1970’s, Kuhn sought to clarify his view about the relation between world and experience by employing the notion of a stimulus (Hoyningen-Huene 1993, section 2.2). The stimulus is the external cause of sensory experience. While the stimulus that a scientist receives from the world does not fully determine the content of perception, it does play a causal role in producing the sensation. Thus, both in Structure and in subsequent work undertaken to clarify some aspects of Structure, Kuhn attributes a causal role to the world in the production of experience. Given the causal role played by the world in the production of experience, it seems clear that scientists receive input from the world. Hence, they are not, as the input objection suggests, cut off from the world with no possibility of empirical input into their belief-systems from the world around them. The coherentist must tread carefully around the input objection. To meet the objection, a role must be granted to the world, so that the system of beliefs is not cut off from the world. At the same time, input into the belief-system from the external world must not be granted a role in the justification of belief. To allow empirical input to contribute to justification would be a departure from a coherentist account of justification on which justification must have a doxastic source. To allow empirical input to play a justificatory role would be to incorporate a foundationalist element into the theory of justification. Nothing about the two points I have made on behalf of Kuhn in response to the input objection entails that input from the external world must contribute to epistemic justification. As for the first point, the fact that the world plays a role in the generation of problems does not entail that the input from the world contributes to the justification of scientists’ beliefs. As for the second point, the coherentist may allow that the world plays a causal role in the production of sensory experience while insisting that justification consists in coherence relations between beliefs. If I am right about the two points, then I think it is safe to conclude that Kuhn’s model of theory-change has the resources to meet the input objection.

1.6 Kuhn, the Cartesian Paradigm and the Myth of the Given In the second half of the twentieth century, coherentism emerged as part of a reaction against empiricist forms of foundationalism. An important part of the reaction was rejection of what Wilfrid Sellars referred to as the “myth of the given”. Kuhn’s work forms part of the reaction to empiricist forms of foundationalism as well as against the notion of the given. In this section, I will explore this aspect of Kuhn’s work in the context of the suggestion that his position may be interpreted as a form of coherentism.

1  Kuhn, Coherentism and Perception

9

I wish to explore two aspects of Kuhn’s work under this head. The first relates to Kuhn’s rejection of what he described as a “philosophical paradigm initiated by Descartes” (2012, 121). The question I wish to ask is whether rejection of this Cartesian paradigm entails rejection of foundationalism and an associated acceptance of coherentism. The second aspect that I wish to explore relates to the myth of the given. In Sellars’ work, rejection of the myth of the given leads in a coherentist direction. The question I wish to explore on this front is whether Kuhn’s rejection of the given has the same coherentist implications. Kuhn was an advocate of the theory-dependence of observation. He did not just hold that observation is guided by or interpreted in light of theory. He held that the perceptual experience of a scientist is influenced by the theory (or paradigm) that the scientist brings to bear on the observation which gives rise to the experience. In articulating this view, Kuhn speaks of a “philosophical paradigm” (2012, 121) or “epistemological viewpoint” (2012, 125) which he believes to have held sway in the Western philosophical tradition since Descartes. According to this view, perceptual experience is fixed “by the nature of the environment and of the perceptual apparatus” (2012, 120). It is “fixed and neutral” (2012, 125). But, as Kuhn explained in the terms he employed in work after Structure, “we now know (as Descartes did not) that the stimulus-sensation correlation is neither one-to-one nor independent of education” (1970, 276). In other words, Kuhn took the traditional philosophical view since Descartes to be that the character of a perceptual experience is fixed by the circumstances of observation in such a way that two people in exactly the same situation would have exactly the same experience. Against this view, Kuhn holds that scientists’ experience may vary with respect to paradigm. Hoyningen-Huene describes Kuhn’s view as the “plurality-of-phenomenal-worlds thesis” or the thesis of “the nonuniqueness of the relationship between the world-in-itself and the many phenomenal worlds” (1993, 37).12 I regard this as a mistaken thesis. It trades on confusion between what one perceives and what one thinks about what one perceives.13 It does not follow from the fact that what one thinks about an object may vary that the object may vary. What varies with theory (or paradigm) is not the object experienced or even the experience of the object. What varies are the beliefs or theories that one holds about the object. What may also vary, possibly because of such variation, are the concepts, vocabulary and meaning of the terms that are applied to the fixed objects of which one has experience.

 I do not propose to consider here the historical question of whether Kuhn is right to say that the philosophical tradition since Descartes held that there is a one-to-one correspondence between object and experience. As a potential counter-example to Kuhn’s claim, I will simply mention the idea of an inverted spectrum that is found in Locke’s Essay on Human Understanding (II, xxxii, 15). I only offer this as a potential counter-example, however, since Locke does speak of difference in the structure of organs, which might entail difference in what Kuhn calls “perceptual apparatus”. 13  It seems to me that Alexander Bird’s analysis of the theory-dependence of observation in terms of intensional and extensional uses of ‘see’ both brings out the basis for this confusion and resolves it (see, e.g. Bird 2000, 104). 12

10

H. Sankey

But this is not the point at issue here. What is at issue here is whether the thesis of the “non-uniqueness” of the relationship between external object and subjective experience has any implication with respect to the coherentist interpretation of Kuhn. Does Kuhn’s rejection of the idea that there is a one-to-one correspondence between object and experience indicate acceptance of a coherentist as opposed to a foundationalist epistemology? I want to suggest that it does no such thing. For the sake of argument, we may suppose that Kuhn is right about the lack of one-to-one correspondence between object and experience (or “stimulus” and “sensation”). Two perceiving subjects who observe the same object in the same circumstances may have a different experience depending on background paradigm. Equally, one and the same subject who undergoes a paradigm shift may have a different experience of the same object in the same conditions before and after the paradigm shift. What follows about the relationship between perception and the justification of a perceptual belief? On a foundationalist theory of epistemic justification, the two perceiving subjects who have different experiences in the same circumstances (or the same subject before and after a revolution) are still able to have basic beliefs justified by the experience that they have in those circumstances. One subject may have a belief that is justified by the experience that they have of the object. The other subject may have a different belief that is justified by the different experience that they have of the same object. In both cases, the subject is able to appeal to their experience of the object as the justification for the basic belief which they hold about the object on the basis of their experience. What is challenged by the lack of one-to-one correspondence is not the foundational structure of the relation between perception and basic perceptual belief. What is challenged is the invariance, neutrality or paradigm-independence of the experience. Hence, the alleged non-uniqueness of the relationship between perceived object and perceptual experience fails to provide support for the coherentist interpretation of Kuhn. What about the second point, Kuhn’s rejection of the given? Kuhn does speak somewhat critically of the given on several occasions (e.g. 2012, 125–7). But it is not clear that Kuhn understood the notion of the given in the way that Sellars did when he wrote of the myth of the given. Because of this it is not clear that Kuhn’s rejection of the given has the same implications with respect to coherentism as Sellars’ rejection of the myth of the given. Kuhn does not consistently employ the expression ‘the given’. He also speaks inter alia of “immediate experience”, “raw data”, “brute experience” and “given data” (2012, 125), as well as “observation” (2012, 120). Because of the lack of consistency and precision in the vocabulary Kuhn employs, it is not always perfectly clear what he is talking about or what he means. Indeed, it is not even clear whether he rejects the existence of the given rather than merely holding that the given plays a limited or negligible role in the sciences. Nevertheless, it seems clear that there is a sense of “the given”, in which Kuhn can be shown to reject it. As we have previously seen, Kuhn takes the traditional philosophical view of perception to be one that may be traced back to Descartes. According to Kuhn, the traditional view is that there is a unique experience that corresponds to any

1  Kuhn, Coherentism and Perception

11

particular set of observational circumstances. All human perceivers have qualitatively the same experience when presented with the same observational circumstances. On the traditional view, a distinction is to be made between observation and interpretation. Observation is “fixed once and for all by the nature of the environment and of the perceptual apparatus” (2012, 120). Observation itself is invariant between perceivers. But it may be subject to alternative interpretations by different scientists in the context of different theories or paradigms. The idea that observation is fixed, and that it is entirely independent of the interpretation placed upon it by theory (or paradigm), provides a clear sense for the notion of “the given”. In particular, “the given” is the sensory experience which a perceiving subject undergoes when they observe some state of affairs. This experience is invariant between perceivers with respect to shared observational conditions. The experience is prior to and independent of the interpretation placed upon it in the context of a theory (or paradigm). Kuhn rejects the traditional view that there is a given in the sense just specified in favour of the view that different perceivers may have a different experience when they perceive the same thing in the same circumstances. So, in this specific sense Kuhn does reject the idea of the given. The question I will now raise is whether this is the same notion of the given that Sellars (1963) had in mind in rejecting the myth of the given. Rather than engage in Sellars scholarship, I will work with a version of Sellars’ argument that is found in the epistemological literature. The argument is sometimes referred to as the “Sellarsian dilemma” (e.g. Lyons 2008). To set up the dilemma, we need to bear in mind the role that the given is meant to play within the context of an empiricist epistemology with a foundationalist justificatory structure. Within the context of such an epistemological theory, a distinction is made between basic beliefs which receive non-inferential justification from experience, and non-basic beliefs which receive inferential justification from other beliefs. In this context, the role of the given is to provide basic beliefs with epistemic support. The question is how an experience may provide a belief with such support. Here it is important to note that a belief is a propositional attitude state. As such, a belief has a propositional content. The question now becomes that of what kind of state a sensory state must be for it to stand in a justificatory relation to a belief state which has such content. There are basically two options. Either a sensory state has a propositional content or it does not. If a sensory state has a propositional content, then it will be able to stand in an appropriate relation of justification to the belief state. But the problem is that, given that the sensory state has a content, the sensory state itself stands in need of justification. In order to justify the sensory state, appeal must be made to some other state. Because the sensory state requires further justification, it cannot play a regress-terminating role in a foundationalist epistemology. The other option is that the sensory state does not have a content, and, as such, does not stand in need of justification. This might seem to be the right way to think about sensory states. But the problem now is that, if a sensory state does not have a content, then it is unable to stand in an appropriate justificatory relation to a belief state which does have a content. So, it cannot play the role that it is required to play in the

12

H. Sankey

justification of beliefs. Thus, either a sensory state cannot play its foundational role because it must be justified or it does not need to be justified but cannot play its justificatory role.14 This argument against foundationalism provides support for a coherentist theory of epistemic justification. On the one hand, the argument suggests that there may be no non-doxastic source for the justification of basic beliefs. On the other hand, the argument suggests that justificatory relations may only hold among beliefs because beliefs have propositional content, and so may enter into justificatory (inferential) relations with each other. The resulting position is, in effect, a coherentist view on which there are only non-basic beliefs which receive justification via inference from a doxastic source, namely, other beliefs within a belief-system. Now, to return to Kuhn, the question is whether Kuhn’s rejection of the given leads in the way just sketched to a coherentist position. The answer, I think, is negative. The reason is that the notion of the given that Kuhn rejects does not have built into it the capacity to justify basic beliefs. The notion of the given that Kuhn rejects is the notion of perceiver-invariant experience that is independent of interpretation by paradigm. It is not the notion of a sensory state that has the capacity to provide justification for a basic belief. Thus, Kuhn’s rejection of the given does not entail endorsement of a coherentist view. His rejection of the given does entail rejection of the idea that experience is neutral or invariant. But, as we saw, that is consistent with a foundationalist theory of justification.

1.7 Conclusion In this paper, I have sought to further explore the suggestion originally due to Kuukkanen that Kuhn’s account of science may be understood in coherentist terms. My focus has been on coherentism as a theory of epistemic justification rather than on the relationship between coherentism and the correspondence theory of truth, which is one of Kuukkanen’s chief concerns. I think that Kuukkanen is right in claiming to have identified coherentist or coherentist-tending themes in Kuhn’s philosophy of science. However, I think that one crucial element of coherentism is lacking from Kuhn’s work, since he fails to explicitly deny the existence of basic beliefs which have a non-doxastic source of justification. Correlatively, Kuhn does  In my formulation of the argument, I loosely follow BonJour (1985, 69; cf. 78). However, BonJour uses the expressions “cognitive” and “judgment” in speaking about sensory states, whereas I put the point explicitly in terms of propositional content. In this, I follow Sosa (1980, 6–7), who formulates a similar argument in terms of propositional attitudes and content. The argument presupposes that relations of justification must be relations that are either inferential or inference-­like, and that they can only obtain between states with propositional content. Sosa characterizes this as an “intellectualist model of justification” (1980, 8), the key feature of which is that justificatory relations are “parasitic” on logical relations. Of course, one way to avoid the dilemma is to reject the intellectualist model, for example, by adopting a reliabilist account of the warrant of basic beliefs. 14

1  Kuhn, Coherentism and Perception

13

not explicitly assert that epistemic justification may only derive from inferential relationships which obtain between (non-basic) beliefs. Despite this shortcoming in the coherentist interpretation of Kuhn, I regard the interpretation as highly promising. For this reason, I have attempted to further develop the coherentist interpretation of Kuhn. In order to do this, I raised the question of whether Kuhn’s account of science is able to deal with a major objection to coherentism (the “input objection”). I argued that the role played by problems in Kuhn’s theory of science ensures that there is input from the external world into the belief-systems of scientists. Moreover, I followed Hoyningen-Huene in pointing to the causal role played by the external world in the determination of perceptual states. Having argued that Kuhn’s account may withstand a major objection to coherentism, I turned to the question of whether Kuhn’s rejection of foundationalism contains elements of coherentism. Against this, I argued that Kuhn’s rejection of the one-to-one relation between object and experience is not incompatible with a foundationalist account of justification. Nor does Kuhn’s rejection of a notion of the given carry with it the same coherentist implications as Sellars’ critique of the myth of the given. There is one remaining point to make. There is no doubt that Kuhn’s work came to prominence during a time-period in which foundationalism was under attack and coherentism was on the rise. It is natural to read Kuhn as having views that are consonant with philosophers working at the same time who had coherentist leanings. It is not overly surprising to find that there is a coherentist tendency in Kuhn, since coherentist tendencies were widespread in the time-period during which he worked. What is surprising to me, as a philosopher of science turned epistemologist, is to find that coherentism is currently on the run and that foundationalism has made a come-back, at least within epistemology.

References Bird, Alexander. 2000. Thomas Kuhn. Chesham: Acumen Publishing. BonJour, Laurence. 1985. The Structure of Empirical Knowledge. Cambridge: Harvard University Press. Feyerabend, Paul. 1970. Consolations for the Specialist. In Criticism and the Growth of Knowledge, ed. I. Lakatos and A. Musgrave, 197–230. Cambridge: Cambridge University Press. Haack, Susan. 1995. Evidence and Inquiry. Oxford: Blackwell. Hempel, Carl G. 1935. On the Logical Positivists’ Theory of Truth. Analysis 2 (4): 49–59. Hoyningen-Huene, Paul. 1993. Reconstructing Scientific Revolutions: Thomas S.  Kuhn’s Philosophy of Science. Chicago: University of Chicago Press. Kuhn, Thomas S. 1970. Reflections on My Critics. In Criticism and the Growth of Knowledge, ed. Imre Lakatos and Alan Musgrave, 231–278. Cambridge: Cambridge University Press. ———. 1977. Objectivity, Value Judgment, and Theory Choice. In The Essential Tension, ed. T. Kuhn, 320–339. Chicago: University of Chicago Press. ———. 2012. The Structure of Scientific Revolutions. 4th ed. Chicago: University of Chicago Press. Kuukkanen, Jouni-Matti. 2007. Kuhn, the Correspondence Theory of Truth and Coherentist Epistemology. Studies in History and Philosophy of Science Part A 38 (3): 555–566.

14

H. Sankey

———. 2009. Closing the Door to Cloud-Cuckoo Land: A Reply to ŠeŠelja and Straßer. Studies in History and Philosophy of Science Part A 40 (3): 328–331. Lyons, Jack C. 2008. Evidence, Experience, and Externalism. Australasian Journal of Philosophy 86 (3): 461–479. Sellars, Wilfrid. 1963. Empiricism and the Philosophy of Mind. In Science, Perception and Reality, ed. W. Sellars, 127–196. Atascadero: Ridgeview. ŠeŠelja, Dunja, and Christian Straßer. 2009. Kuhn and Coherentist Epistemology. Studies in History and Philosophy of Science Part A 40 (3): 322–327. Sosa, Ernest. 1980. The Raft and the Pyramid: Coherence versus Foundations in the Theory of Knowledge. In Midwest Studies in Philosophy V 1980: Studies in Epistemology, ed. P. French, T. Uehling, and H. Wettstein, vol. 5, 3–25. Minneapolis: University of Minnesota Press.

Chapter 2

Sankey on Kuhn and Epistemological Coherentism: A Commentary Juan V. Mayoral

2.1 Introduction In his paper, “Kuhn, Coherentism and Perception,” Howard Sankey shows that Thomas Kuhn’s epistemology does not fit squarely in the coherentist framework, though some aspects of his theory are akin to that position. Sankey examines and then criticizes a previous work by Jouni-Matti Kuukkanen that argues that Kuhn’s work fits in well with a coherentist approach and with a realistic position in which the correspondence theory of truth is still defensible.1 I will not examine Kuukkanen’s argument here. I will only comment on Sankey’s qualifications on Kuhn’s alleged coherentism. In the first section of this commentary I shall add some more details to Kuhn’s relation to the concept of “the given” that reinforce Sankey’s doubts concerning the thesis that the author of The Structure of Scientific Revolutions fits in with a coherentist account of scientific knowledge—though I also agree with Sankey (and Kuukannen) that a study of coherentist ingredients in Kuhn’s epistemology would bring insight. Even so, labeling Kuhn’s epistemological views as either foundationalist or coherentist goes beyond his original point of view. I shall devote the second section of this commentary to argue in favor of this thesis on the basis of an unpublished lecture by Kuhn in which he discusses his convictions in the theory of knowledge—which agree with his attack to convergent views of scientific progress—and their connection to Ludwig Wittgenstein’s On Certainty.  See Kuukkanen (2007). Kuukkanen thus explores a possibility Laurence BonJour (1985, 88) mentions, that is, that “there is no manifest absurdity in combining a coherence theory of justification with a correspondence theory of truth.” 1

J. V. Mayoral (*) Universidad de Zaragoza, Zaragoza, Spain e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 L. Giri et al. (eds.), Perspectives on Kuhn, The Western Ontario Series in Philosophy of Science 84, https://doi.org/10.1007/978-3-031-16371-5_2

15

16

J. V. Mayoral

2.2 Kuhn and Lewis’s “the Given” For Sankey, there is an important sense in which Kuhn is not a coherentist (see Sankey 2018). The main coherentist thesis is that there are no “basic beliefs,” as they are often called, which are responsible for non-inferential justification of the rest of our empirical knowledge. Following Laurence BonJour, the coherentist answers a classic problem of epistemic justification, namely epistemic regress, by means of that thesis. The coherentist epistemologist explores the possibility that, in BonJour’s words, this “regress might circle back upon itself, so that … beliefs which have already appeared as premises (and have themselves been provisionally justified) earlier in the sequence of justificatory arguments are again appealed to as justifying premises.”2 As this circle can readily be considered vicious, BonJour says, a “holistic and nonlinear” alternative is preferable.3 So, according to this view, even though the coherentist argues that a finite sequence of beliefs (our “system”) “can only be justified from within, by virtue of the relations of its component beliefs to each other,” coherentism turns out to be for BonJour the “holistic view” that “it is [the] … system of beliefs which is the primary unit of justification; particular beliefs are justified only derivatively, by virtue of membership in such a that system.”4 In any case, the lack of basic beliefs is crucial in defining a given position as coherentist and Sankey criticizes Kuukkanen because he does not show that Kuhn rules out the option that some beliefs can be justified non-inferentially, non-doxastically.5 Sankey points out that a coherentist position as that of Wilfrid Sellars’ involves ruling out the idea of “the given,” and that, admittedly, Kuhn is credited with rejecting that notion in section X of The Structure of Scientific Revolutions. Kuhn mentions “the given” in the book at some points and, as Sankey also says, he does so when he refers to a concept that belongs to the Cartesian and empiricist traditions in epistemology, according to which the empirical foundations of our knowledge about the world must be found in a universal, neutral vocabulary (of “givens,” as it were) that we interpret according to our theoretical concepts. As Kuhn says: “is sensory experience fixed and neutral? Are theories simply man-made interpretations of given data? The epistemological viewpoint that has most often guided Western

 BonJour (1985, 21). For the concept of “basic beliefs” and the above picture of foundationalism in this paragraph, see also BonJour (1985, 17 ff.). A similar exposition can be found in Williams (2004, 249–250). 3  See BonJour (1985, 24–25, 90–92); the quotation is from 24. 4  BonJour (1985, 88 and 24, respectively). BonJour goes on saying that there are some evident questions this conception must answer: (a) how to choose between alternative coherent systems?; (b) how coherent systems are related to the mind-independent world?; and (c) how is coherence related to truth? Sankey devotes a central section of his paper to question (b), whereas Kuukkanen deals with question (c). For BonJour’s full version of coherentist epistemology, see BonJour (1985, Pt. 2) (he devotes ch. 8 to (c), pp. 143–146 to (a), and ch. 6 and the first pages of 7 to (b)). His “systematic” version of coherentism (1985, 90 ff.) is richer than I can expound in this short commentary, however. 5  See Sankey (2018, 6), and section 6 for a discussion of the option itself. 2

2  Sankey on Kuhn and Epistemological Coherentism: A Commentary

17

philosophy for three centuries dictates an immediate and unequivocal, Yes!”6 As it is well known, Kuhn employs that section X of Structure to show that the divide perception/interpretation does not work well because our paradigm-based education intervenes in the experience we have as soon as we have it. It thus seems as if Kuhn were indeed criticizing the same concept of the given that the coherentist attacks. However, Sankey argues that, although it is not always entirely clear what Kuhn is ruling out in his perspective on perception, the idea of the given that Sellars has in mind is probably not Kuhn’s target in his criticism of the Cartesian tradition in epistemology. And consequently, if Kuhn is not rejecting the same concept Sellars criticizes in order to defend coherentism as a legitimate alternative to foundationalism (and if his argument is compatible with foundationalism), then it follows that Kuhn is not joining the ranks of coherentism (see Sankey 2018, 14–16, 18). If we examine Sellars’s likely source for the idea of the given, C. I. Lewis, we see that, as Sankey says, Kuhn’s and Lewis’s references for the given do not completely overlap. In Lewis’s work, the given is a sort of “relation to the world,” as it were, that helps to avoid skeptical considerations about our empirical knowledge. There are two features that grant that the given in Lewis is not the basis of a neutral, universal language. First, the given is just a commitment that, as noted, takes us away from skepticism (or idealism, for that matter), but as soon as we try to make a report on it, or simply “point to it,” as it were, verbally, we let interpretation enter in.7 “Interpretation” is not here understood as an activity we carry out once perception reports are in place, but a medium in terms of which perceptions—and not only perceptions—are reported. Second, interpretation is variable, so that the way two different persons speak about their respective givens may be different from one another.8 So reports about sensations by two different individuals may be as different as their interpretative resources are. There is a reduced probability of finding a neutral, universal language all over the history of science. On that point, I am obliged to reproduce a passage from a paper by Lewis that, I am sure, will resound in the reader’s ears because of its similitude with something s/he has read before: In any given period, there is some body of generally accepted concepts in terms of which men describe and interpret their experience. Later, these may be all strange. If we go back to the Middle Ages or to the civilization of ancient Greece, and try to view the world as men then saw it, only by an effort can we do so. We might expect that fundamental things—life,

 See Kuhn (1962, 126). As he says in the “Postscript—1969,” “What I have been opposing in this book is therefore the attempt, traditional since Descartes but not before, to analyze perception as an interpretive process, as an unconscious version of what we do after we have perceived.” (Kuhn 1962, 195). 7  “Pointing to it,” is inspired in C. I. Lewis’s phrase “point to the given”; see Lewis (1926, 249). See also Mayoral (2017, 84). 8  Lewis also provided a theory of the a priori—the “pragmatic a priori”—that let categories vary in time according to pragmatic considerations. A good analysis of his version of that classic concept in epistemology (and a comparison with other recent perspectives on the a priori) is Chang (2008). 6

18

J. V. Mayoral mind, matter and force, cause and effect—would be conceived in the same way. Yet it is exactly here that we find the greatest differences. (…) Words such as ‘life,’ ‘matter,’ ‘cause,’ and so on have been used since thought began, but the meanings of them have continuously altered. There is hardly a category or principle of explanation which survives from Aristotle or the science of the Middle Ages. Quite literally, men of those days lived in a different world because their instruments of intellectual interpretation were so different.9

The given is, in Lewis, the relation to the world that we can find in ourselves when examining our experience. However, experience itself, or at least the way we give it shape and then report about it, may vary as our categories do as well. In that sense, I do not see how Kuhn would disagree completely with this concept of the given, leaving aside that he surely prefers talking about how stimuli impinge on us, thereby causing us to produce an output of paradigm-based and variable sensory experiences.10 Lewis’s and Kuhn’s explanations of perceptual experience could be put in agreement and the latter’s attack to the given does not seem to me an insurmountable obstacle.11 Therefore, I agree with Sankey’s thesis that, if we try to place Kuhn in one of the two options of mainstream tendencies in epistemology, foundationalism and coherentism, it is difficult for us to do so in either of them, and, in any case, it is hard to put him in the coherentist tradition alone. The reason why this is so difficult is because Kuhn did not want his work to be labeled that way, either foundationalist or coherentist. This is the position I now wish to explore, and which I think Sankey does not pursue in his paper.

2.3 Kuhn and Wittgenstein’s on Certainty 2.3.1 Kuhn’s New Questions A good parallelism can be established, in this sense, with another philosopher whose works Kuhn used to read and, to some extent, follow: Ludwig Wittgenstein. In this case, the work by Wittgenstein I wish to mention is On Certainty because of Kuhn’s professed influence of this work’s epistemological views on his own.12 In Wittgenstein’s case, the question concerning where to place his work when the epistemological positions are divided as we have done before—between coherentism and foundationalism—is still unsettled. Some authors put his latest work in a foundationalist trend, but that he might have been a coherentist is also a possibility  Lewis (1926, 252–253).  On this ontology of stimuli and its relation to sensory experience, see Hoyningen-Huene (1993, Ch. 2, esp. sect. 2.2). 11  For a comparison of their respective positions, see Mayoral (2017). 12  See Kuhn (1976), where he explicitly confesses and addresses the connection of his epistemological views with On Certainty (and also on J. L. Austin’s Other Minds). For a study of that conference, see Mayoral (2015). 9

10

2  Sankey on Kuhn and Epistemological Coherentism: A Commentary

19

to take into account.13 Perhaps a more careful consideration is to explain his position in its own terms because he does not seem to be trying to find his place in one of those parties, but rather to be dealing with a particular problem from a radically different point of view.14 So, just as Kuhn is similar in some thesis to Wittgenstein, it would be helpful to approach the former’s position in epistemology as we sometimes do with the latter.15 Kuhn’s position about epistemology has been often presented in terms of his defense of a dynamic perspective that opposes a static one concerning the proper task of the philosopher of science. He supports the idea that the task for epistemology of science is to show that the option between two partially overlapping bodies of belief was based on good reasons and that, on those grounds, a fresh body of beliefs displaced the previous one.16 If the ultimate result of that activity is to speak out for or against the evidence or beliefs that support a given theory choice, then epistemological analysis is reduced to a double question: (1) What is it that justifies just any kind of belief? (2) Is that justification a guide to truth? Meanwhile, Kuhn tries to reformulate the right questions as follows: (a) What are the proper grounds for different kinds of belief in a body of them? (b) According to their kind, how do those beliefs change? The step from (1) and (2) to (a) and (b) is not, as it were, from a more basic position to a derivative one, but rather a step to a different kind of question completely (see Kuhn 1976/1980, 18). I will address these two latter questions in turn.

2.3.2 Kinds of Belief Concerning (a), Kuhn shows that, in a given body of knowledge, some beliefs might be submitted to scrutiny in epistemological terms, whereas some others are simply taken for granted. BonJour said that there is a “holistic view” for the coherentist epistemologist, according to which justification of a given belief is granted in terms of “membership” in a given “system of beliefs” which is now the more basic unit of justification. Maybe Kuhn would agree with this latter point, but what he is truly  A good discussion and critique of those alternatives can be found Hamilton (2014, 98–102).  This is also Hamilton’s point in that regard; see Hamilton (2014, 102–103). 15  I am not saying, for sure, that Kuhn’s epistemological views are the same as Wittgenstein’s, but rather that his attitude to the epistemological tradition at large is critical and he does not look for a place in it, which is similar in both cases. There are explicit influences, for sure, but a complete overlap of both positions is not what I am here arguing for. I am not saying that Kuhn is an orthodox follower of Wittgenstein’s writings, either. To what extent Kuhn reads Wittgenstein “correctly,” so to speak, or even carefully, is a matter for a different kind of essay. 16  See Kuhn (1984, lect. I; 1991, 95–96, 102; 1992, 111–112). For the difference between the static and the dynamic picture in Kuhn see also Melogno (2019). 13 14

20

J. V. Mayoral

investigating is how we pass from a given body of beliefs to another one. This is no longer the classic justification problem; so, maybe we should start by approaching Kuhn’s perspective itself (see Kuhn 1991, 102). For Kuhn, the internal composition of the body of beliefs is important when we study its kind of change. Some more details must be taken into account. I will comment on two of them. First, beliefs at large are known to belong to a given body of them and semantic considerations act as necessary conditions for those beliefs to be “candidates for true/false” pronouncements.17 Second, the way in which they are considered justified or not and, accordingly, the kind of epistemological scrutiny to which they are subject depend on the role they play in that body of beliefs. Kuhn devoted an unpublished lecture in the late 1970s and early 1980s, “Does Knowledge ‘Grow’?”, to explain that point further. I shall briefly comment on its contents in that respect. Some beliefs, Kuhn says there, are considered knowledge from the point of view that their status may not be altered without changing the body of beliefs itself. Our attitude towards them—to put it in the Wittgensteinian terms Kuhn also uses—is that they are the grounds on which other beliefs are supported and their epistemological status is examined. These latter beliefs, the less sure ones, conform the rest of them within our systems or bodies of beliefs. By contrast with the previous group (the core, as Kuhn terms it), not every member of the human group that shares the former also shares all of them. As they are thus not considered infallible, any kind of scrutiny of the evidence (or other beliefs) supporting them is in order. This activity is absent when we take the former group into consideration: the former are equivalent to what Wittgenstein considered “certainties,” a connection that Kuhn exhibits. So, the former are not subject to epistemic scrutiny, whereas the latter ones are, on the basis of the grounding ones.18 This classification of beliefs within any body of them makes it clear that the epistemological analysis of belief takes place within an epistemic system, but that not every (kind of) belief is analyzed in the same way as every other. So the “holistic view” of justification we saw before does not apply straightforwardly and in a coherentist manner. In the next section we will see that the epistemological inquiry does not get as deep as the lexical core, and this is not something a coherentist (as BonJour) would accept.19 Perhaps a foundationalist philosopher would argue that the first group of beliefs—the one within the core—is composed of basic beliefs. 17  See Kuhn (1991, 99–100); as he admits there, in “The Road since Structure” (and other places), he takes inspiration from Hacking (1982). 18  On all this, see Kuhn (1976, 9–10, 18). See Mayoral (2015) for further details. 19  See, e.g., BonJour (1985, 22). Kuhn is not committed to coherentism in the same sense that Michael Williams (2004) says Wittgenstein is not (or not completely), either. That is, even though they assume the existence of some beliefs that are, as it were, essential to the normal functioning of a whole system of beliefs, whenever finding evidence for the beliefs one holds starts out, those former beliefs that are more essential provide support to the latter but are left out of that epistemological scrutiny themselves. Kuhn finds inspiration in Wittgenstein’s On Certainty, as noted below, but he adapts Wittgenstein’s insights to his own perspective and vocabulary. For this version of Wittgenstein’s On Certainty, see Williams (2004, 250–257).

2  Sankey on Kuhn and Epistemological Coherentism: A Commentary

21

But that is not the case in Kuhn’s perspective either because there is not any more specifically non-doxastic connection of those beliefs to the world (as given, as it were) than can be found in the rest—the peripheral area, as Kuhn talks about it (see Kuhn 1976, 10, 18). Kuhn’s picture here is intended to resemble Wittgenstein’s views in On Certainty (despite the differences). And, as in Wittgenstein’s case, it is not easy to find a place in the mainstream epistemological traditions for a picture like this. Kuhn, actually, does not seem interested in doing so.20

2.3.3 Modes of Change Now, concerning question (b)—how these different kinds of beliefs change—Kuhn has a classic answer that he now applies to this epistemological issue. In order to answer that question as Kuhn does, I shall mention two of his key thesis, mainly in his later thinking. The first one has to do with the relation of a lexicon to the world. A lexicon is not the access to only one world, but rather to a set of possible worlds. Accordingly, normal science can be considered as the search of the actual lexicon-­ relative world among a populated set of slightly different lexicon-relative possible worlds.21 The second thesis I would like to mention is that the lexicon is accessible by means of a combination of those beliefs that most members of a community share together with situations in which those beliefs are applied. If there is a core, as Kuhn says there is, then this is transmitted in typical situations when learning the lexicon. That is why Kuhn states that, when learning a lexicon, the individual not only acquires membership in a community and a vehicle of communication and understanding within it, but also knowledge of the world as shared by members of the group.22 “Knowledge” is thus identified with the core of beliefs every member of that community shares. So exploring the lexicon and how it works for the members of a community shows what can be altered within the lexicon and what cannot—that is, what is a guide for exploring the set of possible worlds and what is a way out of that set. So a semantic inquiry into the structure of the lexicon exhibits epistemological consequences as well. Normal science is understood, thus, as exploring the semantic periphery of the kind terms in a lexicon and, at the same time, as trying to prove or disprove elements in the peripheral area of the corresponding body of beliefs that conforms the core of knowledge, and by means of an inquiry in which the classic  See again Kuhn (1991, 95–96, 102). His first Thalheimer Lecture (Kuhn 1984) shows how Kuhn follows a different path; see Kuhn (1984, lect. I); a good and useful Spanish edition of Thalheimer Lectures is Kuhn (2017). 21  See Kuhn (1989, 76); Kuhn (1987, lect. II, 63). 22  Kuhn finds support in J. L. Austin to say that, when inquired about how a given individual may know something about a certain phenomenon, a way—perhaps the most decisive and, at the same time, the most fundamental way—to answer is to invoke membership in a community of experts (“credentials,” in a more Austinian expression). See Kuhn (1976, 15–16). Again, see Mayoral (2015). 20

22

J. V. Mayoral

epistemological apparatus play a role and which involves the search for new beliefs or for evidence in support of the more cherished ones among the less stable. This latter activity—which, as noted, is readily assimilated to the classic epistemic activity visible in normal research—is one of the best examples of change of beliefs (see Kuhn 1976, 10). The other kind of change is the one that affects the core itself. This is the kind of change Kuhn emphasizes in his late thinking as the proper place for the philosopher of science to apply his/her epistemological method, which must be comparative besides. Here Kuhn mentions values Sankey also talks about—greater scope, problem-­solving potential, simplicity, fruitfulness, and so on—in order for a body of beliefs to supersede another (see Sankey 2018, 4–5). Once the community accepts the satisfaction of one or more values, the new core supersedes the old one.23 Yet, it is not a matter of empirical support of the new core that the epistemologist must search for in this process, because, once the core is in place, it shall be considered a set of groundwork beliefs, on the basis of which the rest of beliefs (the new peripheral area, so to speak) shall be sought for or epistemically scrutinized. The conversion of the whole community to the new body of beliefs shall take some time and it shall be a matter of a new community of converts and newcomers that supplants the previous one. Conversion, here, is not a matter of proving a core and disproving another, whether by virtue of a doxastic or non-doxastic basis. The classic epistemological inquiry that supports a (dynamic) coherentist reading of Kuhn is not readily found here.24 Conversion is rather a matter of judging that another set of beliefs is worthwhile as a guide to control and understand a new phenomenal world. As Wittgenstein says, in a paragraph of On Certainty that Kuhn quotes at length, 92. However, we can ask: May someone have telling grounds for believing that the earth has only existed for a short time, say since his own birth?—Suppose he had always been told that,—would he have any good reason to doubt it? Men have believed that they could make rain; why should not a king be brought up in the belief that the world began with him? And if Moore and this king were to meet and discuss, could Moore really prove his belief to be the right one? I do not say that Moore could not convert the king to his view, but it would be a conversion of a special kind; the king would be brought to look at the world in a different way. Remember that one is sometimes convinced of the correctness of a view by its simplicity or symmetry, i.e., these are what induce one to go over to this point of view. One then simply says something like: “That’s how it must be.”25

Some criteria may help to judge that, after all, a new way to look at the world is the right way to approach the study of nature. However, the access to that new set of beliefs is not a matter of satisfying coherence26 or finding new basic beliefs as new foundations, but of embracing the new context—the new world—as we (or others  See the details in Kuhn (1984, lect. I, esp. 18 ff).  See, e.g., BonJour (1985, 99–100, 140, 144–146). See also Kuukkanen (2007) and Sankey (2018). 25  Wittgenstein (1969, 92); as quoted by Kuhn (1976, 17). 26  See references in fn. 24, above. 23 24

2  Sankey on Kuhn and Epistemological Coherentism: A Commentary

23

like us) did in the past. Once more, a quotation from Wittgenstein by Kuhn is appropriate here: 93. (…) Everything that I have seen or heard gives me the conviction that no man has ever been far from the earth. Nothing in my picture of the world speaks in favour of the opposite. 94. But I did not get my picture of the world by satisfying myself of its correctness; nor do I have it because I am satisfied of its correctness. No: it is the  inherited background against which I distinguish between true and false.27

2.3.4 Kuhn’s Attitude Towards Epistemology In short, the set of questions Kuhn addresses as the main target of an epistemologist of science—(a) and (b) rather than (1) and (2)—belongs to a conception of epistemology that tries to leave previous questions behind, or else, at least, reformulate them as part and parcel of a metaphilosophical reform of the epistemological enterprise. In his Thalheimer Lectures, Kuhn started by showing how his own perspective of philosophy of science moves away from the previous bases of epistemology—the priority of propositions, the appeal to foundations and deductive reconstruction of theories, a holistic perspective, and the mathematical reconstruction of scientific method with its implicit solipsism.28 From that point of view, it is difficult to accept that a full coherentist interpretation of Kuhn’s epistemological views rightly labels his perspective.

2.4 Conclusion In this commentary I have tried to show that there is room for Howard Sankey’s critique to a full coherentist reading of Kuhn, on the basis of the possibility of basic beliefs in the latter’s depiction of science. Although I share with Sankey and Kuukkanen the opinion that there are coherentist ingredients in Kuhn’s thought and that it would be interesting to bring them to light as they have done, Kuhn is, in my view, scarcely akin to identifying his position as either coherentist or foundationalist. I have supported my view by virtue of Kuhn’s confessed proximity to Wittgenstein’s position in epistemological matters, which does not squarely fit the traditional divide between coherentism and foundationalism.

27 28

 Wittgenstein (1969, 93–94); as quoted by Kuhn (1976, 17).  See Kuhn (1984, lect. I, esp. 3 ff.).

24

J. V. Mayoral

References BonJour, Laurence. 1985. The Structure of Empirical Knowledge. Cambridge, MA: Harvard University Press. Chang, Hasok. 2008. Contingent Transcendental Arguments for Metaphysical Principles. Royal Institute of Philosophy Supplements 63: 113–133. Hamilton, Andy. 2014. Wittgenstein and On Certainty. London: Routledge. Hoyningen-Huene, P. 1993. Reconstructing Scientific Revolutions. Foreword by T.S. Kuhn, Trans. A.T. Levine. Chicago: The University of Chicago Press. Kuhn, T.S. 1962. The Structure of Scientific Revolutions, 3rd ed., 1996. Chicago: The University of Chicago Press. ———. 1976. Does Knowledge ‘Grow’? Foerster Lecture, Berkeley. TSKP 5.13. ———. 1984. Scientific Development and Lexical Change. Thalheimer Lectures. Johns Hopkins University, 12–19 November. TSKP 23.21. ———. 1987. The Presence of Past Science. The Shearman Memorial Lectures. London: University College, 23, 24 and 25 November. TSKP 23.32. ———. 1989. Possible Worlds in History of Science. In The Road since Structure, ed. J. Conant and J. Haugeland, 58–89. Chicago: The University of Chicago Press. ———. 1991. The Road since Structure. In The Road since Structure, ed. J. Conant and J. Haugeland, 90–104. Chicago: The University of Chicago Press. ———. 1992. The Trouble with the Historical Philosophy of Science. In The Road since Structure, ed. J. Conant and J. Haugeland, 105–120. Chicago: The University of Chicago Press. ———. 2017. Desarrollo científico y cambio léxico. Conferencias Thalheimer. Baltimore: The Johns Hopkins University, 12–19 November 1984. Foreword by Paul Hoyningen-Huene. Pablo Melogno and Hernán Miguel (eds.), Translated by Leandro Giri. Montevideo: Universidad de La República, Uruguay/SADAF. Kuukkanen, Jouni-Matti. 2007. Kuhn, the Correspondence Theory of Truth and Coherentist Epistemology. Studies in History and Philosophy of Science Part A 38 (3): 555–566. Lewis, C.I. 1926. The Pragmatic Element in Knowledge. University of California Publications in Philosophy 6: 205–227. Reprinted in Goheen, J.D., and J.L.  Mothershead Jr., eds. 1970. Collected Papers of Clarence Irving Lewis, 240–257. Stanford: Stanford University Press. Mayoral, J. 2015. Significado, Conocimiento y Creencia En Kuhn. La Influencia de Wittgenstein y Austin. In Wittgenstein: La Superación Del Escepticismo, ed. D. Pérez Chico and J.V. Mayoral, 177–227. Madrid: Plaza y Valdés Madrid. Mayoral, Juan V. 2017. The Given, the Pragmatic A Priori, and Scientific Change. In Pragmatism in Transition: Contemporary Perspectives on C. I. Lewis, ed. P. Olen and C. Sachs, 79–101. London: Palgrave. Melogno, Pablo. 2019. The Discovery-Justification Distinction and the New Historiography of Science: On Thomas Kuhn’s Thalheimer Lectures. HOPOS: The Journal of the International Society for the History of Philosophy of Science 9 (1): 152–178. Sankey, Howard. 2018. Kuhn, Coherentism and Perception. Montevideo, Uruguay (unpublished). Williams, Michael. 2004. Wittgenstein, Truth and Certainty. In Wittgenstein’s Lasting Significance, ed. M. Kölbel and B. Weiss, 247–281. London: Routledge. Wittgenstein, Ludwig. 1969. On Certainty. G.  E. M.  Anscombe and G.  H. von Wright (eds.). New York: Harper and Row.

Chapter 3

A Defense of Structure in Structure of Scientific Revolutions K. Brad Wray

3.1 Introduction Kuhn’s Structure of Scientific Revolutions has been attacked for many reasons. Key analytic terms, most notably “paradigm,” were widely regarded as poorly defined (see Shapere 1964/1980; Stopes-Roe 1964; Sahlins 1964; Masterman 1970).1 For many readers, Structure seemed to suggest that the process of theory change is irrational, or at least non-rational (Lakatos 1970; Scheffler 1967). Imre Lakatos, for example, suggests that on Kuhn’s account, “scientific change is a kind of religious change” (Lakatos 1979, 93). Even Kuhn’s characterization of normal science seemed to some readers to paint a very unflattering picture of scientists as excessively dogmatic and uncritical (see Popper 1970; Watkins 1970). John Watkins suggests that, given Kuhn’s characterization of normal science, it is unclear how paradigm change is even possible (see Watkins 1970). More recently, as scholars from various fields began to the reflect on the 50th anniversary of the publication of Structure, Lorraine Daston has argued that the notion of “structure” that figures in the title of the book, as well as in the analysis of the history of science discussed throughout the book, is “dusty and dated” (Daston 2016, 116).

This is a substantially modified version of a chapter that appeared in my book Kuhn’s Intellectual Path (Cambridge University Press, 2021).  But also the concept of “incommensurability.” Many readers thought that Kuhn was claiming that competing theories were not even comparable (see, for example, Scheffler 1967, 81–83). Even his concept “scientific revolution” was subjected to criticism (see, for example, McMullin 1992). 1 

K. B. Wray (*) Centre for Science Studies, Aarhus University, Aarhus, Denmark e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 L. Giri et al. (eds.), Perspectives on Kuhn, The Western Ontario Series in Philosophy of Science 84, https://doi.org/10.1007/978-3-031-16371-5_3

25

26

K. B. Wray

Daston acknowledges that the concept was quite popular in the social sciences and history in the early 1960s. But she insists that, in light of our more developed historical sensibilities, these days the term “structure” is widely regarded as irrelevant to historical analysis. Daston argues that “most historians of science no longer believe that any kind of structure could possibility do justice to their subject matter. The very idea of looking for overarching regularities in the history of science seems bizarre” (Daston 2016, 117; emphasis in original). Instead, historians of science tend to focus on the particular, in recognition that the sorts of events they study are unique. My aim in this paper is to defend Kuhn’s appeal to the notion of structure as it is used in The Structure of Scientific Revolutions. First, I argue that Kuhn’s book is not primarily a contribution to the history of science, despite the fact that it cites many historical sources and articles in the history of science. Consequently, it is a mistake to regard Kuhn’s appeal to the notion of structure as a case of badly written history of science. Second, I argue that insofar as the book is a contribution to philosophy, the sort of thing Kuhn means by “structure” is perfectly respectable, and is often presupposed in many philosophical studies, especially in general philosophy of science. Third, I defend Kuhn’s analysis of science, especially the enterprise of identifying the structure of scientific revolutions.

3.2 Is Structure a Contribution to the History of Science? The book Structure of Scientific Revolutions can be challenging to classify. Indeed, even the author of the book can be challenging to classify. Kuhn had his formal training in physics, from Bachelors to Ph.D., completing his doctoral dissertation under the supervision of John Van Vleck. But he worked in a variety of different departments and programs, including the General Education program at Harvard, and then concurrently in the History Department and the Philosophy Department at U.C. Berkeley.2 During the time Kuhn was at Harvard, there were some significant radical experiments going on in the organization of the various disciplines. The social sciences, in particular, were exploring novel institutional structures. Joel Isaac has discussed this period in Harvard’s history extensively (see Isaac 2012). The Harvard Society of Fellows was one part of this “interstitial academy” that cut across disciplinary lines and Kuhn was a product of it. And, in the process of completing Structure, Kuhn also spent time at the Center for Advanced Study in

 He would later work in an interdisciplinary program in the History and Philosophy of Science at Princeton, and then in a Department of Linguistics and Philosophy at the Massachusetts Institute of Technology. Arguably, he seldom felt at home in any department or institutional arrangement. Kuhn notes, for example, that when he was leaving Harvard in the mid-1950s, “the philosophers at Berkeley wanted to hire a historian of science. They didn’t know that they didn’t want one, they didn’t know that this was not a philosophical discipline” (Kuhn 2000, 294). I discuss Kuhn’s departure from Berkeley in more detail in Chapter 7 of Kuhn’s Intellectual Path (see Wray 2021). 2

3  A Defense of Structure in Structure of Scientific Revolutions

27

the Behavioral Sciences, at Stanford University (see Wray 2021, 7). The interdisciplinary nature of the Center played a formative role in helping him develop his view of science. He was especially struck by the fact that the social sciences were so different from the natural science, insofar as the former do not even agree about the fundamentals in their field. On the one hand, it is understandable that some people have regarded Structure as a contribution to the history of science. The book does draw heavily on then-­ current historical scholarship. In fact, as I have reported elsewhere, over three quarters of the sources cited in Structure are either works in the history of science, scientific biographies, or scientific classics, like Darwin’s Origin (see Wray 2015). Indeed, the manuscript was initially commissioned to be part of the Encyclopedia of Unified Science, as a volume devoted to the topic of the history of science (see Kuhn 2000, 292; 300). And in fact historians of science did read the book with enthusiasm when it was initially published (see Kuhn 2000, 286). Charles Gillispie, a distinguished historian of science, reviewed the book in Science (see Gillispie 1962). It was also reviewed in Isis and The American Historical Review. But many others read the book as well, including economists, psychologists, sociologists, anthropologists, educational theorists, philosophers, and scientists (see Wray 2017; Kaiser 2016, 84, Table 4.1). A book that sells over 1.3 million copies will, no doubt, have a wide and diverse readership. In fact, Kuhn notes that he “used to say that if you go through college in science and mathematics you may very well get your bachelor’s degree without having been exposed to the Structure of Scientific Revolutions. If you go through college in any other field you will read it at least once.” (Kuhn 2000, 282–283; emphasis in original). Indeed, when he wrote Structure, Kuhn was a recognized and published historian of science.3 A number of his early publications were unquestionably contributions to the history of science, discussing very specific issues in the history of science, like “Newton’s ‘31st Query’ and the Degradation of Gold,” and “The Caloric Theory of Adiabatic Compression.”4 And his historical work is clearly situated in that genre. In fact, his major book-length contribution to the history of science, Black-Body Theory and the Quantum Discontinuity, 1894–1912, is clearly written with the historians’ aims in mind (see Kuhn 1978/1987). His intention was to understand the role and impact that Planck’s research on the black-body problem had on the revolution in early Twentieth Century physics that led to the development of quantum

 Robert Merton was somewhat underwhelmed by Kuhn’s early publication record. Merton notes that “by the age of thirty-two, when [Kuhn] made his application [for a Guggenheim fellowship], he had published few articles: principally, one with Van Vleck in physics on a simplified method of computing the cohesive energy of monovalent metals which appeared in Physical Review (1950: 382–88) and the other, a historical piece on Boyle and structural chemistry in the seventeenth century which appeared in Isis (1952: 12–36)” (see Merton 1977, 91–92). 4  Kuhn’s paper “Energy Conservation as an Example of Simultaneous Discovery” (see Kuhn 1959/1977) could be regarded as a contribution to the sociology of science, as Robert K. Merton had made multiple discoveries a key topic in the sociology of science, first, in Merton (1957/1973), and then more fully in Merton (1961/1973) and Merton (1963/1973). 3

28

K. B. Wray

mechanics.5 Indeed, that book was so historical that it disappointed his readers in sociology who were hoping for a critical analysis in the terms and concepts employed in Structure (see, for example, Pinch 1979). Instead, what they found was old fashion internalist history of ideas. Further, Kuhn was an active early participant in the History of Science Society, even contributing to the operations of the organization (see Kuhn 2000, 282). On the other hand, though, I think Kuhn makes clear that his aims in Structure are not those of a historian. Granted, he does claim that if we study the history of science it will transform our vision of science (see Kuhn 1962/2012, 1).6 But that is not a typical aim of historical scholarship. And Kuhn was quite aware of this. In his own reflective writings on the differences between history of science and philosophy of science, he makes clear that the two fields have very different objectives or aims (see, for example, Kuhn 1976/1977, 4–5).7 Whereas historians aim to develop a narrative, philosophers aim for a general theory. In this respect, philosophers are much more like sociologists, at least as the latter field was widely understood in the 1960s and 1970s. And, Kuhn’s understanding of history seems more in line with Daston’s understanding; history of science, and history more generally, is concerned with the particular. Whatever Kuhn’s intentions were, contemporary historians of science generally have a bleak view of the relevance and impact of Structure on the history of science. As noted above, Daston makes it clear that the notion of “structure” that figures so centrally in Kuhn’s analysis is outdated, and has no place in the history of science. She suggests that though the book provided a blue print of sorts for the professionalization of the history of science, the book fails to measure up to the contemporary standards of historical scholarship.8 Kuhn thus seemed to be announcing the coming of a more professional history of science in a book that would not measure up to the new standards to which it drew attention. Peter Galison’s assessment is equally damning. He describes the book as “a valiant and productive analysis of the physics of the 1930s done in the 1940s about the science of the seventeenth, eighteenth, and nineteenth centuries” (Galison 2016, 66). And Joel Isaac suggests that Kuhn’s theory of science does not properly characterize “the contemporary world of ­biotechnology, information science, and computer simulation” (see Isaac 2013,  Contrary to the standard narrative in physics, Kuhn believes that Planck did not fully understand his “discovery”. That is, he did not initially believe that energy was emitted in discrete units. He merely treated it as such in order to make his problem computationally tractable (see Kuhn 1978/1987). 6  The precise quotation is as follows: “history, if viewed as a repository for more than anecdote or chronology, could produce a decisive transformation in the image of science by which we are now possessed” (Kuhn 1962/2012, 1). 7  Kuhn was insistent that you cannot do both at once. One either worked as a historian of science or one worked as a philosopher of science (Kuhn 1976/1977, 4–5). Though Kuhn has inspired many to work in the history and philosophy of science, an integrated field, he did not see the fields as so intimately related to each other. 8  Other disciplines also saw Structure as providing a map of sorts on how to become scientific (see Wray 2017, 67–68). 5

3  A Defense of Structure in Structure of Scientific Revolutions

29

659). And Peter Dear notes that Kuhn “was never really a guide to historical research except by association” (Dear 2012, 425). Further, Dear suggests that “any historian of science who sits down to reread SSR will be struck by its almost archaic historiographical sensibilities” (Dear 2012, 426). At least one influential contemporary historian of science sees past the many citations to historical sources in Structure. David Kaiser argues that Kuhn drew the chief lessons in Structure from then-recent research in psychology, not the history of science (see Kaiser 2016). Both (i) Brunner’s and Postman’s anomalous playing cards experiment, and (ii) Theodor Erismann’s inverted glasses experiment play a crucial role in Kuhn’s presentation of his provocative theses.9 And Kuhn also discusses various gestalt images, like the duck/rabbit image. Kaiser even suggests that they may have been the source of Kuhn’s ideas about science, and that the various historical cases discussed were merely intended to illustrate his points (see Kaiser 2016).10 So, even though Kaiser does not criticize Kuhn’s historical scholarship in Structure, he does imply that the book is not a book in the history of science. Kaiser’s assessment of Kuhn seems to align more closely with Kuhn’s self-­ understanding than does either Daston’s or Galison’s assessment. At the end of his career, Kuhn makes clear that his interests in history of science were largely instrumental. Studying the history of science was a means to advancing his philosophical project. As Kuhn explains, “I could read texts, get inside the heads of the people who wrote them … I loved doing that. I took real pride and satisfaction in doing it. So being a historian of that sort was something I was quite willing to be and got a lot of kicks out of being … But my objectives in this, throughout, were to make philosophy out of it” (Kuhn 2000, 276).11 Kuhn not only aspired to do philosophy. He also identified as a philosopher. This is evident from the fact that Kuhn was very excited when he initially secured an appointment in a philosophy department at Berkeley. In his words: “I jumped at the change, because I wanted to do philosophy” (see Kuhn 2000, 294). And he saw the book, Structure, as a contribution to the philosophy of science. Kuhn explains: “my ambitions were always philosophical. And I thought of Structure … as being a book for philosophers” (Kuhn 2000, 276 and 307). Indeed, late in his life, when he describes the struggles he was having writing the book, it is clear he was explicitly engaging philosophers of science. When he spent the year at the Center for Advanced Study in the Behavioral Sciences at Stanford, he “tried to write a chapter on normal science.” In his words:  Daston discusses the inverted glasses experiment by Theodor Erismann in some detail in her contribution to the edited volume for the 50th Anniversary of the publication of Structure published by University of Chicago Press (see Daston 2016). 10  Kaiser also notes that many psychologists corresponded with Kuhn about Structure (see Table 4.1 in Kaiser 2016, 84). 11  Obviously, Kuhn’s later reflections on what he was thinking when he wrote Structure could be mistaken. In fact, there is some evidence that Kuhn’s memories of the past with respect to his use of the term “paradigm” changed (see Wray 2011, 49). Kuhn, though, seems to be aware of this. 9

30

K. B. Wray I kept finding that I had to  — since I was taking a relatively classical, received view approach to what a scientific theory was — I had to attribute all sorts of agreement about this, that, and the other thing, which would have appeared in the axiomatization either as axioms or as definitions. And I was enough of a historian to know that that agreement did not exist among the people who were [concerned].12

Indeed, even in the Preface to Structure itself, he is quite explicit about his philosophical ambitions. Kuhn notes that his involvement in teaching the history of science at Harvard with Conant profoundly altered his understanding of science. His “first exposure to … out-of-date scientific theory and practice radically undermined some of [his] basic conceptions about the nature of science and the reasons for its special success” (see Kuhn 1962/2012, xxxix). He notes that “those [pre]conceptions were ones [he] had previously drawn partly from scientific training itself and partly from a long-standing vocational interest in the philosophy of science” (see Kuhn 1962/2012, xxxix). Further, in the Preface to Structure Kuhn describes his “shift in … career plans, a shift from physics to history of science and then, gradually, from relatively straightforward historical problems back to the more philosophical concerns that had initially led [him] to history” (Kuhn 1962/2012, xxxix-xl). He thus characterizes Structure as a contribution to the philosophy of science, not the history of science. So it seems that those historians who have seen Structure as an example of poor historical scholarship are mistaken. It is not historical scholarship at all.13 Reflecting on the impact of Structure after 50 years, the historian of science Peter Dear remarks that “in 1979 Kuhn showed his cards by leaving Princeton to become a philosopher at MIT. I think that historians of science had long, by then, tended to conceive of Kuhn’s significance in the terms that the philosophers (and Kuhn himself) had established as central: incommensurability and relativism” (Dear 2012, 425).

3.3  Structure and Structures in Philosophy of Science Though Structure is a contribution to the philosophy of science, it was not a typical one, especially for its day. I now want to consider the extent to which an analysis of science in terms of structure is an apt contribution to philosophy of science. That is, I want to consider the extent to which the concept “structure”, a concept that Daston

 Kuhn (2000, 296). Alan Richardson questions whether Kuhn has the received view right. Richardson rhetorically asks: “is Kuhn arguing against logical empiricism in the Structure…?” Richardson answers as follows: “He is certainly not arguing against a sophisticated understanding of the mature work of Carnap, for example. He is arguing against an image of something he takes to be logical empiricism” (see Richardson 2007, 361). 13  It is noteworthy that Kuhn’s own student, the historian science John Heilbron, understood that Structure was a philosophical book (see Heilbron 1997, 505). Heilbron worked closely with Kuhn on the History of Quantum Physics Project, a project that was clearly historical. 12

3  A Defense of Structure in Structure of Scientific Revolutions

31

claims has no place in contemporary historical scholarship, can have a constructive role to play in the philosophy of science. The term “structure” has many connotations in philosophy. In contemporary philosophy of science, it is most often associated with a family of positions under the label “structural realism” (see, for example, Worrall 1989; and Ladyman 1998). Structural realists are concerned with the structural aspects of reality that are both (i) captured by and (ii) reflected in the structural aspects of our most successful scientific theories. Structural realism was developed as an attempt to reconcile the intuitions of scientific realism with the seemingly indisputable fact that the development of science is marked by radical changes of theory. This is not the sort of structure that concerned Kuhn in The Structure of Scientific Revolutions. The sort of structure that concerned Kuhn has its roots in sociology. The structures that interest sociologists are most often deep and elusive structures. And sociologists often purport to find structure in unexpected places. In this respect, the concept “structure” is similar to other key analytic terms in sociology, including “social construction” and “function.” As Robert K. Merton notes, it is latent functions, not manifest functions, that sociologists are in the business of revealing (see Merton 1949/1996). Latent functions are often related to unintended consequences. A social practice persists because of the effects it has, even if achieving the effect is no part of the motives of those who perpetuate the practice. For example, the practice of extended lactation common amongst hunter-gatherers may persist as a practice in these communities even if hunter-gatherers are not aware of its effect on suppressing fertility, and thus controlling population growth (see, Little 1991, 96–97, and Wray 2002, for discussions of this example). Similarly, more recently, sociologists have often appealed to the notion of “social construction” in order to evoke some wonder at something that one might otherwise not detect or expect (on social construction, see Hacking 1999, especially Chapter 1). Ian Hacking thus suggests that much of the sociological literature on social construction is concerned with unmasking, revealing something that would otherwise elude our attention. For example, when we see that gender is merely a social construction (assuming it is), we see that it is possible to change it and the social practices affected by prevailing gender concepts. Social constructionists aim to bring these otherwise elusive facts to our attention. This, I believe, is the sort of thing that Kuhn wanted to emphasize in his title: The Structure of Scientific Revolutions. If there is a structure to scientific revolutions, then it may not be an obvious fact. Kuhn, though, was not concerned with unmasking so that we could change science. Rather, he wanted to help us understand why science works so effectively. He did, though, believe that the common understanding of why science works so well was fundamentally mistaken. In invoking this notion of structure, Kuhn wants to show that scientific revolutions do not happen in some random, chaotic, or unpatterned way. Rather, they take on a particular form. In fact, according to Kuhn, it is not only scientific revolutions that have a structure. The development of a scientific field as whole has a pattern or form. Once a field has emerged out of the pre-paradigm stage, once all or most of the scientists working in a field accept the same theory, a scientific field takes on a

32

K. B. Wray

certain patterned structure. Roughly, the structure is as follows. A period of normal science, in which scientists take the fundamentals of a field for granted, leads to a period of crisis, caused by persistent anomalies that resist solutions. The crisis causes a slackening of the disciplinary norms and standards, which leads to the generation of new theories. The slackening of standards is essential, as the norms and standards of a normal scientific research tradition are quite rigid, and their rigidity discourages radical innovations. In fact, their rigidity is what makes scientists so effective in periods of normal science, ensuring the steady growth of knowledge that many associate with scientific progress. This is the essential tension in science, the tension between (i) a commitment to a tradition and (ii) a propensity for radical innovation. Finally, a new theory proves to meet the challenges the field faced and it becomes the dominant theory, which leads to a new normal scientific research tradition.14 This description of the development of the empirical sciences may strike some as audacious, just as it does contemporary historians of science. One might challenge Kuhn’s claim that he can see such regularities across so many disciplines, and over the course of hundreds of years. But I think the view Kuhn puts forward is less audacious than one might initially think. Consider the alternative. The alternative to an account like Kuhn’s is that there is no pattern or predictability in the development of a scientific field. Of course some things may happen with no predictable regularity, or follow no set pattern. But sociologists have uncovered surprising patterns in many sorts of phenomena. Émile Durkheim, for example, discovered robust patterns in suicide rates in a number of European countries throughout the second half of the Nineteenth Century (Durkheim 1930/1951). This was a tumultuous period in European history, with radical changes due to industrialization and a significant growth in population. But Durkheim did find resilient seasonal patterns in suicide rates; people tend to commit suicide more in the summer months than in the winter months. There are also resilient patterns with respect to marital status and suicide rates; marriage reduced the suicide rates among men, but increased the suicide rates among women. And there are patterns with respect to religious affiliation and suicide rates as well; Protestants commit suicide at a higher rate than Catholics. Studies like Durkheim’s should heighten our sensitivity and openness to finding a pattern of some kind in social phenomena. So there is some prima facie plausibility to Kuhn’s claim that revolutionary changes of theory in science happen according to some sort of pattern. Scientific research, after all, is conducted by social groups, and thus prone to the same sorts of dynamics that affect other social groups. But perhaps science is different, and there is no pattern or structure to it. Perhaps it is not like other sorts of social phenomena, like suicide, that have a structure. This suggested exceptionalism in science is somewhat doubtful. And it is worth noting

 That Kuhn was concerned with “general patterns of scientific development” was noted by Karl Hufbauer (2012, 450). 14

3  A Defense of Structure in Structure of Scientific Revolutions

33

that the project of looking for a structure in scientific change neither began with nor ends with Kuhn. Henri Poincaré, for example, suggested that as scientists make increasingly more penetrating investigations of nature a structure or pattern emerges in the way they conceive of nature. Specifically, Poincaré suggested that as scientists seek to develop a better understanding of the world they “discover the simple beneath the complex, and then the complex from the simple, and then again the simple beneath the complex, and so on, without ever being able to predict what the last term will be” (see Poincaré 1903/2001, 114). Even Karl Popper, who disagreed with Kuhn on many matters, believed that the sciences develop according to a pattern (see Popper 1975/1981). And, more recently, Andrew Abbott has argued that the social sciences, unlike the natural sciences, have a distinctive developmental structure that resembles a fractal pattern, driven by the recycling or reinvention of older concepts and distinctions (see Abbott 2001). In sociology, for example, a period in which conflict theorists are the dominant group gives way to a period in which consensus theorists dominate. And this period is followed by a period in which conflict theorists dominate, etc., with each new iteration absorbing something from its predecessor (see Abbott 2001, 17). So the case for there being a structure to scientific change is far less dubious than Daston implies. Granted, it may not be the sort of thing that interests historians or that has a place in a proper historical analysis of science, but that such a structure is there to be found is at least prima facie plausible. It was not just the macro-structure of scientific change that Kuhn thought had a pattern or structure. He insists scientific discoveries also have a structure (see Kuhn 1962/2012, Chapter VI). As Kuhn explains, “discoveries …are not isolated events but extended episodes with a regularly recurrent structure” (see Kuhn 1962/2012, 53; emphasis added). Here he is quite explicit that he is looking for a regular recurrent structure or pattern. He describes the structure in details: “[1] discovery commences with the awareness of anomaly … [2] it then continues with a more or less extended exploration of the area of anomaly. And [3] it closes only when the paradigm theory has been adjusted so that the anomalous has become the expected” (Kuhn 1962/2012, 53; numerals added). Kuhn is quite explicit that this is a general account of discovery. He notes, he is providing “an elucidation of the nature of discoveries” and “an understanding of the circumstances under which discoveries can come about” (Kuhn 1962/2012, 57; emphasis added). Kuhn also discusses the structure of scientific discoveries in an article published in the journal Science just prior to the publication of Structure. There Kuhn is even more explicit about there being a structure to scientific discoveries. “It is … just because [discoveries of the sort that concern Kuhn] demand readjustments [to the accepted theory] … that the process of discovery is necessarily and inevitably one that shows structure and that therefore extends in time” (Kuhn 1962, 764; emphasis added). The notion of structure is thus integral to his project. A general account of scientific change or scientific discovery will necessarily aim at identifying a pattern. I think there are very good reasons supporting Kuhn’s theory of scientific change. First, as Kuhn notes, every theory is a partial representation of reality. That is, every

34

K. B. Wray

theory directs attention to some variables and not others. The physicist who believes that light is a wave will attend to things that the physicist who believes that light is a particle will not, and vice versa. This, according to Kuhn, is why scientists working in a normal scientific tradition are generally so efficient in addressing their research goals (see, for example, Kuhn 1962/2012, 36). The theory they accept, the theory through which they have been socialized to see the world, will lead them to attend to only those features of the world that are regarded as salient, where what is “salient” is determined by the theory  that they accept. But insofar as a theory is partial, it is bound to be extended or applied to phenomena for which it is ill fit to model. Contrary to what some of Kuhn’s critics suggested, scientists are not at liberty to ignore anything that conflicts with their theoretical expectations. At one point, anomalies will surface. This seems to be an inevitable result of applying a theory too broadly, that is, to phenomena for which it was neither designed to take account nor suited to take account. Because of the partial nature of theories, every theory is bound to fail at some time (see Wray 2019). Consequently, the sort of cycle of change that Kuhn describes is to be expected. New theories, theories that make fundamentally different assumptions about reality than the theories we currently accept, will replace our current theories. Only then will scientists be able effectively to investigate the phenomena they seek to understand. Kuhn’s focus on the structure of scientific revolutions has not been completely lost by commentators. In the Introductory Essay to the 50th Anniversary edition of Structure, Ian Hacking notes that “structure and revolution are rightly put in the book’s title” (see Hacking 2012, x). As Hacking explains, “Kuhn thought not only that there are scientific revolutions but also that they have a structure. He had laid out the structure with great care, attaching a useful name to each node of the structure” (Hacking 2012, x).15 And, as noted above, Hacking recognizes that historians do not think of history as having a structure. Hacking attributes Kuhn’s penchant for looking for a structure to his early training in physics (see Hacking 2012, xi and xxxiii). As Hacking explains, “Kuhn’s instinct as a physicist … led him to find a simple and insightful all-purpose structure” (see Hacking 2012, xi). I am inclined to disagree with Hacking about the source of Kuhn’s thinking about structures in science. As Daston points out, when Kuhn was writing, the notion of structure was a popular concept in sociological analysis. Indeed, in America, in the 1950s and 1960s, structuralism (and functionalism) may have been the dominant paradigm in sociology. Robert K. Merton, the father of American sociology of science, developed a functionalist analysis of science that explained much in terms of the structure of scientific research communities, including, for example, priority disputes, and multiple discoveries. Kuhn was very familiar with Merton’s work, including Merton’s dissertation, Science, Technology, and Society in Seventeenth Century England, which Kuhn had read when he worked with James B. Conant on the History of Science courses at Harvard (see Kuhn 2000, 279; 287;  Hacking is not the only one to see the role of structure in Kuhn’s Structure. Alexander Bird (2015) also discusses this dimension of Kuhn’s project (see Bird 2015, 37). I discuss Bird’s view of Kuhn in more detail in the next section. 15

3  A Defense of Structure in Structure of Scientific Revolutions

35

see also TSK Archives: Box 1: Folder 7, Notes and Ideas, 1949). Kuhn was also in contact with Merton during the late 1950s and early 1960s (see Cole and Zuckerman 1975, 159). In fact, there are a number of letters back and forth between Kuhn and Merton in the Thomas S. Kuhn Archives at M.I.T. These include an exchange about the concept of an anomaly, simultaneous discoveries, and Kuhn’s measurement paper (see TSK Archives: Box 22: Correspondence: Merton, Robert K.). Merton even expressed a willingness to write a supportive letter on Kuhn’s behalf to University of Chicago Press if Kuhn encountered any difficulties in getting Structure published (see Merton and Barber 2004, 267; also Cole and Zuckerman 1975, 159). Further, in the Preface to Structure, Kuhn explicitly notes the influence of sociology on his thinking. He explains that, while he was a fellow at the Harvard Society of Fellows, two events led him to realize that his ideas about scientific change “might [need] to be set in the sociology of the scientific community.” The first was reading Ludwik Fleck’s book, Genesis and Development of a Scientific Fact, and the second was a discussion with a sociologist, Francis Sutton (see Kuhn 1962/2012, xli). Kuhn repeated similar sentiments in the Forward to the English translation to Fleck’s book. As he explains there, “Fleck’s text helped me to realize that the problems which concerned me had a fundamentally sociological dimension” (see Kuhn 1979, viii). Thus, Kuhn’s penchant for finding structures was more likely due to his familiarity with sociology of science rather than his training in physics.

3.4 What Structure Is Not It is worth distinguishing Kuhn’s view from the view that Popper calls “historicism.” Historicism is the view that “there [is] a trajectory to history, and that the task of the social sciences [is] to understand this trajectory scientifically” (see Shearmur 2017, 57; see also Popper 1944, 86; and Birner 2018).16 Popper was critical of historicism. More precisely, he argued against “the possibility of predicting historical developments to the extent to which they may be influenced by the growth of our knowledge” (see Popper 1957/1991, vii). That is, we cannot anticipate the future, when the future depends on the growth of our knowledge. The growth of knowledge will, after all, alter the world we live in, and in ways that we cannot anticipate. So  Andrew Reynolds notes that Popper’s characterization of historicism is at odds with another, popular characterization of historicism associated with Vico, Dilthey and Collingwood (see Reynolds 1999, 276–277). According to Reynolds, Popper characterizes historicism as the view that “there are to be found in history general laws, rhythms, or patterns. And with these the social sciences can make predictions about the future” (see Reynolds 1999, 277, emphasis in original). Thus the social scientist and historian, on this view, will employ the same methods as the natural scientist. In contrast, Vico and others characterize historicism as the view that “history has its own methods which are distinct from those of the natural sciences” (Reynolds 1999, 276, emphasis in original). Reynolds also suggests that “it would not be entirely strange to [classify] Kuhn’s thesis of scientific revolutions” as a form of what he calls Popperian historicism (see Reynolds 1999, 277). I argue against this interpretation, below. 16

36

K. B. Wray

making predictions about this world is problematic. Popper clarifies that he does not “refute the possibility of every kind of social prediction; on the contrary, [his argument against historicism] is perfectly compatible with the possibility of testing social theories  — for example, economic theories  — by way of predicting that certain developments will take place under certain conditions” (see Popper 1957/1991, vii). General trends are predictable. And social scientific theories that make such predictions can be tested. We merely have to either bring about the appropriate conditions and look for the appropriate effects, or look to the past to see if the predicted trend is manifest. What is not predictable, Popper claims, are specific future outcomes. I suspect that underlying Daston’s resistance to appeals to the notion of structure is a presumption that explanations about scientific change in terms of structure are inevitably linked to implausible historicist claims about science and scientific change. But this need not be the case, and Kuhn certainly does not link the two. Though Kuhn thinks that there is a pattern or structure to the development of science, he is adamant that science is not aptly described as going anywhere (see Kuhn 1962/2012, 169–172; see Wray 2011, Chapter 6).17 As far as Kuhn is concerned, there is no trajectory to the history of science. That is, there is no final goal, like the “true” account of reality, to which science aims, that explains the development of science. So, for this reason it would be a mistake to think of his view as a form of historicism, at least as Popper understood the term. Kuhn insists that we can explain the growth of scientific knowledge without appeal to teleology. Indeed, he thought that this would be one of the most contentious claims made in Structure. Instead, it was largely neglected, until relatively recently (see Kuhn 2000, 307). Significantly, the sorts of patterns that Kuhn purports to have identified are not attached to any sort of specific conceptual developments in any of the sciences. Thus, one cannot expect to predict precise changes of theory. One cannot even expect to predict the timing of the next change of theory. What Kuhn purports to be able to predict is the development of scientific fields in general. He makes no pretentions to being able to anticipate future conceptual developments. In this respect, Kuhn endorses what Hacking has called the contingency thesis. That is, the conceptual development of science is not predetermined (see Hacking 1999, 72–73). But, significantly, if Kuhn’s theory of science is correct it certainly tells us a lot about

 Shearmur notes that Popper was “also critical …  about their being a ‘plot’ to history, and of historical periodization” (see Shearmur 2017, 57). In one sense, Kuhn is offering a periodization of sorts in his account of the development of scientific disciplines or fields. But there is good reason to believe that Popper did not regard Kuhn as offering the sort of periodization that he objected to in his attack on historicism. After all, Popper never discusses Kuhn’s view in these terms, and he had plenty of opportunity to do this, including in his contribution to the Bedford College conference organized by Imre Lakatos, when he chose to take on Kuhn’s notion of normal science (see Popper 1970/1972). Comte clearly offers a historical periodization of the sort that Popper objects to. 17

3  A Defense of Structure in Structure of Scientific Revolutions

37

science and scientific knowledge. It highlights a pattern of change in science.18 And it suggests that the theories we currently accept will be replaced by different theories, theories that make significantly different assumptions about reality. What exactly these assumptions are Kuhn rightly does not say. That is precisely the sort of thing that his theory of science does not predict. Recently, Alexander Bird (2015) has argued that there are two historicist strains in Kuhn’s historiography of science. First, Kuhn is committed to “a conservative strand of historicism” according to which “the evaluation of a theory is relative to a specific tradition of puzzle-solving” (see Bird 2015, 25 and 26). This was quite contrary to the longstanding tradition in philosophy of science that sought to identify timeless and universal criteria for theory evaluation. Second, Kuhn is committed to a “determinist strand” of historicism, according to which “there is a … fundamentally cyclical pattern [to the development of science] with … alternating phases of normal and extraordinary (revolutionary) … science” (see Bird 2015, 25 and 27). Central to Kuhn’s cycle are disruptive revolutionary changes of theory that involve incurring some loss of explanatory power relative to the long-­ accepted theory. Bird claims that this, too, is at odds with the longstanding tradition in philosophy of science that regards the progress of science as strictly cumulative, with no setbacks. Popper, for example, claimed that scientists will only ever replace a long-accepted theory with another theory if it can explain more than the long-­ accepted theory can explain. What has come to be called “Kuhn-loss” was anathema to philosophy of science before Kuhn. Significantly, Bird’s characterization of the two dimensions of Kuhn’s historicism are not the same as the historicism to which Popper objected. Indeed, though Popper had many objections to Kuhn’s view none of them related to accusations of historicism. In summary, my aim has been to vindicate Kuhn’s appeal to the notion of structure in his account of science. First, I have argued that Kuhn did not write Structure as a contribution to the history of science. Though he draws on the history of science quite extensively, his aims are not those of the historian of science. Second, I have argued that the notion of structure has a legitimate place in the philosophy of science. As in sociological studies that appeal to structure, a philosophical investigation that appeals to the notion of structure signals a concern to get at a pattern in the phenomena. And there is no compelling reason to think that there is not a structure or pattern in the way that scientific fields change. Social scientists, as noted above, have found remarkable unexpected patterns in social phenomena. Third, I have tried to provide some reason to believe that Kuhn’s appeals to structure in his analysis of the dynamics of theory change is plausible. Given the partial nature of our theories, it is no wonder that they are ultimately rejected. So, given

 F. A. Hayek makes a similar claim about the predictive power of theories in the social and biological sciences (see Hayek 1967/1994). Hayek refers to theories in these fields as theories of complex phenomena, in contrast to the sorts of theories developed in physics. And Hayek certainly thought that even this degree of predictive power was useful and insightful. 18

38

K. B. Wray

what we know about scientific theories, we should expect that the theories we accept today will be replaced in the future by theories that make significant different assumptions about reality.

References Abbott, Andrew. 2001. Chaos of Disciplines. Chicago: University of Chicago Press. Bird, Alexander. 2015. Kuhn and the Historiography of Science. In Kuhn’s Structure of Scientific Revolutions-50 Years On, ed. R.J. Richards and L. Daston, 23–38. Dordrecht: Springer. Birner, Jack. 2018. Karl Popper’s The Poverty of Historicism after 60 Years. Metascience 27 (2): 183–193. Cole, Jonathan R., and Zuckerman Harriet. 1975. The Emergence of a Scientific Specialty: The Self-Exemplifying Case of the Sociology of Science. In The Idea of Social Structure: Papers in Honor of Robert K. Merton, ed. Lewis A. Coser, 139–174. New York: Harcourt, Brace and Jovanovich. Daston, Lorraine. 2016. History of Science without Structure. In Kuhn’s Structure of Scientific Revolutions at Fifty: Reflections on a Scientific Classic, ed. R.J.  Richards and L.  Daston, 115–132. Chicago: University of Chicago Press. Dear, Peter. 2012. Fifty Years of Structure. Social Studies of Science 42 (3): 424–428. Durkheim, Emile. 1951. Suicide: A Study in Sociology. Trans. J. A Spaulding and G. Simpson. New York: The Free Press. First published 1930. Galison, Peter. 2016. Practice All the Way Down. In Kuhn’s Structure of Scientific Revolutions at Fifty: Reflections on a Scientific Classic, ed. R.J. Richards and L. Daston, 42–69. Chicago: University of Chicago Press. Gillispie, Charles C. 1962. The Nature of Science: Normal Science Is Succeeded by a Creative Phase of Revolution Out of Which New Concepts Emerge. Science 138 (14): 1251–1253. Hacking, Ian. 1999. The Social Construction of What? Cambridge, MA: Harvard University Press. ———. 2012. Introductory Essay. In T. S. Kuhn’s Structure of Scientific Revolutions, 4th ed., vii– xxxvii. Chicago: University of Chicago. Hayek, Friedrich A. 1994. A Theory of Complex Phenomena. In Readings in the Philosophy of Social Science, ed. M. Martin and L.C. McIntyre, 55–70. Cambridge, MA: MIT Press. First Published 1967. Heilbron, John L. 1997. Thomas Samuel Kuhn, 18 July 1922-17 June 1996. Isis 89: 505–515. Hufbauer, Karl. 2012. From Student of Physics to Historian of Science: TS Kuhn’s Education and Early Career, 1940–1958. Physics in Perspective 14 (4): 421–470. Isaac, Joel. 2012. Working Knowledge: Making the Human Sciences from Parsons to Kuhn. Cambridge, MA: Harvard University Press. ———. 2013. Review of Thomas S.  Kuhn, The Structure of Scientific Revolutions, Fiftieth Anniversary Edition. Isis (3): 104, 658–159. Kaiser, David. 2016. Thomas Kuhn and the Psychology of Scientific Revolutions. In Kuhn’s Structure of Scientific Revolutions at Fifty: Reflections on a Scientific Classic, ed. R.J. Richards and L. Daston, 71–95. Chicago: University of Chicago Press. Kuhn, Thomas S. 1962. Historical Structure of Scientific Discovery. Science 136 (3518): 760–764. ———. 1977a. The Relations between the History and the Philosophy of Science. In The Essential Tension: Selected Studies in Scientific Tradition and Change, ed. T.S. Kuhn, 3–20. Chicago: University of Chicago Press. First published 1976. ———. 1977b. Energy Conservation as an Example of Simultaneous Discovery. In The Essential Tension: Selected Studies in Scientific Tradition and Change, ed. T.S. Kuhn, 66–104. Chicago: University of Chicago Press. First published 1959.

3  A Defense of Structure in Structure of Scientific Revolutions

39

———. 1979. “Forward.” In L.  Fleck, Genesis and Development of a Scientific Fact, Edited by T. J. Trenn and R. K. Merton, Translated by F. Bradley and T. J. Trenn., vii–xi. Chicago: University of Chicago Press. ———. 1987. Black-Body Theory and the Quantum Discontinuity, 1894-1912. Chicago: University of Chicago Press. First published 1978. ———. 2000. A Discussion with Thomas S.  Kuhn. In T.  S. Kuhn’s The Road Since Structure: Philosophical Essays, 1970–1993, With an Autobiographical Interview, ed. J.  Conant and J. Haugeland, 255–323. Chicago: University of Chicago Press. ———. 2012. The Structure of Scientific Revolutions. 4th ed. Chicago: University of Chicago Press. First published 1962. Ladyman, James. 1998. What Is Structural Realism? Studies in History and Philosophy of Science 29 (3): 409–424. Little, Daniel. 1991. Varieties of Social Explanation: An Introduction to the Philosophy of Social Science. Boulder: Westview Press, Inc. Masterman, Margaret. “The Nature of a Paradigm.” In Criticism and the Growth of Knowledge: Proceedings of the International Colloquium in the Philosophy of Science, London, 1965, I.  Lakatos and A.  Musgrave, (Eds.). First published 1970. Printed with corrections., Vol. IV:59–89. Cambridge: Cambridge University Press, 1972. McMullin, Ernan. 1992. Rationality and Paradigm Change in Science. In World Changes: Thomas Kuhn and the Nature of Science, ed. P. Horwich, 55–78. Cambridge, MA: MIT Press. Merton, Robert K. 1973a. Multiple Discoveries as Strategic Research Site. In The Sociology of Science: Theoretical and Empirical Investigations, ed. N.W.  Storer, 371–382. Chicago: University of Chicago Press. First published 1963. ———. 1973b. Singletons and Multiples in Science. In The Sociology of Science: Theoretical and Empirical Investigations, ed. N.W.  Storer, 343–370. Chicago: University of Chicago Press. First published 1961. ———. 1973c. Priorities in Scientific Discovery. In The Sociology of Science: Theoretical and Empirical Investigations, ed. N.W.  Storer, 286–324. Chicago: University of Chicago Press. First published 1961. ———. 1977. The Sociology of Science: An Episodic Memoir. Carbondale/Edwardsville, IL: Southern Illinois University Press. Merton, R.K. 1996. Manifest and Latent Functions. In On Social Structure and Science, ed. P. Sztompka, 87–95. Chicago: University of Chicago Press. First published 1949. Merton, Robert K., and Elinor Barber. 2004. The Travels and Adventures of Serendipity. Princeton: Princeton University Press. Pinch, Trevor J. 1979. Paradigm Lost? A Review Symposium. Isis 70 (3): 437–440. Poincaré, Henri. 2001. Science and Hypotheses. In The Value of Science: Essential Writings of Henri Poincaré, ed. S.J. Gould. New York: Modern Library. First published 1903. Popper, Karl R. 1944. The Poverty of Historicism, I. Economica 11 (42): 86–103. ———. “Normal Science and its Dangers.” In Criticism and the Growth of Knowledge: Proceedings of the International Colloquium in the Philosophy of Science, London, 1965, I. Lakatos and A. Musgrave, (Eds.). First published 1970. Printed with corrections, Vol. IV:51–58. Cambridge: Cambridge University Press, 1972. ———. 1981. The Rationality of Scientific Revolutions. In Scientific Revolutions, ed. I. Hacking, 80–106. Oxford: Oxford University Press. First published 1975. ———. 1991. The Poverty of Historicism. London: Routledge. First published 1957. Reynolds, Andrew. 1999. What Is Historicism? International Studies in the Philosophy of Science 13 (3): 275–287. Richardson, Alan. 2007. ‘That Sort of Everyday Image of Logical Positivism’: Thomas Kuhn and the Decline of Logical Empiricist Philosophy of Science. In The Cambridge Companion to Logical Empiricism, ed. A. Richardson and T. Uebel, 346–369. Cambridge: University of Cambridge Press.

40

K. B. Wray

Sahlins, Marshall D. 1964. Review of The Structure of Scientific Revolutions. Scientific American 210 (5): 142–144. Scheffler, Israel. 1967. Science and Subjectivity. Indianapolis: The Bobbs-Merrill Company, Inc. Shapere, Dudley. 1980. Review of The Structure of Scientific Revolutions. In Paradigms and Revolutions: Applications and Appraisals of Thomas Kuhn’s Philosophy of Science, ed. G. Gutting, 27–38. Notre Dame: University of Notre Dame Press. First published 1964. Shearmur, Jeremy. 2017. Popper’s Influence on the Social Sciences. In The Routledge Companion to Philosophy of Social Science, ed. L. McIntyre and A. Rosenberg, 55–64. London: Routledge. Stopes-Roe, H.V. 1964. Review of The Structure of Scientific Revolutions. British Journal for the Philosophy of Science 15 (58): 158–161. TSK - Kuhn, Thomas S. “Archives: Box 22: Correspondence: Merton, Robert K.,” n.d. ———. “Archives: Box 1: Folder 7, Notes and Ideas.” (Notebook), 1949. Watkins, John W.  N. “Against ‘Normal Science.’” In Criticism and the Growth of Knowledge: Proceedings of the International Colloquium in the Philosophy of Science, London, 1965, I.  Lakatos and A.  Musgrave, (Eds.). First published 1970. Reprinted with corrections., Vol. IV:25–37. Cambridge: Cambridge University Press, 1972. Worrall, John. 1989. Structural Realism: The Best of Both Worlds? Dialectica 43 (1–2): 99–124. Wray, K. Brad. 2002. Social Selection, Agents’ Intentions, and Functional Explanation. Analyse & Kritik 24 (1): 72–86. ———. 2011. Kuhn’s Evolutionary Social Epistemology. Cambridge: Cambridge University Press. ———. 2015. Kuhn’s Social Epistemology and the Sociology of Science. In Kuhn’s Structure of Scientific Revolutions ̶ 50 Years On, ed. W.J. Devlin and A. Bokulich, 167–183. Dordercht: Springer. ———. 2017. Kuhn’s Influence on the Social Sciences. In The Routledge Companion to the Philosophy of Social Science, ed. L. McIntyre and A. Rosenberg, 65–75. London: Routledge. ———. 2019. Discarded Theories: The Role of Changing Interest. Synthese 196 (2): 553–569. ———. 2021. Kuhn’s Intellectual Path: Charting The Structure of Scientific Revolutions. Cambridge: Cambridge University Press.

Chapter 4

A Vindication of Structure in Structure of Scientific Revolutions: A Comment to K. Brad Wray Pablo Melogno

4.1 Introduction In “A Defense of Structure in Structure of Scientific Revolutions” and in Wray (2021), Prof. K.  Brad Wray discusses the critical interpretation introduced by Lorraine Daston on the importance of Kuhn for historians of science, and specifically the usefulness of the Kuhnian notion of structure. According to Daston (2016), the notion of structure introduced by Kuhn in The Structure of Scientific Revolutions (Kuhn 1970, hereinafter Structure) became important due to the popularity of the term in social sciences when Kuhn wrote the book. Daston argues that the very idea of a structure has been abandoned both in history and social sciences, and therefore its use in Structure has become unacceptable. In Daston’s discussion, the meaning of ‘structure’ appears intricately linked to the search for historical regularities and the idea that the historian’s work entails unscrambling the mechanisms, trends, or laws underlying singular historical processes. On the one hand, Daston reminds us that historians have completely abandoned this kind of enterprise, and she also points out that Kuhn’s idea of a structure is directly inspired by this project: “Most historians of science no longer believe that any structure could do justice to their subject matter. The very idea of looking for overarching regularities in the history of science seems bizarre, a kind of leftover Hegelianism seeking a hidden, inexorable logic in the apparent vagaries of history.” (Daston 2016, 117). This article has been written in the framework of the project “The unpublished writings of Thomas Kuhn: 1980–1994” (CSIC-UdelaR), and with the support of a Posdoctoral Fellowship of the Fundación Carolina. P. Melogno (*) University of the Republic, Montevideo, Uruguay e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 L. Giri et al. (eds.), Perspectives on Kuhn, The Western Ontario Series in Philosophy of Science 84, https://doi.org/10.1007/978-3-031-16371-5_4

41

42

P. Melogno

This kind of critical assessment is not new in the history of science. As Wray (2011) has emphasized, similar challenges have been posited by historians like Shapin (1996) and Dobbs (2000). They have mainly targeted the notion of a scientific revolution and the value that Kuhn ascribes to it in his historiographical program. Arabatzis also reminds us that the historians’ reluctance to adopt the Kuhnian framework is not recent, “Kuhn’s grand narrative of scientific development was not well received by historians of science, who have been skeptical of his generalizations and have not adopted his terminology and conceptual apparatus (paradigm, normal science, crisis, revolutions, etc.) to describe and explain how the sciences have developed.” (2016, 196).1 Daston’s position is challenging because it focuses on a dimension of Structure that has been insufficiently discussed—surprising as it is that we still have one—in Kuhnian scholarship: the notion of structure and the role Kuhn assigned to it in the dynamic of scientific revolutions. Daston not only questions Kuhn’s arguments but also attempts to shed light—following a path similar to Fuller (2000) and Reisch (2005)—on the contextual factors that influenced Kuhn’s choice of the notion. Previously, Wray (2011) had indicated that the primary purpose of Structure was to identify some type of structure in scientific revolutions. In this frame, “A Defense of Structure…” intends to establish—against Daston’s position—the status and scope of Kuhn’s notion of structure. Wray tries to counteract Daston’s position by showing that: (1) Structure is not primarily a study in the history of science, so the notion of structure does not qualify as a tool for historical narrative; (2) Kuhn’s appeal to ‘structure’ is philosophical rather than historical, and (3) the Kuhnian notion of structure is philosophically interesting and legitimate. In this comment, I intend to vindicate the notion of structure in Kuhn’s work through a discussion with objectives analogous to Wray’s, but with differences in method and scope. First, I shall examine some implicit assumptions in Wray’s discussion. Secondly, I shall attempt to situate the discussion of Kuhn’s notion of structure in the historiographical realm, and not only within philosophy.

4.2 Wray vs. Daston First, I will identify the form of Daston’s argument. Second, I aim to determine which premises of the argument are challenged by Wray and which ones are not. Next, I shall venture some objections to Wray’s proposal, and in the following section, I shall develop an alternative strategy. Daston claims that historians of science have abandoned the notion of structure as a search for historical regularities. It has been displaced by a contextual and  We are not going to develop this here. Nevertheless, if Bird (2012) and Arabatzis (2016) are right, historians of science did not abandon the Kuhnian frame because of a break with the idea of a structure, nor because of a growing interest in specific cases—as Daston asserts. They did not abandon the notion of structure or the Kuhnian frame because they had never adopted them. 1

4  A Vindication of Structure in Structure of Scientific Revolutions: A Comment…

43

particularistic view of history, next to which Kuhn’s view seems obsolete. Paradoxically, this trend towards the singular processes is Kuhn’s direct legacy: “History of science has never been more resolutely historical in its methods (archival), and modes of explanation (contextual)… the historicism Kuhn prophesied and welcomed has ultimately dismantled the structures he sought…” (Daston 2016, 118). How does Daston arrive at this formulation? Her argument can be broken down as follows: 1 . The notion of structure implies the search for historical regularities. 2. Currently, this notion is no longer applicable to the work of historians of science. 3. In Structure, Kuhn aims to explain the history of scientific revolutions through the notion of structure. Conclusion  The Kuhnian notion of structure is not acceptable. Wray’s strategy to reject the conclusion consists in accepting—or at least in not challenging—the first and second premises and focusing his criticism on the third one. Wray does not open the debate about the relationship between the notion of structure and the search for historical patterns, neither does he question if this is outmoded for historians of science. His focus against the third premise is that Kuhn’s use of ‘structure’ is in the philosophical domain and not in the historical one. To defend this position, Wray looks for support in several pieces of the Kuhnian corpus. First, Kuhn himself claimed that Structure did not have historical purposes (see Wray, this volume). Also, Kuhn considered that history and philosophy of science were distinct disciplinary fields with different objectives, so that “Whereas historians aim to develop a narrative, philosophers aim for a general theory.” (Wray 2021, 138). Wray’s diagnosis is highly consistent with the fact that Kuhn did not attempt to identify long-term historical regularities (i.e., structures) in his recognized historical works (1957, 1978). Moreover, Wray highlights that the theses in Structure resort to the results of psychological empirical studies as their main source, and not those of the history of science. In Wray’s terms, all this makes Kuhn’s position immune to the problems raised by Daston’s first and second premises and to the situation within the history of science. Wray’s strategy is based on removing the discussion from the field of the historiography of science and confining it to philosophy. Therefore, it becomes possible to settle the status and legitimacy of Kuhn’s notion of structure. This issue will not be discussed in this paper. However, if we were to endorse Wray’s proposal, the role of the many historical cases included in Structure is not clear, and neither is the relationship between history and philosophy of science in Kuhn’s work.2 Nevertheless, Wray’s position is welcome by those who consider that the Kuhnian notion of structure has philosophical interest and relevance. Having said that, the emphasis Wray places on the concept’s philosophical dimension seems to leave a gap in its historiographical potential. We may wonder if a defense so  Different answers to these questions are available in Hoyningen-Huene (1993, 2008, 2012), Sharrock and Read (2002), Bird (2002, 2015), Mladenovic (2007), and Pinto de Oliveira (2012). 2

44

P. Melogno

slanted to philosophy does not imply withdrawing forces. Although Wray claims— in opposition to Daston—that Kuhn’s notion of structure is not strictly historical, he also seems to accept—as does Daston—the infertility of the notion for historians of science, as well as its connection with the search for historical regularities. Accordingly, the Kuhnian structure is—philosophically—preserved, but at the high cost of being stripped of any value in the historiographical plane.3 To avoid misunderstanding Wray’s arguments, we can present the problem differently: if the defense of Kuhn’s notion of structure is designed exclusively in philosophical terms, is there any scope to assign it some value in the field of history? In what follows, I will sketch an alternative strategy to address these difficulties.

4.3 Kuhnian Historiography We need to concede that Prof. Wray is right to claim that Structure is not primarily a work in the history of science, nor historiography of science. Neither its discussion agenda nor its methodological design is specific to historians of science. Historical examples are only included to ground or illustrate philosophical theses, and in most cases, historical material is supplied by well-established episodes. The book’s focus and style are clearly different from Kuhn’s works in the history of science written before and after Structure. Therefore, we can accept that Structure is not a historical work stricto sensu. However, even if we accepted that SSR contains neither new historical discoveries nor a proper historical narrative, we can still consider that the book offers some guidance on how to work on the history of science and how historians of science should organize their work. It also includes several general lessons about the nature of scientific revolutions, based on and inspired by historical cases. This entails considering Structure not as historical but as a work including some well-defined historiographical theses about how scientific revolutions occur. Kuhn’s historiographical commitments are founded on the idea that paradigms, scientific communities, anomalies—and other concepts of the like—act as constitutive elements of scientific revolutions. This allows us to claim that scientific revolutions are not irreducible and random processes, since they develop because of the interaction between relatively stable elements. It is unnecessary to take up the reader’s time with textual support for this interpretation. For brevity’s sake, I will only highlight its consequences. Conceding that Structure involves some theses about the historical dynamics of scientific revolutions impacts Wray’s proposal, since the scope of the Kuhnian notion of structure is not restricted to the philosophy of science but also shows a historiographical dimension.  Vasso Kindi (2005) also claimed that the history of science does not have any relevant role in Kuhn’s philosophical position. However, she did so by following a hugely different road to the one chosen by Prof. Wray. 3

4  A Vindication of Structure in Structure of Scientific Revolutions: A Comment…

45

However, it is unclear whether Prof. Wray’s position manages to limit the issue entirely to the philosophical domain. While describing Kuhn’s use of ‘structure,’ Wray points out that. In invoking this notion of structure, Kuhn wants to show that scientific revolutions do not happen in some random, chaotic, or unpatterned way. Rather, they take on a particular form. In fact, according to Kuhn, it is not only scientific revolutions that have a structure. The development of a scientific field as a whole has a pattern or form.4

If we consider that Kuhn claims that scientific revolutions do not occur randomly but follow a pattern, and if this pattern is repeated throughout history, we are attributing to Kuhn a historiographical thesis about scientific revolutions. I agree entirely with Prof. Wray in that this is precisely how Kuhn uses ‘structure’ in Structure. However, I see no way to consider Kuhn’s position as exclusively philosophical and lacking in historiographical implications. Understanding Structure in this fashion enables us to accept Daston’s third premise. Namely, that Kuhn offered a historical explanation of scientific revolutions appealing to the notion of structure. In these terms, the possible options are to deny the first premise, namely that the notion of structure implies the search for historical regularities, or to deny the second one, namely, that this notion is currently useless for historians of science. My argumentation entails accepting the second premise and questioning the formulation and the scope of the first one. I will question the negative perception historians of science have of the notion of structure. However, I will attempt to put up a fight over the connection between the appeal to structures and the search for historical regularities.

4.4 Structures as Historical Regularities The search for historical regularities has been a heterogeneous process, not endowed with univocal conceptual content. Many historians have conducted it in diverse and sometimes irreconcilable ways. Some posited rather formal tendencies; others identified a few essential elements as the causes of the entirety of historical development, and some advocated the existence of ineluctable historical laws. Examples are countless. Burke (1790) took from Aristotle the idea that corrupt democracies lead to tyrannies. Will Durant (1939) highlighted structural similarities between slavery structures in ancient Greece and North America. Braudel (1949) introduced the concept of long term—longue durée—to subsume the complete history of civilizations in a stratified series of historical times. Toynbee (1987) understood universal history as a set of successive dialectical cycles. As Arabatzis (2016) recalls, Comte, Whewell, and Mach were responsible for similar doctrines in the history of science. Daston seems to have these kinds of regularities in mind when she dismisses the notion of structure as bizarre and obsolete. The Hegelian image (cf. Daston) of  Wray (2021, 143).

4

46

P. Melogno

history as a mechanism with inexorable laws that allow us to predict the future was still common among the historians mentioned above. As Wray points out, I suspect that underlying Daston’s resistance to appeals to the notion of structure is a presumption that an explanation about scientific change in terms of structure, that is, an explanation of the sort that Kuhn offers, is inevitably linked to implausible historicist claims about science and scientific change.5

If the rejection of neo-Hegelian historiography inspires Daston’s objections to Kuhn, then we need to specify to what extent Kuhn’s position can be framed in this tendency. Alexander Bird claims that Kuhn adopted a Hegelian historicism, at least in two strands. A ‘conservative’ one, referring to the relative and tradition-­dependent status of the criteria for theory choice, and a ‘deterministic’ one, related to suggesting historical patterns.6 We will not examine the so-called conservative strand, since it does not affect the domain of our discussion. Regarding determinism, Bird posits that Kuhn’s Hegelian root falls within the search for historical patterns. In this broad sense, Kuhn has a Hegelian heritage, in the same way as Toynbee, Braudel, or Marx. However, as I will show below, Kuhn’s historical patterns are far more formal and lacking in content than the patterns introduced by Hegel or later Hegelian historians. We can accept Bird’s view of Kuhn as a Hegelian determinist, but we still need to refer to the different ways of adopting such determinism. All in all, Daston is right when she says that most historians—both inside and outside the history of science—no longer search for historical patterns in the style of Toynbee or Braudel. We can still inquire whether this is enough to trigger all possible searches for historical regularities and whether this is precisely Kuhn’s position. To elucidate this point, we must remember that the Kuhnian notion of structure implies identifying specific patterns in the historical dynamics of scientific revolutions. But the regularities proposed by Kuhn are formal, or at least they have a high degree of formality. In a Kuhnian frame, general assertions such as “Every crisis is caused by an anomaly” or “Scientific work is organized in communities” are acceptable. However, less general assertions such as “Every anomaly rises from a crucial experiment” are unacceptable. Unlike Braudel, Burke, or even Comte, Kuhn does not state that historical processes necessarily have a limited set of causes. Scientific revolutions have specific causes that include a high diversity of processes and historical contents. In Kuhn’s terms, every scientific revolution is preceded by an anomaly. But the reasons that give rise to the anomaly and the strategies displayed by scientists are singular, diverse, and specific to each scientific revolution. The same applies if we ask how long a paradigm can resist an anomaly in such a way that precludes any historical prediction. Similarly, every scientific community is defined by the consensuses reached, but their content and mechanism vary in each case.

 Wray (2021, 149).  “According to determinist historicism, the historian is not limited to describing and explaining particular events but may also hope to see an underlying pattern in the many particular events. In this respect, history has one affinity with the sciences.” (Bird 2015, 25) 5 6

4  A Vindication of Structure in Structure of Scientific Revolutions: A Comment…

47

If Kuhn had affirmed that anomalies necessarily appear after failures in a paradigm’s mathematical foundations, or because of the inability to predict new phenomena, he would have been appealing to the same type of regularity adopted by Burke when he says that the dispersion of power in a democracy necessarily leads to the emergence of an oligarchy. Nevertheless, the kind of regularity Kuhn attributes to the history of science does not entail that the same causal content can be ascribed to every scientific revolution. The regularities introduced by Kuhn only aim to clarify the formal mechanisms that define the revolutionary process. In this fashion, Kuhnian historiography enables the integration of the general and the particular, given that the constitutive causal mechanisms of scientific revolutions have different content in each historical episode. According to this, Kuhn’s notion of structure does not assume homologous content reasons—for example—for the transition from Ptolemy to Copernicus and those that explain the transition from Newton to Einstein, or from Priestley to Lavoisier. If we want to explain why and how each of these revolutions took place in the historical narrative domain, we have no choice but to reconstruct each case. There is no room for searching for regularities at this level, but to only for understanding the distinctive features of each revolutionary episode—as Daston puts it. But in a Kuhnian view, specific cases are reconstructed under the premise that there are common formal patterns underlying all scientific revolutions.7 Therefore, Kuhnian historiography does not allow us to introduce predictions about the historical dynamics of science. Furthermore, this is a relevant difference from the views of Braudel and Toynbee et al. After detecting the formal patterns involved in the notions of paradigm, crisis, anomaly, and scientific community, only the case-by-case reconstruction remains. This was also pointed out by Wray: What Kuhn purports to be able to predict is the development of scientific fields in some sort of general form. But, significantly, if his theory of science is correct it certainly tells us a lot about science and scientific knowledge. It highlights a pattern of change in science, and provides us with significant insight into what we can expect from the natural sciences. For example, it suggests that the theories we currently accept will be replaced in the future by theories that make significantly different assumptions about reality. But what exactly these new assumptions will be, no one can know in advance.8

Therefore, we can claim that a crisis precedes every scientific revolution, but we cannot predict when or why the next crisis will take place. In this sense, Kuhnian thought coincides with the Popperian anti-historicist dictum (Popper 1994) according to which we cannot predict the future development of our knowledge. Up to this point, our discussion does deny Daston’s first premise that the notion of structure is based on the search for historical regularities. Instead, it allows us to question how the premise is built and its role in the argument. The positions that include historical regularities are neither of the same kind nor can they all be  We will not examine whether these patterns represent necessary laws of historical dynamics, or just methodological tools for guiding historical reconstruction. Bojana Mladenovic (2007) has opened this debate, and she defends the second alternative. 8  Wray (2021, 150). 7

48

P. Melogno

criticized for the same reasons. The historical structures posited by Kuhn are more formal, less ambitious, and clearly weaker than those predominant among historians at the beginning of the twentieth century—and among some historians of science during the nineteenth century. Proposing structures and introducing historical regularities are elements of a project that can adopt different expressions, showing different degrees of formality and rigidity. Therefore, it is doubtful that the detection of historical regularities can be cleared off in all its possible forms with just one blow. In other words, Daston’s rejection of the historiographical use of the notion of structure requires a specific criticism of the historical patterns assumed by Kuhn, rather than a generic delegitimization of the notion as he used it.

4.5 Conclusion At the beginning, I anticipated possible agreements and disagreements with Prof. Wray’s text. As a result of this discussion, I agree with Wray that: (a) the Kuhnian notion of structure is legitimate and defensible, and (b) the notion responds to the search for constitutive patterns of scientific development. However, I disagree with Prof. Wray in taking (b) as an exclusively philosophical thesis, and I understand that its appropriate vindication comes into open battle within the field of historiography. My proposal clearly has the same objectives as Prof. Wray’s, although I use different tools and attempt to place the debate in the historiographical realm. How does this discussion impact Daston’s proposal? In principle, it requires that the criticism of Kuhnian historiography be the result of a specific examination of Kuhn’s historical regularities, not the product of a general rejection of the notion of structure mostly grounded on the prevailing consensus among historians of science. However, even beyond Kuhn, the search for formal patterns in the history of science is not an obsolete project among philosophers. It has been a constant feature in postKuhnian philosophy of science, as attested by the works of Toulmin (1972), Laudan (1984), Kitcher (1993), and Chang (2012).9 However, the strength and topicality of these philosophical views can hardly change the negative perception of most historians of science on the notion of

 Kitcher himself complained about the lack of systematicity in the historical works based on particular study cases. At the same time, he vindicated the search for general principles: “the contemporary attention to a wide range of examples is typically eclectic: practitioners borrow what general philosophical categories they need for the purposes of their individual studies, attempting only to rely on concepts and distinctions that seem to have survived Legend’s demise. Perhaps these individual studies represent the proper role for the discipline. Perhaps there is no general picture of the sciences that philosophers can give. Perhaps individual scientific disciplines and achievements are just that—individual, bound together by nothing more than ‘family resemblance.’ Although these suggestions may offer sound advice, I am not yet ready to abandon the search for generality.” (1993, 8) 9

4  A Vindication of Structure in Structure of Scientific Revolutions: A Comment…

49

structure and the search for historical patterns.10 Historians have mostly been reluctant to adopt the general explanations provided by philosophers, and historians of science are no exception. Even between the end of the nineteenth and the beginning of the twentieth century, when historians were extremely enthusiastic about structures, they preferred to build their models instead of borrowing philosopher’s elucubrations. As Hoyningen-Huene pointed out, “For reasons that have a long and indeed plausible history, historians of science as historians are often not particularly interested in what philosophers have had to say about science and its history” (2012, 281). This mismatch between historians and philosophers is neither new nor problematic. The logical empiricists were contemporaries of Toynbee and Braudel. They concentrated their efforts on the logical reconstruction of scientific theories and did not show the slightest interest in searching for historical patterns in science. The fact that an opposite trend prevailed among most historians did not lead them to change their views. In the same fashion, these days, most historians of science have no interest in historical patterns, and they are assuredly not going to change their minds on account of philosophers’ opinions. As an exact counterpart, and under the same principle of corporative health, philosophers can arrogate the right to keep cultivating the art of searching for historical patterns, even if pro tempore that practice is assessed by historians as bizarre and outmoded. It has been pointed out (Bird 2015; Arabatzis 2016) that Kuhn’s work entails a breakdown with the ahistorical philosophy of science of the early twentieth century. This was then followed by a return to the interest in historical patterns typical of the history of science in the second half of the nineteenth century. Kuhn’s historiographical theses about the dynamics of scientific revolutions have turned out to be extremely fertile for shedding light on the historical nature of science. Structure enabled the development of a philosophy of science with a better historical sensibility and introduced game rules followed even by Kuhn’s staunchest detractors. When considered in detail, the specific theses of the Kuhnian historiographical frame can be more or less acceptable. However, after Kuhn, it is no longer possible to address the problems of philosophy of science without considering the historical dimension of science. Indeed, this business must look sterile and rusty to any historian disappointed with Hegelianism, Marxism, the Annales School, and with any invocation of historical patterns. It would perhaps seem more useful and fruitful if the historians of the future—once again—became more interested in the skeleton of history and less attentive to its cells. Meanwhile, philosophers can vindicate—in Kitcher’s words—their determination not to abandon the search for generality.

 This kind of search occurs in the debates about philosophical problems (truth, progress, conceptual change), from which historians of science tend to keep a safe and cautious distance (cf. Arabatzis 2016). 10

50

P. Melogno

References Arabatzis, Theodore. 2016. The Structure of Scientific Revolutions and History and Philosophy of Science in Historical Perspective. In Shifting Paradigms: Thomas S. Kuhn and the History of Science, ed. A. Blum, K. Gavroglu, C. Joas, and J. Renn, 191–201. Bird, Alexander. 2002. Kuhn’s Wrong Turning. Studies in History and Philosophy of Science Part A 33 (3): 443–463. ———. 2015. Kuhn and the Historiography of Science. In Kuhn’s Structure of Scientific Revolutions-50 Years On, ed. W. Devlin and A. Bokulich, 23–38. Springer. Braudel, Fernand. 1949. The Mediterranean and the Mediterranean World in the Age of Philip II. Berkeley & Los Angeles: University of California Press, 1997. First published. Burke, Edmund. 1790. Reflections on the Revolution in France. Whitorn: Anodos Books, 2017. First published. Chang, Hasok. 2012. Ivs Water H2O?: Evidence, Realism and Pluralism. BSPS. Vol. 293. Dordrecht: Springer. Daston, Lorraine. 2016. History of Science without Structure. In Kuhn’s Structure of Scientific Revolutions at Fifty: Reflections on a Scientific Classic, ed. R.J.  Richards and L.  Daston, 115–132. Chicago: University of Chicago Press. Pinto de Oliveira, José Carlos. 2012. Kuhn and the Genesis of the ‘New Historiography of Science’. Studies in History and Philosophy of Science Part A 43(A) (1): 115–121. Dobbs, Betty Jo Teeter. 2000. Newton as Final Cause and First Mover. In Rethinking the Scientific Revolution, ed. M. Osler, 25–39. Cambridge: Cambridge University Press. Durant, Will. 1939. The Life of Greece. New York: Simon and Schuster. Fuller, Steve. 2000. Thomas Kuhn. A Philosophical History for Our Times. Chicago: University of Chicago Press. Hoyningen-Huene, Paul. 1993. Reconstructing Scientific Revolutions: Thomas S.  Kuhn’s Philosophy of Science. Foreword by T.  S. Kuhn, Translation by A.  T. Levine. Chicago: University of Chicago Press. ———. 2008. Thomas Kuhn and the Chemical Revolution. Foundations of Chemistry 10 (2): 101–115. ———. 2012. Philosophical Elements in Thomas Kuhn’s Historiography of Science. Theoria 75 (3): 281–292. Kindi, Vasso. 2005. The Relation of History of Science to Philosophy of Science in the Structure of Scientific Revolutions and Kuhn’s Later Work. Perspectives on Science 13 (4): 495–530. Kitcher, Philip. 1993. The Advancement of Science: Science Without Legend, Objectivity Without Illusions. New York: Oxford University Press. Kuhn, Thomas S. 1957. The Copernican Revolution. Cambridge, MA: Harvard University Press. ———. 1970. The Structure of Scientific Revolutions. 2nd ed. First published 1962. Chicago: University of Chicago Press. ———. 1978. Black-Body Theory and the Quantum Discontinuity, 1894–1912. Oxford/New York: Clarendon and Oxford University Presses. Laudan, Larry. 1984. Science and Values. Berkeley: University of California Press. Mladenović, Bojana. 2007. ‘Muckraking in History’: The Role of the History of Science in Kuhn’s Philosophy. Perspectives on Science 15 (3): 261–294. Popper, Karl. 1994. The Poverty of Historicism. London: Routledge. First published 1957. Reisch, George A. 2005. How the Cold War Transformed Philosophy of Science. Cambridge: Cambridge University Press. Shapin, Steven. 1996. The Scientific Revolution. Chicago: University of Chicago Press. Sharrock, Wes, and Rupert Read. 2002. Kuhn: Philosopher of Scientific Revolutions. Cambridge: Polity Press. Toulmin, Stephen E. 1972. Human Understanding, The Collective Use and Evolution of Concepts. Vol. 1. Oxford: Clarendon Press.

4  A Vindication of Structure in Structure of Scientific Revolutions: A Comment…

51

Toynbee, Arnold J. 1987. A Study of History: Volume I: Abridgement of Volumes I–VI. First published 1939. New York: Oxford University Press. Wray, K.  Brad. 2011. Kuhn’s Evolutionary Social Epistemology. Cambridge: Cambridge University Press. ———. 2021. Kuhn’s Intellectual Path: Charting The Structure of Scientific Revolutions. Cambridge: Cambridge University Press.

Chapter 5

Kuhn’s Reconstruction of Structure: The Theoretical Background Juan V. Mayoral

5.1 Introduction In The Structure of Scientific Revolutions, Kuhn expounded his main beliefs on the nature of science in a way that seemed for him, at least for a while, definitive.1 Yet, his writings before and after Structure, a significant part of them unpublished, do not support the claim that Structure exhausted those aspects of science and its development Kuhn had planned to entertain. Philosophical issues of scientific development attracted his attention a decade before a first draft of Structure was finished and around 8 years before he first started to write the first chapter. As a Harvard junior fellow, Kuhn explored those issues for the first time.2 Similarly, when he considered that his historiographical work on the archival and oral sources of quantum mechanics and on the history of the old quantum theory had been finished, he turned his attention again to philosophical aspects of scientific development from a new point of view. In this chapter, I shall try to show that the I am indebted to Paul Hoyningen-Huene, Pío García, Pablo Melogno, Howard Sankey, K. Brad Wray and other persons in the audience of the Buenos Aires and Montevideo workshop in which this paper was presented for their interesting commentaries on it. I am also grateful to Megan Watkins and Ashley Musick for editing its English. Research for this paper has been supported by the Research Project FFI2017-84781-P, co-funded by the Spanish Research Council and the European Regional Development Fund.  See Kuhn’s letter to Charles Morris, 13 Oct. 1960, TSKP 25.53, 1–2.  See Mayoral (2017a, esp. Chs. 5, 8), on this process of writing Structure. See also Pinto de Oliveira (2017) on the unfinished first chapter of Structure from around 1958. 1 2

J. V. Mayoral (*) Universidad de Zaragoza, Zaragoza, Spain e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 L. Giri et al. (eds.), Perspectives on Kuhn, The Western Ontario Series in Philosophy of Science 84, https://doi.org/10.1007/978-3-031-16371-5_5

53

54

J. V. Mayoral

background to that new point of view was as old as his oldest texts on philosophy of science. That shall be one of my main goals in this paper. Two related ideas are involved in that point: first, there is a common background –  a theoretical background – that unifies Kuhn’s earlier and latter stages of his intellectual development; and second, Kuhn draws explanatory resources from that background at different moments of his intellectual development. Although the theoretical background remains stable throughout those years, Kuhn never was completely satisfied with the presentations of his theory at large. The Structure of Scientific Revolutions, a work whose public and academic influence is without question, was less definitive than Kuhn had supposed it to be when he finished it. Certainly, in his (so far) unpublished book, The Plurality of Worlds, Kuhn tried to make that theoretical background explicit, as a whole, for the first time ever. In Sects. 5.2, 5.3 and 5.4 of this chapter, I shall reconstruct that background from the point of view of the lexical theory Kuhn developed in the last stage of his life. In that stage, Kuhn leaned toward a linguistic treatment of the philosophical problems of scientific development. Yet, this latter point must be qualified as follows: first, Kuhn’s involvement with the linguistic aspects of scientific theories is not new at all. His concern is very old, actually, as we find theses on language and scientific change in the late 1940s and early 1950s.3 And second, when the linguistic approach comes to the foreground, it is usually accompanied by considerations on perception, learning, and representation. Again, this latter qualification applies to his approach to scientific development in the early fifties as well as in later stages. Therefore, in Sects. 5.2, 5.3 and 5.4 I shall start by expounding Kuhn’s (well-known) lexical theory, then I will finish by showing how a similar theoretical background is present in a preparatory work for Structure, Kuhn’s Lowell Lectures, 1951 – The Quest for Physical Theory. That is one of the main goals of this paper, but there is another explicit goal. There are (roughly) a couple of versions of that theoretical background, and, whereas the first version in the early fifties helped him to see what was still left for Structure to become a well-grounded argument, the second version was the basis for reconstructing that argument so that some of the problems critics had pointed out in between could be solved. The second version, then, became a source of solutions for long-standing problems in Kuhn’s theory since Structure. My reconstruction of Kuhn’s late theory in Sects. 5.2 and 5.3 shall be used as a basis to show how he addressed two main issues in his developmental theory, namely, the problem of our historical identity and the related problem of relativism. These two problems are discussed later on, in Sects. 5.5 and 5.6, respectively. A summary of the chapter is as follows. In Sect. 5.2, I show Kuhn’s account of the task ahead for the individual in charge of studying scientific development. As in section I of Structure and other works, Kuhn defends that the history of science plays a key role in evaluating the rationality of scientific change: an ethnographic  See Mayoral (2017a, 299 ff.; 2017c, 126, fn. 33). This critical remark about Kuhn’s alleged “linguistic turn” is also visible in Reisch (2019, 355) (see also 416, fn. 1, for further references about studies on that “turn”). 3

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

55

role. Once the results of that activity are available, Kuhn argues, the philosophical examination begins, though it is not the previous epistemological analysis of the justification of individual beliefs anymore. In Sect. 5.3, I show something well known: Kuhn’s model of lexicon by means of which those two related activities are fulfilled. The lexicon is a source of meaning for our kind terms and also the vehicle for the transmission of scientific knowledge. On that basis, Kuhn criticized the causal theory of reference by Saul Kripke and Hilary Putnam, and I will also summarize his point regarding that theory in this section. In Sect. 5.4, I show the link between the lexical theory so far expounded and the background to Structure in the fifties. In Kuhn’s earliest depiction of conceptual schemes – a notion he inherits from the philosophical tradition at Harvard – he uses linguistic (and nonlinguistic) arguments to support his pragmatic version of their function.4 In that version, two main poles are visible. One of them is the individual and, as it will be argued here, it plays a key role (with its variability, rationality and judgmental abilities intact) in the social perspective of scientific knowledge Kuhn conveys in Structure and further works. The other pole is the world, without which there is no science at all – or else science becomes a purely idealistic game. These ingredients emerge in Kuhn’s early background to Structure and their representation is enhanced and improved in the post-Structure stage. I will show that in this section. Finally, before Sect. 5.7 concludes the paper, Sects. 5.5 and 5.6 elaborate on two issues whose solution is important to the late Kuhn: how we forge a historical scientific identity despite the relativistic consequences of Structure (sect. 5.5) and how Kuhn’s relativism should be understood in order for it to be an acceptable position.

5.2 Ethnography and Reconstruction In The Presence of Past Science (The Shearman Memorial Lectures, delivered at London in 1987), Thomas Kuhn stated that history gives shape to scientific knowledge. So, when we talk about the latter, we do not refer to a permanent core that is gradually augmented by adding new statements, but rather to a whole that is in a continuous state of change and evolution.5 Scientific knowledge is, thus, a process and this view has consequences for the philosopher of science’s epistemological perspective. Kuhn supports the idea that, against a “static” tradition in philosophy of science, his point of view should be considered a new “dynamic” program, whose main issue is explaining the justification of belief change, not belief

 Concerning Kuhn, the notion of “conceptual scheme” and its origins (as regards Kuhn’s theory alone), see Mayoral (2017a, esp. chap. 4). 5  PPS, I, 2. 4

56

J. V. Mayoral

itself (more on this later). Scientific Development and Lexical Change, another series of lectures, makes precisely this same point.6 For Kuhn, development takes place between two sets of beliefs, which I will call “theories” for convenience (in sect. 5.5 I will qualify this term), T and T’, where T’ comes after T historically. For Kuhn, there is usually a T that serves as a foundation for a subsequent T’ in two senses: (a) T is the conceptual and instrumental foundation for any discovery that comes with T’, and it also provides the pool of data that T’ accounts for; (b) the background empirical knowledge that helps to find justification for statements in T’ is its predecessor T.7 Discovery and justification of theories (or rather, of their components, beliefs or statements) are therefore historical processes in which beliefs coming from different theories are put together. If we take a further step from the theoretical level to the epistemological level, we should say, following Kuhn, that intentional views of the world are also put together. For Kuhn, these two aspects of the same historical development are examined from two different points of view, namely, the historian of science’s point of view and the philosopher of science’s point of view. Kuhn says that these perspectives can be combined, but also that the historian’s and the philosopher’s stages must not be confused. We should rather distinguish between two different activities: 1. Ethnographic Phase. In this phase, the historian or philosopher behaves like an ethnographer, who tries to reconstruct a certain language and the form of life in which the former is included, and also tries to make sense of that setting for a particular audience. If we considered scientific theories as the example of the language to be reconstructed, we would be talking about reconstructing T, for instance. 2. Developmental Phase. The philosopher or historian tries to display the road from T to its successor, T’, i.e., the theory that emerges from T’s cognitive foundations.8 Kuhn emphasizes that activity 1 is not merely a series of quotations (in literal terms), insofar as it is not possible to make sense of certain statements in T or T’ without taking into account the background composed of theories themselves and the semantic network created by the terms involved and their use (all of which gives access to the community members’ intentional point of view). For instance, a quotation from Aristotle or Ptolemy without further remarks – a mere statement – may be understandable from a purely grammatical point of view but might seem nonsense or simply false for us today.9 The historian’s ethnographic activity is  See SDLC, I; Kuhn (1987, 14, n. 2; 1991a, 91; 1993, 226). See Melogno (2019), for this latter series and the distinction between static and dynamic perspectives. 7  See PPS, I, 3. See also Kuhn (1993, 226), where he shows the connection with C. G. Hempel’s (1965) substitution of the observational/theoretical vocabulary distinction with, respectively, “antecedently available terms” (in T) with the new vocabulary (in T’). Kuhn alludes to Hempel in that regard in Kuhn (1977c, 12, 282, fn. 29; NCC, II, 39). 8  See PPS, I, 3–4. 9  This point is made by Kuhn in an unpublished lecture (Kuhn 1976). 6

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

57

meant to make sense of such quotations by providing beliefs that are assumed by, or implicit in, the historical actors’ statements. The kind of beliefs Kuhn talks about is something to be explained later. A further thought on the same theme is that it makes no sense, therefore, to start phase 2 without having accomplished activity in 1 satisfactorily.10 Seen in this way, it seems clear that, in Kuhn’s theory, nature is examined by categorizing with regard to certain cognitive categories and that discovery processes include a re-examination and revision of those categories, which are sometimes triggered by the emergence of an anomaly acting as an obstacle to normal research.11 Phase 1, the ethnographic one, includes a hermeneutical reconstruction of those categories and, accordingly, an adequate exposition of the lexicon employed in T that responds to the beliefs (whether empirical or conceptual) that are common in that historical stage.12 For Kuhn, on the other hand, those lexicons are not enough to control every aspect of nature; so sooner or later they are replaced by alternative ones that have to account for old observations and help to unearth and term newer ones, anomalies in T included. Thus, a new lexicon arises and accordingly a new setting for beliefs arises, too, in which the new lexicon is the proper guide for the community’s scientific behavior. Phase 2 aims at showing how we pass from one lexicon to the next: how the individuals included in the new community choose the new lexicon, adopt it, and give shape to it; how the set of beliefs changes; and how categorial classifications are replaced with new ones in order for the community to progress from T to T’. If the resulting effect when comparing T and T’ is that lexicons do not completely fit one another, then T and T’ are locally incommensurable theories. Incommensurability is thus explained in terms of their respective lexicons.13 The historian and philosopher of science’s ultimate task is to reveal the specific limits of incommensurability for the example of revolutionary transition from T to T’ that has been selected in each case. The historian of science must reconstruct the basic lexical classification and the philosopher must explore the foundations of change in terms of comparison of T’ with T – and on the basis of common aspects of those theories (more on this below). From Kuhn’s point of view, philosophy of science is no longer a reconstruction of a static, mostly permanent body of beliefs,

 See PPS, I, 30.  Although I may be talking, in what follows, about the classification of individuals in natural kinds, Kuhn criticizes Ian Hacking’s 1993 nominalistic interpretation of his work (and thus referring to individuals alone) and prefers referring to “substances” and “lifelines”. See NCC, III, as well as RSS (more on this in sect. 5.3). See also Kuhn (1993, 229; 1990c, 10 ff.). 12  On that phase and its foundations, see SDLC, II, esp. 30. In the same place (see 30–52), Kuhn expounds the well-known examples on scientific change that come from Aristotle’s physics, Alessandro Volta’s work on batteries and on Max Planck’s work in the old quantum theory, all of which conform to the text of Kuhn (1987, see esp. 13). As Pablo Melogno (2017, 16 ff.) says, there are few differences between both texts. 13  On changes in the concept of incommensurability from Structure to his late works see Sankey 1997, and Andersen et al. (2006, esp. 106–107). 10 11

58

J. V. Mayoral

but rather an examination of a dynamic process in which rational decisions between bodies of beliefs are made.14

5.3 What Constitutes a Lexicon? In The Presence of Past Science, a lexicon is composed of natural-kind terms that label the classes of objects in the categorial classification of the world. The elements in these classes are subject to scientific laws and their terms appear in these too. They also appear in counterfactual conditionals.15 There are, for Kuhn, two main features for natural-kind terms: 1 . They satisfy the no-overlap principle (NOP). 2. They are labeled in a special manner. Satisfaction of NOP (feature 1) is easy to explain. In Kuhn’s version, given two natural-kind terms, F and G, their respective extensions do not overlap unless one is fully included in the other as the genus contains the species (see Kuhn 1991a/2000, 92). As regards the labels (feature 2), if we take two identical terms that belong to two different languages, e.g., ML1 and ML2, respectively ascribed to languages L1 and L2, both being characterized by the same word (e.g., “motion”), one of them is preserved as an obsolete term within the other’s language – a sort of relic that helps to understand the past  – and the other term is the natural-kind term in its own language and is thereby used when stating the laws, expressing counterfactual conditionals, and so on.16 It is, as Kuhn used to say, following Nelson Goodman, “projectable.”17 So, let us say again that ML1 and ML2 symbolize “motion” in Aristotle and Newton, respectively. For Kuhn, ML1, the label that belongs to the language of Aristotelian physics, L1, could be preserved in L2, the language of Newtonian physics. Yet, if L2 rules a mechanical understanding of physical phenomena, then only “motion,” understood as ML2, provides the meaning and applies to referents appropriate to the use of that term in L2; meanwhile, “motion” – as ML1 – would only serve for historiographical purposes (and with some cautionary notes).18 In his “Afterwords”, 6 years after Presence, Kuhn qualifies that, instead of talking about “natural kinds,” he should rather be talking about “kinds and kind terms  See SDLC, I, 2.  See PPS, II, 40. 16  Kuhn’s example (“motion”) is not completely appropriate here, insofar as (as Kuhn acknowledges), as regards the terms Aristotle uses (“kinesis” and “metabole”) and their meanings, we cannot assure that we are talking about the same term (“motion”) that we use in Newtonian physics. See PPS, I, 7 fn. 5. 17  See Kuhn (1989, 74–75; 1993, 230–232). See also PPS, II, 40–41; III, 80. 18  See PPS, II, 41. Kuhn acknowledges J. S. Mill’s A System of Logic (Mill 1858/2011) as the main influence on his view of these terms; see Kuhn (1987, I, 30–31; II, 41, fn. 8). Buchwald (1992, 39) says that Kuhn went ultimately beyond that influence as the next paragraph of this paper shows. 14 15

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

59

in general,” a usage he adopts in his subsequent work (see Kuhn 1993/2000, 229). In that case, NOP and projectability remain key features of those terms and kinds. Labels for these latter are distinguished, Kuhn says now, by virtue of our grammatical habits with them (see the way, e.g., we use expressions such as “a duck” or “a gold ring”) and also because the pragmatic context is key, because their usefulness is key as well in the way we learn them and capture their meaning, thereby creating “expectations”.19 Our lexicon mirrors our classification of the world, whether natural or social: it mirrors the way we characterize incoming information. For Kuhn, this kind of classification is not limited to our species – though the fact that it is linguistic is exclusive of our species. By virtue of natural-kind terms and property terms (the latter help to define the former) we convey our beliefs concerning the surrounding world, present, past and future. For example, as regards the natural-kind term “wolf,” by means of some known properties like fur, or social, reproductive and food behaviors, it is possible for us to use the term properly and even to give a quasi-­ definitional description of it. As Kuhn says, the lexicon embodies our knowledge about the world – about its inhabitants – by way of the correct use of language.20 Kuhn says something like this in The Quest for Physical Theory, in the early 1950s, and years later in “Second Thoughts on Paradigms” (more on this below, sect. 5.4).21 Parts of the lexicon may be more general than others, Kuhn says. Some of its terms might be considered innate, because they are usually present; e.g., terms for general reference to objects, colors and so on. Other terms can be found in some human groups – but not necessarily in all – and are exclusive to their particular use of language. This particularity helps to understand the limits that Kuhn defines as barriers to our access to the historical past.22 This lexical relationship lets us talk about something I shall deal with in Sect. 5.5 of this chapter – our link with the past (i.e., with our past) and the way we forge our scientific identity.23 That link, Kuhn says in Scientific Development and Lexical Change, is based on a common biology and cognition and on what we could call a causal background, context or, simply, world. This latter is a source of stimuli and is (or rather we should assume it to be) stable.24 Kuhn shows, in this way, some inclination toward a commonsense realism that would be useful if we look for some “brotherhood” between diverse scientific cultures.

 See Kuhn (1993, 229–232). I will not describe here Kuhn’s well-known distinction between kinds of general statement – “normic” versus “nomic,” i.e., with or without exceptions – on which their specific disposition to be projectable depends, and so the way they are present to us in language. On all this, see Kuhn 1993, 230–233. 20  See PPS, II, 42, 47, 53, 59. 21  See QPT, VIII, 7 ff.; GR 139 ff.; Kuhn (1977a, 309–312). 22  See PPS, II, 43. 23  There are some significant changes in view for Kuhn as regards this issue; more on it in Sect. 5.5. 24  On this causal background of stimuli in Kuhn’s work, see Hoyningen-Huene (1993, chap. 2, esp. sect. 2.2). 19

60

J. V. Mayoral

In scientific research, the search for terms includes the vocabulary that is common to us from an intuitive or commonsensical point of view. I am referring to those terms that name animals, plants, minerals, etc. That search, however, includes other less usual terms. An example for this latter kind of term would be Kuhn’s own case on Alessandro Volta: “battery.”25 That search also includes those other terms whose recognition grants us access to the previous ones. This idea of recognition involves previous acquaintance with a set of properties we know how to recognize and some others that we learn to recognize. For Kuhn, as we know, our knowledge of kind-­membership does not depend on the satisfaction of necessary and sufficient conditions, like definitions, properly speaking, of kind terms, but rather on recognition of a group of characteristics that can be observed in the object in question, which are sufficient for us to pronounce on the object at issue and its kind-­membership  – i.e., on its F-ness or its G-ness (given any two kind terms F and G).26 That a certain group of characteristics is sufficient for an individual to recognize kind-membership cannot be considered more general than that. It cannot be considered the resulting group of characteristics a community of speakers have agreed upon. This is, at most, a likely result of an extended practice. In short, therefore, kind-membership depends on a sufficient condition whose composition is context-sensitive and whose sufficiency is given in terms of pragmatic conditions. Let us talk a bit more on this issue. I shall start by picking out two sets of terms that shall help us to make the case easier to understand. In one of those sets, group 1, we will find natural-kind terms – or rather kind terms, as the later Kuhn corrected himself. In the other set, or group 2, we shall find property terms. Kuhn would say that terms in group 2 allow stating kind-membership to group 1. So, for Kuhn, given a group-1 term, “planet” for instance, there will be a series of characteristics that belong to group 2, such as “brightness,” “apparent motion,” and other qualifications for this latter property, such as “circular path,” “rectilinear path,” etc., which constitute the “feature space” in virtue of which we set up kind-membership, e.g., of the planets, stars, comets, etc.27 Once group-2 members are given, group-1 members are easily gathered in terms of their similarities and differences. Kuhn alludes to Wittgenstein as his source as regards meaning.28 I said above that for Kuhn the sufficient condition to affirm kind-membership is context-sensitive. That is, the composition of that condition depends on the context in which the kind (or group-1) term is meant to apply. Some contexts require a  PPS, II, p. 45.  See QPT, VIII, 7 ff.; GR 139 ff.; Structure, 45; 1977a; PPS, II, 45. This perspective on meaning is as old in Kuhn as his earliest considerations on scientific languages, the meaning of their terms, and their cognitive import. The early view traces a long way back to the early 1950s, when he was a Junior Fellow of the Harvard Society of Fellows, a stage in which the main lines of The Structure of Scientific Revolutions were set up. More on this in sect. 5.4. 27  On the concept of “feature space” and its origin, see Kuhn (1979, 201; PPS, II, 45; 1987, 31). Andersen et al. (2006) are a good introduction to this aspect of Kuhn’s lexical theory. 28  PPS, II, 46, fn. 11. 25 26

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

61

certain number of features while others require a different number of them. Kuhn is particularly interested in a specific characteristic of these contexts: in them, picking out a certain object x as an F does not depend on satisfying a minimum number n of features that are a sufficient and necessary condition for the truth of “This x is an F.” It rather depends on checking that the features that are obtained are sufficient to adjudicate whether x is an F or a G. Kuhn places special emphasis on the idea that features in group 2 are not usually learned one by one in order to pick out items in group 1. As a rule, most of them are learned at once and by means of discriminatory heuristics – that is, by means of obvious (or paradigmatic) cases in which a given object x is definitely an F, while, say, that y is clearly a G.29 Besides that, kinds do not belong to a continuous feature space, but rather to a discontinuity in which we find marked inter-categorial interstices.30 This idea is common to texts of Kuhn’s late phase, such as The Presence of Past Science (1987) and to writings of a pre-­ Structure stage such as The Quest for Physical Theory (1951), whose two last lectures show a similar version, though less detailed, of that model for language learning (more on this in sect. 5.4).31 Kuhn shows that collective learning helps to clear up the distance between categories. On the other hand, the fact that Kuhn does not talk about meaning in terms of a constant disjunction of properties helps to show, in his view, to what extent his position is different from the cluster-meaning theorists.32 Actually, the whole position helps in that respect. There are natural kinds because of that possibility of being differentiated from one another by a speaker of a given language by means of a mixture of learning the language terms (group 1 or 2) and observation. Kuhn already spoke about this in Structure, in The Essential Tension (“Second Thoughts on Paradigms”) and it is also a recognizable feature of his whole theory of reference. This way of learning lets us do without the old ideas of semantic rules and methodological rules. From Kuhn’s point of view, every normative aspect of scientific behavior, whether referring to language or to scientific research, are learned by means of two simultaneous processes: (1) a labeling process, which is pragmatically determined and context-sensitive; and (2) a discriminating process, in which learning to distinguish what belongs to a given category and what does not is learned simultaneously. As regards (1), the presence of a sample, prototype, or paradigm, i.e., a representative of a certain kind, helps to define the kind itself and to find other members. I would like to add another characteristic that Kuhn also emphasizes. Though the lexicon is singular, unique for a whole community, feature spaces can be different from one individual to another within the community, because feature spaces vary

 Kuhn’s paradigmatic case of learning in this way was presented in Kuhn (1977a).  For a first presentation and defense of this idea, see Structure, 45. 31  On this model of language in 1951 and its connection to texts in the 1980s and early 1990s, see Mayoral (2017a, 299–320). 32  See PPS, II, 46, 49–50. 29 30

62

J. V. Mayoral

according to the situation(s) in which they have been acquired.33 This is another aspect of context-sensitivity (or maybe subject-sensitivity), which shows that the ways in which a term is applied vary with the subject that has learned to apply them, i.e., with the specific situation in which the speaker has learned to do it. In this way, once again, the knowing subject plays a central role in Kuhn’s work. Not only is it the link to constructing a historical identity – those kinds of lexical variations (or variations within a lexicon) show a fundamental source for problem-solving: the likely divergence of the individual’s access to the lexicon. Despite the lexicon not being open to endless variations, it does allow some of them and these are sometimes a source of solutions (see PPS, II, 49, 62-63). Now, even though lexicon acquisition makes room for individual variations as regards group-2 terms, the hierarchical organization of kind terms must be respected. Similarly, if terms for two objects, x and y, share the same hierarchical level, similarity and dissimilarity relations that feature spaces create must also be identical.34 So, as Kuhn also says, variations are, after all, scarce.35 They may happen, of course, but the lexical structure is preserved anyway. Interrelations and hierarchies, as well as similarities and dissimilarities between objects, are preserved even though the feature space might be different. Furthermore, even though two individuals within the same community have learned the lexicon through different paths, their access to other empirical consequences of the usage of that lexicon (e.g., knowledge that some of their colleagues have gained by other routes) are at their disposal and shall be coherent with the part of the lexicon to which each of them has had access.36 Lexicons are learned when their terms are employed in true statements in observational contexts. The case of Greek astronomy refers to objects that can be classified in a feature space with few specialized terms; its vocabulary almost counts as observational in that example.37 In other cases with more specialized terms, things are not much different. That is the case with learning the term “electricity.” In this new example, we see that, together with the term “electricity” – and in the same observational contexts – other lexical terms for electric phenomena are acquired as well, e.g., “insulator,” “charge,” or “conductor.” The rest of terms, such as “discharge” or “attraction,” are included in the feature space by means of which the lexicon may be acquired.38 Although the lexicon affects terms related to commonsense observation and practice as well as to those in an expert vocabulary, according to Kuhn we must be careful when we use a specific group of them in order to formulate a theory of reference. This is one of his criticisms to the causal theory of reference in The

 PPS, II, 49.  See PPS, II, 50. “They must, one might say, share kinship relations,” Kuhn says there (ibid.). 35  See PPS, II, 49–50. 36  See PPS, II, 54–55; Kuhn (1993, 239). 37  PPS, II, 52. 38  See PPS, II, 53–54. 33 34

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

63

Natures of Conceptual Change and other texts, and more particularly to Hilary Putnam. In Putnam’s case, Kuhn says, the examples he chooses exhibit a permanence and uniformity that are not always present in the history of science. In The Natures of Conceptual Change, Kuhn presents a similar idea to that which he had already presented in “Metaphor in Science” 3 years before. For Kuhn, the causal theory of reference needs to reconstruct the individuals’ “lifelines.” In order to do so, that theory finds support in two kinds of evidence: (1) the evidence that reveals a continuous existence over time; and (2) the evidence that shows independence or singular character of every line during a certain chronological period.39 Although Kuhn says that this strategy works well with some natural-kind terms, for instance, “gold” or “water,” it does not fit with everything else.40 To begin with, as he also said in 1977, “individuals which constitute natural families do have lifelines, but the natural family itself does not.” (Kuhn 1977/2000, 199). Natural families are subject to a possible variation when the lifelines of their components stop gathering under the same group-1 term. At that point, a lexical change takes place and so does a scientific revolution. So, the idea of necessary identity would reveal, as it were, its contingent, relative character. Statements that speak about such identities preserve, accordingly, their necessary character only in transitions between the worlds Saul Kripke talks about – “normal” transitions. In those worlds Kuhn talks about, on the other hand, individuals and substances have been redistributed, and transformations are abrupt and may affect the truth of this kind of statement. Tautological statements, Kuhn says, are affected by a scientific revolution.41 This kind of modification is not visible in the same fashion when it affects some kind terms instead of others. In particular, it does not affect those terms Putnam used to talk about in the same way, but it does affect those Kuhn mentions here and in “Possible Worlds in the History of Science,” an extract of which helps to finish this section: What remains special about ‘gold’ is simply that, unlike ‘water’, only one of the underlying properties recognized by modern science – having atomic number 79 – need be called upon to pick out members of the sample to which the term has continued through history to refer … But not all everyday terms are of this sort. ‘Planet’ and ‘star’ now categorize the world of celestial objects differently from the way they did before Copernicus, and the differences are not well described by phrases like “marginal adjustment” or “zeroing in.” Similar transitions have characterized the historical development of virtually all the

 See NCC, III, 44–49; Kuhn (1979, 198–199; 1993, 229).  NCC, III, 58. 41  NCC, III, 55. On this distinction between Kripkean and Kuhnian kinds of worlds – I am adapting Kuhn’s vocabulary here (“Kripke world” and “Kuhn world”) – see NCC, III, 54. Kuhn claims there that the main way to distinguish between them is to focus on the difficulties that emerge in the transition from a Kripkean to a Kuhnian world, difficulties that are absent in transitions from one Kripkean world to another. In the first kind of transition, in particular, terms do not refer to the same objects or substances anymore and variation affects some interrelated terms. The ways to classify these terms in a hierarchical lexicon – i.e., the similarity and dissimilarity relations – have traversed a process of reorganization in which metaphor is present. 39 40

64

J. V. Mayoral referring terms of the sciences, including the most elementary: ‘force’, ‘species’, ‘heat’, ‘element’, ‘temperature’, and so on.42

5.4 A Recurring Outline Paul Hoyningen-Huene says that the contact with the phenomenal world that we observe in The Structure of Scientific Revolutions – as well as in writings from the late 1960s  – is overall “visual,” and that this emphasis changes significantly in further writings in the decades of the 1970s, 1980s and early 1990s, where there is a greater emphasis on language. However, Hoyningen-Huene qualifies that claim by saying that, despite this latter change, Kuhn’s dependence on perception is still patent. The thesis that defends a “plurality of phenomenal worlds” – as reconstructed by Hoyningen-Huene  – is transformed accordingly, though it is essentially preserved, particularly as regards the link between the phenomenal world and the language used to refer to its components (see Hoyningen-Huene 1993, 61–62). As regards world-change, it could be said that the transformation is the same. Its expression in Structure is troublesome in the sense Ian Hacking and Hoyningen-­ Huene have shown, that is, by reference to the “new-world problem” – according to Hacking’s denomination – and because of the use of a setting based on the change of the world in which experience takes place. Its exposition in terms of lexical change is easier to accept as far as it is the lexical taxonomy itself that changes, with a causal background of stimuli as the transcendental, permanent element, ineffable anyway (see Hoyningen-Huene 1993, 201–206; Hacking 1993). Yet, not even in Structure is reference to language accidental, because, already in earlier moments, the discontinuist explanatory model is similar in many ways to the later one. So, for instance, before Structure, for Kuhn, the main consequences of a revolution  – not only scientific, also in politics  – can be felt in language. In this respect, a few words by the German historian Eugen Rosenstock-Huessy  – who portrayed the idea of revolution in history as a transition that leaves its mark in language  – seem to have been appealing to Kuhn. For Rosenstock-Huessy, a revolution promotes a change in the way of expression and also in the way of living, not only in institutions. His notes on Rosenstock-Huessy can be found among other preparatory documents for Structure in his archives.43 As Hoyningen-Huene says, the issue is that, in published texts (I would add that in unpublished texts, too) of around the 1980s, Kuhn provides arguments based on his idea of the lexicon, which is (obviously) of a linguistic nature, but that they do not move him away from his previous approach in which the roles of perception and of what we could term “relation to the world” – or with the “world-in-itself,” to use  Kuhn (1989, 84–85). See also RSS (Kuhn’s 2000 interview), 312–313.  See Rosenstock-Huessy (1931, 5, 18, 23–30). Kuhn’s notes are in TSKP 25.53, with correspondence and other notes for Structure. Very recently, George Reisch has written about this historian’s book and these notes by Kuhn; see Reisch (2019, 222–224). 42 43

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

65

Hoyningen-Huene’s expression44  – are central. Kuhn is aware that, with his new emphasis on lexicons, he brings in a new “style,” so to speak, in his presentations, which, as he said in December, 1980, at Denison University, maybe got him closer to the philosophy of his age; even more so if, as he said to philosopher Emmanuel Mesthene some time before, in 1979, he wished to compare aspects of his work with that of Kripke’s, of Putnam’s, and of Quine’s.45 Yet, there is little in Kuhn at that time that, though in a less specific format, we have not found time and time before. This latter statement includes, in temporal terms, the decade before Structure. In 1980, Kuhn considers that there are three characteristics of scientific revolutions he wishes to talk about. The three of them refer to language issues. The first characteristic is holism with the qualification that, for Kuhn, it only affects those parts of language whose change involves an alteration in the way we observe the world and interact with it – i.e., whenever NOP is violated. Holism, in other words, is local and revolutionary change only affects some parts of our classification system: our lexical taxonomies.46 The second characteristic is that every revolutionary change involves meaning change. Not every meaning change is revolutionary, of course, but the converse is true. Kuhn distinguishes, for instance, between small changes within a working lexicon that remains stable – its structure is preserved – and changes that involve a complete categorical reorganization.47 For Kuhn, this identification of scientific revolution and linguistic recategorization is key in his new work. A similar identification of revolutions with meaning change, Kuhn says, is present in Structure, though for him it is not the same notion of meaning. The previous notion of meaning was weakened because of Quine’s criticism.48 The third characteristic, finally, is the metaphoric redefinition of objects, substances, and phenomena at large.49 This is an important characteristic for Kuhn, but I will focus on the second one in what follows.50 The notion of “meaning” that Kuhn expounds in The Natures of Conceptual Change, second lecture—making way for his criticism of the causal theory of nouns in Kripke and natural-kind terms in Putnam (third lecture)—refers to the example of learning of commonsense terms in “Second Thoughts on Paradigms.”51 Kuhn  See Hoyningen-Huene (1993, 31 ff.).  See Kuhn 1980 (“After Paradigms What”), p. 1; letter from Kuhn to E. G. Mesthene, 17 May, 1979, TSKP 5.26, 1. His third lecture in The Natures of Conceptual Change testifies that Kripke and Putnam are among his arguments’ targets at that time. 46  See NCC, I, 17–18; II, 21, 30. See also Kuhn (1983, esp. 52). 47  NCC, II, 33–35. 48  See NCC, I, 18–19, 21; II, 22, 31. 49  NCC, I, 19; II, 31. 50  As far as I will not pursue this third point in what follows, it might be relevant to say that Kuhn is already talking about metaphor in The Quest for Physical Theory in 1951. Moreover, he ascribes metaphor an important role in the activity of “orientations,” or “points of view,” a precedent of paradigm-based worldviews. See QPT, V, 9–10; GR 80–81; Mayoral (2017a, 277–278). 51  See NCC, II, 28–35; Kuhn (1977a). 44 45

66

J. V. Mayoral

illustrates the acquisition of a (part of a) lexicon in observational contexts by means of the well-known example of Johnny learning terms to be applied in the restricted domain of waterfowl – even though it is not scientific vocabulary.52 His conception of meaning does not refer to descriptions or fixed collections of them, but rather to a hierarchical lexical structure as described in the previous section. The meaning of terms in it are not learned one at a time but as part of such structure. The learner always depends on the observational, practical context in which the term shall be applied and usually in conjunction with other terms. I have already said that meaning is context-sensitive in that sense and that it is also subject-sensitive, so to speak, as it depends on the particular feature space out of which the structure emerges. This conception of meaning is visible in The Quest for Physical Theory – Kuhn’s Lowell Lectures – 30 years before, although, in that case, there is no mention of lexicons and taxonomies. That difference, however, could easily be considered as merely nominal. In 1950–1951, the conception of meaning in Kuhn also refers to a structure of terms that is built on pragmatic considerations – in two senses which I am going to describe by means of the expressions I have shown in previous sections of this chapter. First, the natural-kinds vocabulary is useful to label certain groups of individuals that are formed according to learning and future use. We learn to differentiate between objects by virtue of our necessities in behavior. Maybe those differences are not enough later, and call for new differences – a redistribution, and a more finely-tuned labeling. Second, as long as I am talking about behavior, it is obvious that there shall be individual variations in all conceptual vocabulary; as Kuhn says, there shall be variations at least as regards group-2 terms (the individual arranges his or her own conceptual space), but these variations shall not exclude the individual from the group. Logical and linguistic rules cannot be broken without, at the same time, bringing about social exclusion consequences.53 There are some features in this earlier conception, that precede Structure by 10 years – a bit more, in fact, as there are some extant notes of 1949 revealing the core ideas later presented in The Quest for Physical Theory – that I would like to mention here. The first feature is this appreciation of the importance of studying how language is configured, having in view the relationship the individual establishes with his or her social and natural environment (the world). The second feature is the context-sensitivity and the subject-sensitivity that I talked about at the end of the previous paragraph. These are two basic poles on which Kuhn’s social, historical and developmental view of science is built. The third feature is the idea that we need not hold a unified background (or theory) of meanings of those language terms in order for us to use them adequately in the context we apply them. The fourth feature is that this absence of a unified background is not only admissible but also necessary, because discriminating novelty necessitates a form of plasticity that a well-defined semantic unification helps to remove  – this argument is repeated in “Second

 In NCC, II, 39, Kuhn grants that this example is perhaps too far away from a scientific case, though it reveals central aspects of the acquisition of a more specialized lexicon anyway. 53  See QPT, VIII, 7–11; GR 139–141. 52

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

67

Thoughts on Paradigms.” The fifth feature is that this image of language includes areas in which vagueness prevails. Moreover, this latter is not eliminable without, at the same time, removing adaptive plasticity from language.54 Finally, in a sixth feature, I would include two critical versions of philosophy of language that arise from the vision we have already seen in these early years (1949–1951). One of these is Kuhn’s criticism toward positivistic and operationalistic conceptions of scientific method, because of their tendency to set limits where there is no point in having them. This tendency, Kuhn says, reduces what I have just called adaptive plasticity of language – which is a pragmatic vehicle of expression and classification for facing the future with success. The other critical version anticipates reasons Kuhn shall submit against the causal theory of reference; for instance, that our language should be ready for changes in those objects to which we apply a term.55 Perceptual stability is not warranted and much less is it – Kuhn said years later – beyond common sense, as we saw in the previous section.56 In this conception of meaning of pre-Structure Kuhn there are several recognizable elements of post-Structure Kuhn. In the second phase (1980s-early 1990s), that older characterization of language and meaning is clarified more precisely, as we saw in Sect. 5.3. The significant point here is, in any case, its presence as explanation of the way in which language is a tool, or vehicle, for dealing with the world. And it is not presence alone that is relevant: it is also significant that some ingredients, such as those referred to above, remain in place until later stages of his thought. In The Quest for Physical Theory, language – whether ordinary or scientific – needs to be as vague and flexible as it usually is, because it is preserved in that way as a medium, or instrument of adaptation to the surrounding social and natural environment. It helps to acknowledge ambiguous cases in which applying a kind term (our group-1 term) is warranted if our only resources for doing so are the usual (group-2) terms we have been exposed to in the past. Four more connections are significant. The first one is that, in 1951, linguistic change and scientific change were straightforwardly related. Besides that, talking about normal science – or rather “classical” science at the time, as Ludwik Fleck (see, e.g. Fleck 1979, 9) used to do – is equivalent to exploring the vague areas of language in search for a better usage of terms. The second connection is Hoyningen-­ Huene’s thesis referred to above  – namely, it does not make sense to talk about language if this is not considered a medium of interaction with the world. To be

 For the points in this paragraph and the next, see Kuhn’s eighth Lowell Lecture; see esp. QPT, VIII, 17 ff.; GR 143 ff. 55  See QPT, VIII, 6; GR 139. 56  For an exploration of The Quest for Physical Theory on which these features are based, see Mayoral 2017a, chap. 5. My exposition in the three previous paragraphs and the rest of this section is based on it (see it for further references). See also Mayoral 2015 for the relationship between The Quest and later texts; in particular, for the relationship between that series of lectures and his late-1970s Foerster Lecture, “Does Knowledge ‘Grow’?” (1976, TSKP 5.13), a lecture in which the influence of Wittgenstein’s On Certainty is explored and exploited by Kuhn. On that lecture, see also my commentary on Howard Sankey’s chapter, in this volume. 54

68

J. V. Mayoral

more precise, we are talking about the interaction with an “external world,” so to speak, causally efficacious and active, by way of a psychological world or environment that, as is already well known, Kuhn calls in those years “behavioral world,” using in this way a term we can grasp in Gestalt psychology.57 Configuring a language is a parallel activity to that of fitting out a behavioral world with the individuals and substances with which we shall interact in the future. In both cases, the language (later on, the lexicon) allows us to identify objects or samples of substance as tokens of kinds. The third connection is the relationship Kuhn establishes between the dynamics of scientific change – which in The Quest is already the idea we can later see in Structure, albeit still dependent on Fleck’s and J. B. Conant’s views – and Whorfian linguistic relativism (more on this in Sect. 5.6, below).58 Last, the fourth connection refers us again to the role of the subject. It is the subject who fits out his or her own understanding of the world and of kind-terms vocabulary in use according to his or her psychological, behavioral or phenomenal world  – I shall take these terms as interchangeable for the sake of the argument. This is the fundamental background of our understanding of the genesis of shared conceptual schemes, according to Kuhn. In 1951, our author already conveyed this view in The Quest and worked on it in later texts (and I am not only referring to the latest papers or lecture series as The Presence of Past Science). The area of term-learning and practice-learning, which is socially bounded but admits some degrees of liberty, is an integral part of the argument in Structure. Kuhn’s attack on the idea of rule-based behavior and its replacement with a paradigm-based one was born out of these roots that explain the constitution of the active subject in the sciences.

5.5 Reconstructing Our History The connection with the past is clear when there are beliefs that are preserved, including their truth-value, from a previous period in the history of humankind to our own. It is not surprising, therefore, that we set up a historical line connecting Aristarchos of Samos with Nicholas Copernicus and with us, since there is a common belief for the whole group: “The planets, the earth included among them, move around the sun.” The same is true if, even though it is not possible to establish that link taking a true belief like that as a guide, we can turn an outdated (and so false) belief into a true belief that is today accepted with only one change: a translation of some term that allows preserving the truth of the statement thus

 See, e.g., Koffka (1935, 31). See also Mayoral (2017a, 267), on that concept in the early Kuhn. Marcum (2015) is a good source of information for all these concepts and ideas in the early Kuhn and also in the later. 58  See Mayoral (2017a, chap. 5). In addition to Fleck (1979), see also Conant (1947) and Whorf (1956). 57

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

69

transformed.59 So, when conserving a true belief, whether in its original format or in an allegedly possible truth-preserving translation – i.e., in a procedure of reducing a theory T to another one T’ – we do more than simply find a way to establish the progress: if we look back, those beliefs serve us as a guide to find out what our past is. The past of our theory, expounded in terms of the beliefs that compound it, is our past as practicing scientists of that theory. There is no estrangement with regard to our theory.60 Kuhn assumes that this historical reconstruction is often provided and that, moreover, it plays a role for us (more on that role below).61 Yet, if we take into account Kuhn’s idea of belief change as explained in terms of alterations in the lexical structure of theoretical language, the reliability of that link is open to question. Recall that there is a local incommensurability between two lexicons, which leads to the idea that the sets of possible worlds that these two lexicons open are not easily identifiable as the same set.62 Nor is the quest that the scientist starts out from each of them the same quest. Even in the case of beliefs that are preserved when we come from T to T’  – whether it is the same beliefs or the result of a translation – we are not really considering a reference to the same objects. These include features and properties that allude to different sets of possible worlds and so to a wide variety of counterfactual situations. The scientist inquires about them in different manners, trying to find an accurate representation of the actual world. If the truth of those statements is possible to state at all, it is by reference to those sets of worlds, to those clusters of features and natural-kind terms – to the objects there referred to. Those worlds, however, and the reference to them in language, are not preserved when coming from one lexicon to the next one, from T to T’. Not, at least, in such a way that the translation of every term is guaranteed.63 In Kuhn, therefore, incommensurable lexicons are separated by means of a frontier that is not possible to save by truth-preserving translation. It involves a rupture between theories T and T’ that impedes the ability to understand immediately that T’ is a successor to T. To do so, we might need to attend to other factors, for instance the continuity of institutions, of material resources, a certain domain of phenomena (roughly understood), maybe some calculating tools, and so on.64 This option was within our reach in the previous reductionistic model.

 A good example of that possibility is provided by Philip Kitcher (1978, esp. 534 ff.). For a criticism of that strategy see Bishop and Stich (1998, 241 ff.). 60  See PPS, III, 66, 87 ff., and esp. 89–91, where Kuhn expounds the nature and function of that kind of (Whig) historical account of our science’s past. 61  See PPS, III, 89–91. 62  See PPS, II, 62. 63  See PPS, III, 67. Kuhn does grant the translation of some terms. 64  Peter Galison has explored in Image and Logic the developmental model in which some aspects – e.g., experimental or instrumental – of scientific research may help to keep scientific practice continuous whenever the theoretical aspects go through a revolutionary change. See Galison (1997, esp. chap. 9). 59

70

J. V. Mayoral

Nevertheless, Kuhn qualifies that idea. There are points in common between two lexicons or theories. Actually, Kuhn says, those commonalities – he uses a term that Martin Hollis often alludes to as well, “bridgeheads” – allow us to learn a second lexicon.65 We would talk, thus, about the possibility of acquiring an extended lexicon, starting from the one we already possess. We take one of those lexicons (e.g., ours) as a source vocabulary and the new lexicon as a new vocabulary to add to the former. This relationship is set up between any two lexicons or theories – i.e., whether it is between the scientist in T (the source vocabulary, in that case) who wishes to learn and understand T’, or it is the historian of science who tries to understand T after having been trained in T’ (the source vocabulary, now). Let’s talk about lexicons instead of theories; and not only because of what I will say next, but also because for Kuhn we cannot consider theories and lexicons as synonymous: the former help in the determination of a lexicon.66 Given two lexicons, A and B, where A precedes B in historical time (as T did for T’) we can depict those relationships as follows: either the historian starts out from B to arrive at – i.e., to understand, speak about and classify the world according to – A, or a scientist begins with A and, going through a scientific revolution, adopts a new lexicon B. For Kuhn, in both cases we extend the lexicon we already know by means of different resources (e.g., glosses) that allow us to understand how a term in our language (say, B) corresponds to a term in another one (say, A) (See Kuhn 1983/2000, 38 ff.) Purely terminological closeness or equivalence should not confound us. If we return to the example of Aristotle and Newton on motion, we noticed that even though “kinesis” and “motion” are taken as equivalent in our language (which they are not, or not completely), the kind of empirical regularities, counterfactual statements, etc., which using them commits us to is pretty different in each case.67 Their symbolic representation (e.g., ML1 and ML2, respectively, for each term, in direct relation to L1 for Aristotle’s physics and L2 for Newton’s physics) might be of help to avoid confusion of committed equivalence with truth-preserving translation. Kuhn says that, if two lexicons are historically close, lexical structures widely overlap and there are many chances that their semantic closeness leads to a translation of many terms in A to B (or vice versa).68 These chances diminish when the interpreter is more historically or culturally distanced from the target lexicon and truth-preserving translation is not granted throughout – local incommensurability between them emerges. That is the case for instance when the interpreter tries to understand Aristotle’s physics from the lexicon that the Newtonian picture of the world opens (this is Kuhn’s own experience with Aristotle’s physics in 1947).69

 PPS, III, 66–70. I thank Howard Sankey for pointing out to me the significance and connections of that notion (“bridgehead”). See Hollis (1970, 215 ff.), for further details. 66  See PPS, III, 75, fn. 1. 67  For this point in Kuhn, see PPS, III, 66–71. 68  See PPS, III, 70. 69  Kuhn narrates this experience many times during his career. See, e.g., Kuhn (1989, 15 ff.; PPS, I, 5–6). There are other places in his bibliography where the reader can find a similar narration. 65

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

71

Whenever this happens, we cannot indeed say that there is a truth-preserving translation, and the interpreter can only become bilingual.70 As an example, let us examine the following statements71: (pA): “The sun is a planet.” (pN): “The sun is not a planet.” Clearly, excluding other qualifications concerning their lexical origins, pA is true if and only if pN is not (pN is just “not pA”) and vice versa; their conjunction is a contradiction. The first one, pA, is a true belief in Aristotle’s physics and pN is a true belief in Newton’s physics (actually, since Copernicus’ astronomy). No theory admits both beliefs at once  – it would be a clear inconsistency otherwise. Our current theory (say, T’) would suppress pA, which is a part of a previous theory, say, T, together with the rest of elements in T that are identically false  – i.e., that contradict (or are inconsistent with) other statements or beliefs in our current theory. This would be the traditional, reductionist position. For the reductionist, pA is the past of the theory that includes pN as a true belief. Kuhn’s position provides an alternative. For him, there is a rupture between both theories (or lexicons), which means that one of these theories makes one of those statements true while the other makes only the other one true. Now the question is, how can pA (or its corresponding lexicon, let us call it “A”) be the past of pN (or its lexicon, let’s call it “N”) if their original systems are closed and incompatible? I shall speculate by alluding to a sort of possible “bridge” between those lexicons. Let us add a third lexicon in between, C, which will be the Copernican one. It is possible to assume that lexical changes from A to C would be easier to understand than changes from A to N, by virtue of a greater proximity and, thus, also greater structural overlapping of A and C. The relationship between C and N would be explainable just as well and in similar ways. Then, assuming a transitive property, which would be admissible in that case, the relationship between A and N could be settled. A could be quite far from our current form of life as expressed through N (or a likely successor of it), but it could be correct to say that A is our predecessor.72 To my mind, this seems to be what Kuhn can conclude from his lexical theory as regards historical connection and heritage. Actually, in Scientific Development and Lexical Change, 3 years before the Shearman version, there is a declaration of intent on the relativistic thesis that each lexicon imposes an independent categorial framework – “encapsulated,” I would say – that blocks communication, comparison, and understanding (consequences that, as it is well known, Kuhn does not draw from his main thesis on incommensurability). As I said in Sect. 5.3, in the fourth lecture of the series referred to above, Kuhn mentions the existence of common  See Kuhn (1983, 61; PPS, III, 66–67; 1989, 76–77; 1991a, 93).  This is Kuhn’s example; see PPS, III, 72 ff. 72  I find support for this speculation on PPS, III, 67 ff., esp. 70. An interesting issue would be if this picture could be connected to the continuist perspective that Dudley Shapere expounds (Shapere 1989), according to Andersen et al. (2006, 174), and that these authors support. I shall not pursue that point here, however, but the possibility is worth mentioning. 70 71

72

J. V. Mayoral

aspects of human beings; not only their obvious common biological heritage, but also their common cognitive background.73 Despite lexical and belief change, and leaving aside that – as he said 4 years before – that kind of transformation allows us to discover new patterns in nature,74 we draw on the past to talk about our ancestors. That identification is possible in the first place because – and obviously, in spite of what such arguments as those referring to world change seem to convey – we share a common world with whose activity in form of stimuli we interact. This interaction is, we should assume, identical for the whole human species in every moment.75 In the second place, if we assume that shared background, our current scientific theory may convey relevant results on the world that make historical understanding easier. Finally, in the third place, shared aspects of lexicons – that is, those aspects Kuhn considers innate or returning in every scientific culture – allow us to set kinship relationships with which we can build something pretty valued for Kuhn: a modern “identity” that relates (and is faithful) to aspects  – extraneous, despite it all  – of its own past.76 This difficult “historical tension,” so to speak, conveys the complex link involved in the search for and finding of our ancestors, a search that, prima facie, is unaffected by anachronism, ethnocentrism or, in short, the Whig approach to history. Yet, Kuhn does not solve that tension in this way. He deals with this problem in his later The Presence of Past Science. He devoted the very last pages of that 1987 series to that problem. He does not consider it a minor issue: the title of the third lecture, “Embodying the Past,” refers to it, he says, and the series largely – though perhaps, at many points, only implicitly – deals with it.77 Kuhn’s conclusion there is that his own approach to the past is not as useful a tool to forge a historical scientific identity as the Whig perspective is. This latter assumes that the reference for the terms involved is and always has been the same and that humanity has emerged from error owing to the efforts of a few stubborn scientists who have indefatigably pursued the truth  – which, obviously, is not lexicon-relative. Science has always gone after the same objects and the same world and, whereas something like Kuhnian lexicons may reveal a conceptual superstructure only partially (and therefore, in some sections, mistakenly) connected to the world – which explains the source of some widely held erroneous beliefs  – it is just by alluding to the former discoverers of truth that we find our true historical past.78 Kuhn expresses abhorrence for that kind of historiography in all stages of his writings. Yet, he is unable to contest its undeniable role in forging our modern scientific image and self-identity; moreover, he argues that his own brand of historical interpretation would be counterproductive when it comes to training

 SDLC, IV, 112.  See NCC, I, 6. 75  As Paul Hoyningen-Huene shows in his book; see Hoyningen-Huene (1993, 49–50). 76  On what is told in this paragraph, see SDLC, IV, 112–116. See also Structure, 201. 77  See PPS, III, 87, 92. 78  See PPS, III, 89. 73 74

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

73

would-be scientists, insofar as the historical inquiry might weaken the student’s trust in the current theory.79 In short, Kuhn opts for removing the historical account that created a historical tension, as I said before, and for promoting a difficult coexistence of two historical explanations of the science’s past: one which sees it as an encapsulated world that is completely extraneous to us and another which sees it as an imperfect, unfinished, and widely false depiction of the same world we have currently come to understand somewhat better.80 The winners’ tale is, after all, the correct narrative in proper circumstances.81 Clearly, the main menace at this point is relativism. Kuhn assumes the challenge as something serious, though the form of relativism that mostly concerns him is relativism about truth. Kuhn is clear at this point: he is not defending relativism about truth. He supports relativism, but of a specific kind that is not related to anything like relative truth. He confesses that he is a relativist about effability, or sayability, or expressability – any label will do – and thereby to an old picture of language. He says that “what I am relativizing is not truth value, but effability.”82 This kind of relativism  – though he does not like the label for his position  – is admissible.83 Let us find out more about it.

5.6 Lexicons, Relativism and Truth Let us expound the problem Kuhn meets as follows: pA and pN are two incompatible statements. If we do not consider that the option that we defend today on behalf of the truth of pN and against that of pA gets us closer to theoretical truth, then we are considering that the truth of pA, though unacceptable today, is acceptable nonetheless insofar as it was adopted in a past period (i.e., that in which the lexicon A was in place). This leads to a conceptual relativism with well-known epistemic consequences and open to the usual self-refutation criticism. Yet, Kuhn says that this is not his point. One can be relativistic or not and it depends on what one takes pA to be. pA (i.e., “The sun is a planet”) may be either (a) a statement of our current lexicon (N), properly speaking our way to state a belief that is part of a theoretical setting marked by other lexicon (A, in this case), or (b) a statement in the other lexicon A. In this latter case (b)  – leaving aside that the statement should be written in classical Greek language  – the statement is true, whereas as we rephrase it in lexicon N we produce a false statement. This is completely admissible, Kuhn says, without incurring truth relativism. The point is  See PPS, III, 90–91.  For this conclusion, see the last pages of the Shearman Lectures (PPS, III, 91–92). 81  Thus, the Orwellian depiction of the past seems to have a positive role; see Structure, 167, for that well-known reference to 1984. See also its Sect. XI and Kindi (2005) for an extended commentary on those passages. 82  PPS, III, 72. 83  See PPS, II, 38, 51; III, 72, 77. On effability in philosophy of language, see Katz (1978). 79 80

74

J. V. Mayoral

more modest here: in lexicon A (individual feature spaces leading to it, the set of possible worlds and the kind of regularities it produces included) pA makes sense and, furthermore, corresponds to how the world is.84 The set of possible worlds A opens is different from the set that N opens, nor has it anything to do with an extended lexicon N that we have enriched to understand statements in A. Some terms like “sun,” “planet,” “motion” and so on are taken as one and the same term in A and N (or, for that matter, in T and T’), but they refer to different individuals and/or extensions. Only one of them shall be employed in laws, for instance. In that case, the translator, who (say) works from lexicon N (let’s assume that it is our own lexicon, too) reconstructs an older statement (from A, for example) with a scheme that is similar to A’s but that is actually N’s – or rather, an extended N. This is the “translator’s solution,” as it were.85 These solutions represent, on some occasions, beliefs that have come to be false in the new lexicon – otherwise they would be instances of truth-preserving translation – but that are useful in order to understand what beliefs were true in the past of our science. We should bear in mind, however, that we are considering statements that are made within the historian’s lexicon (an extended meta-language, so to speak), not in the lexicon that is being reconstructed (an object language, metaphorically speaking, again). However, former statements are not those made in the lexicon in question. This language had as references for terms in its statements those objects and substances whose physical behavior range from what is evident in an observable world to what is possible in a conceptual structure, an area to which the diverse modal options, expounded by means of scientific laws and counterfactual statements, give access. From this point of view, we can state that pA was true in lexicon A, not that it was falsely considered true in Aristotle’s time in spite of its falsity. pA is no longer a part of N as a true belief – our N. It is reconstructed in N, and the pseudo-pA in it is simply false.86 Kuhn is trying to make it clear that statements do not float freely among languages. Rather, they belong to lexicons and refer to worlds that are classified according to them. Is it therefore possible to state that for Kuhn the same belief could be true and false at the same time (e.g., according to the “point of view”)? Not at all – if a given statement indeed goes from one lexicon to another, then it takes its truthvalue with it. The, as it were, “trans-lexical availability” of a statement is not granted

 We could even say that it corresponds with a belief in it that we consider to be quite basic. For Wittgenstein, pA might be a certainty. Kuhn talks about the details we have examined in this paragraph in his third lecture of The Presence of Past Science. See also previous sections of this chapter. 85  See PPS, III, 72 ff. 86  On the idea conveyed so far here see also SDLC, IV, p.  98. If we examine some passages of Clarence Irving Lewis’s work, Mind and the World Order, together with other texts by the same author, we observe a position that is similar to the one we have just observed in Kuhn. The relativity of truths in a science – or, more generally, of a collective activity mediated by a language and a conceptual scheme, is not new at all in epistemology. On that comparison of Kuhn’s and Lewis’ positions, see Mayoral (2017b) and my companion commentary to Sankey in this volume. 84

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

75

in every case. It does not happen to pA, for instance, which for Kuhn is not a simple case of denying pN. Our understanding of pN is more than that, just as our grasp of pA from our lexicon N is. If we try to do so properly, we should enter the world – or set of possible worlds – that A opens; and thereby we become bilinguals, as Kuhn liked to say (see the previous section of this chapter). Contrary to a common assumption, Kuhn is not assuming a change in truth-value for the same statement. Truth is not relative to the viewpoint.87 Two questions are now in order. (1) How do we predicate truth of a statement in that case? And (2), if truth is not relative, what is relative? Concerning (1), for Kuhn, truth is predicable of a statement only when there is a lexicon that opens a set of possible worlds, and also a justifiable “true” world, that allows us to do it. Lexicons involve often incompatible ontological commitments – inconsistency between them is granted as soon as the NOP is broken – so there is not any application of “true” under the assumption of more than one lexicon, because it will cause contradictory beliefs, which are taken as simultaneously true. Truth, on the other hand, for Kuhn, does not adopt a relativist or pragmatist representation. For Kuhn, truth is something more than a sort of agreement or a final outcome (and goal), a perspective with which he rejects positions that go from C. S. Peirce’s and John Dewey’s conceptions to that of Richard Rorty’s.88 Without committing to foundationalism, which he rejects out of hand and considers outdated, truth is predicated in every language unambiguously – it is not born out of a consensus. The question is that there is no truth beyond language, but there is truth within it.89 Concerning (2), what is relative is effability, or sayability, or speakability. The “speakable and unspeakable” – to use a famous phrase by J. S. Bell – in each theory is lexicon-relative.90 What is expressed in A, for example, may not be stated in N. We can feel comfortable with a “translator’s solution”, so to speak, of pA into N. Yet it is not granted that it is going to be a truth-preserving translation. Kuhn thus breaks with J. J. Katz’s Principle of Effability, according to which “Each proposition can be expressed by some sentence in any natural language.”91 Kuhn is thereby defending a breaking with this principle, which involves a rejection of the realistic theories of scientific progress and a relativistic anti-realism that is not ipso facto in danger of being accused of relativistic self-refutation.92 In this new sense, this late work by Kuhn contributes to clarifying his point of view in Structure. So, once the assessment framework is established, as conveyed by the lexicon, the epistemic agent, the subject, is able to apply his or her evaluative baggage (i.e., those values that conform the normative background that marks his rational  See PPS, II, 38–39; III, 72; Kuhn (1990b, 14; 1993, 244, 249; 1999, 35).  See PPS, III, 76; SDLC, IV, 118. 89  See Kuhn (1991a, 95). 90  See Bell (1987, chap. 18). 91  See Katz (1978, 209). See also PPS, II, 38–39; III, 72; Kuhn (1990b, 14; 1993, 244, 249; 1999, 35); Oberheim and Hoyningen-Huene (2013). All these references are relevant to the position Kuhn shows with regard to effability, relativism, and truth. 92  See references in previous footnote 69. 87 88

76

J. V. Mayoral

behavior) and, accordingly, judge which theory he or she is going to choose – if T or T’. In Scientific Development and Lexical Change, Kuhn resorts to the idea that, in the new perspective in philosophy of science that examines belief change and, thus, scientific development (and not the static justification of beliefs), the important point is to capture how new discoveries are accepted and transform the lexicon, or are absorbed by the lexicon in place.93 In Kuhn’s main view since Structure, the relevant outcome is the way in which theory changes and knowledge is updated, either by enhancement or by replacement. It does not change in his mature stage. In this late phase of his thought (e.g., in his “Comments on Shimony”), he supports the Bayesian view, for instance, as far as its way to talk about the update of belief by virtue of the evidence available would fit in with his developmental perspective.94 In that phase, of course, science is also a group, communitarian activity.95 However, in this mature phase he re-emphasizes that the unit is a subject that weighs pros and cons of theory choice.96 Nothing in the more recent presentation modifies the idea that science is a collective activity, but it does allow analysis of which perspective the individual depends on. For Kuhn, the distribution of judgment and decision within the community is also a distribution of risks in a whole community, even though the final outcome comes from the average direction the group adopts – the “group decision,” as it were. This perspective is absent in the static tradition, in which every individual, without possible variation in that sense, is submitted to methodological rules that, in every case of rational behavior, force a choice. If, as Kuhn does, we take into account individual variation, thus giving presence and significance to the individual, the social perspective in epistemology of science comes to the forefront.97

5.7 Conclusion I started with two objectives in mind at the beginning of this essay. One of them was to reflect on Kuhn’s reconstruction of his argument in The Structure of Scientific Revolutions during a stage of his intellectual development that came after his historiographical project on the sources of quantum theory and the early quantum revolution in physics. As noted in the introduction, it took four sections in this essay (Sects. 5.2, 5.3, 5.5 and 5.6) to talk about it properly. In those sections, Kuhn’s model of lexicon has been expounded (Sect. 5.3), mainly by means of his exposition in The Presence of Past Science. These Shearman Lectures (1987) were not his last word on lexicons, and actually – as noted here and in other authors’ works – some

 See SDLC, I, esp. 17 ff.  See Kuhn (1990a, 2). 95  SDLC, I, p. 17. See also Structure, 178, 200. 96  SDLC, I, 18–19. See also, of course, Kuhn (1977b). 97  For the perspective conveyed in these last lines, see SDLC, I, 9, 17–19. 93 94

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

77

background beliefs assumed there, and sometimes conveyed, are outdated shortly after. However, the main structure remains in essence untouched and shows what ideas on the individual and his or her relationship to the world and his or her social environment  – two relationships mediated by language and the background of cognitive categories the individual contributes to design – underlie the version we see in Structure. This model for the lexicon is, at the same time, a key piece of a perspective Kuhn usually conveys in opening texts for his books and series of lectures; in particular, in Scientific Development and Lexical Change, in available versions of chapter 1 of The Plurality of Worlds  – his unfinished book  – and, in schematic fashion, in Structure, section I. That central piece is the role granted to a reconstruction of scientific development – not in an empiricist or positivistic key. That reconstruction necessitates at least two essential phases: (1) the reconstruction of a science in its precise time (on the basis of the explanatory model of the lexicon); and (2) the reconstruction of the belief-change process, also based on the model of lexicon – in this new case, in terms of a comparison between lexicons. Both phases create Kuhn’s developmental alternative in philosophy of science, as opposed to the previous static tradition, where justifying each belief by itself and not belief change was the central piece.98 The dialogue between lexicons takes place between communities, not between individuals; a community adopts a certain group of beliefs and the following adopts another one. Change, however, as seen by Kuhn, does not focus our attention on a standard epistemic agent – as in the previous empiricist tradition that, since Structure (much earlier) criticizes  – but rather on subjects that are accepted in such communities thanks to common training, and that also preserve their biographic and idiosyncratic features. As Kuhn points out in the second edition of Structure, those features do not become as important in his philosophy of science as the relationship between current beliefs in a community. Yet, his perspective remains social, I have claimed, because individuals with occasionally divergent training, perspectives, and virtues make the decisions. Now, the community is not reducible to a standard epistemic agent, as in the previous tradition Kuhn criticizes. Now it is a population – I use this term by analogy with its standard application in philosophy of biology (see Lewens 2009, 324 ff.)  – with individuals with potentially divergent training whose average row in a certain definite direction. This importance granted to the individual is visible in the later as well as in the earliest Kuhn  – that of the early 1950s, who prepared for the Lowell Institute, Boston, an outline of what later on would be The Structure of Scientific Revolutions, namely, the series of lectures whose title was The Quest for Physical Theory. Already, at that time, in that relationship between the modes of individual education – whether common or specialized – and its further membership of the social and the expert groups, the nature of language and the scientific development in  In addition to the references already examined in this chapter, see also Kuhn’s drafts of the first chapter of The Plurality of Worlds (“Scientific Knowledge as Historical Product”) in TSKP 24.2. On Kuhn’s last book, see Hoyningen-Huene (2015, 190–193); Marcum (2015, 135–137). 98

78

J. V. Mayoral

history are manifest. This model, once improved and detailed, is used again when Kuhn tries to reconstruct the argument of Structure in order to better answer old problems, or to solve new problems for the first time. That model is reproduced and improved in that of the lexicons in the 1980s. Section 5.4 in this paper has been devoted to expounding that form of continuity, or rather, of one explanatory project Kuhn undertakes several times, a different point emerging every time, during his whole life. Finally, Kuhn wishes to explore, particularly in his series of lectures in the 1980s, how this discontinuist history of scientific thought lets us talk about our history, our science. Establishing that identity necessitates making lexical diversity and human community compatible. The penultimate section of this chapter has explored if something like that is possible. We have seen that it is possible to obtain a solution, which Kuhn nevertheless excludes, that is similar to the combination of a commonsense realism, in which access to everyday objects and substances does not find many obstacles to unity in  local lexical differences (Kuhn’s local holism is key here), with an anti-realism about what Kuhn would have called “scientific objects” in the early 1950s.99 Kuhn opts for a different solution. Finally, two related themes are relativism and truth, the subject-matter of Sect. 5.6. Truth plays a key role in Kuhn’s anti-realism, but it is not expounded according to the correspondence theory, or according to a foundationalist epistemology. Besides that, relativism is also admitted, though it is not about truth but rather about the expressibility of certain beliefs. In his later phase, Kuhn uses a former model, now improved, in order to show that the apparent consequences of relativism, subjectivism, and irrationality that have occasionally been ascribed to his work are erroneous.

References Archival Sources Kuhn, Thomas S. “After Paradigms What.” Denison University, December 1980. TSKP 5.34. ———. “Comments on Shimony.” Boston University Colloquium, January 1990a. TSKP 24.7. ———. “A Function for Incommensurability.” UCLA Philosophy Colloquium, April 1990b. TSKP 24.8. ———. “An Historian’s Theory of Meaning.” UCLA Philosophy Colloquium, April 1990c. TSKP 24.9.

 See Mayoral (2017a, 150–153) (and other parts of the same book), on that concept of “scientific objects,” recently studied – not necessarily in Kuhnian fashion – by Lorraine Daston and others (see Daston (ed.) 2000). I am following Mario De Caro’s terminology of realism in this paragraph (see De Caro 2016, 2–9). The combination of realism and anti-realism that I am defending here is close, in that particular sense, to De Caro’s views on B. C. van Fraassen. Though it is worth further study, I shall not explore that connection here (see De Caro 2016, 7). 99

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

79

NCC = Kuhn, Thomas S. “The Natures of Conceptual Change.” 1–3, Perspective Lectures, University of Notre Dame, November 17, 1980. TSKP 5.36. PPS = Kuhn, Thomas S. “The Presence of Past Science.” I–III, The Shearman Memorial Lectures, University College, London, November 23, 1987. TSKP 23.32. QPT = Kuhn, Thomas S. “The Quest for Physical Theory: Problems in the Methodology of Scientific Research.” Lowell Lectures, I–VIII, n.d., 1951. TSKP 3.11. See also GR. SDLC = Kuhn, Thomas S. “Scientific Development and Lexical Change.” Thalheimer Lectures, Johns Hopkins University, November 12, 1984. TSKP 32.21. TSKP = Thomas S. Kuhn Papers, 1922–1996. MC 240. Institute Archives and Special Collections. Massachusetts Institute of Technology. Cambridge, Massachusetts. (TSKP X.Y = TSKP, box X, file Y.).

Published Texts and Secondary Literature Andersen, Hanne, Peter Barker, and Xiang Chen. 2006. The Cognitive Structure of Scientific Revolutions. Cambridge: Cambridge University Press. Bell, John S. 1987. Speakable and Unspeakable in Quantum Mechanics. Cambridge: Cambridge University Press. Bishop, Michael A., and Stephen P. Stich. 1998. The Flight to Reference, or How Not to Make Progress in the Philosophy of Science. Philosophy of Science 65 (1): 33–49. Reprinted in S.  Stich, Collected Papers, Volume 1: Mind and Language, 1972–2010, Oxford: Oxford University Press, chap. 12. (q. v.). Buchwald, Jed Z. 1992. Kinds and the Wave Theory of Light. Studies in History and Philosophy of Science Part A 23 (1): 39–74. Conant, James Bryant. 1947. On Understanding Science. New Haven: Yale University Press. Daston, Lorraine. 2000. Biographies of Scientific Objects. Chicago: University of Chicago Press. De Caro, Mario. 2016. Introduction: Putnam’s Philosophy and Metaphilosophy. In Philosophy in an Age of Science: Physics, Mathematics, and Skepticism, ed. Mario de Caro and David Macarthur, 1–18. Cambridge, MA: Harvard University Press. Fleck, Ludwik. 1979. Genesis and Development of a Scientific Fact. Foreword by T.  S. Kuhn, T.  J. Trenn and R.  K. Merton (eds.), Translation by F.  Bradley and T.  J. Trenn, Chicago: University of Chicago Press. Galison, Peter. 1997. Image and Logic. Chicago: University of Chicago Press. Hacking, Ian. 1993. Work in a New World: The Taxonomic Solution. In World Changes: Thomas Kuhn and the Nature of Science, ed. Paul Horwich. Cambridge, MA: The MIT Press. Hempel, Carl G. 1965. Chapter 8: The Theoretician’s Dilemma: A Study in the Logic of Theory Construction. In Aspects of Scientific Explanation and Other Essays in the Philosophy of Science. New York: Free Press. First published 1958. Hollis, Martin. 1970. Chapter 10: The Limits of Irrationality.” Archives Européennes de Sociologie 7 (1967): 267–271. Reprinted in B. R. Wilson (ed.), Rationality, Oxford: Blackwell. Hoyningen-Huene, Paul. 1993. Reconstructing Scientific Revolutions: Thomas S.  Kuhn’s Philosophy of Science. Foreword by T.  S. Kuhn, Translation by A.  T. Levine. Chicago: University of Chicago Press. ———. 2015. Chapter 13: Kuhn’s Development before and after Structure. In Kuhn’s Structure of Scientific Revolutions  – 50 Years On, Boston Studies in Philosophy of Science, 111, ed. W.J. Devlin and A. Bokulich. Dordercht: Springer. Katz, Jerrold J. 1978. Effability and Translation. In Meaning and Translation, ed. F. Guenthner and M. Guenthner-Reutter. New York: New York University Press. Kindi, V. 2005. Should Science Teaching Involve the History of Science? An Assessment of Kuhn’s View. Science & Education 14: 721–731.

80

J. V. Mayoral

Kitcher, Philip. 1978. Theories, Theorists and Theoretical Change. The Philosophical Review 87 (4): 519–547. Koffka, Kurt. 1935. Principles of Gestalt Psychology. New York: Harcourt, Brace and World. Kuhn, T. S. 1976. “Does Knowledge ‘Grow’?” Foerster Lecture, University of California, Berkeley, TSKP 5.13. ———. 1977a. Chapter 12: Second Thoughts on Paradigms. In The Essential Tension, ed. Thomas S. Kuhn. Chicago: The University of Chicago Press. First published 1974. ———. 1977b. Objectivity, Value Judgment and Theory Choice. In The Essential Tension, ed. Thomas S. Kuhn. Chicago: The University of Chicago Press. ———. 1977c. The Essential Tension. Chicago: The University of Chicago Press. ———. 1979. Metaphor in Science. In The Road since Structure, ed. J. Conant and J. Haugeland. Chicago: The University of Chicago Press, 2000, chap. 8. ———. 1983. Commensurability, Comparability, Communicability. In The Road since Structure, ed. J. Conant and J. Haugeland. Chicago: The University of Chicago Press, 2000, chap. 2. ———. 1987. What Are Scientific Revolutions? In The Road since Structure, ed. J. Conant and J. Haugeland. Chicago: The University of Chicago Press, 2000, chap. 1. ———. 1989. Possible Worlds in the History of Science. In The Road since Structure, ed. J. Conant and J. Haugeland. Chicago: The University of Chicago Press, 2000, chap. 3. ———. 1991a. The Road since Structure. In The Road since Structure, ed. J. Conant and J. Haugeland. Chicago: The University of Chicago Press, chap. 4. ———. 1991b. The Natural and the Human Sciences. In The Road since Structure, ed. J. Conant and J. Haugeland. Chicago: The University of Chicago Press, chap. 10. ———. 1992. The Trouble with the Historical Philosophy of Science. In The Road since Structure, ed. J. Conant and J. Haugeland. Chicago: The University of Chicago Press, chap. 5. ———. 1993. Afterwords. In The Road since Structure, ed. J. Conant and J. Haugeland. Chicago: The University of Chicago Press, chap. 11. ———. 1996. The Structure of Scientific Revolutions, 3rd ed. Chicago: The University of Chicago Press. ———. 1999. Remarks on Incommensurability and Translation. In Incommensurability and Translation: Kuhnian Perspectives on Scientific Communication and Theory Change, ed. R. Rossini Favretti, G. Sandri, and R. Scazzieri, 33–37. Cheltenham: Edward Elgar. ———. 2000. A Discussion with Thomas S. Kuhn. In The Road since Structure, part 3. London: Palgrave Macmillan. ———. 2017. Thomas S. Kuhn. Desarrollo científico y cambio de léxico. [SDLC] Foreword by Paul Hoyningen-Huene. Pablo Melogno and Hernán Miguel (eds.). Montevideo: FIC-Udelar/ ANII/SADAF. ———. 2021. The Quest for Physical Theory: Problems in the Methodology of Scientific Research. Thomas S.  Kuhn’s Lowell Lectures of 1951. I–VIII.  G. A.  Reisch (ed.), Amazon paperback. https://dome.mit.edu/handle/1721.3/189338. Lewens, Tim. 2009. Chapter 17: The Origin and Philosophy. In The Cambridge Companion to “The Origin of Species”, ed. Michael Ruse and Robert J. Richards. Cambridge: Cambridge University Press. Marcum, James A. 2015. Thomas Kuhn’s Revolution. London: Bloomsbury. Mayoral, Juan V. 2015. Significado, Conocimiento y Creencia En Kuhn. La Influencia de Wittgenstein y Austin. In Wittgenstein: La Superación Del Escepticismo, ed. D. Pérez Chico and J.V. Mayoral, 177–227. Madrid: Plaza y Valdés Madrid. ———. 2017a. Thomas S. Kuhn: La Búsqueda de La Estructura. PUZ: Zaragoza. ———. 2017b. The Given, the Pragmatic A Priori, and Scientific Change. In Pragmatism in Transition: Contemporary Perspectives on C.  I. Lewis, ed. P.  Olen and C.  Sachs, 79–101. London: Palgrave. ———. 2017c. Mundos fenoménicos y léxicos científicos: el relativismo lingüístico de Thomas Kuhn. Revista de Filosofía 42 (1): 117–134.

5  Kuhn’s Reconstruction of Structure: The Theoretical Background

81

Melogno, Pablo. 2017. Retornando al Pasado. Un Estudio Preliminar de Las Conferencias Thalheimer. In Thomas S. Kuhn. Desarrollo Científico y Cambio de Léxico, ed. Pablo Melogno and Hernán Miguel, 11–50. Montevideo: FIC-Udelar/ANII/SADAF. ———. 2019. The Discovery-Justification Distinction and the New Historiography of Science: On Thomas Kuhn’s Thalheimer Lectures. HOPOS: The Journal of the International Society for the History of Philosophy of Science 9 (1): 152–178. Mill, John Stuart. 2011. A System of Logic Ratiocinative and Inductive. In Collected Works of John Stuart Mill, John M.  Robson (ed.), Vol. VII.  London: Routledge. First published 1858. Oberheim, Eric, and Paul Hoyningen-Huene. 2013. The Incommensurability of Scientific Theories. The Stanford Encyclopedia of Philosophy, no. (Spring 2013 Edition), E. N. Zalta (ed.) https://plato.stanford.edu/archives/spr2013/entries/incommensurability/. Accessed 2 February 2015). Pinto de Oliveira, J.C. 2017. Thomas Kuhn, the Image of Science and the Image of Art: The first Manuscript of Structure. Perspectives on Science 25: 746–765. Reisch, George A. 2019. The Politics of Paradigms. Albany: SUNY Press. Rosenstock-Huessy, Eugen. 1931. Die Europäischen Revolutionen. Jena: E. Diederich. Sankey, Howard. 1997. Chapter 2: Kuhn’s Changing Concept of Incommensurability. In Rationality, Relativism and Incommensurability. Aldershot: Ashgate. Shapere, Dudley. 1989. Evolution and Continuity in Scientific Change. Philosophy of Science 56 (3): 419–437. Whorf, Benjamin Lee. 1956. Language, Thought, and Reality: Selected Writings of Benjamin Lee Whorf. Foreword by S. Chase, J. B. Carroll (ed.). Cambridge, MA: MIT Press.

Chapter 6

A Role for Cognitive Agents from a Kuhnian Point of View: A Comment to Juan Vicente Mayoral Pío García

6.1 Introduction Mayoral’s paper has two primary goals. First, Mayoral argues that “there is a common background” throughout Kuhn’s thoughts. This common background can be traced back to Kuhn’s early works. Second, Mayoral suggests that “Kuhn draws explanatory resources from that background”. Mayoral sees The Structure of Scientific Revolutions as a place where Kuhn expressed his convictions about the nature of science. Further, he also suggest that we can see The Structure as a source of problems, in the sense that this particular work does not exhaust the philosophical issues that Kuhn wants to discuss. Accordingly, previous and subsequent works fulfill an important function, both as a way to posit key problems and as an attempt to answer them. Following this idea, Mayoral then proposes that there was a kind of framework in Kuhn’s work that revealed a philosophical common space of problems even though the specific configuration and the answers to these problems may not always coincide. Furthermore, this framework was stable over the years. Another way to express those ideas is in terms of a common theoretical background that unifies different stages of Kuhn’s thought. In his paper, Mayoral offers good evidence to support these claims. First, Mayoral presents some consequences from how “history gives shape to scientific knowledge”. Knowledge is now considered a dynamic process. “Discovery and justification” are now historical processes where the tasks of historians and philosophers can be described in terms of an Ethnographic and a Developmental Phase. The work in those phases has to consider “a background composed of theories” and the “semantic network created by the terms involved”. Second, Mayoral presents a basic picture of a Lexicon which on some accounts is composed of natural-kind terms. A no-overlap P. García (*) Universidad Nacional de Córdoba, Córdoba, Argentina e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 L. Giri et al. (eds.), Perspectives on Kuhn, The Western Ontario Series in Philosophy of Science 84, https://doi.org/10.1007/978-3-031-16371-5_6

83

84

P. García

principle, a kind of labeling, and the property of projectability, among other aspects, can characterize a Lexicon. At the same time, “innate” and more specific terms, with different explanations, can highlight the general aspects of a Lexicon. Mayoral reminds us that, for Kuhn, kind membership does not depend on the satisfaction of necessary and sufficient conditions and that it is “context-sensitive”. Here it is important to specify how to describe the acquisition of a “features space”. A collective learning process can mitigate the discontinuity of feature space. Then Mayoral considers the critique that Kuhn made to the causal theory of reference. The differences between Kuhnian and Kripkean kinds of worlds reveal that behind necessary identity there is contingency. The next step in Mayoral’s paper is to discuss the perceptual and the linguistic approach in different stages of Kuhn’s thought. Here the strategy of Mayoral is, first to show linguistic and philosophical aspects in early Kuhnian work and second to show perceptual aspects of the later Kuhn. Discussions about holism and meaning, including their dynamical aspects, are the kind of philosophical problems most tied to linguistic topics. Regarding a conception of meaning, Mayoral sees similar problems—with a different vocabulary—in the 1950–1951 work (The Quest for Physical Theory). In the final part of Mayoral’s paper, he tackles two problems: our historical identity and the problem of relativism about “Truth”. Both problems are tied to each other. Local incommensurability, in terms of lexical structure alterations, is a way to introduce the first problem. The fact that there is no “truth-preserving translation” shows a link with the second problem. But Kuhn “qualifies” this idea, for example, by introducing resources that permit us to extend a Lexicon (e.g. glosses). Those qualifications are a way to answer the question about when some past is our past (or our historical identity). Along the same lines, Mayoral notes that in Scientific Development as Lexical Change, there is a reference to a “common biological heritage” and a “common cognitive background”. Mayoral finalizes his paper by taking into account the problem of relativism about truth. Each Lexicon opens up a set of possible worlds that allows constituting references for sentences. If statements can be true in different lexical, then it is not easy to avoid relativism. But we have to distinguish between relativism about Truth and relativism about effability or sayability. This second kind of relativism is the only one that Kuhn accepts. To accept this type of truth we have to admit first that “there is no truth beyond language but there is truth within it”.

6.2 The Relevance of Individuals in Kuhn Besides the interesting approach that Mayoral proposes for several problems, in this commentary I am going to address only the relevance of the individual in Kuhn. This point can be considered as a secondary aspect in Mayoral’s work, but I think it will allow us to address at least some of Mayoral’s reading strategies. Even when Mayoral makes few references to the role of individuals, in those places their importance and their problematic nature are underlined. I will suggest a way of

6  A Role for Cognitive Agents from a Kuhnian Point of View: A Comment to Juan…

85

emphasizing the relevance of the individual and defending what we are going to call a “compatibility account”. By a “compatibility account” I mean a way to obtain a Kuhnian coherent conceptual picture for individuals and communities. The contrast between what we will call a rational and an interpretative individual will be a means to underline a “compatibility account”. While, as we will see, the activity of individuals in a Kuhnian account is mainly tied to variability, I will present a potential extension of this perspective to a, particularly problematic area: the context of discovery. Pointing out the relevance of individuals, in a philosophy and history of science as focused on the role of communities as Kuhn’s is, might sound like a somewhat controversial proposal. For example, Hoyningen-Huene said that “science was seen as a one-person game but for Kuhn, it is, at its heart, a social enterprise” (Hoyningen-­ Huene 1998, 6). In the same direction, when Hoyningen-Huene tries to explain why Kuhn changes from a visual, perceptual or Gestaltic metaphor to a “description by means of a linguistic or conceptual framework” (Hoyningen-Huene 1998, 7), he points out that “although perception is socially conditioned, it is possessed by individuals; whereas, although the language is possessed by individuals, it is essentially social by nature” (Hoyningen-Huene 1998, 8). And, of course, the explanation of theoretical change relies strongly on communities. Thus, for example, how Kuhn understands the “justification” of theoretical change is from a “set of communal cognitive values that determines the outcome of the theory choice situation” (Hoyningen-Huene 1992, 147–154; 239–245). So “for Kuhn the principal agents of science are communities, not individuals” (Hoyningen-Huene 1998, 5). In Mayoral’s account, as we said above, the contribution of the individual is mainly tied to the variability. This variability does not necessarily have a problematic side for knowledge and it can also be seen as a step to knowledge generation. However, variability as a step to knowledge generation and discovery could be understood as a plain novelty (a kind of blind variability) or as a sort of strategy with its dynamics (a kind of ‘positive’ or not blind variability). Those questions take on a stronger form when they are settled in the context of discovery. But, when we ask about which aspects of the discovery context are relevant to the context of justification in the Kuhnian account, Hoyningen-Huene answers, again, by pointing out the epistemic values of a community. And here the individual differences can be, at least initially, an obstacle. The first challenge to an account that tries to underline a positive individual activity is the fact that consensus and agreement are properties that emerge from communities in the sense that they allow theory choice. It does not seem easy, then, to highlight the role of the individual in the context of innovation, at least not in a Kuhnian account. In particular, there are two other strong difficulties. The first one is to clarify, at least schematically, how to understand the context of discovery or innovation. The second is to specify what kind of cognitive role can be assigned to the individual in this space. Regarding this last challenge, we can recall all those research programs that try to revive a sort of “logic of discovery”. Most of those research programs have a common assumption: the unit of analysis is the individual agent. So, we can rephrase our question in terms of the place that those research programs can have in a Kuhnian account. But here we

86

P. García

have to make an important clarification: what is called discovery in those research programs includes normal science activities. And in a Kuhnian account, normal science could be described as a problem-solving activity, of some sort, that can be carried out by individuals, with important restrictions. Even though this extension of the innovation context removes some problems, there are other difficulties. What type of task is involved in an innovation context? When Kuhn comments on this issue concerning theory change, he said “it should be clear that the explanation must, in the final analysis, be psychological or sociological. It must, that is, be a description of a value system, an ideology, together with an analysis of the institutions through which that system is transmitted and enforced.” (Kuhn 1970, 21). And then Kuhn adds the following remark: “I doubt that there is another sort of answer to be found.” (Kuhn 1970, 21). All those problems seem to rule out a positive account of individuals in an innovation context, at least in a traditional logic of discovery. Nonetheless, we can look for, in a more general way, the kind of positive activity that a Kuhnian account would allow for individuals. This reading involved a kind of abstraction because as we have just seen, the analysis of communities is inescapable. But, as we will argue, it is far from being irrelevant. Of course, I want to remark again that this analysis, taking into account some of Mayoral’s reading keys, goes beyond his explicit assertions. However, I think that, as a complementary strategy, it can be helpful for putting together a full picture of the innovation context.

6.3 A Compatibility Account: Rational and Interpretative Individual Despite the considerations that we make above, Mayoral shows that several aspects of Kuhn’s philosophy could be seen, at least partially, from a complementary perspective that underlines the importance of individuals. Thus, for example, Mayoral’s analysis of the notion of Lexicon could be used as a way to highlight not only the common and, to some extent, stable aspects of Kuhn’s account but it could also be used to show the possibility of relative variability, through the influence of individuals. This would be a way of showing that there are forms of understanding the individual that make units of analysis compatible that at first glance were not so. However, it is clear that there are ways to understand such units of analysis that involve incompatibility, and this form of understanding, as Kuhn has said several times, has been quite common in the philosophy of science. Making an incompatible account explicit, regarding individuals and communities, can shed light on a compatibility account. One way to carry out this strategy is to contrast not only the notion of individuals with communities but also the concept,

6  A Role for Cognitive Agents from a Kuhnian Point of View: A Comment to Juan…

87

using Kuhn’s expression, of “rational individual” (Kuhn 2017, 66).1 If we consider the rational individual from what he can do as a cognitive agent, the contrast with what we might call the “interpretive” agent or individual, in other words, an individual that can be described from a Kuhnian perspective, appears in another dimension. The dimension to which we refer here is the cognitive one in the sense of an individual that produces and evaluates knowledge. In different texts, Kuhn characterizes the core and distinctive aspects of the “rational individual”. In general, this rational individual is understood from a characterization of knowledge or, more generally, from the type of activity that he can develop in a scientific context. Thus, for example in “The Trouble with the Historical Philosophy of Science” the cognitive activity of what we call “rational individual” characterizes a positivist or “traditional” type of science (Kuhn 2000, 106–107). In this way of understanding science, “facts are given”, therefore the individual does not have any interesting active role. But the task of “inventing laws and theories” requires interpretation, which entails a potential source of variations because “different individuals can be expected to interpret the facts differently, to invent characteristically different laws and theories” (Kuhn 2000, 107). However, given that these variations arise from the activity of the rational individual, then his activity is restricted as long as refers “to the facts” as “a court of final appeal” (Kuhn 2000, 107). “Propositionalism” and “Justificationism” provide a conceptual framework and a complementary way to present this “coercive” character that evaluation should have from this traditional perspective. All these aspects make “science become a game of one player”. In this traditional account, the community has a role to play, but it is only a resource that allows us to go beyond the limits of individual capacities (Kuhn 2017, 60), rather than a substantive input to the cognitive task itself. On the other hand, it is worth repeating that an individual in a Kuhnian account is always understood in the context of communities. And probably this is one of his most distinctive features. So, when we refer to the activities that an individual interpreter made, we are making, as we said above, a kind of abstraction. The justification of this abstraction, by the way, is one that I hope becomes clear by the end of this commentary. What is implied in the characterization of an “individual-interpreter” can be made explicit from different perspectives. I am just going to mention two aspects. First, there is the activity carried out by the historian and the philosopher of science when they evaluate theories. Thus, the influence of history in a way of doing philosophy of science can be shown in the shift from the task of evaluating beliefs to the task of evaluating the change of beliefs. This movement makes explicit the passage from a static account to a dynamic one. The phases involved in the dynamic account can be synthesized in terms of an Ethnographic Phase and a Reconstructive Phase of historical development. The first phase is typical of the historian and the

 To follow this path, we will take into account, as Mayoral does, Thalheimer conferences (Kuhn 2017) and Kuhn (2000). 1

88

P. García

second is usually understood from the “static” tradition. However, while this last task requires studying a process that is always comparative, the static tradition is not an appropriate description. There is a second aspect that is relevant for understanding what is implied by the individual-interpreter. Now, the individual’s cognitive activity includes the “supposedly solid facts of observations”. Facts lose the epistemic status that they previously had (Kuhn 2000). Of course, the given character of “facts” is also questioned by their relationship with beliefs and theory which is also shown in the process of producing such “facts” by instruments. But, beyond the new ways of characterizing data, the relevant issue for us is how the scope of epistemic activity of the subject is broadened. A further specification of the activity of the subject appears in the context of the discussion of the Lexicon.

6.4 Learning a Lexicon: Variability and Their Limits Probably the place where Mayoral sees with the greatest clarity the relevance of the individual, with all the provisos that we already express, is in the characterization of the Lexicon and, mainly, in the form in which it is generated and modified. Now the central issues are those related to Lexicon appropriation, construction, and modification; and in several texts, as it is well known, Kuhn appeals to learning processes as a privileged way to approach those contexts. The activity of the individual-interpreter now moves to this particular space. We can contrast the variability and the limits that an individual faces in a lexical learning context. Regarding the limits they can be summarized as follows. In the first place, we can present, as Mayoral does, the structural resistance of the lexical hierarchies. Individual variations in the acquisition of a Lexicon must take into account the hierarchical organization of natural class terms. In this sense, the variations are, from a broader perspective, few. Second, we can consider observational or “common sense” contexts. Other aspects could be added to this list,2 but for the purpose of this comment, those are enough. On the positive side of variability, we can quote the second Thalheimer conference. This is one of the places where Kuhn outlines some features of the lexical learning process. Here, acquiring a Lexicon also means acquiring a taxonomy, which entails referential expressions whose referent “plays a role in the world of the linguistic community that uses it” (Kuhn 2017, 95).3 At the same time, this task of acquiring a Lexicon is done by knowing the features that distinguish members in different categories, “but not all of them nor the same features for all” (Kuhn 2017, 94). The fact that feature spaces are so tied to specific situations or contexts, makes

 Kuhn (2017) mentions those resources that allow us to become members of a tribe.  Kuhn does not assume that there are the only three terms to be considered as referential. He simply summarizes some examples (Cf. Kuhn 2017, 94). 2 3

6  A Role for Cognitive Agents from a Kuhnian Point of View: A Comment to Juan…

89

them have a degree of variability that a Lexicon does not have. As Mayoral points out, this is also a way of specifying a kind of “subject-sensitivity”.4 Then, a learning process includes “selecting” and “discovering” features. In turn, similarity, analogy (internal and external), and something like “metaphor” can also play a relevant role. At the same time, we must highlight the interrelated nature of feature learning, a kind of “locally holistic learning” (Kuhn 2017, 96). Almost all of these learning aspects suppose a nontrivial interpretative activity from the individual, but, again, always in a community context. However, as I will defend in the next section, those aspects can be considered as the other side of the non-dependence of a “coercive” rule. Some of those learning aspects, like the internal and external analogies, also seem to go in the direction of knowledge construction. This double role, variability and discovery, is also emphasized by Mayoral in a Kuhnian learning process: “this way of learning lets us do without the old ideas of semantic rules” to obtain the meaning of the terms. And we also could dismiss methodological rules for determining the aims (what we just called a coercive rule).

6.5 Reconsidering the Activity of an Interpretive Individual: The Context of Discovery After summarizing the activity of individuals in a community in terms of variability, it is time to come back to the questions that we presented at the beginning of this commentary. In the Structure, there was already a question about the distinction between contexts. Although, as Hoyningen-Huene said, Kuhn made this aggregate only in the latest versions and by recommendations of ‘academic strategy’ (Hoyningen-­Huene 2006, 125), the important question is that the distinction between contexts was not clear anymore (Hoyningen-Huene 1987). This situation seems natural if we remember the role that the distinction had for the positivist account. Anyway, do we have something left from that distinction? In the first place, the dynamic conception of science, according to Kuhn, has substantially modified the epistemological task of the philosopher, turning it into a comparative one. This modification is the result of questioning the assumed tasks sequentially in the contexts (Hoyningen-Huene 2006; Kuhn 2017). Secondly, this new perspective has the consequence that “the considerations that were previously relevant only for the context of the discovery now become critically important for the context of justification” (Kuhn 2017). This last consideration involves, among other things, that the contextual or descriptive

 In addition, the context of application could be presented here as a space of variability that can be added to the learning context. 4

90

P. García

aspects of knowledge, can be now seen as part of the epistemologist’s enterprise.5 Questioning the sequential character of context and the relevance of a descriptive aspect can be seen as an improvement of what has to be considered as part of an individual activity rather than a total rejection. This is where the compatibility account shows its relevance. It is clear that the cognitive activity of the “rational individual” would collide with this traditional way of understanding the contexts. But, even from the point of view of an individual interpreter, as we have seen, we find an important difficulty in the attempt to construct a positive version of the discovery process, in the sense that we stated above, mainly because the choice between theories should be a collective activity, not a task made by individuals. One way to raise this issue is by appealing to the distinction we have made above between plain or negative–blind–and positive variability. This distinction may seem trivial but this evaluation depends on how it is proposed. What kind of space of inquiry are we suggesting? To use the language we have adopted so far, what kind of activity is the one that allows the movement of the rational individual to the individual interpreter in Kuhn? I present the question in terms of what it “allows” because I want to suggest a space of inquiry that probably has not been explicitly stated by Kuhn but that fits, up to a certain point, with his proposals. Returning to the question about the type of variability–whether it is blind or not–an objection to the possibility of presenting this scope of inquiry is that it would be a way of returning to the context of rules—for discovery in this case—that was rejected together with those inadequate assumptions implied by the rational subject. The following quote could be interpreted in this way: We shall not suppose any longer, according to Kuhn, that rational scientific decisions are rule-governed, that is that they follow algorithms. Rather, these decisions are influenced by the cognitive values to which the respective community is committed. The important point here is that a decision that is influenced by values is not necessarily determined by them. This implies that different individuals influenced by the same values may come to different decisions.6

So, the central issue here is the coercive nature of strategies or their algorithmic form. In other words, this means that we have to distinguish, to make a fair evaluation, between algorithms and other types of strategies–that could not be coercive. Heuristics, in the classical or contemporary tradition,7 could be an example of those other kinds of rules or strategies.

 Hoyningen-Huene (1992, 2006) points out that although Kuhn has not questioned the difference between descriptive or factual and normative or evaluative statements, this instance does not imply a Kuhnian commitment to the nature of facts or norms, nor does Kuhn said anything about how those kinds of statements could be related. 6  Hoyningen-Huene (1992, 492). 7  George Polya (1945, 1954) could be cited in the realm of the mathematical tradition, Allen Newell and Herbert Simon (Newell and Simon 1972) in a computational approach and Amos Tversky, Daniel Kahneman (Tversky and Kahneman 1973) or Gerd Gigerenzer (Goldstein and Gigerenzer 2002) in a psychological or rational choice approach. 5

6  A Role for Cognitive Agents from a Kuhnian Point of View: A Comment to Juan…

91

However, even in this version, we could make a stronger objection: What is the point of having rules or strategies if they are not coercive? This is where greater clarity emerges about what is implied in the abandonment of the assumptions that a rational individual presupposes. The search for “rationality” in theoretical change requires different levels of analysis and different sources. Then, when we ask why we have to analyze the role of the individual if the decision is made in a community context, it can be answered by pointing out that these aspects, if they indeed can be part of the innovation process, will help us understand the whole process. Now, we have offered a justification for the abstraction about talking of individuals that we present above. And this can be seen as a similar argument to the one that Mayoral presents at the end of his paper in which the collective activity in theory choice can be analyzed, until a certain point, in terms of individuals’ perspectives. I want to emphasize that both, asking this question–about why to study individual context–and its answer in algorithmic terms have a similar pattern of “rationality” in which coercion is a distinctive feature. In other words: expecting that coercive rules are the relevant answer to research about the context of discovery is to situate the relevant features of innovation in the realm of a rational subject–a positivistic approach. A “rationality” centered on the individual-interpreter does not have to have those restrictions. The only restriction that interests us here is the historical or philosophical relevance of that reconstruction. The cognitive role of the individual, in an appropriate context and articulated with the cognitive values of the community, far from being irrelevant has to be studied thoroughly. A final remark regarding the research programs of discovery that I mentioned above. There were research programs of discovery related to logic, heuristics, computational process, psychology of different kinds–or even a sort of evolutionary explanation. A Kuhnian account that takes the communities as a central unit, questions the successive character of discovery and justification, makes descriptive issues relevant and includes other disciplines in our judgment about why innovations are made, can be considered as a kind of acid test for research programs about discovery. The relevance of the individual realm does not mean that it is the only place to look for when one tries to investigate the scientific discovery. On the other hand, even when a community can be the place of a final answer about a change of theory it does not means that it becomes the only place to look.

References Goldstein, Daniel G., and Gerd Gigerenzer. 2002. Models of Ecological Rationality: The Recognition Heuristic. Psychological Review 109 (1): 75. Hoyningen-Huene, Paul. 1987. Context of Discovery and Context of Justification. Studies in History and Philosophy of Science Part A 18 (4): 501–515. ———. 1992. The Interrelations Between the Philosophy, History and Sociology of Science in Thomas Kuhn’s Theory of Scientific Development. The British Journal for the Philosophy of Science 42: 487–501.

92

P. García

———. 1998. On Thomas Kuhn’s Philosophical Significance. Configurations 6 (1): 1–14. Johns Hopkins University Press. Retrieved November 22, 2018, from Project MUSE database. ———. 2006. Context of Discovery and Context of Justification. In Revisiting Discovery and Justification: Historical and Philosophical Perspectives on the Context Distinction, ed. Jutta Schickore and Friedrich Steinle, vol. 14, 119–131. Arquimedes. Kuhn, Thomas S. 1970. Logic of Discovery or Psychology of Research? In Criticism and the Growth of Knowledge: Proceedings of the International Colloquium in the Philosophy of Science, London, 1965, ed. I.  Lakatos and A.  Musgrave, 1–24. Cambridge: Cambridge University Press. ———. 2000. The Road since Structure, ed. J. Conant and J. Haugeland. Chicago: The University of Chicago Press. ———. 2017. Thomas S.  Kuhn. Desarrollo científico y cambio de léxico. Conferencias Thalheimer, Universidad Johns Hopkins, Baltimore, November 12–19, 1984, ed. Foreword by Paul Hoyningen-Huene. Melogno, Pablo, and Hernán Miguel. Montevideo: FIC-Udelar/ ANII/SADAF. Newell, Allen, and Herbert Alexander Simon. 1972. Human Problem Solving. Vol. 104. 9. Englewood Cliffs: Prentice-Hall. Polya, George. 1945. How To Solve It. Princeton, NJ: Princeton University Press. ———. 1954. Mathematics of Plausible Reasoning, Volume I: Induction and Analogy in Mathematics. Princeton University Press. Tversky, Amos, and Daniel Kahneman. 1973. Availability: A Heuristic for Judging Frequency and Probability. Cognitive Psychology 5 (2): 207–232.

Chapter 7

Incommensurability and Metaincommensurability. Kind Change, World Change and Indirect Refutation Eric Oberheim

7.1 Introduction The idea that there is incommensurability in science has been controversial since its popularization by Thomas Kuhn (1962, hereafter Structure) and Paul Feyerabend (1962, hereafter ERE). At that time, however, neither notions about incommensurability in science, nor the use of the term ‘incommensurable’ (or some cognate of it) were particularly new in the context of discussions about scientific method. Karl Popper (1935) had developed a formal definition of “inkommensurable” as a characterization of the relative sizes of sets of potential falsifiers of two theories whenever one set is not a proper subset of the other, and Albert Einstein (1949) had already cryptically remarked that theory comparison can, in a way that he was explicitly unable further to articulate, involve “a kind of reciprocal weighing of incommensurable qualities” (1949, 23). Moreover, many of the basic components of incommensurability in science as developed by Kuhn and Feyerabend in 1962 and later works, although not the term itself, can already be found in Pierre Duhem (1954). While incommensurability in science does have significant implications for understanding science and its development, much of the controversy about incommensurability appears to be at least in part due to a lack of clarity about exactly what is being claimed, what that claim implies, and how the claim is justified. This can easily be seen in recent literature, which has continued to struggle to delineate a coherent account of incommensurability in science, with some opponents attempting to undercut its legitimacy entirely, and others attempting to accept it in some weakened form, so as to downplay its

E. Oberheim (*) Humboldt-Universität zu Berlin, Berlin, Germany e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 L. Giri et al. (eds.), Perspectives on Kuhn, The Western Ontario Series in Philosophy of Science 84, https://doi.org/10.1007/978-3-031-16371-5_7

93

94

E. Oberheim

significance for our understanding of science. For this reason, a clear, precise and unambiguous statement of incommensurability, and its causes, consequences and implications may help to demystify something that has proven to be especially difficult to pin down. The remainder of the paper attempts to explain much of the confusion surrounding incommensurability on the basis of metaincommensurability, in an attempt to expose the exact nature of the relationship between incommensurability and scientific realism more clearly. To these ends, this paper addresses the following questions as directly and unambiguously as possible, while replying to Sankey (2018). The questions are: 1. What causes incommensurability? 2. What are the consequences of incommensurability? 3. What are the differences between Kuhn and Feyerabend on incommensurability? 4. What does incommensurability imply about theory comparison? 5. What does incommensurability imply about scientific progress? 6. What does incommensurability imply about scientific realism? 7. What is the justification for incommensurability? 8. What is metaincommensurability? 9. What causes metaincommensurability? 10. What are the consequences of metaincommensurability? 11. What is the justification for metaincommensurability? 12. What can we conclude about incommensurability and metaincommen surability? According to scientific realism, scientific theories should be interpreted realistically (that is, as referring to the entities, both observable and unobservable, postulated by the theory that are independent of our thoughts about them or theoretical commitments to them), and are confirmable, and in fact confirmed as approximately true (by ordinary scientific evidence interpreted in accordance with ordinary methodological standards), so that progress in mature sciences is largely a matter of more accurate approximations to the truth, where later theories build on the knowledge embodied in previous theories (see Boyd 1984; Sankey 2001). Incommensurability in science undercuts some implicit assumptions of common sense images of science and scientific realism. As we shall see, while incommensurability is first and foremost a methodological claim (about scientific method), it has epistemological and ontological components, as it bears directly on the justification of choice of theories about the kinds of things we take to be real. For our purposes, it will be helpful before we begin to try to disambiguate different stages in the development of the ideas about incommensurability within both Kuhn and Feyerabend’s philosophies. I will first briefly sketch their philosophical developments, locating the generation of their conceptions of incommensurability with them. While Kuhn insisted that the central idea driving his philosophy was always incommensurability, he had not been trained as a professional philosopher when he first encountered it in the mid to late 1940s. At the time, he was a graduate student struggling with what appeared to be nonsensical passages in Aristotelian physics,

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

95

parts of which initially seemed preposterous to him. Kuhn could not accept that someone as talented as Aristotle could have written something that appeared to be so confused and nonsensical, so that it seemed to him as if Aristotle had known almost no mechanics at all. Some time after his initial perplexity eventually patterns in the disconcerting passages began to emerge, and then all at once, the texts that he had been struggling with finally made sense to him. He saw this process of meaning change as a method of historical recovery. He realized that in his earlier encounters, he had been projecting contemporary concepts back into his historical sources and that in order to succeed, he had to discard them so as to remove the distortion, thereby allowing him to understand the Aristotelian system in its own right. But to get more closely at Aristotle’s original usage, he had to expand the meaning of motion to cover a much broader range of phenomena that includes various other sorts of change, such as growth and diminution, alternation, and generation and corruption, making an object’s motion in space just a special case of a more general kind of change, and then Aristotle all started to make sense. Kuhn realized that these sorts of conceptual differences indicated breaks between different modes of thought, and he suspected that such breaks must be significant both for the nature of knowledge, and for the sense in which the development of knowledge can be said to make progress. Having made this discovery, Kuhn changed his career plans, to pursue this strange phenomenon. Some 15 years later the term ‘incommensurable’ first appears in his work in his classic The Structure of Scientific Revolutions (1962). When he wrote Structure, professional philosophers were explicitly not the audience that Kuhn was targeting, although they often shared the image of science as progress towards truth that he did target. Kuhn’s aim was to transform society’s image of science, and by many accounts, he was extraordinarily successful in that aim, although not always in ways that met everyone’s approval, not even his own. The popular image of science he wanted to transform is primarily promoted by the superficial, cursory historical introductions in the textbooks used to teach natural sciences, and by cumulative histories of science that present scientific progress as an inevitable march toward truth, which were pervasive in the earlier ‘whiggish’ historiographic tradition, but which is now widely frowned upon among professional historians. Kuhn characterizes the popular image of science that he is targeting as scientific progress as “development-by-accumulation” (Structure, 2) and “the piecemeal process” of adding new “facts, theories and methods” to the “ever growing stockpile that constitutes scientific technique and knowledge” (Structure, 1–2). With Structure, Kuhn tried to sketch a different conception of science; one that emerges from the historical record of research activity itself (Structure, 1); one that includes scientific revolutions (in addition to normal science), and he may indeed have helped reshape the image of science to an enormous extent, given that terms like ‘paradigm-shift’ have become commonplace. The key to reaching this goal, for Kuhn, was coming to terms with a challenging fact: the incommensurability of paradigms separated by scientific revolutions. From a Kuhnian perspective, instead of some cumulative and continuous development toward truth, the historical record shows a pattern. For science to get started in the

96

E. Oberheim

first place there needs to be a paradigm. Paradigms are communities of scientists who share agreement about exemplary problem solutions (‘exemplars’). Agreement about these fundamentals makes the transition from pre-normal science to normal science possible, and it makes cumulative scientific progress within normal science possible. Normal science eventually runs into anomalies that may cause ‘crisis’ phases, research phases that lack consensus about the fundamentals in a field. Kuhn’s account stresses both that “Anomaly appears only against the background provided by the paradigm” (Structure, 65); and that crisis is a precursor to scientific revolution. In crisis phases, existing exemplars fail to solve some problem, or show how some problem that should be solvable isn’t, eventually undermining confidence in the paradigm’s future potential. In these phases, new candidates are proposed and developed to solve the problem as potential replacements for the existing paradigm. As members of the scientific community are gradually converted to the new paradigm, the result is a scientific revolution, in which the preceding paradigm is eventually abandoned and replaced by the new one, starting the cycle anew. This is supposed to be a slow, generational, cyclical process, which can have far reaching consequences, resulting in new realities based on new standards of what counts as science. It is important to emphasize that according to Kuhn’s historical perspective (and in contrast to the cumulative image of science), there is a repeating pattern in how scientific communities develop through phases of normal science (which is cumulative) and then crisis (which may or may not spread while anomalies may or may not accumulate depending on both internal and external factors shaping actual scientific practice) which may lead to a scientific revolution (which is discontinuous), establishing one or more new paradigms and the ensuing normal scientific research in place of the old paradigm. Kuhn (unlike Popper and Feyerabend) was focused on “the dynamic process by which scientific knowledge is acquired rather than with the logical structure of the products of scientific research” (Structure, 1). The actors in this historical process are ‘paradigms’ as understood to be scientific communities, or what Kuhn later called a ‘disciplinary matrix’, which are just groups of individuals who share common exemplary problem-solutions (‘exemplars’). These paradigms (communities sharing exemplars) are multi-purpose, quasi-dogmatic units that hold scientific communities together by securing agreement on basic principles that guide future research, making normal science and normal scientific progress possible. Feyerabend’s background, by contrast, was indeed professional philosophy of science, but that was before it had become professionalized. He had been heavily influenced by Karl Popper from around the time he earned his Ph.D in philosophy under Viktor Kraft’s supervision in Vienna, and Feyerabend’s Popperian roots, in many respects, directly set the stage for his development of the idea about incommensurability in science. One way of looking at it is that with incommensurability, Feyerabend pushed Popper’s maximum to increase testability to an extreme, and thereby undermined the falsificationism from which it was born, transforming Popper’s critical rationalism into Feyerabend’s critical irrationalism. Throughout his early papers, Feyerabend explicitly emphasized and distinguished his normative philosophical claims from his descriptive historical claims,

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

97

using a two-pronged argumentative strategy (normative and descriptive). For example, Feyerabend argued that any universal methodological rule is bound, at some point, to inhibit progress, as there are clear and obvious exceptions of progress in flagrant violation of every universal rule that has ever been proposed. Feyerabend captured this conclusion with the infamous slogan ‘anything goes’, by which he meant both that there are no universally applicable rules to the pursuit of science, and that methodological rules should not be universally enforced in the pursuit of progress. For each specific historical situation, ‘anything goes’ implies ‘whatever works’ (and not everything works). For any proposed methodological rule, there are episodes of progress that were made by breaking that rule (descriptive), and so if we want further progress (and to protect the possibility of further progress), no methodological rules should be universally enforced (normative). For Feyerabend, science is an attempt at a realistic interpretation of experience with no fixed method. He did not see patterns in the history of science like Kuhn did, nor did he think that there were historical laws guiding its historical development (like Kuhn did), nor did he endorse Kuhn’s suspiciously transcendental project of explaining how and why science necessarily must develop according to them. The provocative philosophical pluralism he proposed and practiced suggests that we should: “Invent and elaborate theories which are inconsistent with the accepted point of view, even if it should happen to be highly confirmed and generally accepted.” (1965a, 223–224). The perennial problem in science, in Feyerabend’s view, is dogmatic resistance to improvements that would replace the existing point of view on the basis of undue confidence in it. The driving idea seems to be that imaginative new ideas could more easily flourish if current ideas, no matter how successful they may seem to be, were more easily suspendable, at least temporarily for the sake of argument. Feyerabend’s general aim, as consistently explicitly stated throughout his early papers (that then morphed from Feyerabend (1970) into Against Method (1975), was to try to disinfect various forms of empiricism from the mistaken assumption that theories separated by scientific revolutions share a common observation language that provides a neutral set of facts that act as an arbiter for theory choice. This pernicious assumption, according to Feyerabend, is shared by logical empiricists accounts of reduction and explanation, and as we shall see, anyone who assumes that the hypothetico-­ deductive method of testing theories (which seems to be the only one we have) used in normal science also applies to revolutionary developments without further complications. Understanding incommensurability, for Feyerabend, is the key to protecting progress in science from overly simplistic models used to support dogmatic philosophies (including Popper’s). In Kuhn’s case, things are a little overly-complicated, partly due to a lack of precision in his initial presentation, and partly due to several subsequent missteps on the way to his mature views. His initial idea of incommensurability, as it appears in Structure, is holistic in at least three relevant respects: (a) paradigm specific language of laws and theories used in exemplary problem solutions is holistic in that the central terms are inter-defined; (b) there are co-occurring and interrelated changes in the concepts, methods and standards shared by paradigms in scientific

98

E. Oberheim

revolutions, and (c) he characterizes conceptual change as world change. Incommensurability is also the cause of communication difficulties, as Kuhn emphasized with the dramatic claim that history of science reveals proponents of competing paradigms failing to make complete contact with each other’s views, so that they are always talking at least slightly at cross-purposes, while characterizing the collective reasons for these limits to communication as the incommensurability of pre and post-revolutionary scientific traditions. In Structure Kuhn does not initially apply ‘incommensurable’ (or some cognate of it) directly to ‘theories’, and he makes almost no mention of ‘translation’ at all. This shift of focus directly to theories happens after 1962, as Kuhn struggled to professionalize his historical views philosophically, while Feyerabend was increasingly focused on professionalize his philosophical views historically. Feyerabend also moved toward Kuhn with respect to their intended audiences primarily not being professional philosophers, but rather scientists and the general educated public, which was complete at least by Against Method and included sharp criticism of ‘professional’ philosophy of science. Kuhn attempted to clarify his views in Structure with the addition of “Postscript–1969”. It is here that he began to suggest that “men who hold incommensurable viewpoints be thought of as members of different language communities and that their communication problems be analyzed as problems of translation” (Structure, 175) and that “what the participants in a communication breakdown can do is recognize each other as members of different language communities and then become translators” (Ibid, 202). This would be fine if we replace “translators” by “bilinguals”, as ironically, a particular type of translation turns out to be exactly what Kuhn later claims cannot be done in cases of incommensurability, before realizing that bilingualism makes the issue of translation somewhat moot (at least for scientists comparing theories), and that communication difficulties, like talking slightly at cross-purposes, arise because of ineffability, not because of untranslatability, which is merely an indicator of incommensurability that should not be confused with it. Kuhn’s development of these ideas about incommensurability and translation were sort of a cul-de-sac, which begins with misguided considerations about the indeterminacy of translation and then becomes increasingly fixated on vague and misleading ideas about translation failure, before developing his mature views, which are more specific than his original views, but as we shall see, are more specifically delineated in terms of reference (and not meaning) and ineffability (and not translation failure). Translation failure may continue to be a good indicator of incommensurability and thus useful to historians trying to make sense of antiquated theories, but is only relevant methodologically (for comparing theories empirically) to the extent that it may be needed to distinguish different theories as different theories (and not just two different ways of saying the same thing) in the first place. After this series of missteps about the meaning of meaning increasingly lost in translation, Kuhn refocused on his original emphasis of holistic conceptual change as world change culminating in his 1990s notion of incommensurability as characterized by the no-overlap principle with respect to reference (not meaning) and

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

99

ineffability (not untranslatability), on which we will primarily be focused. As in Structure, Kuhn’s mature views on incommensurability are set into an evolutionary framework (like Popper and Einstein, but not Feyerabend). Moreover, the role of incommensurability in Kuhn’s evolutionary account of scientific progress is also more clearly specified as analogous to the function of reproductive barriers as isolating mechanisms involved in speciation that includes a modification to Kuhn’s original linear, cyclical phase model of scientific advance, so as to include instances of speciation by specialization that are best represented by scientific communities branching into two or more contemporaneous scientific communities that use incommensurable theories to pursue different sets of questions about the same things. The specification of incommensurability on the basis of the no-overlap principle is now sometimes called “taxonomic incommensurability” as distinguished from “methodological incommensurability”. It is this ‘taxonomic’ idea of incommensurability that we will be primarily concerned with here, as the methodological idea is easily understood and presumably widely accepted, although often in different terms, although its implications for our ability to understand scientific progress appear to me to remain largely ignored and generally not well-understood. In his mature philosophy, Kuhn unpacks his idea of incommensurability with tools from linguistics, such as the no-overlap principle and the function of the taxonomic structure of a lexicon, or what I will sometimes call a ‘lexical taxonomy’, which are all just seemingly fancy ways of making a rather simple point. Theories are stated according to mutually exclusive kinds. If you just look at theories, you immediately can see that they state general relations between kinds and that the kinds must be structured taxonomically for the theories to make sense. According to the no-overlap principle, for any theory, no two kinds terms used to state it may overlap in their referents, unless they are related as genus to species, that is, unless all of the referents of one kind term are a proper subset of the referents of the other kind term. For example, there are no dogs that are also cats (while both all dogs and all cats are animals), and that is what makes the terms ‘dogs’ and ‘cats’ kind terms. According to Kuhn, mutually exclusive kind terms are necessary in order to state laws and theories which must be learned together through the experience of using those laws and theories to solve particular puzzles (2000) that can serve as exemplars guiding further research. It is worth emphasizing that what Kuhn is suggesting is that the no-overlap principle is a necessary precondition for scientific explanation, just as breaking it is a necessary condition for revolutionary scientific progress. Here I am suggesting that since 1962, Kuhn had been trying to elucidate incommensurability as an essential mechanism of scientific advance, and with the no-­ overlap principle he was finally able to specify unambiguously necessary conditions for normal science and revolutionary scientific advance (using, and then breaking, the no-overlap principle). Feyerabend had been developing ideas about incommensurability from as early as his Ph.D. thesis submitted in 1951, and perhaps even earlier. Feyerabend had a comparatively stable account of incommensurability in theory testing throughout his early philosophical papers (i.e. pre-Against Method) and sometimes even well into his later philosophy.

100

E. Oberheim

If there were significant changes to his ideas about incommensurability, it was because his departure point concerned Popperian questions about the role of logic in theory testing in science; and then became increasingly historical and anthropological as the details of the logic of testing theories rescinds into the background and Feyerabend shifted focus to transitions between worldviews. In Against Method, he begins to extend the application of the idea beyond scientific theories in anthropological, cognitive, perceptual and developmental directions. Here he uses Worff’s notion of ‘covert classifications’ to incommensurability. Even later, Feyerabend’s use of ‘incommensurability’ appears to have taken on different meanings altogether, and may even have been completely given up in his cultural philosophy, according to which potentially every culture is all cultures.

7.2 What Causes Incommensurability? Incommensurability is caused by reclassification that breaks the no-overlap principle. This occurs in the course of revolutionary advance when scientists propose theories stated according to a new lexical taxonomy that cross-classifies some of the same things into a new incompatible set of kinds with respect to the lexical taxonomy used to state the reigning theory, which the new theory may eventually replace. A lexical taxonomy is incompatible (and thereby mutually exclusive) with another lexical taxonomy, if the kinds of one lexical taxonomy cannot be introduced into the other lexical taxonomy without breaking the no-overlap principle. If that were allowed to happen, the things to which the kinds of the two incommensurable theories refer would be subject to different sets of natural laws, resulting in conflicting expectations about the same things and kinds of things, loss of logical relations between statements made about those things and kinds of things, and ultimately incoherence. In this way, the no-overlap principle limits the scope of logical implications on the basis of referential relations of kind terms to within a single lexical taxonomy, keeping conceptual incompatible theories logically separated. Here is a schematic overview of Kuhn’s favourite example of theories separated by a revolution cross-classifying the same things into two mutually exclusive sets of mutually exclusive kinds breaking the no-overlap principle: the transition from the Ptolemaic to the Copernican worldview. Cross-classification breaking the no-overlap principle. Ptolemaic theory: All Planets travel around the Earth Copernican theory: All Planets travel around the Sun

7  Incommensurability and Metaincommensurability. Kind Change, World Change… Ptolemaic kinds Earth Planet Star

Objects Earth Moon Mars Sun Fixed stars

101

Copernican kinds Moon Planet Star

In this example, we can see the Copernican lexical taxonomy precludes the Ptolemaic lexical taxonomy by breaking the no-overlap principle with respect to it. The Ptolemaic theory distinguishes three mutually exclusive kinds of things (Earth, Planets and Stars) and describes the laws that govern them. For example, all Planets travel around Earth. Here, ‘Earth’ is used as a kind with just one member: the Earth, which according to the theory, is stationary and located near the center of the universe, and is not a Planet or Star (as Planets and Stars travel around it). The Copernican theory distinguishes exactly the same set of things (Earth, Moon, Sun, Mars and Stars) into three different kinds of things (Stars, Planets and Moons), which cross-classify some of the same objects into mutually exclusive sets of mutually exclusive kinds with respect to the Ptolemaic lexical taxonomy. For example, the Moon is a Planet according to one taxonomy, and it is a Moon according to the other, where Moons can no longer be Planets, as Moons travel around Planets. The two kind terms called ‘Planet’ (Ptolemaic Planets and Copernican Planets) cross-­ classify some of the same objects (Earth, Moon, and Sun are all cross-classified by ‘Planet’), breaking the no-overlap principle.

7.3 What Are the Consequences of Incommensurability? Incommensurability causes ineffability. With the conceptual resources of either theory, the other point of view cannot be coherently formulated. A direct consequence of introducing a rival theory whose lexical structure breaks the no-overlap principle with respect to the reigning theory is that the new theory is literary unstateable on the basis of the taxonomic lexical structure imposed by the older theory, just as the reigning theory is literally unstateable on the basis of the taxonomic lexical structure imposed by the new theory. I will call this mutual ineffability, which has two components: forward and backward facing ineffability. This happens because a taxonomic lexical structure determines the mutually exclusive kinds of things that should obey the laws as stated by the theory, and something cannot be two mutually exclusive kinds of kinds of things at the same time without resulting in incoherence. In order to overcome the ineffability, the older vocabulary must be temporarily suspended for the sake of learning and coherently stating the new theory, just as understanding older theories requires temporarily suspending newer ideas.

102

E. Oberheim

Both ineffabilities (forwards and backwards facing) have direct practical implications that can be easily illustrated by the transition from Ptolemaic to Copernican theory. When scientists make progress by introducing a revolutionary new theory that eventually replaces the reigning theory, they make statements such as the correct Copernican statement: “All Planets travel around the Sun”, which is nonsensical from the older Ptolemaic perspective, as the Sun is a Planet that travels around Earth, and anyway, a Planet cannot travel around itself (incoherence). From this simple example, we can see how Copernican statements preclude the use of Ptolemaic kinds. They preclude them by breaking the no-overlap principle so that mixing terminology results in incoherence. From the new perspective, the ineffability of older theories is not really much of a problem for scientists. As the claim to truth of the older theory is relinquished, so too is belief in the kinds of things that the theory postulates. Ptolemaic kinds are not real kinds. There are no such things as Ptolemaic Planets, so for contemporary scientists, it is not really a problem that they can no longer be coherently discussed in our contemporary vocabulary. Scientists don’t need to talk about them because they do not exist. This is not to say that incommensurability does not imply anything about the potentially serious challenges posed to scientists trying to propose a revolutionary new theory in the first place, which at least initially will appear incoherent  (for example, bending space), nor the difficulties in trying to justify the new theory on the basis of the results of measurements, nor difficulties that may arise due to inability or unwillingness to suspend firmly held beliefs already in place for the sake of argument. So scientists moving forward must be able and willing to suspend and eventually unlearn their current taxonomy, even given, “the difficulties of conversion [that] have often been noted by scientists themselves” (Structure, 151). For the historian of science, on the other hand, there are direct practical challenges due to the incommensurability of older theories. Understanding antiquated, seemingly incoherent, scientific theories requires suspending current kind concepts that may distort those antiquated views. The historian of science has to try to learn the old theory in its own kind terms, and this requires suspending (temporarily unlearning) the kind terms used to state current theories. This is why Kuhn recommended focusing on the seemingly incoherent parts when trying to break into antiquated texts. The practical implication for historians moving backward as they try to find their way into the conceptual apparatus of archaic theories that results from confronting the use of kinds that are incompatible and mutually exclusive with the current lexical taxonomy can also easily be illustrated by the Copernican Revolution. Historians of the natural sciences moving backward are confronted with statements such as ‘All Planets travel around the Earth’ which are patently nonsensical from a Copernican perspective because the Earth is a Planet and a Planet cannot travel around itself. From this, we can see how correct Ptolemaic statements require suspension of the taxonomic lexical structure of the Copernican theory, which identifies the kinds of things that obey the laws of that theory. Understanding Ptolemaic theory and making correct Ptolemaic statements requires the suspension of the taxonomic lexical structure of Copernican theory that is incommensurable with it.

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

103

Scientists and historians can overcome the practical challenges of the mutual ineffability caused by incommensurability in exactly the same way. To overcome the challenges both scientists and historians have to become bilingual. Once one has learned both theories, (and this involves learning to keep them separate), the ineffability (and any challenges posed by it) that results from the incommensurability becomes irrelevant.

7.4 How Does Feyerabend’s Conception of Incommensurability Differ from Kuhn’s? According to Feyerabend, two theories are incommensurable (1) “if a new theory entails that all the concepts of the preceding theory have zero extension or if it introduces rules that change the system of classes itself” (1965b, 268). So in order for there to be incommensurability (2) “The situation must be rigged in such a way that the conditions of concept formation in one theory forbid the formation of the basic concepts of the other” (1981b, 154, cf. 1975, 269). Or put another way (3) “When the meaning of their main descriptive terms depend on mutually inconsistent principles” (1965a, 227), so that incommensurability means (4) “deductive disjointedness, and nothing else” (1977, 365). In (1), we see that Feyerabend and Kuhn agree on a very central and decisive point for distinguishing themselves from logical empiricist conceptions of theory reduction and cumulative conceptions of scientific progress. Revolutionary new theories imply that the concepts used to state the old theories refer to kinds of things that are no longer thought to exist. The older theory is replaced, not improved, as new types of kinds and their relations replace older types of kinds and their relations. On that view, older theories are explained by (or reduced to) more general new theories by exposing how the old theory is a special case of the new, so, new theories are improvements of older ideas, not wholesale rejections of them. But in (1) we see that Kuhn and Feyerabend agree that incommensurability involves changing ‘the system of classes’, which Kuhn went on to identify specifically as changing the ‘taxonomic structure of the lexicon’ so that it breaks the no overlap principle by cross-classifying some of the same things into incompatible sets of mutually exclusive kinds. It may be worth noting that “entails” in (1) does not actually mean logical entailment. It just means that if one theory is correct, the other one is talking about kinds of things that have ‘zero extension’. They do not refer to real kinds. There are no Ptolemaic Planets. In other words, their mutually exclusivity is not due to their logical inconsistency. It is caused by conceptual incompatibility, as evidenced by their breaking of the no-overlap principle with respect to each other. In (2) we see that incommensurability is not just any kind of conceptual change. It only occurs when the conditions of concept formation (as determined by the lexicon shared by the new paradigm) of a successor theory forbids the formation of the basic concepts of its predecessor. So conceptual changes that do not have this

104

E. Oberheim

specific and very special result are not episodes in which incommensurability arises. This actually provides quite a wide latitude for conceptual change in the course of normal science. It is also important to note here, “forbids” in (2), like “entails” in (1) is obviously somehow metaphorical. It should be read as intended to mean “makes impossible”, where the impossibility obviously assumes simultaneity (that is, within the same theory, statement or argument), as the coherent formulation of the concepts of each of both theories separately is presumed possible. In (3) we can see why and how Feyerabend’s suggestion about the shortcomings of the logical empiricist account of the logical structure of theory testing got bogged down in irrelevant considerations about meaning. Moreover, we are confronted by loose and misleading talk by Feyerabend with the term “inconsistent” as the principles from incommensurable theories are exactly not logically inconsistent but ‘deductively-disjoint’, see (4). Incommensurable theories are conceptually incompatible, which precludes that they are logically inconsistent. In (4), Feyerabend tries to isolate the basic idea of incommensurability in a single phrase “deductive-disjointedness”, which seems as mysterious and even more awkward than ‘incommensurability’’. As we have seen, general statements (like laws and theories) use kind concepts structured into mutually exclusive sets in order to state general relations that hold between those kinds. Logical implication can only be valid if those kinds are fixed within the theory, statement, measurement procedure, or argument to which they belong, otherwise logical implication will be unreliable and lead to incoherent statements. Logic reaches its limits in the sense that it works fine for theory testing in normal science, but in revolutions, when new incommensurable theories are introduced, a logical gap opens between the two theories that cannot be bridged by logical implication, because the two theories organize their basic concepts into mutually exclusive lexical taxonomies by cross-classifying some of the same things into incompatible sets of mutually exclusive kinds. What is denied is that anybody can consistently continue using this more primitive concept [of temperature] and at the same time believe in the molecular theory [of temperature]. Again, this does not mean that a person may not, on different occasions, use concepts which belong to different and incommensurable frameworks. The only thing that is forbidden for him is the use of both kinds of concepts in the same argument (Feyerabend (1962/1981a, 83).

Kuhn’s taxonomic elaboration of incommensurability on the basis of the no-­overlap principle is basically the same as Feyerabend’s conception of incommensurability as developed in his early papers. Feyerabend does add some restrictions and a qualifier to his discussion of incommensurable theories, and his conception of scientific progress ends up somewhat different than Kuhn’s, but the basic idea of incommensurability is the same for both. As for the qualifier and restrictions, in 1962, Feyerabend explicitly restricts his discussion of incommensurability to non-instantial or ‘universal’ theories in physics and adds the qualifier: if they are both realistically interpreted. If the theories are not interpreted realistically, then there is nothing barring us from using separately either of two incommensurable theories to solve particular problems at any given time, but it still remains that only one of them can be interpreted realistically at a time. We

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

105

still can and do sometimes use theories instrumentally, that is, without believing that the kinds of things that they imply are real. For example, we still use Newtonian theory to solve problems for which its handy, even after it has been replaced by relativity theory for the purposes of a realistic interpretation of experience (there is no absolute mass). Feyerabend’s restrictions to ‘non-instantial’ or ‘universal’ theories in physics make it seem like the scope of Kuhn’s application of incommensurability is much larger than Feyerabend’s, as with this restriction, he seems to be ruling out Kuhn’s favourite example of the Copernican revolution (as well as many of Kuhn’s other examples), which involved a reconceptualization of astronomical laws, not non-­ instantial, universal theories. While Feyerabend (1975), in retrospect, even once made this point himself, it may be somewhat misleading, as can be seen upon closer inspection of Feyerabend (1962/1981a, 29): “scientific theories are ways of looking at the world; and their adoption affects our general beliefs and expectations, and thereby also our experiences and our conception of reality”. The idea seems to be that if a universal theory determines our entire conception of reality, then it is hard to see how facts that do not comply with it could ever be discovered, especially if the only facts that can be used to test it have to be deduced from it, and we have no direct access through ostentation to the things that it relates, so that we are left with a theoretical description of the cause of what we can measure as the only indication of what we are talking about. Feyerabend’s restriction of incommensurability to universal theories is motivated by his investigation (and solution) of this particular problem, as only universal theories have such all-­ pervasive ontological implications that determine the very nature of thinghood (what kinds of things there are) itself. Lastly, Feyerabend differs from Kuhn in that he uses incommensurability to explain his views on the dynamics of the relation between everyday language of common sense and current scientific theories. According to Feyerabend, incommensurability affects the relationship between commonly understood language and current scientific theories, because languages frequently contain principles that are not explicitly formulated, but rather implicit in the way in which terms are used, which are remnants of outdated principles on which former universal theories were once based. The idea, in Kuhnian terms, is that current scientific theories use lexical taxonomies that are incompatible with the lexical taxonomies of the older incommensurable theories and the empirical consequences used to test them as expressed in everyday language. This leads Feyerabend to conclude that in the course of scientific advance, the language of the older theories must therefore be abandoned and replaced by the language of the new and better theories, even in the most common everyday situations, as otherwise, because they must be kept logically separated from each other to make sense, we would end up with a situation in which we have relativism with respect to different truths within the same community (common sense truth versus scientific truth), and presumably that would not contribute to the advancement of science, and would certainly seem to hinder its dissemination through society.

106

E. Oberheim

7.5 What Does Incommensurability Imply About Theory Comparison? Incommensurability implies that there is no shared observation language or neutral set of facts for comparing incommensurable theories with respect to the results of the measurements used to test them empirically, because the lexical taxonomies used to state these rival theories break the no-overlap principle with respect to each other by cross-classifying some of the same things (or relations, or properties) into incompatible sets of mutually exclusive kinds (of things, relations or properties), which must be kept separated on pain of incoherence. Incommensurability does not imply that incommensurable theories cannot be compared empirically. They can be tested on the basis of the results of the measurements that they each predict and compared with respect to how well they pass those tests. However, the normal procedure for testing a single theory on the basis of the results of measurements is insufficient for comparing rival incommensurable theories in crisis phases and across the resulting revolutionary divide. In normal science, a single theory is tested empirically on the basis of the results of measurements by way of deducing, from the theory together with auxiliary assumptions and initial conditions, statements that can be treated as predictions, and then by checking if these empirical consequences correspond to the results of measurements. This method allows us to show how facts, as stated by observation statements, can be used either to refute or to corroborate a theory. When two incommensurable theories are compared empirically, they do not compete against each other on the basis of a single shared observation language or a neutral set of facts. They compete on the basis of two conceptually incompatible observation languages that state two ontologically incompatible sets of facts. This is the case even if the actual sentences that result from the measurements designed to test the respective theories are identical, and regardless as to whether they appear to directly contradict each other or not. Each theory still provides its own independent interpretation of the results of some of the same measurement procedures that can be used to corroborate both of them, but because the lexical taxonomies used to state the competing theories break the nooverlap principle by cross-classifying some of the same objects into incompatible sets of mutually exclusive kinds, the conceptual incompatibility generated cannot be bridged by direct logical implication. Because they are conceptually incompatible, each theory must be tested independently against its own empirically measurable consequences, and only then can the two theories be compared empirically.

7.6 What Do Kuhn’s Views on Incommensurability Imply About Theory Comparison? Kuhn made many claims about incommensurability and theory comparison that were not always very clear and that sparked much controversy. For example, Kuhn claimed that incommensurable theories have “no language into which at least the

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

107

empirical consequences of both can be translated without loss or change” (1970, 266; Structure Postscript– 1969, 201). After his misadventure in the murky swamp of the meaning of meaning, Kuhn spent a lot of time discussing the impossibility of translating incommensurable theories, claiming that “if two theories are incommensurable, they must be stated in mutually untranslatable languages.” (1983, 669–670). These discussions, while often insightful and provocative, turn out not to be directly relevant to theory comparison, which does not rely on translation. Kuhn also tried to clarify his views on the implications incommensurability has for theory comparison with the cryptic claim that there can be no “point by point” comparison between incommensurable theories (1970, 266). By this Kuhn, seems merely to be suggesting that for every empirical prediction of one theory, there is not always a corresponding empirical prediction by its incommensurable rival. But this could also be interpreted in the stronger sense, so that perhaps Kuhn meant that none of the empirical consequences predicted by one theory will be the same as any of those of the other. Kuhn claimed that “As in political revolutions, so in paradigm choice– there is no standard higher than the assent of the relevant community […] paradigm choice can never be unequivocally settled by logic and experiment alone” (Structure, 94). Testing theories empirically is holistic and therefore never absolutely definitive. When empirical measurements are used to test theories, it is not simply a question of testing theory against the facts. The whole unit is tested simultaneously together. The theories, the auxiliary assumptions used to deduce facts  from them, and the measurable facts that the theories imply are all tested together, so that an apparent refutation of the theory could actually turn out to be a corroboration if we correct a mistake in the auxiliary assumptions, or concerning assumptions about the facts used in the test. Kuhn, like Einstein, also emphasized that the decisions about theory choice are not always made on the basis of how well the theories correspond to the observable facts. Different individuals will make different decisions for different reasons at different stages. Sometimes theory comparison is based on epistemic values, such as simplicity, accuracy, consistency, scope and fruitfulness, which do not function as rules that determine rational theory choice, but as values that merely guide it. According to Kuhn, “When paradigms enter, as they must, into a debate about paradigm choice, their role is necessarily circular. Each group uses its own paradigm to argue in that paradigm’s defense […] whatever its force, the status of the circular argument is only that of persuasion. It cannot be made logically or even probabilistically compelling for those who refuse to step into the circle. The premises and values shared by the two parties to a debate over paradigms are not sufficiently extensive for that” (Kuhn (1962/1996, 94).

Typically, commentators focus on Kuhn’s use of ‘persuasion’, in support of the erroneous view that Kuhn’s account of scientific revolutions makes scientific advance irrational, when in fact it merely exposes the boundaries of logical implication set by enforcing the no-overlap principle in his attempt at a realistic reconstruction of scientific revolutions as socio-historical processes. Kuhn distinguishes the

108

E. Oberheim

transition to bilingualism based on ‘persuasion’ from the usual processes of normal science to highlight how and where these boundaries are set by paradigms. He does not mean to imply that science is irrational, but to show how scientific rationality is more complex in revolutions than in normal science. On the reading I’m suggesting, Kuhn’s claims about ‘persuasion’ are not controversial. They turn out just to mean to convince someone to ‘become bilingual’. Incommensurability, for Kuhn, implies that a single paradigm and its logical implications are not sufficient for deciding between two paradigms, each with its own logical implications, including their own interpretations of the results of measurements. If there is only one theoretical language available and there are no crisis causing anomalies (normal science), there is no need to become bilingual. In crisis phases, however, use of exactly these same measurement procedures to defend the reigning paradigm will appear circular or question-begging to members of a competing paradigm that interprets them differently, because they seem to assume the lexical structure that they use to state their theories and interpret the results of measurements, but it is exactly the lexical structures of the competing theories that are incompatible, and between which one must choose, as candidates for the basis of a realistic interpretation (i.e. the one true theory). If Kuhn is correct, it would follow that we should expect to find evidence of scientists claiming or implying that the arguments being used are circular from the perspective of their competing paradigm in the history of science during cyclical crisis phases, and in Structure Kuhn supports his theory of scientific advance with just such historical evidence by showing that during cyclical crisis phases, scientists do act just like philosophers by disagreeing about fundamentals, resulting in accusations of question-begging and the use of circular arguments by their opponents. To see the implications incommensurability has for theory comparison in more detail requires briefly unpacking Kuhn’s metaphorical and somewhat mysterious sounding talk about refusing to ‘step into the circle’, which involves ‘persuasion’ in contrast to ‘proof’. By stepping into the circle, Kuhn simply means becoming bilingual. Only after scientists have been persuaded to learn the new theory, can they consider which of the theories fairs better with respect to the results of measurements. Becoming bilingual is not always easy for scientists, and not just because a strange new theory has to be learned. The established theory seems to gets in the way, and stepping out of the circle comes with its own set of challenges. The deeper the grip of a theory, the harder it may be to set it aside, especially when its eventual successor may be rather new, will initially seem incoherent, and has not yet had sufficient time to develop its potential. Once scientists are bilingual, they can switch back and forth between theories as long as they are careful to keep them separate. For example, they can use Newtonian theory (instrumentally) to solve a practical problem (like how much thrust you need to get to the moon in some rocket), and then use special relativity to solve that problem too. They must, however, be careful to compare each theory with its empirical consequences as whole units to each other, when assessing how well each theory handles the problem. Moreover, only one of them can be true. If Einstein’s theory is true, then Newton’s theory cannot be. According to Kuhn, this is a more accurate

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

109

picture of how scientists actually decide between conceptually incompatible theories on the basis of empirical measurements, and a more accurate reflection of how science actually develops through revolutions with respect to how empirical measurements are used to compare theories.

7.7 What Do Feyerabend’s Views About Incommensurability Imply About Theory Comparison? The short answer is: ‘Nothing special’. The specific challenges posed to scientists when comparing incommensurable theories are routinely handled by scientists comparing theories and their empirical consequences as separate units. So for scientists incommensurability is not really a problem, just as long as they treat the two theories each with their own implications as separate units (are bilingual) and agreement can be reached on what counts as suitable measurements of what, which of course, may not always be the case. Feyerabend thought that incommensurable rivals, in crisis, may sometimes initially involve a state of transient underdetermination that can only be overcome by what he called an ‘indirect refutation’. An indirect refutation is simply a corroboration of just one of two mutually exclusive incommensurable theories. Feyerabend realized that scientists using incommensurable theories may be using exactly the same operational procedures to corroborate very different theories without, at first, being able to distinguish which is better empirically, as was initially the case in the scientific revolutions he considered. In light of Duhem’s (1954) and Popper’s Logik, he examined the methodological situation very closely, and his investigations led him to the conclusion that two incommensurable theories (impetus theory and Newtonian theory) actually used the same operational procedures (dropping balls into wax from various heights) to measure two different kinds of things: impetus according to Aristotelian theory and mass according to Newtonian theory. What is surprising is that both theories are corroborated by predictions that result from exactly the same measurement procedures as reported in exactly the same observation sentences, just like plotting planetary positions to measure celestial motion was used to corroborate both Ptolemaic and Copernican theories. The trick is to understand how the transient underdetermination is resolved by a crucial experiment that does not rely on a shared observation language that states neutral facts, nor any logical consequences, observable or otherwise, shared by both theories. For now it is enough to emphasize that for Feyerabend, like for Kuhn, it is not that there cannot be crucial experiments between incommensurable theories, they are just more complex than the usual case, when a single theory is tested against its own empirical consequences, leading to either refutation or corroboration.

110

E. Oberheim

7.8 What Do Howard Sankey’s Views on Incommensurability Imply About Theory Comparison? Sankey attempts to reconstruct Kuhn and Feyerabend’s views on incommensurability from a scientific realist perspective, and thereby to undermine the challenges that Kuhn and Feyerabend used it to pose for scientific realism. According to Sankey, Kuhn’s denial of the possibility of ‘point-by-point comparison’ means that “it is not possible to directly compare the contents of the theories with respect to specific points of agreement and disagreement” (2018, 86). If this were true, it would pose serious challenges to scientific realist conceptions of scientific progress (according to Sankey). However, as we have seen, perhaps Kuhn only meant merely that it is not possible to directly compare the contents of the theories with respect to all specific points of agreement and disagreement. In Feyerabend’s case, according to Sankey, there is less wiggle room. For Feyerabend, “incommensurable theories may not possess any comparable consequences, observational or otherwise” (ERE, 94), and for Feyerabend, as Sankey emphasizes, incommensurability thus precludes “content comparison” and means “deductive disjointedness, and nothing else” (1977, 365). Therefore, according to Sankey, at least it is clear that Feyerabend denied that incommensurable theories can be compared empirically, that is according to their empirical consequences. However, Feyerabend, actually repeatedly explicitly stressed that there can be crucial experiments between incommensurable theories in his ERE, where he first used the term “incommensurable” and where he gives a detailed exposition of the logical structure of how, through indirect refutation, crucial experiments are still possible between incommensurable rivals. So it seems that Sankey has not properly captured Kuhn and Feyerabend’s views on incommensurability and theory comparison, views which Kuhn and Feyerabend struggled for decades to articulate with any number of seemingly cryptic remarks, and which were at the core of their respective philosophies. Because of co-reference and extensional overlap, the two theories are able to enter into logical relations with respect to content. More specifically, claims made by one theory may agree or disagree with claims made by the other theory. As a result, the theories may be compared with respect to what they say about the world. Where the theories disagree with respect to a specific state of affairs the disagreement may be adjudicated by empirical means […] if the theories disagree with respect to an empirical prediction, it may be possible to conduct a crucial test to decide the matter (Sankey 2018, 85).

So, for Sankey, Kuhn and Feyerabend are wrong about incommensurability and theory comparison. As long as incommensurable theories share some specific points of potential empirical disagreement, those points can be used to stage crucial experiments between incommensurable theories on the basis of neutral facts stated in a common observation language. Kuhn and Feyerabend were wrong about theory comparison, according to Sankey, because they assumed an erroneous descriptive theory of reference

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

111

according to which the set of objects to which a term applies is determined by the meaning a theory provides it. If we replace Kuhn and Feyerabend’s shared erroneous assumption (a descriptive theory of reference) with a causal-descriptivist theory of reference, according to which the objects to which a term applies is determined by causal relations between those objects and the terms that refer to them, supplemented either with a description of the kind to which the term applies (in cases of ostentation to observables), or supplemented with descriptions of the causal mechanism that specify the causal relation between the object and some observable phenomena caused by it (in the case of theoretical entities), then it becomes clear (according to Sankey) that as long as two incommensurable theories refer to some of the same objects, content comparison (on the basis of their empirical predictions) remains possible, and so “it is entirely possible for science to progress in the sense that later theories yield an increase in knowledge about the same entities that earlier theories referred to”, and so concludes: “Given that the content of untranslatable theories may be compared, incommensurability so construed constitutes no impediment to rational choice between theories” so that “what remains of the claim has been so weakened that even if there are actual cases of incommensurability in the history of science they are of little or no interest […] one wonders what all the fuss was about” (Sankey 2018, 87).

7.9 What Does Incommensurability Imply About Theory Comparison? A Response to Sankey Sankey’s “So what?” response to incommensurability is based on showing that incommensurable theories can be used to predict statements that refer to some of the same objects about which they are both able to make empirical claims, which when you think about it, neither Kuhn nor Feyerabend actually ever denied. After all, they were talking about rival theories. In any case, I will now argue that Sankey has not actually shown that the “claims made by one theory may agree or disagree with claims made by the other theory”, so that “it is entirely possible for science to progress in the sense that later theories yield an increase in knowledge about the same entities that earlier theories referred to”. I will do this by emphasizing a distinction that has been glossed over by Sankey, who equivocates exactly on this point. The distinction is between reference to particular objects by observation statements and reference to kinds and their relations by theories themselves. My claim is that Sankey has only shown that some observation statements used to test incommensurable theories refer to some of the same objects (which was never actually in dispute). He has not shown that theories separated by a revolution refer to the same things. Theories, in contrast to observation statements, use general terms to refer to kinds of things whose relations they state. For example, according to the Ptolemaic theory: “All planets travel around Earth”. In this theory, the kind term “planets” is

112

E. Oberheim

used to refer to the kind ‘planet’ (including all of its members). The kind term “Earth” is used to refer to the kind ‘Earth’, which has but one member, the Earth, and so is easily confused as a particular. And the theory states how the former is related to the latter. All planets travel around Earth. By contrast, the observation statements used to test this general theory empirically refer to particular planets. For example, “That is Mars there” refers to Mars. I will call the standard account of testing theories the ‘hypothetico-deductive method’ of testing theories, and I will try briefly to explain the basic idea. In order to be used to test a theory (or any general statement) on the basis of an observation statement, that observation statement must be a deductive consequence of that theory (or general statement). A valid deductive logical inference is the only reliable way to connect general statements (including laws and theories) to particular statements (including results of measurement) that are used to test them. In order to connect general statements, like theories, to observation statements, it must be shown that the observation statements are deductive consequences of it. Additional (mixed) assumptions are always needed just because theories state relations between general kinds that need to be instantiated by particulars in order for them to have empirical consequences that can be deduced from them and then used to test them. If the observation statement matches the results of measurement, we have used the observation statement, which refers to particulars, to corroborate the general theory, which refers to kinds and states their relations, by way of showing that given (mixed) additional assumptions, the particular observation statement is a logical consequence of the general theory. This hypothetico-deductive method of testing theories explains how a theory is tested against measurements in normal science. To understand the implication that incommensurability has for theory comparison, we will now have to examine how the hypothetico-deductive method of testing theories works when comparing competing incommensurable theories empirically, that is with respect to the results of measurements. In such cases, two incommensurable theories are stated on the basis of incompatible lexical taxonomies as they cross-classify some of the same objects into mutually exclusive sets of mutually exclusive kinds, and yet they can still both be tested according to the same measurement procedures by the same observation sentences. The problem is not that the observation statements used to test the two theories do not refer to some of the same particulars. The problem is that exactly the same observation sentences can be used to test two conceptually incompatible rival theories on the basis of two conceptually incompatible observation statements that state two different sets of facts as evidenced by their two sets of incompatible implications. In other words, one and the same observation sentence can actually be used to make two incompatible factual statements. So for example, for two incommensurable theories such as ‘All planets travel around the Earth’ and ‘All planets travel around the sun’ to be tested against the results of measurements, observation statements must be deduced from each theory. The two statements deduced ‘That is Mars’ are made with exactly the same sentence “That is Mars”, which appears to be making one and the same statement. But an observation sentence “That is Mars” must be interpreted by a theory in order for it to be used to test that theory. So “That is Mars” can actually be used to make

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

113

two different observation statements. One statement “That is Mars” refers to the Ptolemaic planet Mars and implies the fact “That is Mars traveling around the Earth” and the other observation statement “That is Mars” refers to the Copernican Planet Mars and implies the fact “That is Mars traveling around the Sun”. On this account of the logic of crucial experiments during scientific revolutions, a crucial experiment between incommensurable alternatives does not need to rely on finding a single observation statement that corroborates one theory while simultaneously falsifying the other, as there are no reliable logical relations between the kinds used to state the two theories for making this possible. But as both Kuhn and Feyerabend emphasized, this is not the only possibility for comparing theories empirically. The newer theory is preferable because it solves a problem (by successfully predicting a correspondence between theory and observation), not because the established view is falsified by that same observation statement or any other of its own empirical consequences. The new theory together with the measurements that support it indirectly refute the older theory together with the measurements that were used to support it because the new theory is corroborated by an observation that the established view does not explain. This is not to say that perhaps, combined with the right (potentially as yet unknown auxiliary assumptions) it could explain those phenomena in its own terms, and that is exactly what defenders of the established view tried (unsuccessfuly) to do. But the newer theory is corroborated. The established theory is no, but it need not be falsified either. Adding additional assumptions to handle these specific phenomena may  seem suspiciously like ad hoc maneuvers, even if they could later be made to seem independently plausible. In this way, theories don’t always simply compete against the evidence (like in normal science). Sometimes they compete against each other as incommensurable rivals. The best corroborated theory wins.  This reconstruction of the logical structure of the test procedures used to compare incommensurable theories on the basis of the results of measurement fits both Kuhn and Feyerabend’s favorite examples of how transient underdetermination can eventually be overcome. In the Copernican revolution, the underdetermination was not overcome by more exact measurements of the planetary positions, so that discrepancies between the predictions of the two theories allowed for the same measurements to refute one theory while simultaneously corroborating the other. It ended with new types of observations: the phases of Venus, which eventually clearly tipped the balance in favor of Copernicanism empirically. Copernicanism offered a natural explanation of them (i.e they were easily deduced from, and so corroborated by, them). One theory was corroborated, while the other theory was not. Similarly, in Feyerabend’s case, the underdetermination was not overcome by a direct refutation of the older theory on the basis of measurements stated in a shared observation language that states neutral facts.  When Einstein deduced Brownian motion, he showed how it corroborated  one  theory by being a deductive consequence of it, while its rival offered no explanation for this otherwise  seemingly strange phenomena.

114

E. Oberheim

7.10 What Does Incommensurability Imply About Scientific Realism? Incommensurability implies that when there is kind change that breaks the no-­ overlap principle during scientific revolutions, there is also world change; or more specifically, that world change accompanies kind change during the process of reconceptualization in scientific revolutions. World change happens when established facts are reconceptualized into new facts, which corroborate the new theory. By reclassifying the same objects into a mutually exclusive set of mutually exclusive kinds, the new theory describes and explains new facts, and the old facts are no longer real facts from the perspective of the new theory, which has replaced those facts with new facts that can be deduced from it and thereby be used to corroborate it. Incommensurability implies that kind change that breaks the no-overlap principle accompanies world change (an ontological claim) given the additional assumption that theories have ontological consequences. Theories have ontological consequences because they imply the kinds of things that we actually do take to be real. The facts are just the facts according to the theory that explains them, which is especially clear when we actually do use observation statements to test a theory empirically. Feyerabend used ‘incommensurable’ to mean conceptual incompatibility due to meaning change in theoretical transitions that affect ontological beliefs. He set out his position right from the start exactly in order to identify the assumptions he was going to make before beginning his otherwise largely immanent criticism of Nagel. I have argued elsewhere that Feyerabend’s position boils down to Kuhn’s ideas about the same issues delineated in different terms. Both Kuhn and Feyerabend set out incommensurability on the basis of what I now am calling Kant-on-wheels realism. Feyerabend never talks much about ‘world change’, obviously attempting to avoid any overtly controversial ‘metaphysical’ assumptions. But he does talk about ‘ontological replacement’, which features prominently as the ontological component of his exposition of why incommensurability undermines Nagel’s account of how new theories explain older theories in the course of scientific progress. Feyerabend’s idea is that theories organize experience into sets of facts that can be used to corroborate (or refute) them, and which they can thereby explain (if they are not refuted). For Feyerabend, theories are also clearly mutually exclusive ontologically, because the two incommensurable theories imply incompatible sets of facts, even when those facts are stated according to identical observation sentences that match the results of exactly the same measurement procedures. The new theory has ontological consequences that are incompatible with the ontological consequences of the old theory exactly because it implies that different kinds of things (and their relations) are real, while explaining different facts about them. The facts explained by the new theory are not the same facts as explained by the older theory

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

115

just because a new incommensurable theory results in a new incompatible set of facts, which are the facts as stated and explained by the new theory. For Kuhn, the idea of world change is part of, or more precisely an ‘aspect’ of, incommensurability, or as he admittedly struggled to put it: “(…) the proponents of competing paradigms practice their trades in different worlds” (Structure, 150). This mysterious sounding idea (world change) was not something particularly new. At the time, world change seems to have been as widely recognized as it was poorly understood. In order to unpack the idea of world change as an aspect of incommensurability, Kuhn (following Einstein) also developed what I am calling ‘Kant-on-wheels’ realism. Kuhn claimed “I go around explaining my own position saying I am a Kantian with moveable categories” (2000, 265). According to such a position, scientists do not discover objective truths about a mind-independent reality. They invent speculative theories that may turn out to contribute to how we experience and understand the world. The basic idea is to distinguish theories from empirical reality, where empirical reality is a ‘phenomenal world’ that is ‘co-constituted’ by ‘subject-sided’ and ‘object-sided moments’. This accounts for the seemingly strange ontological aspect of incommensurability by allowing us to say that what changed about the facts during reconceptualisation involving reclassification that broke the no-overlap principle was the subject-sided moments (or contributions) to empirical reality. For our purposes it is enough to emphasize that these phenomenal worlds are the empirical realities that we actually do experience. They are the set of facts as stated in the observation statements that are used to test theories, which are in fact both conceptually and ontologically incompatible, as they depend on which theory we use to interpret them. In this way, Kuhn’s ‘Kant-on-wheels’ realism attempts to explain how scientific revolutions create new empirical realities through scientific revolutions. The idea that theories have ontological consequences is wrong from a scientific realist perspective because, as just mentioned, the facts do not change just because of, or together with, changes in the beliefs of scientists about them. Objective facts are independent from the subjective theories that they describe and explain. What I want to emphasize here, however, is that there would be nothing really mysterious with world change if what counts as the facts were part of an empirical reality that is co-constituted by subject-sided and object-sided moments, and so are not purely objective. But this clashes with scientific realism, according to which current theories of mature natural sciences are at least approximately true, where truth is purportedly correspondence to an objective reality that is fixed independently of our theories. True theories explain the unchanging facts of reality by corresponding to real kinds of things and their relations, as they exist independently of our theorizing about them. So scientific realism cannot really make sense of world change. According to scientific realism, old facts are not replaced by new facts, as facts don’t change, only beliefs about facts change. If we are to make sense of scientific revolutions involving incommensurability when kind change accompanies world change, we need a different view of science, one in which the facts of empirical

116

E. Oberheim

reality are not completely independent of the concepts used to interpret them. Only then we can say that in the course of scientific revolutions, the facts are reinterpreted into different facts according to the lexical taxonomy of the new theory, resulting in world change accompanying kind change, when one theory replaces another as the basis for a realistic interpretation of experience.

7.11 What Does Incommensurability Imply About Scientific Progress? Incommensurability implies that progress in the natural sciences is not cumulative. While progress in normal science (or normal scientific progress) is cumulative, progress through revolutions is not simply the accumulation of the explanation of more facts. It is not cumulative with respect to the theories used to explain the facts either. Progress through revolutions is not made by simply adding new improved theories to the set of those already accepted as true. According to the newer theory, the older theory (once thought to be true) turned out to be wrong and needed to be replaced for good empirical reasons, regardless of the fact that the two theories shared no common observation language with which to stage a crucial experiment between them in which one and the same fact simultaneously directly refutes one theory, while corroborating the other. Scientific progress is sometimes a process in which newer theories do not explain older views. They replace them, and this holds for both the theories and the facts that are explained by them. Kuhn and Feyerabend agreed that scientific progress is not cumulative, because scientific revolutions result in a complete replacement of the ontology of the older theory. During reconceptualisation in the course of scientific revolutions what was taken to be real before the revolution was replaced by what is taken to be real after the revolution. The process is revolutionary, not retentive, both conceptually and ontologically, that is, with respect to both theories and the facts they purport to describe and explain. Incommensurability clashes with scientific realist conceptions of progress as progress toward truth in four interrelated respects. First, progress is not towards something that is fixed in advance. Second, subsequent revolutions in the same field (Aristotle → Newton → Einstein) in physics suggest that there is no clear fixed direction that progress is making in terms of improvements to, or convergence toward some shared underlying ontology. If truth is correspondence and a better approximation to truth implies an improved correspondence to the same kinds of things, then not only can we not yet say what the truth is, we can’t predict which way we should be going. Third, given the possibility of unconceived incommensurable alternatives, nor can we ever really be sure that we have found truth if we were actually ever  able to find it. So by better ‘approximating truth’, if we can’t even specify the direction we should go or even if we have arrived once we are there, it turns out we’re not saying very much. Fourth, there just is no extra-paradigmatic

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

117

(objective) truth towards which progress could be directed. A claim can only be said to be true (and tested to show its empirical superiority) as stated in its own language on the basis of its own deductive consequences. So saying that scientific progress is progress toward truth turns out not to be saying anything at all. For both Kuhn and Feyerabend, scientific progress is open-ended, as it is always open to revision in light of new empirical evidence, which it may only be possible to discover on the basis of a new revolutionary theory that has as yet not even been proposed. Instead of cumulative progress or progress towards truth, in scientific revolutions, scientific progress involves reconceptualization of well-established facts into new kinds of facts, through the reclassifications of objects into mutually exclusive sets of mutually exclusive kinds, so that when the older theories are replaced by newer incommensurable theories, the old facts are replaced by the new facts. It is revisionary in this specific conceptual and ontological sense. The difference between Kuhn and Feyerabend’s accounts of scientific progress mirrors the differences with respect to their views on what it means to understand a theory. For Kuhn, as we have seen, to understand a theory is to understand how to use it to solve exemplary problems within a paradigm. For Feyerabend (following Duhem) to really understand a theory includes something more than just knowing how to use it to solve a range of heuristically fruitful problems. It also includes understanding what other incommensurable alternative theories have been ruled out by it, including their ontology, i.e., the facts that were used to corroborate those alternatives according to the solutions they provided to the problems that they confronted as stated by them. So for Feyerabend, scientists should be historically informed so that they better understand the theories that they use. In sum, because scientific revolutions result in the reconceptualization of what were considered to be established facts into new, incompatible kinds of facts, it does not really make sense to say that scientific progress is an improved approximation to one and the same set of facts. There is no common measure in the form of a shared observation language because there are no shared, neutralfacts as stated by a shared, neutral observation language.

7.12 What Is the Justification of Incommensurability? The justification of incommensurability is an inference to the best explanation (IBE). The facts explained (the premises of the argument) are facts about science and its history, such as (1) the experience of the historian of science trying to break into archaic worldviews so as to understand science and its history, and the regular recurring historical phenomena like talking at cross-purposes in crisis phases and across revolutions this reveals, (2) conceptual change that introduces new kinds that break the no-overlap principle (with respect to the kinds structured by previously established lexical taxonomies) accompanying world change (new realistic interpretations of experience), (3) the resulting ineffability of new concepts on the bases of older theories and the ineffability of older concepts on the bases of newer

118

E. Oberheim

theories, (4) the limits to logical implication with respect to trans-paradigmatic relations between theories that break the no-overlap principle by cross-classifying some of the same things into mutually exclusive sets of mutually exclusive kinds, (5) mutual accusations of the use of circular and question-begging arguments with each paradigm using its own paradigm to support itself, (6) the need for bilingualism and persuasion as prelude to empirical proof (based on deductions, experiments and measurements). The best explanation of these facts (the conclusion of the argument) is that scientific theories that are stated according to lexical taxonomies that break the no-overlap principle with respect to each other are incommensurable: they have no common measure in the form of a shared observation language or neutral set of facts, because of world change that accompanies kind change in the course of revolutionary advance in science. Both Kuhn and Feyerabend emphasized the point that incommensurability was not a thesis arrived at through philosophical analysis of science (inferred from an analysis of philosophical theories of meaning), but rather a fact about science that becomes apparent when you actually look at it, which has implications for scientific progress and scientific realism. Feyerabend emphasized this attitude that incommensurability is simply a fact about science that he was trying to explain, insisting that his “discussion of the relation between impetus and momentum is not an attempt to draw consequences from a contextual theory of meaning […] it simply shows that both facts and the laws of Newtonian mechanics prevent us from using the concept of impetus as part of Newton’s theory of motion” (1981a, x).

7.13 What Is Metaincommensurability? Metaincommensurability is incommensurability on the meta-level, which in this case means between theories about scientific theories, such as scientific realism and Kant-on-wheels realism. They have no common measure in the form of a shared observation language or neutral set of facts against which they can both be compared. This is (purportedly) a fact about how theories of science develop in light of facts about science and theories of science, or to be more specific, in light of particular facts about the reconceptualisation of scientific revolutions in philosophy of science.

7.14 What Causes Metaincommensurability? Just as incommensurability is caused by reclassification that breaks the no-overlap principle with respect to the reigning scientific theory, metaincommensurability is caused by reclassification that breaks the no-overlap principle with respect to the lexical taxonomy used to state the reigning theories about scientific theories. For example, Kant-on-wheels realism is metaincommensurable with scientific realism

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

119

because historical philosophers of science proposed a theory of science, i.e. ‘Kant-­ on-­Wheels realism’, that re-classifies some of the same things (‘theories’, ‘facts’, ‘truth’ and ‘reality’ about which it states relations) on the basis of a lexical taxonomy that is incompatible with the lexical taxonomy used to state scientific realism. Consider ‘theories’, ‘truth’ and ‘facts’ and the realist dichotomy between the different kinds: subjective theories and objective facts. (So if something is a subjective theory it is not an objective fact, and truth is a correspondence between theories and facts). According to scientific realism, true theories correspond to the objective facts of mind-independent reality, while according to Kant-on-wheels realism, true theories correspond to facts of empirical reality that are co-constituted by subject-sided and object-sided moments inseparably bound. For the scientific realist, the facts are objective and mind-independent. They do not contain genetically subject-sided moments. By contrast, according to Kant-on-wheels realism, there are no purely objective facts in the sense of extra-paradigmatic (or mind-independent) facts, as facts themselves can only be established on the basis of some lexical taxonomy that obeys the no-overlap principle in order for them to be explainable by a theory. Use of any particular lexical taxonomy introduces a subject-sided component into the facts so that there are no purely objective (object-sided) facts in the scientific realist sense. And as facts are facts about empirical reality that contains genetically subject-­ sided contributions, from the perspective of Kantian-on-wheels realism, just as there is no phlogiston, no absolute mass, and no Ptolemaic planets, as it turns out, there are no objective facts about mind-independent reality. There are just the facts of empirical reality as organized by the lexical taxonomies of the theories used to interpret them.

7.15 What Are the Consequences of Metaincommensurability? Metaincommensurability causes mutual ineffability. With just the conceptual resources of either theory about scientific theories, the other point of view cannot be coherently formulated. A direct consequence of introducing a rival theory of science whose lexical structure breaks the no-overlap principle with respect to the reigning theory of science is that the new theory of science is literary unstateable on the basis of the taxonomic lexical structure imposed by the older theory of science, just as the reigning theory of science is literally unstateable on the basis of the taxonomic lexical structure imposed by the new theory of science. This mutual ineffability has direct practical implications for theorizing about science. Theories about scientific theories must be learned in their own terms, on the basis of their own lexical taxonomies together with the consequences (empirical or otherwise) that can be deduced from them. If they are not, then facts like incommensurability, world change, and ineffability may seem incoherent, and discussions of them lead to all kinds of talking at cross purposes. Like proponents of

120

E. Oberheim

incommensurable scientific theories, proponents of incommensurable theories of science, such as scientific realism and Kant-on-wheels realism, can only hope to assess the relative merits of their competing theories of science after becoming bilingual, as each theory of science must be understood in its own terms before it can be tested on the basis of facts about science and its history, and only then can it be compared to its rivals. Sankey’s decades long project to develop incommensurability on the basis of scientific realist assumptions struggles to delineate a coherent conception of Kuhn and Feyerabend’s conception of incommensurability exactly because the lexical taxonomy used to state scientific realism breaks the no-overlap principle with respect to the type of Kant-on-wheels realism that Kuhn and Feyerabend used to develop incommensurability as a fact about science and its history. For this reason, it cannot be coherently formulated on the basis of the lexical taxonomy of scientific realism, according to which the things we make factual claims about do not change when we replace the kinds terms we use to make those claims. Just as incommensurability has implications for scientists comparing scientific theories, metaincommensurability has implications for philosophers of science comparing theories about science. For example, metaincommensurability implies that we should expect to find the same kind of talking at cross-purposes when comparing scientific realism and Kant-on-wheels realism that we find when comparing incommensurable theories in crisis phases in science. We should also find the same kind of recurrent accusations of question-begging and use of circular arguments. The strategy of the miracle argument for scientific realism (basically, the success of science would be a miracle if the theories were not true) seems to assume what needs to be shown from the perspective of its rivals, i.e. that empirical success of a theory warrants belief in its truth. The pessimistic meta-induction against scientific realism (basically, as older theories once thought to be true always turned out to be false up until the current ones, it seems likely that current theories thought to be true will turn out to be false) seems equally question-begging from the scientific realist perspective, because it simply assumes that older theories were false, whereas according to scientific realism, they were at least approximately true. In any case, there is abundant evidence of rampant accusations of question-begging and use of circular arguments available in the literature on the realism debate. It is important to see that Kuhn seems to have been quite aware of metaincommensurability, and some of its consequences, having hinted as much on numerous occasions: “This collection of essays [Criticism and the Growth of Knowledge] therefore provides an extended example of what I have elsewhere called partial or incomplete communication, the talking-through- each-­ other that regularly characterizes discourse between participants in incommensurable points of view.” (1970, 231–232). Here Kuhn is explicitly suggesting that philosophers of science discussing incommensurability experience the same kind of communication difficulties that scientists confront when they are confronted with incommensurability in science. He is also suggesting that he is personally acquainted with them.

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

121

What I am suggesting is that metaincommensurability implies that there is no common measure in the form of a shared observation language or neutral set of facts for comparing theories about science, such as scientific realism and Kant-on-wheels realism, because the lexical taxonomies used to state the two rival theories about scientific theories break the no-overlap principle with respect to each other, so that the same kinds of problems with theory comparison that arise in science also arise in philosophy of science, and the important implications incommensurability has for theory comparison are the same in philosophy of science as they are in science, just on the meta-level. There is the same need for bilingualism, without which theories and meta-theories (theories about theories) cannot be compared with respect to how well they describe and explain their subjects on the basis of what they imply. Moreover, although there are limits to the logical implications with respect to metaincommensurable theories about theories, so that there is no shared observation language or neutral set of facts that can be used to stage a crucial experiment between them, there can still be ‘crucial experiments’ between theories about theories that are metaincommensurable, just as there can still be crucial experiments between incommensurable scientific theories on the basis of indirect refutation, by corroborating a new theory that is incommensurable with the reigning theory. In fact, metaincommensurability implies that incommensurability itself indirectly refutes scientific realism. Incommensurability is a fact about science that can only be coherently stated on the basis of a lexical taxonomy that breaks the no-­ overlap principle with respect to the lexical taxonomy used to state scientific realism. So is world change. Sankey’s  scientific realist reconstruction of incommensurability is doomed from the start, because it tries to use a lexical taxonomy to develop a fact that can only be coherently formulated on the basis of an incompatible lexical taxonomy. But once those facts and their relations are coherently stated, they clearly tip the balance in favor of the newer theory, which offers a natural explanation of them, while explaining why the older theory cannot. In this way, scientific realism is indirectly refuted by incommensurability and  Kant-on-­ wheels realism Incommensurability can only be used as evidence against scientific realism if it is interpreted according to assumptions that are conceptually incompatible with scientific realism. It cannot be coherently formulated or explained on the basis of a shared observation language (shared with respect to both scientific and Kant-on-­ wheels realism) that states a neutral set of facts. But once it is coherently interpreted by Kant-on-wheels realism, it corroborates that theory, without actually directly refuting scientific realism, which in light of these facts, seems badly in need of some fundamental revision, and so should be replaced. Universal theories are indirectly refuted when a new fact corroborates an incommensurable proposal that is conceptually and ontologically incompatible with the established theory. If scientific realism is a universal theory, and testing theories against empirical evidence requires a deductive relationship between theory and evidence, then scientific realism can only be refuted indirectly, by new facts as established on the basis of an incommensurable rival, such as incommensurability as established on the basis of Kant-on-wheels realism.

122

E. Oberheim

Metaincommensurability has implications for progress with respect to theories of science and scientific progress. If metaincommensurability is correct, then just as there is no phlogiston, no absolute mass, and no Ptolemaic planets, perhaps the philosophical ideas that theories correspond to objective facts and progress is progress toward truth (either as better approximations to it, or by convergence toward it), like these older scientific ideas, should simply be relinquished as part of a realistic interpretation of experience in the course of progress in philosophy of science. I am not suggesting that metaincommensurability implies that incommensurability shows that scientific realism is definitely wrong and so should be banished or forgotten, or that theory choice on the meta-level is, or must be, irrational. After all, Kant-on-wheels realism may come with a whole host of difficulties of its own. But it does show that, and why, we need to learn how, at least temporarily, to suspend scientific realist assumptions, if we want to compare scientific realism to its rivals according to what can be deduced from them. Of course, learning to temporarily unlearn scientific realism may not always be easy, as it appears to be deeply ingrained in our common sense view of the world, science and progress, just as unlearning universal scientific theories may pose special challenges to scientists. But if metaincommensurability is correct, then it is a necessary first step to becoming bilingual, which itself is a prerequisite to meta-­ theory comparison. So just as Feyerabend argued for theoretical pluralism in science on the basis of incommensurability, I am suggesting that there is a strong argument from metaincommensurability to pluralism in philosophy. But scientific realists can rest assured that in my view, like in Feyerabend’s ocean of alternatives, outdated meta-theories live on exactly as part of the empirical content of their successors, which consists of both the facts that corroborate them and the incommensurable meta-theories, together with their corroborating facts, that it indirectly rules out. In any case, scientific realism seems to me to be in crisis and need of replacement exactly because it cannot cope with  the fact that theories are incommensurable.

7.16 What Is the Justification of Metaincommensurability? Just as the justification of incommensurability is an IBE from facts about science and its history, the justification of metaincommensurability is an IBE from facts about philosophy of science and its history, such as (1) the experiences of philosophers of science trying to understand theories of science that clash with scientific realism, and the regular phenomena like talking at cross-purposes in the course of such discussions, (2) conceptual change introducing kinds that break the no-overlap principle (with respect to established lexical taxonomies used to understand science) resulting in world change (new realistic interpretations of science), (3) the resulting ineffability of new concepts on the bases of older theories about scientific theories and the ineffability of older concepts on the bases of newer theories about scientific theories, (4) the limits to logical implication with respect to

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

123

trans-­paradigmatic relations between theories of science that break the no-overlap principle by cross-classifying some of the same things into mutually exclusive sets of mutually exclusive kinds, (5) mutual accusations of the use of circular and question-­begging arguments with each side using its own side to support itself (such as in the realism debate), and (6) the need for bilingualism and persuasion as prelude to empirical proof (based on arguments as supported by evidence from philosophical practice and its history). The best explanation of these facts is that theories about science stated according to lexical taxonomies that break the no-overlap principle with respect to each other are metaincommensurable.

7.17 What Can We Conclude About Incommensurability and Metaincommensurability? In this paper, I tried to develop and defend the idea that incommensurability is a fact about science. In the process, I tried to clarify the claims that are required to make sense of incommensurability and to clarify the relations between these claims. The main assumption was that (1) observation statements must be deductive consequences of the theories that they are used to refute or corroborate. The second main assumption was that (2) theories must be stated according to lexical taxonomies that do not break the no-overlap principle by cross-classifying some of the same things into mutually exclusive kinds. Lastly, I assumed that (3) theories have ontological consequences, from which it follows that when there is kind change that breaks the no-overlap principle there is a corresponding world change. The first claim is needed to explain how incommensurable theories can and cannot be tested. The second claim is needed to explain how incommensurable theories are conceptually incompatibility (as distinct from logically inconsistent). The third claim is needed to explain how incommensurable theories are ontologically incompatibility. Moreover, the first claim shows how and why the second claim implies the third claim. On the basis of these three claims, I tried to develop incommensurability as a fact about science and identify its causes, consequences and implications for theory comparison, scientific realism and scientific progress. I tried to explain how the hypothetico-deductive method of theory testing shows how transient underdetermination of theory choice on the basis of the observable facts can always arise (given an incommensurable proposal), then be overcome, only to be re-confronted with the problem of unconceived incommensurable alternatives and the resulting potential for a new round of transient underdetermination. I then used the idea of metaincommensurability to try to explain how and why treating incommensurability as a fact about science will seem either incoherent or question-begging to scientific realists, while scientific realism will seem to be wrong and in need of replacement from the Kant-on-wheels perspective that Kuhn and Feyerabend tried to develop to account

124

E. Oberheim

for incommensurability. And I tried to show how these beliefs (that incommensurability and metaincommensurability are facts) are justified in the same way scientists justify many of their beliefs about facts, as the result of an IBE. I am well aware that my strategy of developing metaincommensurability to help establish incommensurability as a fact on the basis of Kant-on-wheels realism may look question-begging, or circular, just as the idea themselves may seem incoherent, or even self-refuting, from the common sense scientific realist perspective. But I’m not worried about that, because that is just exactly what scientific realists always seem to do in support of scientific realism. The accusation of question-begging is expected. It is not a worry, but a corroboration. While much more could, and perhaps should, be said about incommensurability and metaincommensurability, they are definitely something that one should try to experience for oneself, either by struggling to make sense of surprisingly, seemingly incoherent passages from out-dated scientific theories, or by struggling with bizarre and mysterious sounding theories of science, until they finally make sense. Start with the passages that seem most ridiculous and try to figure them out. Once you have managed to do it, with a little self-reflection, you will have confronted the facts I’ve been trying to talk about. In any case, with these remarks I hope to have cast some light on what has, for more than a half-century, proven to be a controversial idea at the very intersection of methodology, epistemology and ontology.

References Boyd, Richard. 1984. The Current Status of Scientific Realism. In Scientific Realism, ed. Jarrett Leplin, 41–82. Berkeley: University of California Press. Duhem, Pierre Maurice Marie. 1954. The Aim and Structure of Physical Theory. Princeton: Princeton University Press. Einstein, Albert. 1949. Autobiographical Notes. In Albert Einstein: Philosopher-Scientist, ed. Paul Arthur Schilpp, 3–95. La Salle: Open Court. Feyerabend, Paul K. 1962. Explanation, Reduction, and Empiricism. In Scientific Explanation, Space, and Time. Minnesota Studies in the Philosophy of Science, ed. H. Feigl and G. Maxwell, vol. III, 28–97. Minneapolis: University of Minneapolis Press. Revised in (1981a). ———. 1965a. Reply to Criticism: Comments on Smart, Sellars and Putnam. In Proceedings of the Boston Colloquium for the Philosophy of Science 1962–64: In Honor of Philipp Frank, Boston Studies in the Philosophy of Science, ed. R. Cohen and M. Wartofsky, vol. 2, 223–261. New York: Humanities Press. ———. 1965b. On the “Meaning” of Scientific Terms. The Journal of Philosophy 62 (10): 266–274. ———. 1970. Against Method: Outline of an Anarchistic Theory of Knowledge. In Analysis of Theories and Methods of Physics and Psychology. Minnesota Studies in the Philosophy of Science, ed. M.  Radner and S.  Winokur, vol. IV, 17–130. Minneapolis: University of Minnesota Press. ———. 1975. Against Method: Outline of an Anarchistic Theory of Knowledge. London: New Left Books. ———. 1977. Changing Patterns of Reconstruction. The British Journal for the Philosophy of Science 28 (4): 351–382.

7  Incommensurability and Metaincommensurability. Kind Change, World Change…

125

———. 1981a. Realism, Rationalism and Scientific Method. Philosophical Papers. Vol. 1. Cambridge: Cambridge University Press. ———. 1981b. Problems of Empiricism. Philosophical Papers. Vol. 2. Cambridge: Cambridge University Press. Kuhn, Thomas S. 1962. The Structure of Scientific Revolutions. [Page references are to the second enlarged edition with new “Postscript—1969” published 1970, and Unchanged in the third edition published 1996.]. Chicago: University of Chicago Press. ———. 1970. Reflections on My Critics. In Criticism and the Growth of Knowledge. Proceedings of the International Colloquium in the Philosophy of Science, London, 1965, ed. I. Lakatos and A. Musgrave, 231–278. Cambridge: Cambridge University Press. ———. 1983. Commensurability, Comparability, Communicability. In PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association 1982, Volume Two: Symposia and Invited Papers, 669–688. Philosophy of Science Association. ———. 2000. The Road since Structure: Philosophical Essays, 1970–1993, with an Autobiographical Interview. Chicago: University of Chicago Press. Popper, Karl. 1935. Logik Der Forschung. Zur Erkenntnistheorie Der Modernen Naturwissenschaft. Julius Springer, Vienna. Ninth edition 1989. Tübingen: Mohr Siebeck. Sankey, Howard. 2001. Scientific Realism: An Elaboration and a Defence. Theoria A Journal of Social and Political Theory 98 (98): 35–54. ———. 2018. The Demise of the Incommensurability Thesis. In The Kuhnian Image of Science: Time for a Decisive Transformation? ed. M. Mizrahi, 75–91. London/New York: Rowman and Littlefield. First published 2016.

Chapter 8

The Landscape of a Metaphysical Battlefield: A Comment on Eric Oberheim Leandro Giri

8.1 Introduction In 1962, Thomas Kuhn and Paul Feyerabend independently and almost simultaneously outlined the concept of incommensurability, a concept that has been widely studied in philosophy of science ever since. In particular, Eric Oberheim’s “Incommensurability and Metaincommensurability” unveils one of the most interesting discussions in contemporary history around this complex concept: the controversy between Howard Sankey and himself (featuring also Paul Hoyningen-Huene and Hanne Andersen). Several essays that were published in the context of this controversy over a span of more than 20 years,1 reveal the vivacity of this friendly but tough philosophical war. Fortuitously, I had the great pleasure of watching its latest battle here in the southern cone of South America. In 2018, Oberheim presented his answer to Sankey’s last attack on incommensurability (Sankey 2018) in Buenos Aires, and we heard extensively from both of them during our kuhnian conference both in Argentina as well as in Uruguay. I had the enormous honor of being Eric Oberheim’s commentator, while both Howard Sankey and Paul Hoyningen-Huene were there and sparked the debate. In the following pages, I will present my conclusions from this amazing and fruitful philosophical conversation.

 These essays were written by Sankey (1994); Hoyningen-Huene et  al. (1996); Sankey (1997, 2008, 2009a, b, 2018); Oberheim & Hoyningen-Huene (1997); Hoyningen-Huene & Oberheim (2009); among others. 1

L. Giri (*) Institute of Philosophical Investigations, Argentinian Society for Philosophical Analysis, National University of Tres de Febrero – CONICET, Buenos Aires, Argentina e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 L. Giri et al. (eds.), Perspectives on Kuhn, The Western Ontario Series in Philosophy of Science 84, https://doi.org/10.1007/978-3-031-16371-5_8

127

128

L. Giri

Even though both Sankey and Oberheim convey very persuasive arguments making neutrality a challenge, I do not intend to side with either of them, but rather only to shed light on every aspect I consider opaque. I will lay out an analysis that may help the reader to understand the topology of the battlefield and how both philosophers are displaying their arguments. Oberheim’s work presenting the history of the notion of incommensurability as presented by Kuhn and Feyerabend is too profound, detailed and even perfect to add any value to it; therefore, as I said, I will focus on the muddy arenas of the controversy between Oberheim and Sankey. My comments are centered around three dimensions which seem to be the main points of divergence between Sankey and Oberheim’s arguments. I will call the first dimension “the exegetic dimension”, as it calls into question how Sankey and Oberheim understand Kuhn and Feyerabend’s versions of incommensurability. The second dimension is “the phenomenological dimension”, as it asks whether or not incommensurability is something which exists in the world (by existing in the world it would be something to be discovered and tagged, as if it were some kind of biological species or another scientific entity). On the other hand, if it does not exist, then it is a philosophical construct which only makes sense inside of a network of meta-scientific postulates; therefore it is not discovered, but elaborated philosophically in order to support some ideas or undermine others. The third dimension is “the metaphysical dimension”, which is concerned with the origin of Sankey and Oberheim’s different points of view. Oberheim’s hypothesis is that Sankey’s dismissal of incommensurability (Sankey 2018) is due to his nonnegotiable position as a scientific realist. If so, Sankey is taking for granted the existence of completely genetically object-sided2 scientific facts (the foundation of scientific realism) and interpreting incommensurability from there. If he is, then he is also dismissing other points of view (for example Kuhn and Feyerabend’s) on the basis of a metaphysical position that is not compatible with what both authors had in mind (which would be another metaphysical position, namely a Kant-on-Wheels3 one). To discuss philosophical statements with someone who supports a different metaphysical position would be, according to Oberheim, analogous to discussing scientific statements with someone who supports an older theory after a scientific revolution (but at a higher epistemic level). This is called “metaincommensurabilty”, and it is the key to my third dimension.

8.2 The Exegetic Dimension In this dimension, I will focus on how Sankey and Oberheim understand Kuhn and Feyerabend’s positions on incommensurability, specifically on what seems to be the main point of divergence: the relation between incommensurability and   That characterization of a metaphysical position follows the vocabulary of HoyningenHuene (1993). 3  Kuhn’s Neokantian metaphysical position was reconstructed in Hoyningen-Huene (1993). The funny name “Kant-on-Wheels” is due to Lipton (2001). 2

8  The Landscape of a Metaphysical Battlefield: A Comment on Eric Oberheim

129

(in)comparability. Clearly, both Kuhn and Feyerabend’s original versions were difficult to grasp, mainly due to some kind of vagueness and changes they made as they reinforced their positions after being severely criticized by many mainstream philosophers of science. However, it is still possible to reconstruct both positions, as Sankey and Oberheim did. So, I will not focus on the elements they both agree on in Kuhn and Feyerabend, but only on the apparently divergent ones. Sankey (2018) takes Feyerabend’s following literal quote: “as a consequence [of semantic variation] incommensurable theories may not possess any comparable consequences, observational or otherwise.” (1981a, 93), and in another place, he takes another quote (from Science in a Free Society) which states that incommensurability means “deductive disjointness and nothing else.” (Feyerabend 1982, 67). According to Oberheim, Sankey chose those quotes in order to state that Feyerabend defends a form of incommensurability-implies-incomparability thesis, “…according to Sankey, at least it is clear that Feyerabend denied that incommensurable theories can be compared empirically, that is according to their empirical consequences (which indeed would be a problem for scientific realism and more generally empiricism)” (Chap. 7, in this volume, p. 110). This must be softened somehow, as Sankey admits that Feyerabend does allow comparability between incommensurable theories, although content or verisimilitude are not legitimate criteria for such comparison. So, how can we compare these two rival theories? Sankey answers that for Feyerabend such a comparison needs “methods which do not depend on the comparison of statements with identical constituents” (Feyerabend 1981b, 115). If both Sankey and Oberheim read Feyerabend as allowing comparison between incommensurable theories, there is no exegetical issue here. Of course, Sankey, according to his tough scientific realism, states that it is perfectly possible to compare such theories through their common content, as they probably refer to a (mostly but not entirely) overlapping domain of entities, where ostensive strategies can be performed in order to achieve the (rational) comparison. But it is important to understand that Sankey is not reading Feyerabend in a way that is (relevantly) divergent from Oberheim’s reading; he is just stating that according to his analysis, Feyerabend is wrong. As for Kuhn, once again, Oberheim states that according to Sankey, Kuhn’s incommensurability states incomparability: “According to Sankey, Kuhn’s denial of the possibility of ‘point-by-point comparison’ means that “it is not possible to directly compare the contents of the theories with respect to specific points of agreement and disagreement” (2018, 86). (Chap. 7, in this volume, p. 110). Even when some of the passages of Kuhn’s early work may not clearly convey his understanding of incommensurability and its relation to comparability, in my opinion, in his late and mature work Kuhn is very sure that one does not entail the impossibility of the other. If any doubt is cast, Oberheim’s reconstruction would exorcise it. What does Sankey think about Kuhn’s position? it is not entirely clear whether the incomparability in question [in Kuhn’s work] is incomparability of content. This is because he [Kuhn] goes on to assert that “there is no common language” in which a “point-by-point comparison” may be made between the theories. This assertion may be read as the claim that it is not possible to directly compare the contents of

130

L. Giri

the theories with respect to specific points of agreement and disagreement due to the lack of a common language in which both may be expressed.4

As far as I can see, the expression “this assertion may be…” should be connected with the following passage where Sankey says “…it seems clear that on the whole Kuhn’s intention was not to deny that the content of incommensurable theories may be compared. Indeed, the position that Kuhn endorsed in the latter part of his career allows significant scope for content comparison” (Id., my emphasis) Oberheim states that Kuhn just wanted to clarify that it is not possible to directly compare the contents of the theories with respect to all specific points of agreement and disagreement. But it is possible indeed to do so with some of them. Once again, I do not see that Sankey’s understanding of Kuhn’s position differs radically from Oberheim in this point. Of course, there is a huge divergence in the consequences each envision based on their understanding of Kuhn, but that is the metaphysical dimension. Basically, both Sankey and Oberheim read Kuhn and Feyerabend’s versions of incommensurability as allowing (some kind of partial and indirect) comparison. Even when they state that it is not so easy to do so, such as in the Popperian or neopositivist versions of the dynamics of scientific progress, the possibility of rationally comparing two incommensurable theories still exists. Sankey, for reasons we will focus on later, seems to think that such a comparison is easier than Kuhn and Feyerabend thought, while Oberheim seems to find the comparison to be as hard as Kuhn and Feyerabend stated. My conclusion on the exegetical dimension is that the differences between their hermeneutical work on Kuhn and Feyerabend have been overstated. There are, of course, differences, but my feeling is that in this dimension there have been more fireworks than real fire. This does not mean that the controversy is non-existent, but it is not rooted in how each has read Kuhn and Feyerabend, but rather in the conclusions they arrive to from a similar starting point.

8.3 The Phenomenological Dimension Sankey dedicates a whole section of his “Demise of Incommensurability” paper (2018) to attack the “so-called phenomenon of incommensurability”. He seems not to be convinced at all that incommensurability is a fact to be discovered in the world. This incommensurability-as-a-fact thesis can be deduced from Kuhn’s expressions in his story about how he finally could understand Aristotle’s physics. It can also be deduced from Feyerabend’s expressions as he analyzes changes in the theory in particular historical examples. Similarly, Hoyningen-Huene and Oberheim adopt the incommensurability-as-a-fact concept in their incommensurability article on the Stanford Encyclopedia (Oberheim & Hoyningen-Huene, 2009) and in their  Sankey (2018, 86) my emphasis.

4

8  The Landscape of a Metaphysical Battlefield: A Comment on Eric Oberheim

131

Hoyningen-Huene & Oberheim (2009) article on Sankey. In the latter, they state that the very idea of incommensurability arises from historical analysis and not as a consequence of an antirealist point of view. Actually, the antirealism (or, more specifically the Kant-on-Wheels metaphysics, which is certainly a kind of realism) arises, according to Hoyningen-Huene & Oberheim, as an Inference to the Best Explanation from incommensurability, and not the other way around. If this is so, using (scientific) realist assumptions about reference to declare the demise of incommensurability would entail a form of the petitio principii fallacy. We will get back to this in the next dimension, but we can already see clearly why Sankey dedicates time to undermining the phenomenological idea of incommensurability. Sankey (2018) first targets Feyerabend’s incommensurability version, which he suggests entails three components. First of all, he explains it entails a realistic interpretation of theories (which implies that the observational terms get their meaning from the theories where they are used). Second, for Sankey, Feyerabend’s incommensurability version also includes a claim in which the term “impetus” cannot be defined in the context of a theory which does not accept the laws on which the said term is defined. Thirdly, he explains that as a result of the first two components, meaning variant theories are incomparable. I can agree that the realistic interpretation of theories is a metatheoretical claim, but the second claim about “impetus” (or about any T-theoretical term, in hempelian vocabulary) is a fact: you cannot define any Newtonian-theoretical term with Aristotelian laws. Maybe it is a fact which only we philosophers of science care about, but it is a fact anyway. As for the third point, we have already stated that in Feyerabend’s philosophy, incommensurability does not entail incomparability, but let’s look at the question of comparison by content or verisimilitude. This seems to contradict what Sankey himself understands of Feyerabend. As far as I can see, if Feyerabend noticed while studying the history of science that a T-theoretical term couldn’t be defined in the context of a theory with incompatible laws, he would have the right to claim that he noticed an incommensurability phenomenon somehow (or the print it left in the incommensurable theories), even when he then appealed to a battery of metatheoretical constructs to explain what it is, why such a phenomenon occurs and what consequences it entails. What about Kuhn? Sankey admits that Kuhn treats incommensurability as a phenomenon to be discovered, as he did while asking himself (in the preface of Kuhn 1977) why a clever guy like Aristotle had been so awful at physics (he was not awful at all!).5 But then Sankey states that “But it is simply not clear that the actual phenomenon itself is best described as a case of incommensurability. The phenomenon is the act of comprehension following the initial failure to understand. The notion of incommensurability is appealed to as an explanation of one such episode.” (2018, 83).

 The famous Kuhn’s epiphany is not that interesting when you notice that Koyré (1943) had already stated very clearly how Aristotle’s physics work in non-whiggish terms. 5

132

L. Giri

Consequently, Sankey believes that it seems “to be a mistake to think of Kuhn’s encounter with Aristotle as the discovery of the phenomenon of incommensurability rather than the experience of understanding an initially incomprehensible text.” (2018, 83). I think that both cause and effect can be legitimately treated as phenomena. If Kuhn’s epiphany (or maybe a gestalt switch) can be treated like a phenomenon, I cannot see why the very fact that it is not possible to understand Aristotelian dynamics with the vocabulary of Newtonian dynamics cannot also be treated as a phenomenon, the one which caused the very phenomenon inside of Kuhn’s mind. Actually, if incommensurability is a phenomenon, it is possible to discuss it without necessarily appealing to metaphysics, which is what has happened. I think Sankey would have a better position by stating that the phenomenon of incommensurability is trivial or uninteresting than in trying to transform it into a mere philosophical construct. So, on the one hand, Sankey’s argument is that Kuhn and Feyerabend were antirealists; so, they developed, through a philosophical process, a concept of incommensurability that was compatible with their metaphysical positions. Then, with that concept in mind, they made their historiographic reconstructions in order to show scientific controversies in an antirealist (Kant-on-Wheels) fashion. On the other hand, Oberheim’s argument is that while studying history of science Kuhn and Feyerabend found their version of incommensurability and then, as an Inference to the Best Explanation, the Kant-on-Wheels metaphysics appeared as a suitable explanation. Sankey tried to undermine the idea that it is possible to find incommensurability while studying history of science, as it is not a phenomenon but a philosophical construct. He does so because, basically, he is trying to state that the Kant-on-Wheels metaphysics cannot be sustained from any piece of empirical (historical) work. But as far as I can see, it is perfectly possible to at least recognize some of the effects of Kuhn and Feyerabend’s notions of incommensurability through a careful historiographic analysis. This does not necessarily make their story true, but at least it shows that it is plausible. As a conclusion, I don’t think that Sankey’s attack on this dimension could cause any damage to Oberheim’s defenses. But now another question comes to my mind: is it legitimate to declare the demise of incommensurability via realist assumptions, as Sankey does? I will explore that question in the next dimension.

8.4 The Metaphysical Dimension Sankey’s final strike to incommensurability is mounted on his deep commitment to scientific realism. For him, such commitment entails a causal theory of reference. So, according to him, rival incommensurable theories refer to some common objects, and that is the reason why content comparison is possible. If theory A makes claims about object x, and theory B also makes claims about object x, you can rationally choose which claims are better (more fruitful, accurate, simple

8  The Landscape of a Metaphysical Battlefield: A Comment on Eric Oberheim

133

claims…). So there is (an easily identifiable) scientific progress. Even when he admits that incommensurability exists, it is not clear why we have spent so much time and ink talking about it, as it has no non-trivial effects on progress or rationality. In other words, even with incommensurability (in the deflated way Sankey understands it) being present, people somehow decided that Newtonian dynamics were better than Aristotle’s, and Einstein’s mechanics better than Newtonian, and they did it for good and rational (and mainly internal) reasons. I find Sankey’s arguments very persuasive, as it seems to me that if I embraced scientific realism and the causal theory of reference, it would lead me to the same results. However, Oberheim is also persuasive when he states that if you do not take such things for granted, conclusions may be different. Observation sentences, when interpreted via different categories (which are modified after scientific revolutions, if you hold the Kant-on-Wheels metaphysical assumption), preclude an easy crucial experiment, as they entail people who argue circularly using their own paradigmatic assumptions to do so. This does not mean that crucial experiments are impossible, but that they are complex and not always convincing for every actor involved (which only makes such experiments plausible in the context of indirect refutations, if the theories being compared are incommensurable). If you believe this part of the story, it is not possible to hold an approximation-to-truth thesis (another important foundation of scientific realism). So, is incommensurability (in Kuhn and Feyerabend’s version) a deadly weapon against scientific realism? Maybe it is, if we believe Hoyningen-Huene & Oberheim on their claim that Kant-on-Wheels metaphysics is an Inference to the Best Explanation from incommensurability. At least if you take Kuhn or Feyerabend’s incommensurability. But at the same time Sankey is very persuasive when showing that it is plausible to embrace a (weakened) incommensurability which is perfectly compatible with a scientific realist world: an incommensurability that does not cause problems as it can be solved by ostensibility. As you can have a notion of incommensurability that is compatible with a Kant-on-Wheels metaphysics (namely that of Kuhn and Feyerabend) or a weakened notion that is compatible with scientific realism (that of Sankey), the incommensurability missile is not necessarily relevant in the field of metaphysics. The incommensurability phenomenon for Oberheim is an Inference to the Best Explanation for the fact that there is conceptual change between (incommensurable) theories that break the no-overlap principle, which causes ineffability of older concepts on the base of the new theories which generates talking at cross-purposes between defenders of old and new theories. Curiously, Sankey agrees with Feyerabend on a very important point: scientists routinely deal with comparisons of incommensurable theories without any problem (the way they do it is different in Sankey and Feyerabend, but both agree in that it usually poses no problem). Talking at cross-purposes seems to be more usual between philosophers (especially when discussing metaphysics) than between scientists. In fact, a close examination of Kuhn’s favorite example, the Copernican Revolution, shows that all actors involved usually understood each other perfectly without agreeing: everybody was multilingual; everybody understood the difference

134

L. Giri

between “that is Mars” in the context of each competing theory (and, of course, they could understand the sentence ostensively). Everybody was sure that their version of the story was correct and the others were wrong. In short, I’m stating that Copernican Revolution is not a good example of talking at cross-purposes. Yes, every actor involved was multilingual, but in the Feyerabendian sense, I think that you could not have been a relevant medieval astronomer without being multilingual, and the same must be happening with most scientists nowadays.6 After my analysis, I’m starting to believe that Oberheim is right: there is fire everywhere, but the real battle is between realism and antirealism—an inheritance of a mainstream discussion in the ‘80s. Kuhn & Feyerabend and Sankey live in different meta-worlds. This is meta-incommensurability: Scientific Realism and Kant-­ on-­Wheels are analogue to Ptolemaic and Copernican cosmologies, but in a higher epistemic level. What implications does meta-incommensurability have in our discussion? It depends on the version of the phenomenon of incommensurability you take for granted, being more or less difficult to be multilingual and to avoid talking at cross-purposes. I do not think that Oberheim’s dismissal of scientific realism on the base of the incommensurability phenomenon is legitimate, because it is grounded in an analogue strategy of Sankeys, which, as said before, entails a kind of petitio principii fallacy. Sankey’s argument could be (childishly) reconstructed as: (a) Scientific Realism is the one and only truth (b) Incommensurability as Kuhn and Feyerabend presented it is incompatible with Scientific Realism (c) Incommensurability as Kuhn and Feyerabend presented it is wrong (therefore Scientific Realism is the one and only truth) Basically, Sankey’s argument against (Kuhn and Feyerabend’s) incommensurability cannot be sustained without stating the correctness of Scientific Realism, which is not self-evident at all. But Oberheim’s argument against Scientific Realism cannot be sustained unless stating the correctness of Kuhn and Feyerabend’s incommensurability which, after reading Sankey I can say, is not self-evident at all. Let me be very clear. I am convinced that incommensurability exists (incommensurability being the phenomenon of conceptual change between theories that break the no-overlap principle, which causes ineffability of older concepts on the base of the new theories). But Sankey’s version of incommensurability does not deny this fact; it denies that it has interesting non-trivial effects on theory comparison. Is it so? I don’t really know, but it seems plausible, which gives Scientific Realism an extra chance in the game of metaphysics. What about talking at cross-purposes? Kuhn convinced me that talking to an Aristotelian guy could be difficult. Maybe the centuries between me and the last  One good example of a hot controversy is the one sustained between Longomontano and Kepler, in a moment in which Aristotelic, Tychonic and Copernican astronomies coexisted. I thank Diego Pelegrin for this great historical example that he presented in the XIX Congress of the Argentinean Philosophical Association (AFRA), in Mar del Plata, December of 2019. 6

8  The Landscape of a Metaphysical Battlefield: A Comment on Eric Oberheim

135

Aristotelian do not help in the communication. Incommensurability seems something historians of science should take into account when studying old theories in the light of present ones (you could be talking at cross-purposes with dead Aristotelians). But is it necessarily a relevant phenomenon when studying scientific revolutions? Is talking at cross-purposes a ubiquitous phenomenon in every controversy during a crisis? From my viewpoint, not at all. It can happen, but seems to be something more related to the people involved in such talking than in incommensurability itself.

8.5 Conclusion Shift of meaning occurs after scientific revolutions are a fact, after which we could say that the old theory and the new one are incommensurable theories. This is a first level phenomena, something that happens at a scientific level. The cause and consequences, scope, significance and the very meaning of incommensurability are some things to be discussed in the second, philosophical level. And discussed they are: many versions of these aspects of incommensurability were developed after its independent introduction by Kuhn (1962) and Feyerabend (1962). The concept has not been stable at all, and I think there are still many places where fruitful work can be done. It is also clear that the concept of incommensurability is pervasive in a metaphysical level, and so, metaphysical assumptions can influence and be influenced by whatever is understood by the very notion of incommensurability. When two versions of incommensurability are contrasted, if the metaphysical assumptions are not compatible, it seems impossible to arrive at an agreement. This is a second-level fact, something that happens at a meta-scientific level, which is why Hoyningen-Huene and Oberheim named it “meta-incommensurability”. Taking that into account, when discussing incommensurability, I think that one healthy thing to do is to be explicit about the metaphysical assumptions embraced to avoid talking at cross-purposes. So, it is plausible to have different versions of realist incommensurabilities, neo-kantian incommensurabilities, and other incommensurabilities related to other metaphysical positions. In this way, we could ask any defender of a metaphysical position how it deals with the incommensurability phenomena, but it does not seem to be legitimate to attack any philosophical notion on the basis of it not being compatible with some metaphysical assumption. To do that, one would need first to prove that some metaphysical position is better than the others. Because of the lack of consensus in this field, it is not a good starting point to state that something is wrong because it is incompatible with my metaphysics. But it is a good starting point to state that if my metaphysics is right, then incommensurability (which, is understood as “shift of meaning” after scientific revolutions occur) should be understood in this way. This is the reason why I think that Sankey’s work is very important: he, in my view, succeeds in providing a version of incommensurability that is compatible with Scientific Realism in a very coherent fashion. A defender of any other metaphysical position will have to do the same,

136

L. Giri

namely, provide a version of incommensurability such that it is compatible with her position while coherent with a plausible interpretation of historical records. Sankey does not bring the demise of incommensurability at all (I would say that the title of his paper follows marketing purposes), he just reinterprets it to achieve his metaphysical purpose. But Oberheim is right in stating that taking Scientific Realism’s premises for granted is not a legitimate strategy to state that Kuhn and Feyerabend’s notion of incommensurability is utterly wrong, taking into account that their metaphysical position was not that one. In conclusion I think that Sankey does a good job, providing a (very weak) version of incommensurability that does not interfere with Scientific Realism. What I am stating is that Sankey’s work is compatible with Hoyningen-Huene and Oberheim’s claim that Kuhn’s (and Feyerabend’s) stronger versions of incommensurability support, through an Inference to the Best Explanation, a wheeled-Kantian metaphysics. So, meta-incommensurability should be an invitation to try to communicate peacefully with other philosophical tribes, but not to try to convince them to follow our metaphysical gods, or the story will end only when the last member of the other tribe passes away. I am glad to see that Oberheim, in his conclusions, follows the healthy attitude of embracing pluralism. I think Oberheim’s claim that Scientific Realism is wrong on the basis of meta-­ incommensurability is not legitimate, as it implies the correctness of Kuhn and Feyerabend’s notion of incommensurability, which, as I said, is not self-evident and, thanks to Sankey, there are plausible alternatives. But he is right in stating that Sankey’s arguments are not enough to show that Kuhn and Feyerabend are wrong, as their position is meta-incommensurable with Scientific Realism. So, it seems to me that a metaphysical pluralism is something healthy to embrace, in order to set the foundations for a program which would develop incommensurability notions as criteria for evaluating the plausibility of a metaphysical position: if the metaphysical position succeeds in explaining the incommensurability phenomenon in its own terms, it passes the test. As both Kant-on-Wheels realism and Scientific Realism already have passed the test, then they can coexist independently and there is no need to argue. But it is not the only alternative. As far as I see it, it would also be healthy (maybe healthier) to get back to the neopositivistic roots and continue to do philosophy of science without appealing to metaphysics at all.

References Feyerabend, Paul K. 1962. Explanation, Reduction, and Empiricism. In Scientific Explanation, Space, and Time. Minnesota Studies in the Philosophy of Science, Volume III, ed. H. Feigl and G. Maxwell, 28–97. Minneapolis: University of Minneapolis Press. ———. 1981a. Realism, Rationalism and Scientific Method: Philosophical Papers. Vol. 1. Cambridge: Cambridge University Press.

8  The Landscape of a Metaphysical Battlefield: A Comment on Eric Oberheim

137

———. 1981b. Reply to Criticism: Comments on Smart, Sellars and Putnam. In Realism, Rationalism and Scientific Method: Philosophical Papers, Volume 1, Paul K.  Fayerabend (comp.), 104–31. Cambridge: Cambridge University Press. ———. 1982. Science in a Free Society. London: Verso. Hoyningen-Huene, Paul. 1993. Reconstructing Scientific Revolutions. Chicago & London, University of Chicago Press. Hoyningen-Huene, Paul, and Eric Oberheim. 2009. Reference, Ontological Replacement and Neo-­ Kantianism: A Reply to Sankey. Studies in History and Philosophy of Science Part A 40 (2): 203–209. Hoyningen-Huene, Paul, Eric Oberheim, and Hanne Andersen. 1996. On Incommensurability. Studies in History and Philosophy of Science Part A 27 (1): 131–141. Koyré, Alexandre. 1943. Galileo and Plato. Journal of the History of Ideas IV (4): 400–428. Kuhn, Thomas S. 1962. The Structure of Scientific Revolutions. Chicago: University of Chicago Press. ———. 1977. Essential Tension: Selected Studies in Scientific Traditions and Change. Chicago: University of Chicago press. Lipton, Peter. 2001. Kant on Wheels. London Review of Books 23: 30–31. Oberheim, Eric and Paul Hoyningen-Huene. 2009. The Incommensurability of Scientific Theories. In The Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta. Metaphysics Research Lab, Stanford University. https://plato.stanford.edu. ———. 1997. Incommensurability, Realism and Meta-Incommensurability. Theoria 12: 447–465. Sankey, Howard. 1994. The incommensurability thesis. Aldershot: Avebury. ———. 1997. Incommensurability: The Current State of Play. Theoria 12 (3): 425–445. ———. 2008. Scientific Realism and the Rationality of Science. Aldershot: Ashgate. ———. 2009a. Scientific Realism and the Semantic Incommensurability Thesis. Studies in History and Philosophy of Science Part A 40 (2): 196–202. ———. 2009b. A Curious Disagreement: Response to Hoyningen-Huene and Oberheim. Studies in History and Philosophy of Science Part A 40 (2): 210–212. ———. 2018. The Demise of the Incommensurability Thesis. In The Kuhnian Image of Science: Time for a Decisive Transformation? ed. M. Mizrahi, 75–91. London/New York: Rowman and Littlefield. First published 2016.

Chapter 9

The Plausibility of Thomas Kuhn’s Metaphysics Paul Hoyningen-Huene

9.1 Introduction Thomas Kuhn’s writings contain many elements and passages that have confused his readers. What is a paradigm, really? Is there normal science? What is incommensurability? Is Kuhn’s understanding of science relativistic? One of the elements of The Structure of Scientific Revolutions (Structure for short) not only confused his readers but even Kuhn himself, namely, his talk about world change. As Kuhn confessed in Structure: In a sense that I am unable to explicate further, the proponents of competing paradigms practice their trade in different worlds (Kuhn 1970 [1962], p. 150)

Thirty-two years later, in his Thalheimer Lectures of 1984, Kuhn asks and declares:1 Is this an idealist position? Perhaps it is. But the idealism is then unlike any other of which I am aware.

However, this characterization of his specific would-be idealism as novel will certainly not persuade those who are critical of all those forms of idealism they are aware of. Many of them think that any form of idealism is necessarily incoherent: reality is, by the very meaning of that concept, objective and as such fundamentally and irrevocably opposed to anything subjective. Whatever Kuhn’s novel form of idealism is, it cannot escape this conceptual coercion if this idealism deserves its  Kuhn (1984, unpublished), p. 122 (page numbers refer to the unpublished ms.). A Spanish translation of the Thalheimer Lectures is contained in Melogno and Miguel (2017). 1

P. Hoyningen-Huene (*) Institut für Philosophie, Leibniz Universität Hannover, Hannover, Germany Department of Economics, Universität Zürich, Zürich, Switzerland e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 L. Giri et al. (eds.), Perspectives on Kuhn, The Western Ontario Series in Philosophy of Science 84, https://doi.org/10.1007/978-3-031-16371-5_9

139

140

P. Hoyningen-Huene

name. Correspondingly, reactions to Kuhn’s idealist hikes have mostly been very harsh, probably climaxing in Peter Godfrey-Smith’s musing: Chapter X [of Structure] is the worst material in Kuhn’s great book. It would have been better if he had left this chapter in a taxi. (Godfrey-Smith 2003, p. 96)

So, Chapter X appears to be so outrageously bad to Godfrey-Smith that he does not even consider the possibility that his reading of it may be deficient.2 Commentators who appear to be more benevolent have tried to downplay Kuhn’s world change talk as psychological or metaphorical, for example Alexander Bird: In summary, a change in paradigm can bring with it a range of important psychological changes that have cognitive (and emotional) consequences […]. It is these psychological changes that Kuhn is referring to with the metaphor of ‘world-change’ (Bird 2012, p. 869)

However, whether one likes it or not, at least from the 1980s on, Kuhn steadfastly refused this way of reading him: I see no alternative to taking literally my repeated locution that the world changes with the lexicon. (Kuhn 1984, unpublished, p. 120, my italics)

In my earlier work, I have tackled the question of Kuhn’s metaphysics from a viewpoint that was informed by Kant’s critical theoretical philosophy.3 Useful as this may be, in this paper I will try a different approach. I will focus on the fact that Kuhn acted mainly as a reflective historian when he wrote Structure.4 He deplores, in the Preface to Structure, the “limitations of space” that have drastically affected my treatment of the philosophical implications of this essay’s historically oriented view of science. Clearly, there are such implications, and I have tried both to point out and to document the main ones. (Kuhn 1970 [1962], p. x)

Thus, he reflected on what sort of image of science the appropriately pursued historiography of science would present to him. Apparently, he assumed that the result of these reflections would be plausible to at least some people, even conceding that Structure “remains an essay rather than the full-scale book my subject will ultimately demand” (Kuhn 1970 [1962], p. viii). As Kuhn correctly anticipated, philosophers would not be happy with his book:  There is some irony in Godfrey-Smith’s attitude. Kuhn never became tired to practice and teach that whenever one encountered a passage in an old scientific text that seemed not to make sense at all, one should absolutely not dismiss the passage. On the contrary, such a passage usually holds clues that may help to change one’s own understanding of the whole text to the better. Kuhn’s respective experience with Aristotle started his whole enterprise (see, e.g., Hoyningen-Huene (1993), p. 21–22, esp. fn. 80). However, such an attitude is rather foreign to many contemporary analytic philosophers, as Godfrey-Smith nicely exemplifies. Roughly, their maxim seems to be: if I don’t understand it, it must be rubbish. 3  See especially Hoyningen-Huene (1993). 4  I have treated this subject in detail in Hoyningen-Huene (1993), Section 1.2. For a lively impression of this kind of historiography, Kormos Buchwald (2019) is strongly recommended, in which historians of science tell many stories about their interaction with historian of science Jed Buchwald, who learned much of his trade during his time as research assistant to Thomas Kuhn at Princeton. 2

9  The Plausibility of Thomas Kuhn’s Metaphysics

141

I have usually refrained from detailed discussion of the various positions taken by contemporary philosophers on the corresponding issues. […] As a result, some of those who know and work within one of those articulated positions may feel that I have missed their point. I think they will be wrong, but this essay is not calculated to convince them. (Kuhn 1970 [1962], p. x)

Kuhn indicates here that the philosopher’s resistance will be due to the missing explicit confrontation of his theses with their positions. He does not, however, deplore that he did not work out the philosophical underpinnings of his own position, which would, potentially, make his position more persuasive. Kuhn thus seems to assume that his essay is intelligible without this unexplored philosophical basis, although he admitted that he was unable to fully explicate his world change talk. It should thus be possible to reconstruct from Structure itself what drove Kuhn, the historian of science, to his world change talk. This is what I try to do in this chapter. I try to reconstruct the experiences of the historian of science, who is somehow driven to talk about world change as something correlated with scientific revolutions, and want to find out whether Kuhn’s world change talk can be made plausible. I shall not deal with the question where such historians might find a hospitable philosophical home.5

9.2 The Problem: Scientists and the World Let me begin by neutrally describing the situation that Kuhn tries to capture with his world change talk. What entities do scientists take for real? Clearly, usually scientists, like everyone else, take for real the directly observable macroscopic objects by which they are surrounded. In addition, they take for real entities that cannot be directly observed (for whatever reason) but for whose existence they think overwhelming evidence exists. The evidence that makes their existence overwhelmingly credible consists to a considerable part of empirical phenomena that can best be understood as causal effects of existing unobservable objects. For instance, black holes are taken for real today first because there are many phenomena that can best be understood as their effects, although black holes themselves have never been directly observed. Second, black holes have a natural home in the pertinent well-­ confirmed theory, the General Theory of Relativity. Roughly the same evidential considerations hold for the unobservable properties that scientists ascribe to real objects, be they observable or unobservable. Either these properties can be more or less directly observed or measured, or they can be credibly inferred from phenomena that are interpreted as their causal effects.6 So, following this line, earlier scientists took for real Aristotelian essences, the retrograde motion of planets, phlogiston,  Apart from my earlier work, this is the subject matter of a companion paper Hoyningen-­ Huene (2021). 6  I am aware that what I am presenting in this section is a simplification. However, it seems to me that for the problem at hand the subtleties that I am skipping do not play a significant role. 5

142

P. Hoyningen-Huene

the constancy of biological species, the electromagnetic ether, the simultaneity of events, the constancy of masses, trajectories of point particles, etc. These scientists did not commit mistakes in taking the entities and properties mentioned for real, although, from today’s understanding, none of these entities and properties exists, and thus these older scientists were deeply wrong. Their persuasion about the existence of those entities and properties was based on the same principles as our persuasions about the existence of entities and properties:7 existing objects are either directly observable or measurable, or they can be persuasively inferred from the evidence that points to them as their causes. What has significantly changed in the meantime is the available evidence (together with our adherence to other theories but theories are, by themselves, typically not decisive in empirical science). On the basis of this evidence, we can either more or less directly show that the entities and properties earlier postulated do not exist. Or we have empirically confirmed theories that claim the existence of other entities and properties that are much better explanations for old and new empirical phenomena, such that the existence of the older entities and properties becomes extremely implausible. If we now compare what scientists from earlier times took for real and what scientists from today take for real, we see besides continuities also discontinuities. Objects and properties that were once taken for real have disappeared, and new objects and new properties of old objects have emerged. This much is uncontroversial. It may also be uncontroversial that cases of far-reaching and deep changes of the given sort can be described as “changes of world view”. In accordance with his, the apparently innocuous title of Structure’s Chapter X is “Revolutions as Changes of World View” (p. 111). However, as we shall see in a moment, immediately after this digestible title Kuhn begins the chapter by telling the reader about the historian’s temptation to describe such changes as world changes. This transition is extremely strange on two counts. First, the world change locution is in itself very odd in the given context, because it does not refer to some historical change of the world in the sense of, say, the cosmological development of the world. It is rather a world change correlated with or even due to the change of a paradigm. Second, the implicit equation of “change of world view” with “world change” appears even more bizarre. In normal circumstances, the (physical) change of some object A and the change of some image of A are far from identical. Of course, a (physical) change of A may be followed by a change of the image of A if this image is updated. The converse, however, seems bizarre: a change of A due to a change of the image of A. On top of that, however these two changes are correlated, this does not eradicate the fundamental difference between an object A itself and an image of object A. It is not completely surprising that readers may lose their confidence in an author’s philosophical acumen who implicitly seems to equate the two. Let us now investigate how Kuhn develops and justifies his world change talk in the course of Chapter X of Structure and how the above puzzles might be resolved.

 Compare Kuhn (1970 [1962]), pp. 2–3.

7

9  The Plausibility of Thomas Kuhn’s Metaphysics

143

9.3 World Change and Shifts of Vision We should first note how extremely carefully and reluctantly Kuhn introduces the world change topic in the first sentence of Chapter X: Examining the record of past research from the vantage of contemporary historiography, the historian of science may be tempted to exclaim that when paradigms change, the world itself changes with them. (Kuhn 1970 [1962], p. 111, my italics)

Note that it is the “contemporary […] historian of science” who may encounter this temptation. By the “contemporary” historian Kuhn means those historians who try to understand past science in its own terms and decisively not from today’s point of view, or, in other words, historians who try to eschew presentism (or “Whiggish” historiography).8 When Kuhn refers here to contemporary historians, he means to exclude scientists9 and also contemporary philosophers, because they typically do not occupy the vantage point of contemporary historiography.10 According to Kuhn, the contemporary historian “may” encounter this temptation (Kuhn being one of them), but this does not necessarily hold for all historians. Furthermore, it is only a “temptation” to exclaim, because of the obvious reluctance to exclaim something so outrageous.11 And finally, those historians giving in to the temptation would “exclaim” the world change, not just soberly stating it, because of their despair with the apparent nonsense of the claim that a paradigm change is correlated with a change of the world. Apparently, this first sentence of Chapter X of Structure is meant to prepare the reader for the really strong stuff that is awaiting him or her in this chapter, implicitly asking for benevolent patience with a historian at the forefront of his discipline who tries to articulate his predicament. After the above quote, Kuhn explains that during revolutions, scientists see familiar objects “in a different light” and, in addition, new objects because “paradigm changes do cause scientists to see the world of their research-engagement differently” (p.  111). Again, this is uncontroversial. However, immediately afterwards, Kuhn adds his first argument for the world change talk: In so far as their only recourse to that world is through what they see and do, we may want to say that after a revolution scientists are responding to a different world. (Kuhn 1970 [1962], p. 111)

The argument seems to be that when scientists refer to the world, they refer to (that part of) the world that is present in their observational and experimental practice; there is nothing else they refer to under the rubric of “world”. And because this kind

 For a detailed exposition of this form of historiography, see Hoyningen-Huene (1993), pp. 14–19.  Kuhn states: “In the sciences, therefore, if perceptual switches accompany paradigm changes, we may not expect scientists to attest to these changes directly”, Kuhn (1970 [1962]), pp. 114–115. 10  See also Chapter XI of Structure, “The Invisibility of Revolutions”. 11  Kuhn himself says that “working in a different world”, i.e., after such a world change, is a “strange locution”, and he “inquires […] about the possibility of avoiding” it (Kuhn 1970 [1962], p. 118). 8 9

144

P. Hoyningen-Huene

of world is different before and after a revolution, and there is no other kind of world to be referred to, “we may want to say” that the revolution driven scientific change is a world change. Who is that “we”? Is that the presentism avoiding historian? Or is this someone reflecting on our use of language who realizes that the given situation seems to drive him or her to this strange world change talk? As this topic recurs later, I will come back to it. Kuhn continues his discussion by trying to mitigate the harshness of his world change talk. He does so by showing that the world change phenomenon is, or is analogous to, something we are already somehow familiar with in a different area, namely gestalt psychology’s analysis of vision: It is as elementary prototypes for these transformations of the scientist’s world that the familiar demonstrations of a switch in visual gestalt prove so suggestive. (p. 111)

He then shows that scientific training builds up certain new gestalts in the students, such that after a number of such transformations of vision does the student become an inhabitant of the scientist’s world, seeing what the scientist sees and responding as the scientist does. (p. 111)

The important point is now that such scientific training of vision is not uniquely determined by “the nature of the environment”, but also by “the particular normal-­ scientific tradition that the student has been trained to pursue” (p.  112). In order words, the resulting “perception of [the] environment” (p. 112) is not only determined by purely object-sided factors, but also by subject-sided factors:12 What a man sees depends both upon what he looks at and also upon what his previous visual-conceptual experience has taught him to see. (p. 113)

Most importantly, at least some of the subject-sided factors are neither constant and universal nor irreversible, but culture dependent and reversible. Now, it is not far-fetched to investigate the possibility that the reversibility of the subject-sided factors of perception come into play in scientific revolutions. However, there is a problem for the transfer of results of gestalt psychology of vision to the history of science. There are three interrelated fundamental differences between the experimental situation in gestalt psychology on the one hand and scientists undergoing a scientific revolution on the other. First, the subject in the gestalt experiment has perceptual access to both the different gestalts and to the material basis of these gestalts, for instance, the lines on paper (p. 114). By contrast, the scientist lacks any access to the paradigm-free basis of phenomena. Second, for the subject of the gestalt experiment, both accessible gestalts have the same status: they are equally valuable possibilities of perceiving the underlying material basis. By contrast, for the scientist the new way of perceiving the world has typically an epistemologically by far superior status: the old view was wrong, the new view is right. Third, this  I have introduced and used the terminology of “object-sided” and “subject-sided” factors in Hoyningen-Huene (1993), pp.  33–36, 45–47, 62–66, 122 fn. 283, 125, 267–271; Hoyningen-­ Huene et al. (1996), p. 139; Hoyningen-Huene and Oberheim (2009), p. 208. 12

9  The Plausibility of Thomas Kuhn’s Metaphysics

145

status change will prevent the scientist from describing this change as a change of perception, because the common descriptions of a change of perception imply a symmetry in value of the two ways of perceiving.13 As a consequence, “we may not expect direct testimony about that shift” in the historical record (p. 115). Therefore, Kuhn continues by trying to make indirectly plausible by a number of examples that such shift in scientific vision occurred in the history of science (pp. 115–120).

9.4 A Philosophical Interlude Before continuing Kuhn’s train of thought, we should note that Kuhn abruptly interrupts his presentation of historical examples for the shifts of vision and presents an argument for his world change talk: At the very least, as a result of discovering oxygen, Lavoisier saw nature differently. And in the absence of some recourse to that hypothetical fixed nature that he ‘saw differently,’ the principle of economy will urge us to say that after discovering oxygen Lavoisier worked in a different world (Kuhn 1970 [1962], p. 118)

Kuhn says that one and the same object was seen differently by the proponents of different paradigms: one saw a compound ore where others had seen an elementary earth. Because there were also other cases of the same sort, one may generalize and say that Lavoisier saw nature differently. In other words, there was a change in world view. The subject matter of this world view – Kuhn calls it the “hypothetical fixed nature” –, however, is claimed to be inaccessible. Kuhn seems to say that when scientists refer to objects of nature, these objects are always already conceptualized in one way or other (that is, by some paradigm). Put somewhat metaphorically, objects scientists refer to are ipso facto already dressed up by a paradigm; they can never encounter them completely naked. Now, without prior warning, a “principle of economy” enters the stage, that urges us to speak in a certain way. This principle seems to be a relative or a variant of Ockam’s razor. The “principle of economy” urges: if you try to talk about something to which you cannot really refer, you should get rid of this something in your talk. If we follow this principle, then we cannot describe the transition from Lavoisier’s compound ore to Priestley’s elementary earth as a change of our view, because a view is a view of something and this something has just been removed from discourse because of its inaccessibility. The only reality we are left with is first an elementary earth, and after the change a compound ore. And this change is, because it is accompanied by other similar changes, a change of a more general nature: a change of the world.

 Kuhn argues for this point by saying “Looking at the moon, the convert to Copernicanism does not say, ‘I used to see a planet, but now I see a satellite.’ That locution would imply a sense in which the Ptolemaic system had once been correct. Instead, a convert to the new astronomy says, ‘I once took the moon to be (or saw the moon as) a planet, but I was mistaken.’” (p. 115). 13

146

P. Hoyningen-Huene

What sort of argument is the argument of the cited passage? The argument discusses our way of talking, our ability or inability to refer to something, and attempts to regiment reasonable ways of talking by referring to some “principle of economy” (whatever this is). Such an argument is clearly not an argument contained in the standard toolbox of a non-presentist historian of science. If it is anything, it is a philosophical argument using a “principle of economy”. Where does this principle come from? What is its scope? How is it justified, that is, where does it get its normative force from? What is its cogency? The way Kuhn uses this principle one may get the impression that this principle is at least compatible with, or even a part of, a verificationist theory of meaning. This theory implies that the meaning of a predicate is (or is strongly correlated with) the method by which one can determine whether the predicate is true of an object or not. Clearly, a verificationist theory of meaning also implies a sort of “principle of economy”, because it eliminates presumably unneeded elements. Historically, it is certainly not impossible that Kuhn had something like a verificationist theory of meaning in the back of his mind, although it was already in the 1960s highly contested.14 However, we do not know what Kuhn exactly meant by his “principle of economy” and what its credentials could possibly be, so we are left in the dark. The whole argument in the above passage is so badly underdeveloped that we cannot assess it.15

9.5 Shifts of Vision, Again We can therefore continue with Kuhn’s train of thought that made plausible so far, by a number of examples, that shifts of vision did occur in the history of science. For Kuhn, shifts of vision seem to be closely connected with, indeed being the basis of, his world change talk.16 However, Kuhn now takes up an objection against his shifts of vision account. The objection claims that it is not vision itself that changes, because vision is determined by purely object-sided input and the culture independent subject-sided contributions by the “perceptual apparatus”. Expressed differently, on this account our “perceptual apparatus” functions like a camera, in which the respective input from the object-side is quasi-mechanically represented in a  See, e.g., Schrenk (2009) and Uebel (2019).  If Kuhn should have left anything in a taxi, it is this “principle of economy”. It is perhaps worth noting that at least towards the end of his life, Kuhn was quite aware that Structure was philosophically seriously defective. During my last visit to Kuhn in September 1995, at one occasion he exclaimed: “My God, was I philosophically naïve when I wrote Structure!” My response was: “Thanks God you were so naïve, otherwise you would have never written that book!” 16  This seems to be implied by this passage: “Did these men [Galileo and Lavoisier] really see different things when looking at the same sorts of objects? Is there any legitimate sense in which we can say that they pursued their research in different worlds?” (p. 120) The immediateness of the transition from the first sentence to the second seems to indicate that the second sentence is implied by the first. This interpretation is supported by the fact that Kuhn is now going to defend the first sentence against an objection. 14 15

9  The Plausibility of Thomas Kuhn’s Metaphysics

147

unique way on the focal plane: the data. All the epistemic subject is capable of doing is to interpret these data in possibly different ways. Thus, instead of a “transformation of vision”, it is “the scientist’s interpretation of observations” that changes (p. 120). Kuhn’s reacts to this (self-inflicted) objection by claiming that it is part of “the traditional epistemological paradigm”, initiated by Descartes, that has come under attack from a number of disciplines, among them “parts of philosophy, psychology, linguistics and even art history” (p. 121). Although an alternative to this paradigm has not yet been fully worked out, Kuhn is “convinced that we must learn to make sense of statements that at least resemble” this one: “though the world does not change with a change of paradigm, the scientist afterward works in a different world.” (p. 121) Kuhn articulates the fundamental point in support of this statements in the immediately following sentence: What occurs during a scientific revolution is not fully reducible to a reinterpretation of individual and stable data. (p. 121)

And then Kuhn continues to argue that the relevant (revolutionary) process does not consist in stable, discernible data plus changing interpretation, but rather in changing data resulting in changed vision for which Kuhn presents further examples from the history of science. By now, Kuhn’s argumentative goal should be clear. He attacks any realism that claims that the experienced reality is, in my terminology, purely object-sided, that is, absolutely independent of any subject-sided contributions, which is often expressed as the “mind-independence” of reality. Clearly, in our everyday life we have overwhelming experience of this mind-independence of reality, especially that extra-mental reality does not respond to our wishes or any other mental content— reality is just what it is. All we can do is to interpret reality in different, but limited ways. Furthermore, Kuhn sees a philosophical paradigm supporting this idea of reality in the background, at least as far as perceived reality is concerned. It is the idea that our fixed perceptual apparatus produces retinal imprints of the “environment” (the purely object-sided). These imprints are thus completely beyond the control of the perceiving subject, because both factors involved in their production are independent of us once the object of our attention is chosen. All the perceiving subject can actively contribute is to interpret these retinal imprints in possibly different ways. Kuhn now tries to argue against this form of realism in various ways. He thinks that gestalt psychology has shown that shifts of vision can occur, that is that our perceptions of the world are not just fixed by the environment and our perceptual apparatus. This would weaken or even remove the support by the mentioned philosophical paradigm. Furthermore, he thinks that there is some indirect evidence “that history of science would make better and more coherent sense if one could suppose that scientists occasionally experienced shifts of perception” (p.  113). Why does Kuhn so strongly insist on these shifts of vision? He seems to believe that the assumption of the constancy of perception is an important pillar supporting the realism that he is attacking. And I think Kuhn is right: in our everyday experience, as

148

P. Hoyningen-Huene

long as it is not in a gestalt switch regime, perceived objects and their properties display this remarkable resistance against us, apparently conveying their complete mind independence to us. Let us now assume that Kuhn’s principal arguments for the shift of vision are correct and that shifts of vision indeed also occur in the history of science. In this way, he would destroy this pillar of the above-mentioned form of realism. Would this realism collapse as a result of the removal of this pillar? A realist would certainly not agree with this: the (putative) fact that there are shifts of vision would possibly be an interesting epistemological insight, but this would leave ontology untouched. That we perceive the world differently does certainly not imply that the world changes; the world as the object of this perception stays the same, notwithstanding the changing perception.

9.6 The Historian’s Viewpoint Now at this point Kuhn tries to argue that in the given context the reference to this stable world is somehow illegitimate.17 He does so by reference to the above-­ discussed dubious “principle of economy” that is argued for by “the absence of some recourse to that hypothetical fixed nature” (p. 118). What Kuhn wants to say seems to be this. When historians investigate a revolutionary situation, they try to see it in the actors’ categories, which means here: in which ways did the scientists themselves describe the objects and problems of their research situation. This is why Kuhn often refers in his world change passages to expressions like “the world of [the scientists’] research engagement” (p. 111), “the scientist’s world” (p. 111), the world scientists “worked in” (pp. 118, 121), and the like. So, for instance, before, during and after the chemical revolution for chemists the only possibility to describe the nature of water was in terms of an Aristotelian element or, alternatively, as a compound, that is, descriptions in terms of the competing paradigms. For these chemists, water was of course real, and the question was, what is the real nature of water? Is it an element or not? So, if a philosopher joined the discussion and claimed: look, the real reality of water is somehow behind these conceptions of water as an element or a compound, it is beyond your conceptualizations in this or that way! Although I cannot tell you what it is, but this is the real reality.18 These chemists would certainly not have agreed with the philosopher: the reality was what chemists (and lay people) described as water—that was real enough, they were in no need

 I am saying “somehow” illegitimate because Kuhn also writes “that we must learn to make sense of statements” like “though the world does not change with a change of paradigm, the scientist afterward works in a different world” (p.  121): the first occurrence of “world” refers to this unchanging world. 18  This is what Kuhn has in mind when he refers to “the absence of some recourse to that hypothetical fixed nature”, p. 118. 17

9  The Plausibility of Thomas Kuhn’s Metaphysics

149

whatsoever of some different or higher kind of reality behind it—and the argument was about whether this real water was an element or a compound. Thus, for these chemists, and therefore also for the non-presentist historian of science, the only alternative was between real water as an element or real water as a compound, tertium non datur. Generalizing now to all the other changes occurring in the Chemical Revolution, it does not seem illegitimate to say that in the world of the prerevolutionary chemists, there were the Aristotelian elements, water being one of them, and phlogiston, and so on. Similarly, in the world of the postrevolutionary chemist, there were now many more elements, water not being one of them, chemical compounds, no phlogiston, and so on. If these locutions are legitimate, one is drawn to the conclusion that the world of the chemists has changed. Now this sounds quite odd. I think it would be much more acceptable to say that before and after the revolution that the conceptualization of the world has changed, or that the chemists before and after the revolutions worked in differently conceptualized worlds. What comes to the fore here is a linguistic fact, namely, that our everyday concept of world has usually inbuilt the idea of “objectivity”, understood as “subject-­ independence”.19 In my terminology, our common world concept contains pure object-sidedness as a meaning component. In other words, our world concept is part and parcel of our everyday common-sense realism: the apparently fully certified pure object-sidedness of the things we usually encounter. This is the reason why the world change talk sounds so utterly odd: it violates the linguistic conventions of the world concept. Thus, philosophers trying to articulate something that deviates from our everyday realism, like world change, have the linguistic conventions against them, and therefore appear to speak nonsense. By contrast, philosophers trying to articulate something more or less in accordance with everyday realism, can recur to a ready-made language, and have no principal problems to express themselves. However, unless one assumes that common sense is usually right, one should not take the linguistic predicament of non-realist philosophers as an indicator that they are wrong or even talk nonsense. Whether one prefers to refer to the changes in chemistry as “world change” or as “change of the conceptualized world” now seems almost a matter of taste. However, more than taste seems involved, namely a sort of presupposition for or against realism: “world change” seems to side with antirealism, “change of conceptualized world” more with realism. However, this is not quite correct. On the one hand, “world change” can be compatible with scientific realism if, for instance, there is some continuity of reference across world change,20 or with structural realism, if there is some continuity of mathematical structures.21 On the other hand, “change of conceptualized world” may at least border an antirealist position, if knowledge can in principle not approximate nor reach the unconceptualized world. So, it seems that  I am saying “usually”, because we may suppress the objectivity aspect of our world concept in phrases like “she lives in her own world”, where the emphasis is on the subjective creation of something that is multifaceted like the real world but distinctly different from it and subjective. 20  See, e.g., Sankey (1994), Chapter 2. 21  See, e.g., Worrall (1996 [1989]). 19

150

P. Hoyningen-Huene

these two ways of talking about scientific change do not have unambiguous metaphysical presuppositions. Kuhn seems to assume that “world change” talk is, despite its oddness, less misleading than “change of conceptualized world”, because the latter somehow suggests that the unconceptualized world plays some role in the historical scientific debate. For philosophers wrestling with questions of realism, it may indeed be the case that the unconceptualized world plays some role in their arguments.22 Kuhn insists, however, that this is not the case for historians, because it is not the case for the scientists involved. These scientists do not speak about any unconceptualized reality. They unproblematically refer to something existing, water in the above example, and ask what its nature is, for which question different paradigms give different answers.

9.7 “Change of World View” or “Change of World”? We must now come back to the problem that Kuhn not only speaks about world change, but also about change of world view. I think that there are two questions involved. First, isn’t the expression “change of world view” much more appropriate and much less problematic than the world change talk? “Change of world view” is an obviously plausible characterization of scientific revolutions, whereas “world change” is not. Second, why does Kuhn implicitly identify world change with worldview change? This latter question appears to indicate an egregious category mistake that is worth investigating. I shall deal with these two questions in turn. Let us first see what speaks for change of worldview talk. The argument that this is a much more appropriate description of the pertinent change would be that the earlier scientists had (partially) erroneous views of the world, by falsely assuming the existence of some entities or falsely describing the properties of existing entities. These false views were later corrected and led, by a scientific revolution, to the later, more appropriate world views. Thus, a change of world view occurred, not a change of the world. This way of describing the change has the additional invaluable advantage of not challenging plausible metaphysical and epistemological convictions: that our worldview may change due to new data etc. is plausible, and that the underlying world is not affected by this change, too. This plausibility is based on the view that epistemology and metaphysics are strictly separated: an epistemic change, happening on the side of the epistemic subjects, cannot possibly influence metaphysics, i.e., the purely object-sided that is the focus of our epistemic enterprise. However, here is a counter-objection to this view. In phases of well-established scientific knowledge, scientists believe that they know what the (specific part of the) world is like, and behave accordingly. For instance, Aristotelians thought they knew that they live in a two-sphere universe with different kinds of matter and dynamics

 For instance, in the Kantian framework noumena may play a role as a “problematic concept”; see, e.g., Allais (2015), pp. 64–65. 22

9  The Plausibility of Thomas Kuhn’s Metaphysics

151

in the sublunary and the superlunary sphere. By contrast, Newtonians thought they knew that they live in a uniform universe of material particles that move on well defined trajectories due to the influence of forces. Clearly, Newtonians thought of the earlier Aristotelian view, where it differed from their own, as being erroneous. This is why Kuhn often expresses the effect of a scientific revolution by saying that after the revolution, scientists worked or practiced science in a different world.23 The relevant difference of “working in a different world” from “having a different world view” is rooted in the immediateness of the former in contrast to the latter. The statement that a scientist experiments with the chemical compound water expresses that this scientist takes for granted that water is a chemical compound, he just knows it. There is no reflection on the part of the scientist involved having the content “I am experimenting with water and according to my scientific world view, water is a chemical compound”. Kuhn wants to express that the effect of a revolution on the scientists is that the objects of research are what the new paradigms says: immediately and without reflective distance.24 This change of immediate commitments to entities and their nature, as it happens in revolutions, is not properly expressed by “a change of world view”, but it is expressed more adequately—however oddly—by “a change of the world”. In order realize how also we today take many claims of science without the slightest reflection or hesitation for granted, both in science and in everyday life, just consider how we express something like “the dinosaurs went extinct some 65 million years ago” or “the Sun is a star similar to many others”. We do not qualify such statements by “according to our current world view” or something similar, because we believe that these statements express facts, not beliefs. Thus, it is the immediateness of scientists’ reference to existing things and their nature that somehow licenses Kuhn’s odd talk of world change. This may also lead us to an answer to the second question: what is Kuhn doing when implicitly equating “world change” and “worldview change”? Under normal circumstances, a thing and an image of that thing are very different, and there seems to be no way to equate the two. However, in the course of scientific revolutions, a kind of blurring of the boundaries between “world” and “worldview” may occur. Clearly, when any of our contemporary views first came up, it was a (perhaps crazy) hypothesis whose truth was yet undecided. For instance, the heliocentric planetary system could have been seen first as a literally false theory of the planetary system (perhaps useful as an instrument for predictions, which was historically not the case). Only later, in the course of further empirical confirmation, it begun to be

 Kuhn (1970 [1962]), pp. 4, 6, 111, 112, 118, 120, 121, 135, 141, 147, 150.  For Kuhn, such a sense of epistemic security is part and parcel of a paradigm: “Normal science […] is predicated on the assumption that the scientific community knows what the world is like”, Kuhn (1970 [1962]), p. 5. However, this is perhaps a little overgeneralized. There are scientists who always have a critical, reflective distance to whatever paradigm, and do not take any theoretical statements for granted. Steven Hawking, who is a radical instrumentalist, is a case in point. He does not believe that physical theories refer to reality, “it is meaningless to ask whether [a physical theory] corresponds to reality” and therefore, they are neither true nor false; Hawking and Penrose (1996), pp. 3-4, see also Hawking and Mlodinow (2010), especially pp. 53–78. 23 24

152

P. Hoyningen-Huene

believed to be an adequate description of the world. Once it was believed to be true, it just described the world as it is. Every astronomer then “worked in this world” by presupposing that planets circle the Sun, that the Moon is a satellite etc. To describe their new stance as a new world view is, of course, correct, but as explained before, slightly too weak. The reason is that “worldview” implies possible multiplicity, especially possible falsity; by contrast, “world” implies absolute uniqueness and a categorical exclusion of falsity: the world is what it is; there is no conceptual space for something like “a false world”. Thus, if one describes the attitude of scientists in a situation of an undisputed paradigm, to say that their worldview was such-andsuch, is too weak a description. “Worldview” implies the real possibility of reflexive distance taking by, for instance, contemplating alternative worldviews. This, however, does not take place when scientists are in the firm grip of a paradigm. The famous evolutionary biologist and historian of biology Ernst Mayr described this process nicely with respect to evolution: “biologists no longer speak of evolution as a theory but consider it a fact—as well-established as the fact […] that the earth is round and not flat”.25 Or consider a more recent example, dark matter. Dark matter was hypothetically introduced in order to explain, among other things, galaxy rotation curves and gravitational lensing. The inference to dark matter is abductive: it is a (hypothetical) entity whose existence is postulated because of its enormous potential to explain a variety of observable effects. As cosmologist Lisa Randall recently put it, cosmologists “believe that [dark matter] is out there because of its manifold gravitational influences. […] We know it exists, but we do not yet know what it is at a fundamental level”.26 Thus, an element of a possible world view may become an element of an adequate world view and, when finally taken for granted, an element of the world. Note that this is a historical description regarding the development of attitudes of scientists, and this is what a historian of science has to describe. Scientists involved in the hole process will probably describe the series of events differently. They experience first the more or less fast shattering of the old worldview, that is the destruction of their certainty regarding the existing objects and their nature. A period of uncertainty will follow, which is finally replaced by the firm and immediate ontological beliefs provided by the new paradigm. Thus, they experience first the transition between the older paradigm state and the new insecurity state, and then the transition from the insecurity state to the new paradigm state; both transitions are dramatic. The do not experience the contrast between the old and the new stable paradigm state, whereas the historian may willfully compare these two states in order to understand what the result of the respective scientific revolution was.

 Mayr (1997), p. 178, similarly on p. 61. - In conversation, Mayr often deplored that he was not aware that philosophers of science have investigated this transition from theory to fact. 26  Randall (2018), pp. S6-S7, my italics; for a survey of dark matter’s history, see Bertone and Hooper (2018). 25

9  The Plausibility of Thomas Kuhn’s Metaphysics

153

9.8 Conclusion The phenomenon Kuhn is struggling with in Chapter X of Structure is this. A scientific community is committed to a certain paradigm with ontological implications, that is, with statements about what the world is like. As long as this paradigm is unquestioned, its ontological implications are taken for granted by most members of the community. These members will unhesitantly refer to the objects that this paradigm describes as existing. Just remember how we refer, for instance, to the fixed stars, or to other galaxies, or to the extinction of the dinosaurs. There is just no doubt that the fixed stars and galaxies are out there, and that there was an extinction event concerning the dinosaurs some 65 million years ago. Clearly, fixed stars, galaxies and the extinction event are parts of our world. If someone countered: yes, in your worldview, there are fixed stars, we would probably object. Stating that something exists in a certain world view at least opens up the possibility that in another worldview, this something does not exist. This is not what we mean by stating that there are fixed stars, other galaxies and extinction event: we mean to say that these things simply are or were out there, independently of our beliefs and worldviews. The phenomenon termed “scientific revolution” that Kuhn wants to describe is the process that leads from one such paradigm to another, incompatible paradigm. This phenomenon is only insufficiently described as a change of worldview, because in this description the change of the immediate ontological commitments is not captured. The description as world change is very odd because it contradicts a fundamental meaning ingredient of our common concept of world, its subject-independence. To make matters even more counter-intuitive, there may be historical transitions from highly hypothetical, even entirely implausible worldviews to what will later unanimously be the world. Given the unfamiliarity of this phenomenon that the non-presentist historiography of science has uncovered, one should not be surprised that we are lacking a language capable of describing this phenomenon in a smooth and immediately plausible way. The plausibility that Kuhn can maximally achieve is the possible end product of a longish and arduous path that not all of his readers are capable and willing to walk.

References Allais, Lucy. 2015. Manifest Reality: Kant’s Idealism and His Realism. 1st ed. Oxford: Oxford UP. Bertone, Gianfranco, and Dan Hooper. 2018. “History of dark matter”. Reviews of Modern Physics 90 (4): 045002. Bird, Alexander. 2012. The Structure of Scientific Revolutions and Its Significance: An Essay Review of the Fiftieth Anniversary Edition. British Journal for the Philosophy of Science 63 (4): 859–883. Godfrey-Smith, Peter. 2003. Theory and Reality: An Introduction to the Philosophy of Science. Chicago: University of Chicago Press.

154

P. Hoyningen-Huene

Hoyningen-Huene, Paul. 1993. Reconstructing Scientific Revolutions: Thomas S.  Kuhn’s Philosophy of Science. Chicago: University of Chicago Press. ———. 2021. The Genealogy of Kuhn’s Metaphysics. In Interpreting Kuhn, ed. K.B. Wray, 9–26. Cambridge: Cambridge University Press. Hoyningen-Huene, Paul, and Eric Oberheim. 2009. Reference, Ontological Replacement and Neo-­ Kantianism: A Reply to Sankey. Studies in History and Philosophy of Science Part A 40 (2): 203–209. Hoyningen-Huene, Paul, Eric Oberheim, and Hanne Andersen. 1996. On Incommensurability. Studies in History and Philosophy of Science 27 (1): 131–141. Hawking, Stephen, and Roger Penrose. 1996. The Nature of Space and Time. Princeton: Princeton University Press. Hawking, Stephen W, and Leonard Mlodinow. 2010. The Grand Design. London: Bantam. Kormos Buchwald, Diana, ed. 2019. Looking Back as We Move Forward. The Past, Present, and Future of the History of Science. Liber Amicorum for Jed Z. Buchwald on his 70th burthday. New York: Ink. Kuhn, Thomas S. 1970 [1962]. The Structure of Scientific Revolutions. 2nd ed. Chicago: University of Chicago Press. ———. 1984, unpublished. Lecture IV -- Conveying the Past to the Present. In Scientific Development and Lexical Change: The Thalheimer Lectures. The Johns Hopkins University, 12 to 19 November 1984. Mayr, Ernst. 1997. This is Biology: The Science of the Living World. Cambridge, Mass: Harvard University Press. Melogno, Pablo, and Hernán Miguel, eds. 2017. Thomas S. Kuhn. Desarrollo científico y cambio de léxico. Montevideo: FIC-Udelar/ANII/SADAF. Randall, Lisa. 2018. “What is dark matter?”. Nature 557 (7704): S6–S7. Sankey, Howard. 1994. The Incommensurability Thesis. Aldershot: Avebury. Schrenk, Markus. 2009. Meaning (Verification Theory). In Encyclopedia of Neuroscience, ed. M.D. Binder, N. Hirokawa, and U. Windhorst, 2253–2256. Berlin: Springer. Uebel, Thomas. 2019. Vienna Circle. In The Stanford Encyclopedia of Philosophy, ed. E. N. Zalta. Metaphysics Research Lab, Stanford University. URL = https://plato.stanford.edu/archives/ spr2019/entries/vienna-­circle/. Worrall, John. 1996 [1989]. “Structural Realism: The Best of Both Worlds?”. In The Philosophy of Science, edited by D. Papineau. Oxford: Oxford University Press, pp. 139–165. (originally in Dialectica 43: 99–124 (1989)).

Chapter 10

Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment to Paul Hoyningen-Huene Hernán Miguel

10.1 Introduction The different problems raised by the terminology “world change” in comparison with “worldview change” in Kuhn’s thesis was a tidal wave reaching the metaphysical shores, as Paul Hoyningen-Huene has addressed in his chapter. Here I present two concerns as we continue unpacking the topics Hoyningen-­ Huene has already been dealing with.1 First, I will address the problem of talking about the world from a non-­ conceptualized framework, or a “paradigm-free basis of phenomena”, as common sense seems to support. This approach has a corresponding second level problem when talking about the scientists’ way of talking about the world. The latter is an unavoidable task in history and philosophy of science. Second, we will examine a way in which world change can be conceived that leaves some features at least partially untouched. Under this conception we can hold onto the remaining threads that seem to be leading us from one world to another across a revolutionary change. Those threads are not strong enough to avoid the incommensurability thesis, but they are necessary for scientists to have the chance to talk about a given parcel of

This article has been written in the framework of the IN401620 Project, UNAM, City of Mexico.  I refer sections and not pages when referring to Hoyningen-Huene’s chapter in this same volume.

1

H. Miguel (*) Argentinian Society of Philosophical Analysis, University of Buenos Aires, Buenos Aires, Argentina e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 L. Giri et al. (eds.), Perspectives on Kuhn, The Western Ontario Series in Philosophy of Science 84, https://doi.org/10.1007/978-3-031-16371-5_10

155

156

H. Miguel

nature. Thus, those scientists might think they are going through a lexicon improvement, instead of realizing they are going through a more dramatic change. This perspective will enable us to estimate how many things have changed and see what others have not, through the lens of a mid-scale conceptualized framework. This framework would not be neutral enough to provide a way to talk about the world regardless of the changes, meaning said changes can be described as a “worldview change”. Similarly, this platform would not be rich enough to give an acceptable description of that parcel of nature as it used to be. I will show that this conception is compatible with the Kuhn’s position on dealing with “worldview change” versus “world change”, and with Hoyningen-Huene’s readings on this topic.

10.2 First Concern: What If the World Doesn’t Change? Hoyningen-Huene analyzes the possible reasons why Kuhn may have begun with the title of chapter X of The Structure of Scientific Revolutions2 (hereinafter, Structure) talking about worldview change, and already in the first paragraph shifts to talking about world change. He is completely right in selecting some passages of Kuhn to point out the problem with not having a platform free of conceptualization from which to see the world. When Kuhn compares the revolutionary change with gestalt change, the lesson is that any kind of switch that changes the way we see the things (and accordingly, the way we talk about them) is similar to another. Hoyningen-Huene underlines that “the subject in the gestalt experiment has perceptual access to both the different gestalts and to the material basis of these gestalts, for instance, the lines on paper.3 By contrast, the scientist lacks any access to the paradigm-free basis of phenomena.” (Chap. 9, this volume, p. 144). Let’s emphasize the second part of the statement. In that same page one can read: The scientist can have no recourse above or beyond what he sees with his eyes and instruments. If there were some higher authority by recourse to which his vision might be shown to have shifted, then that authority would itself become the source of his data […] The same sorts of problems would arise if the scientist could switch back and forth like the subject of the gestalt experiments.4

 Here we use the same edition cited in Hoyningen-Huene’s chapter, namely Kuhn (1970), for the pages to be coincident. 3  Referring to Kuhn (1970, 114): “Aware that nothing in his environment has changed, he directs his attention increasingly not to the figure (duck or rabbit) but to the lines on the paper he is looking at.” 4  Kuhn (1970, 114). 2

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

157

Scientists have no way of seeing even a little part of the world without any subject-­ sided contribution, in Hoyningen-Huene’s terms (1993 and after). That contribution is needed to see any part of the world. So, the conclusion is: nobody can see a part of the world that has kept unchanged without any subject-sided contribution. And, accordingly, if someone thinks that he is seeing some part of the world that seems not to be changing, it is because the subject-sided contribution to see that part of the world has not been affected in the change, or at least in the same magnitude that it is affected when looking at other parts of the world. This difference will be the main concern of the second part of this commentary. The illusion that the world is unchanged out there and the revolutionary change concerns the worldview, can lead to many intuitive ways of talking. But, after more detailed scrutiny, we find ourselves in some perplexing situations not sure where to go, nor how to escape from them. This illusion comes packaged with the promise to keep the reference fixed in the world no matter what is changing in the scientific paradigm, and of course, in the way the scientist sees it. This situation looks like paradise for giving account of the advancement of science in obtaining better and better theories on what there is. From this perspective, accepting a connection between seeing the world and talking about the world is unavoidable. For example, if the meaning of the terms were anchored in their reference, then an argument could support scientific realism by showing that the different theories are referring to the same aspects of a world. Its existence would be independent of those theories, and consequently, we can talk about those things out there by means of the features that remained untouched in changing our theories. In doing so, we can easily appreciate the improvement in the description of the world as a part of a cumulative perspective of the scientific knowledge. It is interesting to observe the problems we will face in trying to do so.5 First of all, one problem is the difficulty of referring to objects in the world from two different platforms: the epistemological and the ontological. Such a distinction tries to separate the discussion on what there is in the world from the discussion on what we know about the world thanks to our best theories. Every theory has an assuming ontology. The theory describes (scientists use it to do that) the world, as if the world were to have such and such things, and these things were correlated, or even governed by some laws that link their properties. It is possible for the scientists (and for us too) to detect some entities, directly perceived, as in the case of stones, the Moon, the Sun, and others that we can see by means of some kind of instrument or reliable method, as in the case of the galaxies, the cells, the presence of a tumor or a pathology, and so on. In Kuhn’s words: “what he sees with his eyes and instruments.” (Kuhn 1970, 114). But theories come along with baggage containing those things that we still cannot detect in the ways mentioned before. Even though, those things will be taken as part of the world given their astonishing ability to explain and predict. These two

 Part of this discussion was my concern in Miguel (2002).

5

158

H. Miguel

tools, explaining and predicting, are really important because they serve as guidelines to action in terms of both what to do to obtain some aimed result, and what to do to avoid some undesirable consequence. When a technician handles a stone containing natural uranium, she knows many things about that, besides the basic and obvious level that that stone contains uranium. First, she knows that she can keep it in her hand without any risk. Second, that it would be possible to bring that stone under a specific technical process to separate the uranium 235 from the 238, to obtain a richer proportion of 235 in that sample. Third, that joining many of these enriched samples would be very harmful for any person near them without special protection. And fourth, that organizing these samples in a certain geometric array, we could control a chain reaction to obtain useful energy. This very short description shows that theories enable much more than just seeing and talking about the world; they also condition how we behave in the world. Kuhn stresses this point saying that (and referred by Hoyningen-Huene): “paradigm changes do cause scientists to see the world of their research-engagement differently. In so far as their only recourse to that world is through what they see and do, we may want to say that after a revolution scientists are responding to a different world.” (Kuhn 1970, 111), and later, “scientists with different paradigms engage in different concrete laboratory manipulations.” (Kuhn 1970, 126). We can see that, when Kuhn talks about “research-engagement differently” or “different concrete laboratory manipulations”, he focuses on the scientific practices from a technical point of view that, of course, is an important part of the behavior of scientists. What I would like to stress in terms of behavior is that which is done to obtain some aimed result as well as that which intends to avoid some other results that are far from desirable. There is a praxis dimension associated with knowledge that has to be underlined because its omnipresence makes it undetectable under first inspection. For instance nobody will try to open the door of a plane during the flight without knowing (and being sure) that he or she will take a mortal risk. And, obviously, that behavior is not the result of personal experience by trial and error, but by means of the way we see the world, and talk about it, which explains how we behave in the world. No technician will look inside a particle collider pipe, no matter whether we can or cannot see the particles… Columbus and all sailors of that time, navigated a certain world and behaved according to their understanding of it. In the same way, we are confident when undergoing chemotherapy or radiotherapy, and making many other decisions only supported by the state of the art involved in being in a world with processes, properties and things that our best theories take as real. What is at stake is that we behave in the world. When the technician handles the uranium stone, we can, and we have to, ascribe some way of seeing the world. When she speaks about that, it is not just an ascribing issue any more. She is talking about the world. But, since the beginning she was behaving in the world, handling things in a certain way, let’s say the A way, and

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

159

never doing it in certain other ways, let’s say B way. Not only that, but even if she were pushed or stimulated to handle the things in the B way, she would always behave in an A way, no matter how much effort required. So, we can say that the subject is avoiding the B courses of action, in the same way we can say that a living being avoids fire. It is a less compromised conjecture than to ascribe a way of seeing the world. She behaves in a world where B courses are avoided, and A courses are promoted. To see, to talk and to behave, are the ways we inhabit the world. Let’s come back to the problem of using the epistemological and ontological platforms to talk about the world. Hoyningen-Huene correctly underlines this situation in relation to some realists: …the (putative) fact that there are shifts of vision would possibly be an interesting epistemological insight, but this would leave ontology untouched. That we perceive the world differently does certainly not imply that the world changes; the world as the object of this perception stays the same, notwithstanding the changing perception.6

And later, he continues This plausibility is based on the view that epistemology and metaphysics are strictly separated: an epistemic change, happening on the side of the epistemic subjects, cannot possibly influence metaphysics, i.e., the purely object-sided that is the focus of our epistemic enterprise.7

At the same time, this two-fold way of talking about objects, properties and processes, seems to require a retrospective view from a current theory towards a theory that has been abandoned or at least improved by the current one. And this is the fundamental reason why the theory change is usually irreversible. When analyzing the theory change, everyone involved in a second level discourse about science, will take the things of the world from some selected framework, since it is impossible for scientists to choose a non-conceptual framework, as we mentioned earlier. Let’s analyze some alternatives.

10.3 Once Upon a Time, a World Let us suppose that one prefers to talk from the perspective of the old framework. One won’t be able to talk about the things the new theory introduced, unless one abandoned one or more constraints of that framework. For instance, to talk about nova star beyond the Moon from the geocentric point of view, implies abandoning the assumption that no change is possible in the superlunary region. But, it is not always easy to identify which constraints should be abandoned. What is worse, in some cases the speaker can take for granted one combination of constraints that allow his assertion, and because of that, they should be abandoned, and, at the same  Hoyningen-Huene (Chap. 9, this volume, p. 148).  Hoyningen-Huene (Chap. 9, this volume, p. 150).

6 7

160

H. Miguel

time, the audience could be thinking of abandoning another set of constraints to make that same assertion meaningful. So, we would often take the risk of entering into an ambiguous way of talking. This is the problem of changing beliefs in light of new empirical data when some part of the new data is not compatible with the old framework. Gathering data can lead, in such cases, to an inconsistency if one adds, without caution, the new data to the whole information available in the earlier stage of the database. The inconsistency comes from the supposed competence of the subjects (or cognitive device) to obtain consequences that are deduced from the combination of any two or more pieces of information available in its storage. This deductive closure threatens every conclusion the speaker or the audience could arrive at, unless a serious warning is raised about what pieces of information should be avoided in the process.8 But, as the reader can foresee, there could be two or more combinations of pieces of information that give support to the same conclusion. So, it would be necessary to make clear which ones the speaker has decided to ban. Aside from this problem—which will continue to be a problem in the next alternatives—it would seem really odd if someone plainly wanted to talk from his framework about things it denies the existence of. This won’t be the case for scientists, unless they want to demonstrate why that supposed thing cannot exist, as in the case of a perpetual motion machine, or a mountain of enriched uranium, or a flat earth. So, if the existence of certain things puts our framework in trouble, it’s really dubious that we choose to talk about those things using our threatened framework. We will use our framework until either these putative things can be separated from the real things, or until we finally give up on defending the framework and accept these things, waving goodbye to our earlier way of talking. Considering not the scientists but the second level discourses, as a reflection about science, I can conceive of some hypothetical instances of this scenario— conversations about a proposed change that doesn’t succeed. For instance, take the Aristarco system, talking about the rotation of Earth in the geocentric era, or the inertia of Galileo rejected by means of the way of seeing and talking of the world we had. Or, in more recent times, the stationary universe where matter is created according to the expansion of space. So, it sounds to me like a hypothetical story—the way the world could have been, though we realized just in time, that it was not.

 For the problem of removing information before letting enter the new one, see Alchourrón and Makinson (1982), Alchourron et al. (1985), and Gärdenfors (1988); and for a comparison of this problem with that of the theory change in a scientific community, see Miguel (2014). 8

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

161

10.4 Navigating Two Worlds If one were to choose to talk about the world from both frameworks (namely I, the older, and II, the newer), one should explicitly mention every time they change frameworks, making it almost impossible to be understood by the audience. For instance, saying that nova star (II) is located in the superlunary region (I). One can end up saying something like this: some object, that exists in II but does not exist in I, has a property, that exists according to I, but that does not exist in II.9 We find a case that serves as a warning to avoid a two world coexistence when Kuhn tells us about the discovery of Uranus: “On at least seventeen different occasions between 1690 and 1781, a number of astronomers, including several of Europe’s most eminent observers, had seen a star in positions that we now suppose must have been occupied at the time by Uranus”. (Kuhn 1970, 115). And, leaving part of the task to the reader in other passages like: “Yet the net result of their experiments was a variety of gas samples and gas properties so elaborate that the phlogiston theory proved increasingly little able to cope with laboratory experience.” (Kuhn 1970, 70). Let’s note that this sentence seems to show the existence of some aspects (or properties) that are not taken into account in phlogiston theory, and at the same time, it is talking about the experiments that were carried out according to that same theory. So, it looks like a mix between the two worlds, and it also has a clear second level discourse function. The reader can imagine the great temptation for historians, philosophers and the like, to talk about the Moon and how it appears to be an existing object in both frameworks. Moreover it preserves its name. So, for the sake of clarity, one must be explicit about which of the two frameworks we are using for every occurrence of the term “Moon” in the text. Unfortunately, the “name’s survival” hides many aspects lost in the transition and makes it harder to see new ones ascribed since that change. So, one common use of this two-sided way of talking takes place when we refer to things, aspects or processes that seem to survive the revolutionary change, as appears to be the case for the Moon, the Sun, and so on. We discuss this kind of “survival” in the second part of this work. Two-sided way of talking also goes a step further when some aspects can be used to keep track of things, suggesting we would be dealing with the same things in some sense, while with many other relevant aspects we would not, suggesting that, strictly speaking, they would not be the same things after all. And of course, these things seem to deserve another name in light of the new perspective. This seems to be the case for the dephlogisticated air and oxygen. For instance, when Kuhn tell us about “the crisis that preceded the emergence of Lavoisier’s  As Mayoral puts it (Chap. 2, p. 15 and Chap. 5, p. 56, this volume): “[…] a quotation from Aristotle or Ptolemy without further remarks—a mere statement—may be understandable from a purely grammatical point of view, but might seem nonsense or a simply false statement for us today.” 9

162

H. Miguel

oxygen theory of combustion” (Kuhn 1970, 69) he crosses the barrier between the the two frames putting it in this terms in a single sentence: “Scheele actually first produced oxygen by an elaborate chain of experiments designed to dephlogisticate heat.” (Kuhn 1970, 70). The problem we face is this: Those believing that something exists that deserves the name “oxygen” do not believe there exists something named “phlogiston” (or its relatives). And conversely, those believing in the latter, do not believe in the former. So, there is no place in no mind where this sentence is strictly acceptable. Strictly speaking, is not possible to understand this sentence if one does not assume, as an implicit premise, that in some way the oxygen corresponds to something related to dephlogisticated stuff. Something similar, and explicit, we find when Kuhn shows that it was impossible for the pendulum regularities to be found in the Aristotelian frame: “Regularities that could not have existed for an Aristotelian (and that are, in fact, nowhere precisely exemplified by nature) were consequences of immediate experience for the man who saw the swinging stone as Galileo did.” (Kuhn 1970, 124). But all of this makes perfectly clear how a second level discourse gives information about the change that occurred in the first level discourse. Sentences of the kind cited for the phlogiston, give us a lot of information about the lexicon change and also reveals clues about correspondences that will allow us to speak of continuities. Both are needed to appreciate the real magnitude of a particular change. Although the analysis of these subtleties fall beyond the scope of this work, it is worthwhile to note that there are many theory changes that seem to happen inside the same big picture, or the same framework, as in the case of Tycho Brahe proposal. That theory introduces some things that are inexistent for the earlier one, namely the movement around the Earth in a “geo-heliocentric” system. Anyway, it seems that this novelty does not dramatically change the things supposed to be out there, at least it wasn’t as revolutionary as the Copernican proposal was. So, the discussion seems to remain open as to whether we can appreciate different grades of changes, and even if we could compare and grade the revolutionary ones against each other.10 Let’s come back to the two-sided way of talking. It is interesting to keep in mind the two uses of language we mentioned before which are unavoidable in Structure. On the one hand, it is necessary to use language to talk about the phenomena. This is the first level function of discourses, used by scientists (referring to things, properties and processes, explaining phenomena, predicting results, setting up experiments, and the like). On the other hand, language is necessary, and sometimes even in the same context, for explaining how the development in science was like, that is, the second level function of discourses, its reflexion function.11 The reflexion

 See Miguel et al. (2003).  Mayoral (Chaps. 2 and 5, this volume, p.) refers to this level as “an extended meta-language” used by historians to do their job. 10 11

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

163

function is used to talk about the way the scientists practice science and refer to the world. It lets us see the position that each scientist takes in the controversy. For instance, Kuhn mentions: “Herschel therefore announced that he had seen a new comet! Only several months later, after fruitless attempts to fit the observed motion to a cometary orbit, did Lexell suggest that the orbit was probably planetary.” (1970, 115), and shortly after “At least that is what seventeenth century observers said they saw, and we have no more reason to doubt their reports of perception than our own.” (1970, 117). In this way, Kuhn is doing two things at the same time, telling us something about what the process was like, and how the scientists were involved in that process in connection with the phenomena. On the contrary, Kuhn is using only the second level discourse when saying “Lavoisier, we said, saw oxygen where Priestley had seen de-phlogisticated air and where others had seen nothing at all.” (1970, 118). Some other times, it is possible to use the second level discourse written as reported speech, bringing the original first level discourse back, without so much load in terms of the two worlds problem: Buridan describes the motion of a vibrating string as one in which impetus is first implanted when the string is struck; the impetus is next consumed in displacing the string against the resistance of its tension; tension then carries the string back, implanting increasing impetus until the mid-point of motion is reached; after that the impetus displaces the string in the opposite direction, again against the string’s tension, and so on in a symmetric process that may continue indefinitely.12

As one could expect, the first level discourse used by scientists (reported above) usually lacks the problem of referring to two worlds, unless they are involved in the second level discourse to compare their practices with those of the ancient scientists, as in Lavoisier’s (1789) Traité élémentaire de chimie: présenté dans un ordre nouveau et d’après les découvertes modernes; avec figures (Vol. 1): Until recently, water has been considered a simple substance and the elders had have no difficulty to qualify it with an element name: it was undoubtedly an elementary substance for them, since they had no way to break it down, or at least because the decompositions of water that operate daily under their eyes, had escaped to their observations…13

Lavoisier is presenting the novelties in the field, so he is comparing the way the different perspectives make us deal with the water as an elementary or composite substance. But, when he describes the “Preparation” for decomposing water, he is using only the first level discourse and the two worlds problem has disappeared: Take an EF glass tube, plate VII, fig. 11, from 8 to 12 lines of diameter, which is passed through a furnace, giving it a slight inclination from E to F. At the upper end E of this tube, we adjust a glass retort A, which contains a quantity of water well-known distilled, & at its end F, a coil SS ‘which adapts in S’ to the neck of a bottle H with two tubes; finally, to one

12 13

 Kuhn (1970, 120).  Lavoisier (1789, 99) my emphasis.

164

H. Miguel

of the two tubes of the bottle is fitted a curved glass tube KK, intended to conduct air-­ forming fluids or gasses in a device suitable for determining the quality & quantity.14

So, it seems that it is possible to keep the scientist discourse free of the problem of referring to two worlds, but, obviously, in these cases the scientists (and also historians or philosophers) cannot give any account of the novelty. Kuhn makes us realize that when looking at passages of this kind, such as the last one, we can say Lavoisier is working and acting in a certain world, different from that where the earlier scientists were behaving. In Hoyningen-Huene words “When Kuhn refers here to contemporary historians, he means to exclude scientists and also contemporary philosophers, because they typically do not occupy the vantage point of contemporary historiography.” (Chap. 9, this volume, p. 682: my emphasis).

10.5 Don’t Forget to Look Back Now, suppose we choose to analyze the change that has occurred, assuming the world composed by things, properties and processes according to what the last theories say (having been those, the more persuading ones). We take as real what is real for us today. Even so, we need to speak about the science (second level discourse) in a way that gives an account of discoveries highlighting the fact that part of what we are taking as existent today was not even imagined before. There was a process to begin to think that these things are real. So, we need to speak of a time when these things were not in our first level discourses, that is, in our scientific descriptions. We want to discuss the way we came to know about some entity that was unknown before. But, after we have discovered it, we think this entity did exist before. We do not think we are bringing this entity into existence if we are talking about discoveries. We can find this kind of natural attitude in science looking at the words of Hoyningen-Huene: In order realize how also we today take many claims of science without the slightest reflection or hesitation for granted, both in science and in everyday life, just consider how we express something like “the dinosaurs went extinct some 65 million years ago” or “the Sun is a star similar to many others”. We do not qualify such statements by “according to our current world view” or something similar, because we believe that these statements express facts, not beliefs. Thus, it is the immediateness of scientists’ reference to existing things and their nature that somehow licenses Kuhn’s odd talk of world change.15

In his Sect. 9.2 Hoyningen-Huene addresses what the scientists take as real, beyond empirical phenomena. In addition [to the observable entities], they take for real entities that cannot be directly observed (for whatever reason) but for whose existence they think overwhelming evidence

14 15

 Lavoisier (1789, 99–100).  Hoyningen-Huene (Chap. 9, this volume, p. 151).

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

165

exists. The evidence that makes their existence overwhelmingly credible consists to a considerable part of empirical phenomena that can best be understood as causal effects of existing unobservable objects.16

And, at the same time, these entities “have a natural home in the pertinent well-­ confirmed theory”. Because these entities have a “natural home” in the pertinent theory, the aforementioned assertion is possible: “a considerable part of empirical phenomena that can best be understood as causal effects of existing unobservable objects”. This attitude was also present in the Kuhn passage already cited above: “…had seen a star in positions that we now suppose must have been occupied at the time by Uranus.” (Kuhn 1970, 115). No one will say that this way of talking makes no sense because Uranus didn’t exist before being discovered. So, we need a way to talk in retrospect, about things that we think exist and had already been detected, and, at the same time, about things that exist but had not been discovered yet (at a certain time). This need will lead us to talk in a way that seems to mix the ontological with the epistemological fields, as already mentioned before. The result is that we talk about things that “have a natural home in the pertinent well-confirmed theory”. So, we are talking about the things we take as real now. The distinctive feature of these objects that serve to discuss discovery processes, conjecture, etc., is that the discussion is possible only after they are postulated in some theory at some moment, even when they always existed or existed prior to this theory. Take dinosaurs for instance. Even when dinosaurs existed before science as we know it, we couldn’t argue about them until their existence was conjectured by some theory (see Miguel 2002). This situation configures an asymmetric way of giving account of the world from the two theories in consideration: the abandoned and the new one. The requisite of information retrieval falls strongly over the latter without affecting the former. New perspectives have to be able to explain why or at least show how it was possible that the world seemed to be as the old perspectives put it. The old ones have no obligation at all in this respect. This asymmetry is present in the change of lexicon. The “challenger lexicon” has to fit some requirements that the earlier “establishment lexicon” does not. Lavoisier seems to need to explain why the scientists until that time were unable to see how the water was decomposing and recomposing constantly in front of their eyes: “It is very extraordinary that it has so far escaped the attentive eye of Physicists & Chemists, & we must conclude that in the sciences as in ethics it is difficult to overcome the prejudices with which we were originally imbued, & to follow another route than the one you are used to walking on.” (Lavoisier 1789, 110). According to this analysis, there is not only a temptation to look back from our new point of view, but a real requirement to give account of how it was possible to see, talk and behave in such a different way, obtaining an acceptable amount of success here and there.

16

 Hoyningen-Huene (Chap. 9, this volume, p. 141).

166

H. Miguel

10.6 Second Concern: Same Moon, Different moon Hoyningen-Huene refers to the continuities and discontinuities in his Sect. 9.2: “If we now compare what scientists from earlier times took for real and what scientists from today take for real, we see besides continuities also discontinuities.”. We want to continue exploring this topic of continuities to inspect how this is possible. Before presenting a way in which one can give account of this characteristic, we will try to explore how much space remains available in Kuhn’s perspective to develop such a description. To do so, we will remember several Kuhn passages relevant to this issue. We will avoid saying something like (as Kuhn denounces): “On this view, Priestley and Lavoisier both saw oxygen, but they interpreted their observations differently;” (Kuhn 1970, 120). Kuhn acutely ascribes this way of talking to “an essential part of a philosophical paradigm initiated by Descartes and developed at the same time as Newtonian dynamics.” (Kuhn 1970, 121) If we do this, we come back to treat the ontology separately from the epistemology, as we have been discussing before. Rather, we want to stress the tension between the possibility of using a common language, with a large part of its meaning shared by scientists of different theories. In this sense, that language is common enough to assure them that they are talking about the same portion of reality, or that they want to explain the same phenomena, although it is not common enough to mean exactly the same for both communities.17 On the one hand, the influence from the paradigm is not so great that terms are only used and understood by one of the parties, and for this reason, they would not allow us to discern from which paradigm these observations have been made. On the other hand, being like this, the fact that the supporters of different theories can use these same terms to make their assertions, is not sufficient to allow us to decide which has an empirical advantage: “Even Copernicus’ more elaborate proposal was neither simpler nor more accurate than Ptolemy’s system. Available observational tests, as we shall see more clearly below, provided no basis for a choice between them.” (Kuhn 1970, 75–76, my emphasis). Our concern is about how it is possible that, after having been warned about all the changes and how the things are different after a world change, scientists keep on using the same words to talk about the same things (in an interesting sense), and to mean different things, strictly speaking. This topic was far from ignored by Kuhn: In Sections X and XIII have argued that the parties to such debates inevitably see differently certain of the experimental or observational situations to which both have recourse. Since the vocabularies in which they discuss such situations consist, however, predominantly of  See how Mayoral points out this problem (Chaps. 2 and 5, this volume, p. 58): “[…] if we take two identical terms that belong to two different languages, e.g., ML1 and ML2, respectively ascribed to languages L1 and L2, both being characterized by the same word (e.g., “motion”), one of them is preserved as an obsolete term within the other’s language –a sort of relic that helps to understand the past– and the other term is the natural-kind term in its own language and is thereby used when stating the laws, expressing counterfactual conditionals, and so on.” 17

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

167

the same terms, they must be attaching some of those terms to nature differently, and their communication is inevitably only partial.18

This subject is clearly addressed by Hoyningen-Huene (1993) showing the great challenge to establish a communication between the two parties when he talks about what sort of obstacles “arise out of the incommensurability of the standpoints involved”: “many of the changed concepts retain the same names as before. In such cases, changes in the meanings of concepts aren’t marked by the employment of new words. And so parties to communication will initially attribute their own respective understanding of such concepts to their counterparts, erroneously.” (Hoyningen-Huene 1993, 255). At the same time, when we remember the words of Lavoisier already cited, we realize he is speaking to both parties in a sense that is completely shared by them: “Take an EF glass tube, plate VII, fig. 11, from 8 to 12 lines of diameter, which are passed through a furnace, giving it a slight inclination from E to F.” (Lavoisier 1789, 99). Kuhn repeatedly discusses the possibility of devising a neutral observable language to show that this will not go far: It might, for example, be conducted in terms of some neutral observation-language, perhaps one designed to conform to the retinal imprints that mediate what the scientist sees. Only in one of these ways can we hope to retrieve a realm in which experience is again stable once and for all—in which the pendulum and constrained fall are not different perceptions but rather different interpretations of the unequivocal data provided by observation of a swinging stone. […] Yet it no longer functions effectively, and the attempts to make it do so through the introduction of a neutral language of observations now seem to me hopeless.19

And, of course, as the enterprise could remain tempting to some philosopher, Kuhn doesn’t discard the possibility that someone may keep on trying it: As for a pure observation-language, perhaps one will yet be devised. […] Psychology supplies a great deal of other evidence to the same effect, and the doubts that derive from it are readily reinforced by the history of attempts to exhibit an actual language of observation. No current attempt to achieve that end has yet come close to a generally applicable language of pure percepts. And those attempts that come closest share one characteristic that strongly reinforces several of this essay’s main theses. […] There can be no question that efforts of this sort are worth pursuing. But their result is a language that—like those employed in the sciences—embodies a host of expectations about nature and fails to function the moment these expectations are violated.20

At the same time he recognizes some way of talking that allows the scientists to share information, but that is not so plainly neutral as to be considered with much enthusiasm: The preceding examples are selected from astronomy because reports of celestial observation are frequently delivered in a vocabulary consisting of relatively pure observation terms. Only in such reports can we hope to find anything like a full parallelism between the  Kuhn (1970, 198) my emphasis.  Kuhn (1970, 125–126). 20  Kuhn (1970, 126–127) my emphasis. 18 19

168

H. Miguel

observations of scientists and those of the psychologist’s experimental subjects. But we need not insist on so full a parallelism, and we have much to gain by relaxing our standard.21

This precaution comes from the fact that proper terms of the theory (oxygen and pendulums) are treated by scientists as “the fundamental ingredients of their immediate experience” (Kuhn 1970, 127). Compared with these objects of perception, both meter stick readings and retinal imprints are elaborate constructs to which experience has direct access only when the scientist, for the special purposes of his research, arranges that one or the other should do so. This is not to suggest that pendulums, for example, are the only things a scientist could possibly see when looking at a swinging stone. (We have already noted that members of another scientific community could see constrained fall.) But it is to suggest that the scientist who looks at a swinging stone can have no experience that is in principle more elementary than seeing a pendulum. The alternative is not some hypothetical “fixed” vision, but vision through another paradigm, one which makes the swinging stone something else.22

Paradigms determine large areas of experience at the same time. It is, however, only after experience has been thus determined that the search for an operational definition or a pure observation-language can begin. The scientist or philosopher who asks what measurements or retinal imprints make the pendulum what it is must already be able to recognize a pendulum when he sees one.23

Given that this train of thought seems to have been well explored, we can try a different way. Let’s go down another road Kuhn has already widely explored, understanding the paradigm as a map for the scientists to know the world they live in: we had principally examined the paradigm’s role as a vehicle for scientific theory. In that role it functions by telling the scientist about the entities that nature does and does not contain and about the ways in which those entities behave. That information provides a map whose details are elucidated by mature scientific research. And since nature is too complex and varied to be explored at random, that map is as essential as observation and experiment to science’s continuing development.24

And shortly after, in the same page: In particular, our most recent examples show that paradigms provide scientists not only with a map but also with some of the directions essential for map-making.25

We can ask ourselves how the paradigm carries out this function of being a map for the scientists, like a chart for a sailor, to guide their actions, their experimental arrangements, and telling where to head for the next step in scientific exploration.

 Kuhn (1970, 126–127) my emphasis.  Ibid (128). 23  Ibid. 24  Kuhn (1970, 109) my emphasis. 25  Ibid. 21 22

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

169

We can pay attention to the terms in which the paradigm carries out this job, that is, its lexicon.26 In this passage, one can find implicitly a part of the answer:27 One way in which such laboratory operations change with paradigms has already been observed repeatedly. But changes of this sort are never total. Whatever he may then see, the scientist after a revolution is still looking at the same world. Furthermore, though he may previously have employed them differently, much of his language and most of his laboratory instruments are still the same as they were before. As a result, postrevolutionary science invariably includes many of the same manipulations, performed with the same instruments and described in the same terms, as its prerevolutionary predecessor. If these enduring manipulations have been changed at all, the change must lie either in their relation to the paradigm or in their concrete results. I now suggest, by the introduction of one last new example, that both these sorts of changes occur. Examining the work of Dalton and his contemporaries, we shall discover that one and the same operation, when it attaches to nature through a different paradigm, can become an index to a quite different aspect of nature’s regularity.28

The passage above shows how Kuhn is interested in the use of the language, before and after a revolution, and the next one says something similar: “different languages impose different structures on the world. Imagine, for a moment, that for each individual a referring terms is a node in a lexical network from which radiate labels for the criteria that he or she uses in identifying the references of the nodal term.” (Kuhn 1982, 682). Hoyningen-Huene notes that Kuhn, since 1982, addresses the problem in terms of a “lexical network”,29 that is, “a particular system of empirical concepts”30 that articulates problems and solutions. In Kuhn (1990) we appreciate in more detail the role a lexicon will play in describing the world:31 “To possess a lexicon, a structured vocabulary, is to have access to the varied set of worlds which that lexicon can be used to describe. Different lexicons-those of different cultures or different historical periods, for example-give access to different sets of possible worlds, largely but never entirely overlapping.” (Kuhn 1990, 300). Let’s turn then to the language the scientists use, in order to explore a proposal about how they can decide whether or not to keep the same term and, in doing so, choose which things will survive the change, and what others won’t.

 Something similar could be analyzed in terms of manipulation processes, but these manipulations sooner or later are treated in the terms of the paradigm, at least in the second level discourse we need to develop here. 27  Beginning with the change in the language, and following with the change in the lexicon, this topic was addressed many times by Kuhn (1962) and ultimately in (2017). 28  Kuhn (1970, 129–130) my emphasis. 29  Kuhn (1982, 52). 30  Hoyningen-Huene (1993, 159). 31  Kuhn (1982, 1990, 2017). 26

170

H. Miguel

10.7 Layers in the Meaning, Layers in the World Much of the effort made clarifying this point of the thesis has been made by Kuhn himself. As we cited in the last section, for a scientist after a revolution “much of his language and most of his laboratory instruments are still the same” and “same manipulations, performed with the same instruments and described in the same terms”. And in his Postscript, several times: both their everyday and most of their scientific world and language are shared.32

Locutions that present no such difficulties may be homophonically translated.33 But each language community can usually produce from the start a few concrete research results that, though describable in sentences understood in the same way by both groups, cannot yet be accounted for by the other community in its own terms.34

Even so, as cited above, “the same operation, when it attaches to nature through a different paradigm, can become an index to a quite different aspect of nature’s regularity.” So, when we see that after a revolution, scientists keep using some of the same terms as before, that behavior does not reveal whether they ascribe those terms with the same relationships they had before with all the other terms in the lexicon in use after the revolution. In The road since structure: philosophical essays,35 introducing what Kuhn will call “local incommensurability” he states: “Most of the terms common to the two theories function the same way in both; their meanings, whatever those may be, are preserved: their translation is simply homophonic. Only for a small subgroup of (usually interdefined) terms and for sentences containing them do problems of translatability arise.” (Kuhn 2000, 36) Incommensurability is about this part of the language that is problematic, but the rest of the shared language is part of the platform to understand one world while standing in the other. We are concerned with a part of the problematic terms, namely those that, being the same term before and after the revolution (like Moon, Sun, Mars, and so on in the case of the Copernican revolution), a substantial part of the way they are “attached to nature through a different paradigm” changes. For instance, Wray put it in these terms: The Periodic Table thus not only organized the elements. It did so in a manner that revealed hitherto unknown features of the structure of the chemical world.36

 Kuhn (1970, 201).  Ibid (202). Referring to the terms that are not the ones that are “used unproblematically within each community, are nevertheless foci of trouble for inter-group discussions.” 34  Ibid (203). 35  Kuhn (2000). “Commensurability, Comparability, Communicability”, the main paper in a symposium of the Philosophy of Science Association, 1982. 36  Wray (2018, 213). 32 33

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

171

Granted, the Periodic Table of Elements retained much the same structure before and after the change of theory. But we should not be misled by this continuity and misunderstand the significance of the change in our chemical understanding.37

Similarly, in Wray and Andersen: With the change in the concept of a chemical element came a change in the conception and understanding of the ordering of the chemical elements in the periodic table.38

It seems to be clear that it is usual to find situations where the lexicon is changing although the terms are not. Let’s begin noting that the term “Moon” is used by both communities to refer to: that body in heavens that has a cycle of phases in about 29.53 days; her phases are correlated with the frequency of tides, something that we will try to emphasize below; it takes part in eclipses when it is aligned with the sun and the earth; its location in the sky follow a precise trajectory, displaced in a known way from the ecliptic. It’s a very notorious body in the sky, more interesting by night, and no matter how two different communities attaches this term to nature through a different paradigm, they are sure they are talking about the same thing, in some sense. They are pointing to the same body in the sky; they are keeping track and calculating when we’ll have the next full moon, and both of them can agree on the Easter date, and so on. Many of these characterizations we were using in the last paragraph, are not affected directly, or at least not in a way that is necessarily to be taken into account, by the change of paradigm.39 Moreover, most of them are characterizations by means of a much more sophisticated knowledge than common-sense (everyday) knowledge. Let’s analyze only one of them in some detail, the topic of the relation between the tides and the Moon, to see how far from common-sense this knowledge is. If this is correct, then, to characterize the Moon by means of this kind of linkages, falls beyond the observable terminology, far from every day knowledge, and so, involves a scientific knowledge shared by the two parties, that is more than expected. The correlation of tides with the phases of the moon, its place in the sky in the moment of high tide or low tide, its pass over the meridian and so on, is part of a sophisticated knowledge acquired along the centuries. First of all, it is rather odd not to realize the correlation, if the tide is some factor to be taken into account, as in the case of mariners for instance, or to locate or build something in the shore. Pierre

 Wray (2018, 216).  Wray and Andersen (2019, 5–6). 39  In his approach to this problem Mayoral (Sect. 5.3, this volume, p. 60) says: “This idea of recognition involves previous acquaintance with a set of properties we know how to recognize and some others that we learn to recognize.” He is concerned about, following Kuhn “[…] a group of characteristics that can be observed in the object in question, which are sufficient for us to pronounce on the object at issue and its kind-membership […]”. These characteristics used to help in the kind membership of the object are properties. We agree on that, and further, we explore different characteristics other than properties that can also be present to do the job. 37 38

172

H. Miguel

Duhem puts it in this terms:40 “an observer could hardly ignore the constant connection which unites the periods of the tide to course of the Moon so he had to be inclined to join the lunar explanation to the [non lunar explanation of tides];” (Duhem 1958, 8). So we can first look at common-sense knowledge but immediately after, we are pushed to accept a much more articulated, planned and detailed way to keep track of the tides, and construct knowledge available to predictions that allow us to behave on lands near the sea. David Cartwright (2001) recovers this passage of Aristotle in De Mundo, chap. 4, 396a: “Many tides and tidal waves are said always to accompany the periods of the Moon at fixed intervals”. Immediately before that passage in the same paragraph, we read these other words: “We find analogous phenomena occurring in the sea. Chasms form in it and its waters often retire or the waves rush in; this is sometimes followed by recoil and sometimes there is merely a forward surge of water, as is said to have occurred at Helice and Bura.” (Lines 17–21). Cartwright also shows how Plinius Secundus registered “… that the tides on the coast of Britain are very large, up to “80 cubits” [about 37 meters]41 in vertical range, and that the ebb and flood follow the rising and setting of the Moon. (Cartwright 2001, 110)” Also, Aëtius reports Pytheas to have said that “the flood advances with the waxing moon and ebbs with the waning moon.” It is doubtful he stayed long enough to check this data, “Most likely, he was told by the natives what they knew from years of experience, and verified for himself only that there were two tides every day and that at a given place successive times of flood occurred later each day, like the Moon.” Moreover, “Eratosthenes further suggested that the twice-daily tides follow the Moon both in its passage through the local meridian and in its reciprocal passage below the horizon 12 ½ hours later. He accepted the notion which Pytheas first suggested, that the flood commences at moonrise.” (Cartwright 2001, 110–111). It is interesting for us to note the relatively low importance of tides in the general knowledge at that time. The salient reports of tides refer to places other than Greece. This is expectable since in Greece the amplitude of tides are below half a meter, and usually around ten centimeters.42 The knowledge available was enough to give account of the Full Moon coincidence for the highest tide in the month and the corresponding lower tides in quadrature (Half Moon). At the same time they registered the solsticial/equinoctial

 Duhem (1958) Chapitre 15: La thèorie des marées, Vol. IX (my translation unless mentioned otherwise). 41  Cartwright notes that this measurement is greater than any known tide amplitude, so he warns us about the way that information could have been taken. 42  Cartwright says: “All historians who touch on the subject state the truism that the peoples of the Mediterranean coasts were slow to appreciate tides on account of their weak local manifestation. Nevertheless, the Greek philosophers produced the first known written references to tides.” (2001, 108) 40

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

173

variation along the year, although the records available today show this relation in reverse. Seleucus of Babylon, known as the sole supporter of the heliocentric theory of Aristarchus, not only made a very precise series of observations showing the difference between inequalities of morning tides and afternoon tides, that he himself described as a strong “irregularity” at the summer and winter solstices, but he also advanced a theory of how the Moon causes the tides.43 Cartwright recovers a passage of the ancient copies to which others have been accessed, showing that, briefly speaking, the ideas of Seleucus go like this: “The Moon’s drag on the outer atmosphere would then create a disturbance which would appear lunar-daily at any terrestrial point and generate ‘waves’ (i.e. a tide) in a large ocean.” (Cartwright 2001, 115). After recognizing that “the theory that tides are depending on the Moon had suffered, in many Scholastics, a kind of decline; other explanations had been proposed which did not invoke any astral cause” (Duhem 1958, 7), Duhem tells us about the recovery of the lunar approach: The total or partial discredit which had struck for a time the lunar theory of the tide came to an end when Christians from the West began to study the Introductorium Magnum in Astronomiam Albumasaris Abalachi after Hermann the Second translated it in 1140. In this treaty, Abou Masar endeavored to relate most of the variations that present the ebb and flow of the sea to the action of the Moon.44

All of this is telling us about a very deep, articulated, and persistently discussed description of the tides in relation to the position, behavior and the action of the Moon, with a greater or lesser amount of speculation or conjecture, beyond, but based on, observable correlations, undoubtedly highly superior to common-sense knowledge. This is a layer of what the scientists see in, and ascribe to, the Moon. It is more than common-sense knowledge about which heavenly body they are referring to, this common-sense being the “first layer of meaning” for the Moon. All these concepts related to tides, are part of something like a “second layer of meaning.” The first layer of meaning is a layer we can associate with the language shared by both communities, and usually serves to understand the other party’s speech. “Most of the terms common to the two theories function the same way in both; their meanings, whatever those may be, are preserved” as we already noted. This is the case for many things like glasses, angle, directions, meridian, ecliptic, essay tubes, scale indicators, and so on. However, it is not the case for the Moon. This second layer functions as a basis for which a term survives a revolutionary change such that “it attaches to nature through a different paradigm”.

43 44

 Ibid (113).  Duhem (1958, 8).

174

H. Miguel

10.8 Give Me the Lexicon and I’ll Give You the Laws The continuities and discontinuities have to be addressed in terms of the language and the lexicon the scientists use to refer to the world. As Hoyningen-Huene points out, Kuhn emphasizes that “recognizing continuity through revolutions has not led historians or anyone else to abandon the notion”.45

Although he also notes that, But in his 1976b,46 we find that Kuhn isn’t satisfied with his previous treatment of the continuities persisting through revolutions. The reason for this dissatisfaction is doubtless that, although he attested to these continuities, he didn’t analyze them in any depth.47

Kuhn warns us that it is not so easy to fit the examples of revolutionary change with those of changes in the core of theories, without any consequences: Also apparent, I take it, are their reasons for suggesting that at least some changes of core correspond to the episodes I have labeled scientific revolutions. As already indicated, I hope and am inclined to believe that claims of this sort can be made out, but in their present form they have an unfortunate air of circularity.48

We follow this warning; so, we will keep it in mind and explore in a different direction. When we come back to the notion of lexicon network with this approach of layers of meaning (whatever this would be), we might presume that scientists use one term, or another, according to the different features connected or associated with that term. Nonetheless, since there are so many features associated with every node, it sounds very odd that all of them have to be checked in advance to be able to mention that node. For instance, when saying “Moon” in scientific speech, it is possible for the speaker and for the audience to adjust their lexicon network to select the node, or to refer to it, without maintaining focus on all the features linked to that node. In doing this, when we mention the Moon, the audience fixes the part of the world we are talking of. They do this by means of several features we share with the audience, and in this way it is possible to talk about the Moon, and even to arrive at a scientific controversy, as we will see below. The lexicon is the way we refer to our world, although it still leaves place for a much more detailed specification, as we can see in this passage of Kuhn: “Different lexicons -those of different cultures or different historical periods, for example- give access to different sets of possible worlds, largely but never entirely overlapping.” (Kuhn 1990, 300).

 Hoyningen-Huene (1993, 222).  Hoyningen-Huene refers to Kuhn, T. S. (1970). Theory-change as structure-change: Comments on the Sneed formalism. Erkenntnis, 179–199. We use in this paper its reprinted Kuhn (1977). 47  Hoyningen-Huene (1993, 222). 48  Kuhn (1977, 298). 45 46

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

175

In a schematic way, we can say that, the node M, has many features associated with it. Some of them are primary properties, others are secondary properties, and others are more complex. There is a great series of assertions that link the node M with others in the lexicon network. Some of the other nodes that enter into connection with M are nodes that correspond to observable language, others to the experimental realm (readings in technical devices such as telescopes, position in the sky, and so on). Finally, some other nodes correspond to the world only when described with the aid of, or from the posit of, theories. When all the theories involved (or connected with M) are shared by all speakers (scientists), we are in a normal science scenario. By contrast, when only some of these theories can be accepted by both parties, and others not, the problem arises of using the same term to refer to something from both sides of the revolution. This second possibility is the one that interests us. So, for the Moon, taken as the M node in the scheme, there is a lot of common-­ sense knowledge plus a more than moderate amount of scientific knowledge shared by scientists, as described in some detail in the earlier section. Now, imagine you have the following features associated to M: a, b, c, … j, k, l, … r, s, t. Suppose a, b, and c are features that belong to the observable or experimental realm; j, k, l belong to the statements several theories ascribe to M, and these theories are shared by all the parties that focus on discussing other theories, namely those that ascribe the features r, s, and t to M. After distinguishing different subsets of features, the question remains about which of them will be taken as defining features (those that have to be present) to identify a particular object as corresponding to that node. This can be made by strategically choosing a few features, enough to identify the class of the object and not so many as to make the identification task hard. Maybe it is recommendable to simply take a, b, and c which use the observable features to identify situations when the node is present, or perhaps it would be better to include some part of the knowledge that is present before and after the change, and so, taking a, b, and k. Or perhaps the node should be associated with the same features affected by the revolution, let’s say by r, s, and t. These different scenarios are present in different nodes. This kind of analysis doesn’t require a non-conceptualized platform as we are not talking about M without using features already present in the lexicon. The only requirement is that all parties make a survey of their own lexicon network to realize which of the features are associated with a chosen node. This approach recognizes its inspiration in Flichman (2001). Briefly, he first states that it is possible to measure force without using Newton’s second principle.49 Yet, this is not to say that the complete meaning of “force” is already furnished. Force is a primitive for the Newtonian theory, but also, it was available before the theory’s construction. The magnitude corresponds to something in the empirical

 This is a very interesting issue, but to go into detail falls beyond the scope of this work. For the sake of the arguments we want to set up here, it is not relevant to accept it or to reject it. 49

176

H. Miguel

realm. So we can establish some bridge principles between the reading in the dynamometer and the magnitude we ascribe to be present in that arrangement. So, force is related to several observable aspects, one for each kind of measurement. Besides that, this other node (we say) “particle”, is linked to others like “force”, “position” and “length”, and in this way its meaning is enriched so it is possible to enter into a nomological relation with many other nodes in the network. According to Flichman there are files containing the meaning of terms. The external file of “particle”, and the same for “force”, for instance, allows us to refer to all the things in the world that fit that meaning furnished before the theory is proposed. The theory will modify that file by means of adding some internal meaning to some of its terms. Nevertheless, the external file identifies its particulars in a complete manner, in the sense that the meaning added, that corresponds to the laws of the theory, does not modify which particulars fall in the file and which do not. Laws of a theory state that the particulars that share the features of, or fall into, the external file, also share an internal feature. This is the main issue we want to focus on. The laws of a theory count with some kind of talking about things in a way that we can reject the laws without rejecting the things. This situation is what is safe in a scientific controversy, and in the case of talking about things and using the same term before and after a revolution. At the same time, Moon node doesn’t connect to, nor is associated with, the same other nodes before and after the revolution. So, in our terms, and in Flichman terms, Moon does not have the same meaning before and after the revolution. This is no obstacle for convincing both parties that they are talking about the same heavenly body and that much of the scientific knowledge about the Moon is not directly affected by the laws of the theories in discussion. It is by means of these threads of meaning that scientists use the same term meaning different things strictly speaking, but they do not lose their bearings by looking at both maps at the same time, that is, the earlier lexicon network and the new one.

10.9 Ways of Lexicon Making Now we can address a collateral issue concerning the way we decide to use these threads of meaning to characterize a node with the fewest features possible, leaving the rest of the knowledge aside to be dependent on the theories, and in this way subject to appear as a discovery. To bring this point home, let’s take the example of the electron. We know, among other things, that the electron has a determined mass, negative charge, and has half-­ integer spin. For the sake of the argument, suppose the electron has only those three features (and their corresponding connections with other nodes in the lexicon network). Suppose now, someone asks what it is like to be an electron and we explain that it has those three features. According to our best theories, those three features

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

177

are only found together in electrons. Nothing else combines those three, and no electron has less than those three. So, we can think that the definition of an electron has to embrace the three features, because the intersection of the three is coextensive with the term “electron”. At the same time, no other particle than electrons shows the first two features, namely its determined mass and negative charge. In arriving at this point it becomes evident that there are at least two ways to define electrons: using the three features to build the definition of the class, or using only the first two. And thus, there are two ways of taking the node corresponding to electrons. So, there are at least two lexicon networks compatible with the available knowledge. But to take both alternatives as eligible would be a great mistake. If we define electrons as having those three features, the electron has a half-­ integer spin by definition. So to say that “electrons have half-integer value for the spin” is analytically true, similar to the case of “every bachelor is not married”. The mistake can arise from the spirit of giving account of electrons in a definition that should be able to combine all the features we know about electrons, and this could come from a sort of factual definition open to being enhanced by the new discoveries, as was expected for the correspondence rules in the old tradition. This seems to be a kind of encyclopedic and condensed knowledge, but it is presented as though it were a definition. Whatever it is, this would be a great mistake because it blurs the difference between identifying things and discovering a correlation that counts as a law for some theory. For some particle to be an electron, it is not necessary that it fits the three features. Every particle that fits the first two, is an electron. So, the two first features are enough to identify electrons. As we mentioned before, a few features are enough to identify them and are not adding any additional redundant task. So, the scientists will reunite some of these features in a complete subset that shall constitute the definition. We say “complete” in the sense that using fewer features would be deficient for determining whether an entity is an electron, and adding some other one, wouldn’t make any difference. After looking at that collection characterized by the first two features (collection of electrons and only electrons), we realize that all those things in that collection, besides the two features that serve us to identify them as belonging to the same family (the same node is applied to them), all of them also show the third feature. So, we find a correlation in the way of “for every x, if x has the properties A and B, x has the property C”. That correlation can be understood to say something empirically interesting, because the antecedent fix the family of electrons, and the consequent tells the novelty, a novelty that was not present in the definition. It is worthwhile to note that each feature present in electrons is the basis for its corresponding causal power. Namely, its mass will be responsible for the momentum the electron can carry and transfer in striking another particle. Its negative charge will be responsible for the direction it will be attracted or repelled in the presence of an electrostatic field, and its spin value will be responsible for the electron behavior in electromagnetic interactions in atoms. Each one of these features are associated with causal powers of the electron that can be tested independently.

178

H. Miguel

All of this is behind the words of Hoyningen-Huene when he tells us about how the scientists take something as real: “The evidence that makes their existence overwhelmingly credible consists to a considerable part of empirical phenomena that can best be understood as causal effects of existing unobservable objects.” (Chap. 9, this volume, p. 141) (my emphasis). Now we can ask the question: Can you choose the definition in a different way? for instance, can you take the second and the third features instead of the first and the second to construct the definition? If this were possible, the laws would be different. For the case of electrons, in the actual state of scientific knowledge, the answer is no. If we try to avoid the first feature (its mass) the other two features cannot distinguish electrons from muons. If we avoid the second feature (its charge) the two remaining features cannot distinguish electrons from positrons. But this result is far from being general. If we take the discussion about how to choose the physical geometry of the universe together with the physics behavior of solids in the context of the theory of relativity we face this same problem. We live in an Euclidean space and the bars in that space contract in the direction of the movement, or we live in a non-Euclidean space with rigid bars. Worse, take the way to identify squares.50 One can identify squares among other polygons in an Euclidean space, in several different ways. One of them is telling that they are figures with four equal sides (same longitude), and all interior angles are right ones. Another way can be to tell that it is a four-sided shape that has two equal sides forming an angle, and that the diagonal connecting the ends of these two sides opposite to that angle has the same length as that of the side times the root of two. There are many other ways to identify squares. So, if we repeat the question of whether you can choose more than one way to construct your lexicon network, this time the answer is “yes, you can”. So, we face the problem of how scientists pick up a few features to constitute the definition, taking into account that there is more than one complete subset of them that is able to do the identification job. The issues addressed in this section help to clarify how huge the task the scientists have is when they have been involved in a change of lexicon network during a revolutionary change. In Kuhn’s words: In short, dubbing and the procedures that accompany it ordinarily do more than place the dubbed object together with other members of its kind. They also locate it with respect to other kinds, placing it not simply within a taxonomic category but within a taxonomic system. Only while that system endures do the names of the kinds it categorizes designate rigidly. A lexicon that embodies such interrelationships between terms necessarily also embodies knowledge of the world those terms can be used to describe, and that knowledge may be placed at risk.51

50 51

 I am grateful to Miguel Fuentes for suggesting that I treat this case as an example.  Kuhn (1990, 315).

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

179

Part of the problem is that the lexicon network construction cannot be taken as a set of alternative ways of talking about the world. The new lexicon stabilized after the transition, is not any more a way of talking about the world. It is the way the scientists refer to the world they live in, in terms of features which the things are made of, and these features are associated with the causal powers the things have in the world. In that kind of process of stabilization of the network, scientists can abandon the complete subset of features that were taken earlier as a definition of a node, and embrace a, partially or entirely, new one. When it is partially new, some of the features that were in the complete subset used to constitute the definition in the past, are not present in the new subset, but others remain as members of this new complete recently accepted subset. This can be the case for some nodes that appear to be the same with respect to some features that have survived the change of its definition, while others have changed. This seems to have been the case for the Moon in the Copernican revolution. As we can foresee, when scientists don’t change the complete subset of features in the definition for some node, that situation doesn’t mean that the node remains the same entirely in terms of its meaning. Some of the other links that are not involved in the features of this particular subset, may have been changed. So, it is not so easy to know ahead of time which nodes count as continuity because of their relative independence from the revolution or because its complete subset of features involved in its definition were not changed. The most salient paradox is to keep using the same term for the cases for which the complete subset of defining features has changed entirely. But the paradox can be softened by looking at the freedom in choosing one of several complete subsets of features to perform the job of defining or identifying. This could be the case for the node “planet”. Take for instance the notion of a planet associated with that of being circling around the earth and compare it with that of circling around the sun, supposing that these two notions are enough to identify planets before and after the revolution, respectively.52 In this way, much more discussion can be unwinded concerning the use of the same terms, even when they are attached differently to nature through a different paradigm. Part of this attaching operation is achieved by means of the features we take as real for every particular object to belong to or correspond to a node in our lexicon network. And, remember that each of these features is tributary for certain causal powers these particular objects have in our world. To close the loop, our behavior in the world is strongly constrained by the causal powers of the things that surround us. So, we live in a world and we interact with the things of the world in that these causal powers rule the interaction. It is not only a matter of seeing the world and talking about it, but also, and mainly, it is a matter of conducting ourselves in the world.  This could also be the case for “pure plant” for some characteristic that before the revolutionary change in the theory of inheritance was identified by means of its lineage, and after, in terms of having two same alleles for that characteristic. 52

180

H. Miguel

10.10 Conclusion In his chapter, Hoyningen-Huene analyzes the plausibility for metaphysics for the Kuhn expression, and position, referred to as “world-change”. As he already mentioned, what is taken as real for the scientists is by means of its causal powers. So, causal powers appear to be one of the clues that we can follow to get to the world, starting with how we talk about and see the world. The comparison between the gestalt change and the revolutionary change has been widely recognized to be a limited metaphor by Kuhn himself, and Hoyningen-­ Huene addresses this when he analyzes the issue in light of his proposal of subject-­ sided contributions. The limitation arises in two different ways. One of them because there is no neutral platform to access the world as in the case of the drawing that is perceived as one thing or another, alternatively. Second, because while the gestalt change can be carried out back and forth at will, the revolutionary change is a one way trip. This asymmetry was found to play some role in the way the second worldview is imposed over the first one in giving account of the phenomena and at the same time to help to build the second level discourse, aimed at describing the scientific practices and results, but from outside the science. The topic of world-change should always be addressed with an eye on the behavior of scientists, and even on that of non-scientists. The behavior takes place in a world, not preceded by assertions such as “supposing the world were to be like this”. The relevant behavior is not only about how the scientists engage in their research. It goes beyond that and expands to all the decisions each person has to make in the world, evaluating the risks, the best strategies to reach a goal and so on. So, the behavior comes to join the seeing of the world and talking about the world to show the way we inhabit the world. Hoyningen-Huene has highlighted the problem of treating the ontological and the epistemological separately. We agree with that and add that in such scenario, the asymmetry of the two worlds pushes us to prefer to describe the things from the newer platform instead of doing it from the earlier one. So, that scenario does not seem to bring any good news at all. The lexicon network is made of nodes, and each node has some features used to characterize it. As Hoyningen-Huene has mentioned, there are continuities and discontinuities in every revolutionary change. When facing a transition through a revolutionary change, some nodes suffer a modification even when their names (the terms) may persist. So, there is a more complex continuity besides just that of the basic and observable terms that were not affected by the revolution. There is a different kind of continuity when the same term is attached differently to nature through different paradigms. That kind of continuity seems to be related with the persistence of a certain amount of knowledge associated with the nodes whose names have survived the revolution. That knowledge greatly exceeds common-­ sense knowledge and can express very articulated and sophisticated results of the earlier research in areas that are not directly connected with those of the revolution that it has undergone.

10  Seeing, Talking and Behaving… Ways of Inhabiting the World: A Comment…

181

Finally, the construction of a lexicon network enables analyzing the way the scientists have to decide the definition of its nodes, what part of the knowledge takes the shape of a law in some theory and in this way adjusts the whole knowledge so it can be put into words that describe the world we live in.

References Alchourron, Carlos E., and David Makinson. 1982. On the Logic of Theory Change: Contraction Functions and Their Associated Revision Functions. Theoria 48 (1): 14–37. Alchourrón, Carlos E., Peter Gärdenfors, and David Makinson. 1985. On the Logic of Theory Change: Partial Meet Contraction and Revision Functions. The Journal of Symbolic Logic 50 (2): 510–530. Cartwright, David. 2001. On the Origins of Knowledge of the Sea Tides from Antiquity to the Thirteenth Century. Earth Sciences History 20 (2): 105–126. Duhem, Pierre. 1958. Le système du monde: histoire des doctrines cosmologiques de Platon à Copernic. Vol. 5. Paris: Hermann. Flichman, Eduardo H. 2001. Newton’s Dynamics, Kuhn, and Incommensurability. The Proceedings of the Twentieth World Congress of Philosophy 10: 89–96. Gärdenfors, Peter. 1988. Knowledge in Flux: Modeling the Dynamics of Epistemic States. Cambridge, MA: The MIT Press. Hoyningen-Huene, Paul. 1993. Reconstructing Scientific Revolutions: Thomas S.  Kuhn’s Philosophy of Science. Chicago: University of Chicago Press. Kuhn, Thomas S. 1970. The Structure of Scientific Revolutions. 2nd ed. Chicago: University of Chicago Press. First published 1962. ———. 1977. Theory-Change as Structure-Change: Comments on the Sneed Formalism. In Historical and Philosophical Dimensions of Logic, Methodology and Philosophy of Science, 289–309. Dordrecht: Springer. ———. 1982. Commensurability, comparability, communicability. In PSA: Proceedings of the biennial meeting of the Philosophy of Science Association 1982 (2): 668–688. ———. 1990. Dubbing and Redubbing: The Vulnerability of Rigid Designation. Scientific Theories 14: 298–318. ———. 2000. The Road Since Structure: Philosophical Essays, 1970–1993, ed. James Conant, y J. Haugeland. Chicago: University of Chicago Press. ———. 2017. In Thomas S. Kuhn. Desarrollo científico y cambio de léxico, ed. Pablo Melogno and Hernán Miguel. Montevideo: FIC-Udelar/ANII/SADAF. Lavoisier, Antoine-Laurent de. 1789. Traité élémentaire de chimie: présenté dans un ordre nouveau et d’après les découvertes modernes; avec figures. Vol. 1. Paris: Cruchet. Miguel, Hernán. 2002. Ontología vs epistemología en “El significado de ‘significado”. In Filosofía e Historia de la Ciencia en el Cono Sur, Selección de trabajos del II Encuentro de Filosofia e Historia de la Ciencia del Conosur, ed. Pablo Lorenzano and Fernando Tula Molina, 135–144. Bernal: UNQ. ———. 2014. Comparando cambios de creencias: inteligencia artificial, comunidad científica y enseñanza de las ciencias. In Cambio conceptual y elección de teorías, ed. Pablo Melogno, 355–372. Montevideo: FIC-UDELAR. Miguel, Hernán, Jorge Paruelo, and Guillermo Pissinis. 2003. Las salvedades (provisos) y la magnitud del cambio teórico (Provisos and the Magnitude of Scientific Change). Crítica: Revista Hispanoamericana de Filosofía 34 (101): 43–71. Wray, K.  Brad. 2018. The Atomic Number Revolution in Chemistry: A Kuhnian Analysis. Foundations of Chemistry 20 (3): 209–217. Wray, K.  Brad, and Line Edslev Andersen. 2019. Reporting the Discovery of New Chemical Elements: Working in Different Worlds, Only 25 Years Apart. Foundations of Chemistry 22 (2): 1–10.

Index

A Abbott, A., 33 Adaptive plasticity, 67 Alchourrón, C.E., 160 Allais, L., 150 Andersen, H., 57, 60, 71, 127, 171 Annales School, 49 Anomalies, 4, 6, 7, 32–35, 44, 46, 47, 57, 96, 108 Antirealism, 131, 134 Arabatzis, T., 42, 45, 49 Aristarchos of Samos, 68 Aristotle, 18, 45, 56, 58, 70, 71, 74, 95, 116, 130–133, 172 Austin, J.L., 18, 21 Auxiliary assumptions, 106, 107, 113 B Barker, P., 57, 60, 71 Behavioral world, 68 Belief change, 3, 55, 69, 72, 76, 77 Bell, J.S., 75 Bilingualism, 98, 108, 118, 121, 123 Bird, A., 9, 34, 37, 42, 43, 46, 49, 140 Birner, J., 35 Bishop, M.A., 69 Black-body, 27 BonJour, L., 4–6, 12, 15, 16, 19, 20, 22 Boyd, R., 94 Braudel, F., 45–47, 49 Bridgeheads, 70 Buchwald, J.Z., 58 Burke, E., 45–47

C Carnap, R., 30 Cartwright, D., 172, 173 Causal effects, ix, 141, 165, 178 Causal powers, 177, 179, 180 Causal theory of reference, 55, 62, 63, 67, 84, 132, 133 Chang, H., 17, 48 Chen, X., 57, 60, 71 Cognitive agents, viii, 83–91 Cognitive values, 85, 90, 91 Coherentism, vi, vii, 1–13, 15–23 Cole, J.R., 35 Collingwood, R.G., 35 Common-sense knowledge, 172, 173, 175, 180 Common-sense realism, 149 Common-sense truth, 105 Comparability, 129 Compatibility account, viii, 85–88, 90 Comte, A., 45, 46 Conant, J.B., 30, 34, 68 Conceptual change, 49, 63, 65, 98, 103, 104, 117, 122, 133, 134 Conceptual schemes, 55, 68 Consensus, 33, 46, 48, 75, 85, 96, 135 Content comparison, 110, 111, 130, 132 Context of discovery, 85, 89–91 Context of justification, 85, 89 Continuist perspective, 71 Copernicus, N., 47, 63, 68, 71, 166 Corroboration, 107, 109, 124 Counterfactual conditionals, 58, 166 Covert classifications, 100

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 L. Giri et al. (eds.), Perspectives on Kuhn, The Western Ontario Series in Philosophy of Science 84, https://doi.org/10.1007/978-3-031-16371-5

183

184 Crisis, 6, 32, 42, 46, 47, 96, 106, 108, 109, 117, 120, 122, 135, 161 Crucial experiments, 46, 109, 110, 113, 116, 121, 133 D Darwin, C., 27 Daston, L., vii, 25, 26, 28–30, 33, 34, 36, 41–48, 78 Dear, P., 29, 30 De Caro, M., 78 Descartes, R., 9, 10, 17, 147, 166 Descriptive theory of reference, 110, 111 Determinism, 46 Developmental Phase, 56, 83 Developmental theory, 54 Dilthey, W., 35 Disciplinary matrix, 96 Dobbs, B.J.T., 42 Duhem, P., 93, 109, 117, 172, 173 Durkheim, E., 32 E Effability, 73, 75, 84 Einstein, A., viii, 47, 93, 99, 107, 108, 115, 116, 133 Empirical measurements, 107, 109 Empirical realities, 115, 116, 119 Empiricism, 2, 97, 129 Epistemic values, 85, 107 Epistemology, vii, ix, 1, 2, 4–6, 10, 11, 13, 15–19, 23, 76, 78, 124, 159, 166 Erismann, T., 29 Ethnographic Phase, 56, 87 Everyday language, 105 Evolutionary explanation, 91 Exemplars, 96, 99 External world, 6–8, 13, 68 Extraordinary science, 37 F Facts, ix, 7–9, 26, 27, 29, 31, 32, 35, 43, 45, 49, 59, 61, 66, 84, 85, 87, 88, 90, 94, 95, 97, 105–107, 109, 110, 112–124, 128, 130–135, 140, 146, 148, 149, 159, 162, 164, 166, 168 Feature spaces, 60–62, 66, 74, 84, 88 Feyerabend, P.K., viii, ix, 3, 93, 94, 96–100, 103–105, 109–111, 113, 114, 116–118, 120, 122, 123, 127–136

Index Fleck, L., 35, 67, 68 Flichman, E.H., 175, 176 Foundationalism, vii, 4, 5, 8, 9, 12, 13, 17, 18, 23, 75 Fuller, S., 42 G Galileo Galilei, 160, 162 Galison, P., 28, 29, 69 Gärdenfors, P., 160 Gestalt, ix, 29, 68, 132, 144, 147, 148, 156, 180 Gillispie, C.C., 27 Godfrey-Smith, P., 140 Goldstein, D.G., 90 Goodman, N., 58 Greek astronomy, 62 H Haack, S., 4 Hacking, I., 20, 31, 34, 36, 57, 64 Hamilton, A., 19 Hayek, F.A., 37 Hegel, G.W.F., 46 Hegelianism, 41, 49 Heilbron, J.L., 30 Hempel, C.G., 2, 56 Historicism, 35–37, 43, 46 Historiography, 37, 43–48, 72, 140, 143, 153, 164 History of science, v–vii, 2, 17, 25–30, 34, 36, 37, 41–49, 54, 63, 85, 97, 98, 108, 111, 131, 132, 144–148 Holism, 65, 78, 84 Hollis, M., 70 Hoyningen-Huene, P., vi, viii, ix, 7–9, 13, 18, 43, 49, 59, 64, 65, 67, 72, 75, 77, 85, 89, 90, 127, 128, 130, 133, 135, 136, 139–153, 155–181 Hufbauer, K., 32 Hypothetico-deductive method, 97, 112, 123 I Incommensurability, v, vi, viii, ix, 30, 57, 69–71, 84, 93–124, 127–136, 139, 155, 167, 170 Incomparability, 129, 131 Indirect refutation, viii, 93–124, 133 Individual interpreter, 87, 88, 90, 91

Index Ineffabilities, 98, 99, 101–103, 117, 119, 122, 133, 134 Inference to the best explanation, 117, 131–133, 136 Innovation context, 86 Input objection, vii, 1, 5–8, 13 Isaac, J., 26, 28 Isolation objection, 5 J Justificationism, 87 K Kahneman, D., 90 Kaiser, D., 27, 29 Kant-on-wheels realism, ix, 114, 118–122, 124, 136 Katz, J.J., 73, 75 Kind change, viii, 93–124 Kind terms, 21, 55, 58, 60, 62, 63, 67, 68, 99–102, 111, 112 Kindi, V., 44 Kitcher, P., 48, 49, 69 Koffka, K., 68 Kormos Buchwald, D., 140 Koyré, A., 131 Kraft, V., 96 Kripke, S., 55, 63, 65 Kuhn-loss, 37 Kuukkanen, J.-M., 1–7, 12, 15, 16, 22, 23 L Ladyman, J., 31 Lakatos, I., 25, 36 Language of observations, 167 Laudan, L., 48 Lavoisier, A.-L. de, 47, 145, 161, 163–167 Laws, 41, 45, 46, 58, 74, 87, 97, 99–102, 104, 105, 112, 118, 131, 157, 174–178, 181 Lewens, T., 77 Lewis, C.I., 16–18, 74 Lexical changes, vi, 56, 59, 63, 64, 71, 76, 77, 84 Lexical network, ix, 169 See also Lexical Structure Lexical structures, 62, 66, 69, 70, 84, 101, 102, 108, 119 Lexical taxonomies, 64, 65, 99–102, 104–106, 112, 116–123 Lexical theory, 54, 55, 71

185 Lexicons, vii, 21, 55, 57–66, 68–78, 83, 84, 86, 88–89, 99, 103, 140, 156, 162, 165, 169–171, 174–181 See also Lexical structure Lifelines, 63 Linguistic recategorization, 65 Linguistic relativism, 68 Linguistic turn, 54 Lipton, P., 128 Little, D., 31 Locally holistic learning, 89 Locke, J., 9 Longomontano, C.S., 134 Longue durée, 45 Lowell Lectures, vii, 54, 66 Lyons, J.C., 11 M Mach, E., 45 Marcum, J.A., 68, 77 Marx, K., 46 Marxism, 49 Masterman, M., 25 Mayoral, J.V., vi–viii, 15–23, 53–78, 83–91, 161, 162, 166, 171 Mayr, E., 152 McMullin, E., 25 Meaning, 9, 18, 41, 55, 58–61, 65–67, 84, 89, 95, 98, 100, 103, 104, 107, 111, 118, 131, 135, 139, 146, 149, 153, 156, 157, 166, 167, 170–176, 179 Meaning change, 65, 95, 114 Melogno, P., vi, vii, 19, 41–49, 56, 57 Merton, R.K., 27, 31, 34, 35 Mesthene, E., 65 Meta-incommensurability, viii, ix, 94, 118–124, 134–136 Metaphor, 63, 85, 89, 140, 180 Metaphysics, ix, 131–136, 139–153, 159, 180 Meta-theories, viii, 121, 122 Methodological incommensurability, 99 Methodology, 124 Miguel, H., vi, ix, 155–181 Mill, J.S., 58 Mind-independent reality, 6, 115, 119 Mladenović, B., 43, 47 N Nagel, E., 114 Newell, A., 90 Newton, I., 7, 27, 47, 58, 70, 71, 108, 116, 118, 175

186 No-overlap principle, 58, 83–84, 98–104, 106, 107, 114, 115, 117–123, 133, 134 Normal research, 22, 57 See also Normal science Normal science, v, vi, 6, 21, 25, 29, 32, 36, 42, 67, 86, 95–97, 99, 104, 106, 108, 112, 113, 116, 139, 175 O Oberheim, E., vi, viii, ix, 75, 93–124, 127–136 Objective reality, 115 Observation statements, 106, 111–115, 123 Ontology, viii, 116, 117, 124, 157, 159, 166 P Paradigms, v, vi, ix, 6–12, 25, 33, 34, 42, 44, 46, 47, 59, 61, 65, 67, 95–98, 103, 107, 108, 115, 117, 118, 139, 140, 142–145, 147, 148, 150, 153, 157, 158, 166, 168–171, 173, 179, 180 Paradigm-shift, 10, 95 Paruelo, J, 160 Perception, vi, 1–13, 15, 17, 45, 48, 54, 64, 85, 144, 145, 147, 148, 159, 163, 167, 168 Persuasion, 107, 108, 118, 123, 142 Pessimistic meta-induction, 120 Petitio principii fallacy, 131, 134 Phenomenal worlds, 7, 9, 22, 64, 68, 115 Philosophy of science, v–vii, 1, 12, 23, 26, 28–35, 37, 43, 44, 48, 49, 54, 55, 57, 76, 77, 86, 87, 96, 98, 118, 121, 122, 127, 136, 155, 170 Pinch, T.J., 28 Pinto de Oliveira, J.C., 43, 53 Pissinis, G., 160 Planck, M., 27, 28, 57 Pluralism, 97, 122, 136 Plurality of Worlds, 54, 77 Poincaré, H., 33 Polya, G., 90 Popper, K., 25, 33, 35–37, 93, 96, 97, 99, 109 Postman, L., 29 Predictions, 7, 36, 46, 47, 106, 107, 109, 110, 113, 172 Presentism, 143, 144 See also Whig history Priestley, J., 47, 145, 163, 166 Principle of economy, ix, 145, 146, 148 Problem-solving activity, 3, 86 Projectability, 59, 84 Propositionalism, 87

Index Psychological world, 68 Ptolemy, 47, 56, 166 Putnam, H., 55, 63, 65 Puzzle-solving, 37 Q Quantum mechanics, 27–28, 53 Quine, W.V.O., 65 R Randall, L., 152 Rational individual, viii, 87, 90, 91 Read, R., 43 Reference, 17, 22, 59, 61, 62, 64, 69, 72, 74, 84, 98, 111, 131, 148, 149, 157, 164, 169 Refutation, 107, 109, 113 Reisch, G.A., 42, 54, 64 Relativism, 30, 54, 55, 73–76, 78, 84, 105 Reynolds, A., 35 Richardson, A., 30 Rosenstock-Huessy, E., 64 S Sahlins, M.D., 25 Sankey, H., vi, viii, 1–13, 15–23, 57, 67, 70, 74, 94, 110–113, 120, 121, 127–136 Sayability, 73, 75, 84 Scheffler, I., 25 Schrenk, M., 146 Scientific change, vii, 7, 25, 33, 35, 36, 46, 54, 67, 68, 150 Scientific community, 35, 44, 46, 47, 96, 99, 168 Scientific controversies, 132, 174, 176 Scientific progress, v, 2, 15, 32, 75, 94–96, 99, 103, 104, 110, 114, 116–118, 122, 123, 130, 133 Scientific realism, viii, ix, 7, 31, 94, 110, 114, 115, 118–124, 128, 129, 132–136, 149, 157 Scientific revolution, v–vii, 15, 16, 25–38, 41–49, 53, 54, 63–65, 70, 76, 77, 83, 95–98, 107, 109, 113–118, 128, 133, 135, 139, 141, 144, 150, 153, 156, 174 Scientific truth, 105 Sellars, W., 8–11, 13, 16, 17 ŠeŠelja, D., 2 Shapere, D., 25, 71 Shapin, S., 42

Index Sharrock, W., 43 Shearman Lectures, vii, 76 Shearmur, J., 35, 36 Simon, H.A., 90 Skepticism, 17 Social construction, 31 Sociology, v, 28, 31, 33–35 Sociology of science, 34, 35 Solipsism, 23 Sosa, E., 12 Stopes-Roe, H., 25 Straßer, C., 2 Structural realism, 31, 149 Structure, v–viii, 1, 3, 8–11, 15–17, 21, 25–38, 41–49, 53–78, 83, 89, 93, 95–99, 102–104, 107, 108, 110, 113, 115, 139–143, 146, 153, 156, 162, 169–171, 174 Subject-sided moments, 115, 119 Sutton, F., 35 T Taxonomic incommensurability, 99 Thalheimer Lectures, vi, 21, 23, 139 Theory change, 8, 25, 37, 76, 86, 159, 162, 174 Theory choice, 3, 19, 46, 76, 85, 91, 97, 107, 122, 123 Theory comparison, 93, 94, 106–113, 121, 123, 134 Theory of relativity, 141, 178 Toulmin, S.E., 48 Toynbee, A.J., 45–47, 49 Translation, 35, 68–70, 98, 107, 170

187 Truth, viii, 1, 2, 12, 15, 19, 61, 63, 68, 69, 72–76, 78, 84, 94, 95, 102, 105, 115–117, 119, 120, 122, 134 Truth-preserving translation, 69–71, 74, 75, 84 Tversky, A., 90 U Uebel, T., 146 Underdetermination, 109, 113, 123 Unobservable, ix, 94, 141, 165, 178 Untranslatability, 98, 99 V Van Vleck, J.H., 26, 27 Vico, G., 35 Volta, A., 57, 60 W Watkins, J.W.N., 25 Whewell, W., 45 Whig history, 69 Whorf, B.L., 68 Williams, M., 16, 20 Wittgenstein, L., vii, 15, 18–23, 60, 67, 74 World change, viii, ix, 64, 72, 93–124, 139–146, 148–150, 153, 155, 156, 159, 164, 166, 180 World-in-itself, 7–9, 64 Worldview, ix, 100, 117, 150, 153, 155–157, 180 Worrall, J., 31 Wray, K.B., vi, vii, 25–37, 41–49, 170