The methods of metaphilosophy. Kant, Maimon, and Schelling on how to philosophize about philosophy [1. ed.] 9783465045984

Table of contents :
Front Cover
Impressum
Contents
0 Introduction
0.1 Metaphilosophy
0.2 A philosophy designed to be metaphilosophy-first
0.3 Scientific tools for metaphilosophy: finding the right procedure
0.3.1 Baconian and Newtonian experimentalism
0.3.2 Galilean idealisations
0.4 Overview
1 Kant’s propaedeutic method
1.1 Propaedeutic philosophy is the study of philosophy
1.2 A two-layered analogy: Baconian and Newtonian experimentalism, and chemical analysis and synthesis
1.3 From empirical experimentation to a priori experimentation
1.4 Kant’s conception of the philosophical experiment
1.5 Chemical analysis and synthesis
1.6 Philosophical procedure in the Aesthetic/Analytic and the Dialectic
1.7 Metaphilosophy as experimentalist practice
2 Maimon’s method of fictions
2.1 Philosophy is the science of the form of all sciences
2.2 Against propaedeutic philosophy: quid facti? and quid juris?
2.2.1 Quid facti?
2.2.2 Quid juris?
2.3 A coalition system: Maimon’s rational dogmatism and empirical scepticism
2.4 The analogy to calculus
2.4.1 Maimon’s philosophy of mathematics
2.4.2 An immanent account of cognition
2.5 The method of fictions
2.6 Philosophical fictions
2.7 Maimon’s metaphilosophy: philosophy as modelling practice
3 Schelling’s method of nature-construction
3.1 Philosophy is the science of the unconditioned
3.2 Fichte’s original insight, Schelling’s observation, and Schelling’s point
3.3 "Naturphilosophie" as science of the unconditioned
3.4 Positing an absolute hypothesis
3.5 The Method of Nature-Construction
3.6 Presenting nature through experiment
3.7 Metaphilosophy as constructive and experimental practice
4 Experiments of reason
4.1 Kant’s experiments of pure reason
4.2 Maimon’s metaphysical modelling
4.3 Schelling’s experimental constructions
4.4 Conclusion
Acknowledgements
References
Abbreviations
Other texts

Citation preview

Jelscha Schmid · The methods of metaphilosophy

Herausgegeben von Gerald Hartung und Alexander Schnell in Zusammenarbeit mit Andrea Esser (Jena) Anne Eusterschulte (Berlin) Rahel Jaeggi (Berlin) Rainer Schäfer (Bonn) Philipp Schwab (Freiburg)

KlostermannWeißeReihe

Jelscha Schmid

The methods of metaphilosophy Kant, Maimon, and Schelling on how to philosophize about philosophy

KlostermannWeißeReihe

Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.dnb.de abrufbar. Originalausgabe © 2022 · Vittorio Klostermann GmbH · Frankfurt am Main Alle Rechte vorbehalten, insbesondere die des Nachdrucks und der Übersetzung. Ohne Genehmigung des Verlages ist es nicht gestattet, dieses Werk oder Teile in einem photomechanischen oder sonstigen Reproduktionsverfahren oder unter Verwendung elektronischer Systeme zu verarbeiten, zu vervielfältigen und zu verbreiten. Gedruckt auf EOS Werkdruck von Salzer, alterungsbeständig ISO 9706 und PEFC-zertifiziert. Druck und Bindung: docupoint GmbH, Barleben Printed in Germany ISSN 2625-8218 ISBN 978-3-465-04598-4

Contents

0 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.1 Metaphilosophy . . . . . . . . . . . . . . . . . . . . 0.2 A philosophy designed to be metaphilosophy-first . . 0.3 Scientific tools for metaphilosophy: finding the right procedure . . . . . . . . . . . . . . . . . . . . . . . 0.3.1 Baconian and Newtonian experimentalism . . 0.3.2 Galilean idealisations . . . . . . . . . . . . . 0.4 Overview . . . . . . . . . . . . . . . . . . . . . . . 1 Kant’s propaedeutic method . . . . . . . . . . . . . . . . . . . . . 1.1 Propaedeutic philosophy is the study of philosophy . . 1.2 A two-layered analogy . . . . . . . . . . . . . . . . . 1.3 From empirical experimentation to a priori experimentation . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 Kant’s conception of the philosophical experiment . . 1.5 Chemical analysis and synthesis . . . . . . . . . . . . 1.6 Philosophical procedure in the Aesthetic/Analytic and the Dialectic . . . . . . . . . . . . . . . . . . . . . . 1.7 Metaphilosophy as experimentalist practice . . . . . . 2 Maimon’s method of fictions. . . . . . . . . . . . . . . . . . . . . 2.1 Philosophy is the science of the form of all sciences . . 2.2 Against propaedeutic philosophy: quid facti? and quid juris? . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Quid facti? . . . . . . . . . . . . . . . . . . 2.2.2 Quid juris? . . . . . . . . . . . . . . . . . . 2.3 A coalition system . . . . . . . . . . . . . . . . . . . 2.4 The analogy to calculus . . . . . . . . . . . . . . . . 2.4.1 Maimon’s philosophy of mathematics . . . .

9 11 16 21 27 34 38 41 44 56 59 65 71 79 94 98 103 111 112 116 121 125 125

6

CONTENTS

2.4.2 An immanent account of cognition . . . . . The method of fictions . . . . . . . . . . . . . . . . Philosophical fictions . . . . . . . . . . . . . . . . . Maimon’s metaphilosophy: philosophy as modelling practice . . . . . . . . . . . . . . . . . . . . . . . . 3 Schelling’s method of nature-construction . . . . . . . . . . . . 3.1 Philosophy is the science of the unconditioned . . . . 3.2 Fichte’s original insight, Schelling’s observation, and Schelling’s point . . . . . . . . . . . . . . . . . . . . 3.3 Naturphilosophie as science of the unconditioned . . . 3.4 Positing an absolute hypothesis . . . . . . . . . . . . 3.5 The Method of Nature-Construction . . . . . . . . . 3.6 Presenting nature through experiment . . . . . . . . 3.7 Metaphilosophy as constructive and experimental practice . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Experiments of reason . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Kant’s experiments of pure reason . . . . . . . . . . . 4.2 Maimon’s metaphysical modelling . . . . . . . . . . 4.3 Schelling’s experimental constructions . . . . . . . . 4.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . Other texts . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 2.6 2.7

134 145 152 165 170 174 183 197 205 212 218 231 234 236 240 245 247 249 252 252 255

Author’s note

Abbreviations used in the footnotes are explained in the bibliography.

0 Introduction

In this work, I investigate the thought of three philosophers — Immanuel Kant, Salomon Maimon, and Friedrich Wilhelm Joseph Schelling — under the hypothesis that all of them contribute a methodological solution to the problem of establishing the scientificity of theoretical philosophy. They all engage in an attempt “to promise […] to metaphysics the secure course of a science” (KrV, Bxviii-xix). While nowadays the project of a ‘scientification’ of philosophy has been abandoned by most philosophers, things looked quite differently for philosophers throughout modernity, and especially so at the end of the 18th century. These philosophers did not pursue such a project out of delusions of grandeur, nor due to gross neglect for scientific practice and its standards. On the contrary, their renewed push for a scientification of philosophy arose from direct engagement with the sciences and their methods. In light of the revolutions in modern science, philosophers began to feel a pressure to secure the scientific status of their own discipline. From Kant’s Critique of Pure Reason (1781/1787) onwards, and throughout the majority of German idealism, we can trace an on-going debate about what theoretical philosophy is, and what it ought to be, framed within the context of science.1 This book is dedicated to investigating one specific methodological programme that evolves from these 1 I agree with Franks who defines German idealism as a “family of philosophical programs” that share “a common origin in Kant’s critical philosophy”, and “seek to complete the revolution begun by Kant”, which can be condensed to the claim that the truth of synthetic a priori propositions such as those in physics or mathematics “consists […] in conformity to the a priori conditions of human knowledge, which constitute the objects of human knowledge, objects that are thus mind-dependent” (2005, pp. 13-4).

10

Introduction

debates: that of devising a methodological solution to the problem of how philosophy can investigate itself. This research programme arises from a specific concern with philosophy about philosophy, which I call “metaphilosophy-first”. And what unites its proponents is their shared view that all philosophy must begin with a methodological investigation into what philosophy as science should look like. In other words, metaphilosophy is not treated as that one branch of philosophy concerned with metaphilosophical theories, but as the fundamental discipline, in which all other philosophical disciplines must be grounded. Moreover, metaphilosophy-first views hold that metaphilosophy, as the fundamental science, must come in the form of a specific method. Beginning with Kant, and continuing throughout German Idealist philosophies, one is bound to encounter a rich variety of philosophical methods that are specifically designed for the purpose of producing metaphilosophical theories. These metaphilosophical theories are concerned not only with “the inquiry into the nature of philosophical question and the methods (to be) adopted in answering them” (Overgaard, Gilbert, & Burwood, 2013, p.4): what characterises metaphilosophy-first as a family of research programmes is exactly that it contains an investigation into the nature of metaphilosophical method, inquiring into the nature of metaphilosophical questions and the methods to be adopted in answering them. My focus in this work will be on Kant’s propaedeutic method, Maimon’s method of fictions, and Schelling’s method of nature-construction. I shall argue that it is through the theoretical lens of this specific understanding of metaphilosophy-first that we can understand them as unified by one research programme. In contrast to other metaphilosophyfirst programmes of the time, all three philosophers are united in their metaphilosophical belief that the philosophical methods used to answer metaphilosophical questions must be developed in continuity with the methods of the sciences. As early philosophers of science, Maimon and Schelling follow Kant in designing their methodological solutions in light of some of their conclusions from the reflective enterprise of analysing the forms of theoretical and experimental practice within certain sciences, and especially their analyses of sciences that were still in formation. It is my contention that their solutions to the problem of

Metaphilosophy

11

metaphilosophical method are continuous with earlier programmes of experimentalist philosophy. In the spirit of the Instauratio Magna, and its calls for a methodological revolution in the sciences, Kant, Maimon, and Schelling sought to overcome the limits of the “speculative”, nonscientific, philosophy of their day through a revolution in philosophical method. Moreover, they integrated the idea that “observations and experiments can ground philosophical claims” (Anstey & Vanzo, 2016, p. 87) into the domain of metaphilosophy, developing experimental methods for the investigation of philosophy itself. However, and this will be the core concern of this book, this led Kant, Maimon, and Schelling to advance their own brand of philosophical experimentalism. As we will see, to establish the nature of a philosophy that is truly scientific, they put forward a method that is both speculative and experimental. In preparation for the study of the individual metaphilosophical methods and their connection to a distinct research programme, I provide a brief introduction to the discipline of metaphilosophy, and show how this discipline is conceptualised within the context of Kant’s broader philosophical framework. I then give a brief outline of how this leads Kant to his commitment to metaphilosophy-first, and why this commitment results in a methodological research programme that is taken up by many Early Post-Kantians. In turn, this will necessitate a preliminary exploration of some of the concepts and methods that characterise the context of Modern science, within which theoretical philosophy needed to assert its status as a proper science.

0.1 Metaphilosophy Since the beginnings of Western philosophy, and up until today, one important area of philosophical research has been dedicated to questions about the nature of philosophical investigation itself. This field of study—commonly referred to as ‘metaphilosophy’2 —deals with questions about what philosophy is, what its purposes are, and how it should be done. Throughout history, there has been a notable divide between 2 Morris Lazerowitz (1940) is usually considered to have coined this term, see Reese (1990, p. 28).

12

Introduction

two ways of understanding the metaphilosophical project—as descriptive or prescriptive—which concern the ways in which metaphilosophy approaches its objects and goals. While prescriptive metaphilosophy tells us what philosophy ought to be, its descriptive counterpart tells us what philosophy is.3 An increasing number of contemporary metaphilosophers subscribe to the view that metaphilosophy is to be understood as a descriptive project.4 Their metaphilosophies typically study which concepts, principles, and methods different philosophical schools or traditions have been employing, and for what purposes.5 Just as there is something to be said about what the sciences do, how they do what they do, and why they do what they do, there is something to be said about what philosophy does, how philosophy does what it does, and why it does what it does. As Pettersson notes, “the distinctive quality in the attitude of descriptive metaphilosophy is that it is not judgmental about what actually is good or bad philosophy. Instead, the practitioners of descriptive metaphilosophy remain as neutral spectators outside philosophical disagreements and focus on describing philosophy as it really functions—or has functioned during its history—warts and all” (Pettersson, 2019, p. 133). Before the increased emphasis on metaphilosophy as a descriptive project, however, the study of the nature of philosophy was mostly conceived of as a prescriptive project. That is, although philosophers such as Hume and Spinoza did study and describe the nature of other philosophical programmes than their own, they also clearly took sides. Most metaphilosophical projects in the past two thousand years have in fact made judgments about what distinguishes good and bad philosophy. For a long time, the main role of metaphilosophical contributions 3

Glock (2008, p.3); (2013, p.35–6). This is how I understand the tenor of Overgaard, Gilbert, & Burwood (2013), Rescher (2014), and most recently Pettersson (2019). 5 One possible starting point of the discipline is the founding of the journal Metaphilosophy in 1970. Other possible origins are connected to Lazerowitz’s Studies in Metaphilosophy (1964), as well as Rorty’s Linguistic Turn: Recent Essays in Philosophical Method (1967). For discussion, see Petterson (2019, pp. 113-16). 4

Metaphilosophy

13

was to take a stand for or against different conceptions and methods of philosophy. Many of these contributions also aimed at the explicit development of “recipes” for what philosophy ought to be, ought to do, and how it ought to proceed. As we will see, Kant, Maimon, and Schelling’s metaphilosophies also fall under this broad prescriptive category of metaphilosophical conceptions. All of them are interested in theories and methods for establishing the right kind of philosophy, and all of them think that this can only be achieved by applying one specific method, which takes up the same status and role as the scientific method in mathematics and natural science. Although I see the appeal of applying these categories to distinguish between methodological differences in metaphilosophy, I shall propose that we accept another distinction in its place: (i) metaphilosophy that has metaphilosophy as its goal, and (ii) metaphilosophy that has philosophy as its goal.6 The first type of metaphilosophy largely overlaps with what has been described under the category of descriptive metaphilosophy: it aims at a neutral stance from which to study different varieties of philosophical theories and methodological approaches, in order to describe their similarities and differences. Their goal is to produce philosophical arguments to support metaphilosophical theses. The second type, on the other hand, engages in metaphilosophical inquiry with a different goal. These inquiries typically produce metaphilosophical arguments to support philosophical theses. Indeed, many philosophers (including the ones mentioned when discussing prescriptive metaphilosophy) engage in metaphilosophical discussions because they claim to have good reasons for doing the kind of philosophy they do. In this case, metaphilosophy can serve to license specific philosophical positions, theories, and methodologies. Now, whatever stance is driving one’s metaphilosophical reflections, it remains philosophy about philosophy. But to say this is not to conceive of metaphilosophy as a second-order inquiry that “look[s] down on philosophy from above, or beyond” (Williamson, 2007, p. ix), but rather as a branch of first-order philosophy which engages with philosophy as its object of study. As Heidgger pointedly remarked, all metaphilo6

I am indebted to Markus Wild for making me aware of this point.

14

Introduction

sophical projects must grapple with a particular circumstance intrinsic to this endeavour: When we ask, “What is philosophy?” then we are speaking about philosophy. By asking in this way we are obviously taking a stand above and, therefore, outside of philosophy. But the aim of our question is to enter into philosophy, to tarry in it, to conduct ourselves in its manner, that is, to “philosophize”. The path of our discussion must, therefore, not only have a clear direction, but this direction must at the same time give us the guarantee that we are moving within philosophy and not outside of it and around it. Heidegger (1956, p. 21) [emphasis added].

This problem lies at the heart of the metaphilosophical inquiries that will concern me in this work. When philosophy turns to studying its own nature, it might at first seems as if it takes an ‘outside’ perspective on its object of study. Yet, the very activity that is employed to study this object, the manner in which we conduct ourselves during this investigation, can only be called philosophising. If metaphilosophy is to be first-order philosophy about philosophy, however, then it gets harder to see how one can take a truly descriptive stance toward their object of study. It seems that metaphilosophical inquiry “cannot and should not be philosophically neutral”, that it is always “just more philosophy, turned on philosophy itself” (Williamson, 2007, p. 6-7). On an uncharitable reading, it looks as if metaphilosophy, whatever its goal, might turn out to be biased and circular after all.7 Let me explain this in some more detail. Suppose, for example, that a metaphilosophical inquiry has as its goal to determine the nature of philosophical method. To do so, it must philosophise about philosophy. That is, it must operate in a certain manner in order to investigate its object of study—philosophical method—and determine its nature. I exclude here non-philosophical methods that are employed for the aim of 7 Petterson discusses two forms of circularity that endanger the metaphilosophical project: that of (i) circulus in defiendo, and that of (ii) circulus in demonstrando. (i) is not problematic because “this circle of self-understanding is exactly what metaphilosophy is about”; the circularity only arises if philosophy and metaphilosophy are understood as two autonomous and clearly distinct projects that require separate definitions (2019, p. 119). Circularity of kind (ii) is more serious, and identical with the one I describe above, as it arises from the fact that “in metaphilosophy’s case the subject matter and the phenomenon studying it are by definition the same (ibid., p. 120 [emphasis added]).

Metaphilosophy

15

producing metaphilosophy from this discussion, e.g., historical, sociological, or anthropological methods. But then what is the nature of the method by which this metaphilosophical inquiry proceeds? Following Williamson’s credo, for example, it would have to be the philosophical method at issue. After all, it is philosophy of philosophy (Williamson, 2007, p.ix). But then, by investigating in this way, aren’t we already partial? Have we not already implicitly decided on a method that we take to be appropriate for philosophical inquiry, namely the one by virtue of which we proceed to examine philosophy? Even worse, have we not already presupposed the nature and legitimacy of the very thing we set out to investigate? Take another example. If I am trying to explicate and argue for the validity of conceptual analysis, and my methodological tool for doing so is nothing other than conceptual analysis, am I not already presupposing what I am trying to demonstrate? We can illustrate this by drawing a contrast with a project similar to metaphilosophy: in philosophy of science, for example, the method of investigation is clearly differentiated from the scientific practice under scrutiny. In contrast, in metaphilosophy (i.e., philosophy of philosophy), what is being studied and what is studying it are by definition identical. How can there be philosophy of philosophy, then, that studies the nature, methods, and role of philosophy in a principled, but non-circular, way? If we were to define yet another distinct science that studies the nature and method of metaphilosophy, i.e., metametaphilosophy, we would end up in an infinite regress. But simply failing to reflect on the very thing by which philosophy is investigated exposes metaphilosophy to the charge of arbitrariness.8 It seems, then, as if metaphilosophies, so understood, cannot justify the validity of their own methods, since they cannot provide further arguments for their methods of argumentation without either accepting their own partiality or becoming circular.9 Metaphilosophies which 8

See Petterson (2019, pp. 119-121). One further option would consist in the adoption of a pragmatic explanation. A metaphilosophy of this kind takes a successful practice as its starting point and then proceeds to extract the norms implicit to it. Still, even with such explanations, we can press the following point: what makes a practice successful? And where do we get those criteria from? 9

16

Introduction

aim at the advancement of, say, a particular way of doing philosophy, do seem to have an advantage in this matter. It is inherent to such projects that they will legitimise the kind of philosophy they encompass, of the one they judge to be “good” or valid. One, if not the, central task of this type of metaphilosophy lies exactly in simultaneously describing what good philosophy is while doing good philosophy. In this case, metaphilosophy can either take up some known conception of philosophy and defend it, or it can articulate its own conception of good philosophy and simultaneously argue for its validity. Nevertheless, this second type of metaphilosophy will have to accept the partiality and/or circularity of its argumentation.10 Returning to Kant: Kant’s metaphilosophical theses initiate a research programme which underwent several variations throughout Early Post-Kantianism. His first thesis consists in the declaration that there simply exists no metaphysical conception which would deserve the label ‘scientific’. For this reason, he sets out to develop a new philosophy, i.e., a metaphilosophy, that will be equipped to determine what a scientific metaphysics must look like. This is the project of the Critique of Pure Reason. What qualifies the resulting research programme is that it commits to the view that metaphysics, to become a proper science, must become metaphilosophy first, and that this metaphilosophy must acquire a methodology that is specifically suited to the specific nature of its task.

0.2 A philosophy designed to be metaphilosophy-first Kant writes the first Critique with a specific purpose in mind: “to promise […] to metaphysics the secure course of a science” (KrV, Bxviii-xix). 10 Certainly, there are various ways to deal with this problem, for example by first admitting that metaphilosophy is circular but then insisting that not all circles must be bad or vicious circles. We can argue that if we put our explanandum “in its rightful place in a big enough tent of related concepts […], then we will succeed in shedding light—from within—on [the notion at issue]” (Della Rocca, 2020, p. 107). See Della Rocca (2020, pp. 107-111, 138) for criticism (that would certainly have been much appreciated by Maimon and Schelling) of this strategy.

A philosophy designed to be metaphilosophy-first

17

His investigation is built upon three central premises. First, Kant submits that science is both possible and actual, its instances being logic, mathematics, and physics.11 Second, he determines that, in contrast, metaphysics does not yet qualify as a science. Hence, philosophical science is not actual, and philosophy must begin by investigating whether it is possible for it to be a science at all. In more detail, Kant’s diagnosis consists in the insight that [m]etaphysics—a wholly isolated speculative cognition of reason that elevates itself entirely above all instruction from experience, and that through mere concepts (not, like mathematics, through the application of concepts to intuition), where reason thus is supposed to be its own pupil—has up to now not been so favored by fate as to have been able to enter upon the secure course of a science, […]. [T]here is no doubt that up to now the procedure of metaphysics has been a mere groping, and what is the worst, a groping among mere concepts. (KrV, Bxv)

Kant’s ‘groping’ metaphor indicates, thirdly, what Kant identifies as the reason for metaphysics’ failure to become a proper science: metaphysics does not yet have a rigorous (read: scientific) procedure. His point is not, as some have claimed,12 that all metaphysics is impossible, and hence must be replaced with epistemology, but that the nature and limits of metaphysical cognition must be investigated to determine the ways in which the procedure of metaphysics must be transformed in order to attain the qualities of scientific cognition.13 By establishing what metaphysical cognition can (and should) consist in, that is, what its conditions of possibility are, it can be shown that metaphysics is possible 11 Kant articulates his conception of what he takes to be a science proper most explicitly in the Metaphysical Foundations of Natural Science (2004b), i.e., (MNS, AA 04, 467-470). 12 Kant is by means the “all-crushing destroyer of Metaphysics” (Mendelssohn 1785, pp. 91, 93) he is sometimes made out to be in, e.g., Ameriks (1992). Kant does not argue for the impossibility of metaphysics; he argues for the impossibility of special metaphysics, but not of general metaphysics, which philosophy ought to produce according to the right procedure, i.e., through experimental but rational cognition from concepts. 13 For a long time, and particularly amongst many of his contemporaries, Kant’s metaphilosophy has been perceived as leading the way to an ‘elimination of metaphysics’ (see e.g., Carnap (1932)). Or, as Hegel notes, “what was hitherto called “metaphysics” has been, so to speak, extirpated root and branch, and has vanished from the ranks of the sciences.” (2010 [1813]: 7).

18

Introduction

as science. Namely, if it adheres to a procedure that remains within the limits of its epistemic possibilities, then metaphysics is possible as a kind of general metaphysics, and can even assume some of the former tasks of earlier metaphysical systems.14 What is often overlooked or downplayed is his insistence that, to do so, the Critique must take the form of “propaedeutic philosophy”.15 Rather than just devising a new philosophy that has the right properties to qualify as science, Kant argues for the necessity of a propaedeutic philosophy that is distinct from metaphysics: Now the philosophy of pure reason is either propaedeutic (preparation), which investigates the faculty of reason in regard to all pure a priori cognition, and is called critique, or, second, the system of pure reason (science), the whole (true as well as apparent) philosophical cognition from pure reason in systematic interconnection, and is called metaphysics. (KrV, A841/B869)

Kant distinguishes between the critical project as propaedeutic philosophy and the metaphysical project as science. Metaphysics is described as a particular kind of cognition, i.e., “a priori cognition from pure reason in systematic interconnection” (KrV, A841/B869). Kant holds that the study of metaphysics, as well as the study of some other sciences, must entail the study of a specific kind of cognitive activity, e.g., metaphysical cognition. He further believes that the cognitive activity of metaphysics can only enter the secure path of a science when it is investigated through another philosophical discipline, namely that of so-called “propaedeutic” philosophy. According to Kant, the task of propaedeutic philosophy is to “investigate […] the faculty of reason in regard to all pure a priori cognition, and is called critique” (KrV, A841/B869).16 More precisely, propaedeutic philosophy must investigate the conditions of possibility of all a priori cognition in order to be able to determine the conditions of scientific metaphysics. We will see that, by contending that propaedeutic philo14

See De Boer (2015). Two interesting contributions in this respect are Serck-Hannsen (2015) and Ferrarin (2015); both discuss the implications of understanding the Critique as propaedeutic. 16 In chapter 1, we will see that Kant sets up this distinction with a less definite difference in mind than I present it to be here. 15

A philosophy designed to be metaphilosophy-first

19

sophy defines the structure of intentionality as such (i.e., the general structure of what human thought and perception can be about), propaedeutic philosophy also defines the scope and limit of any other science. In offering a “treatise of method” (Bxxii), the Critique employs a set of procedures for establishing the cognitive structure fundamental to any rational inquiry and representation of what there is. The correct procedure for scientific metaphysics can only be determined on the grounds of an a priori critique of pure reason. As I claim in the first chapter, the Critique of Pure Reason thereby advances a scientific programme that puts metaphilosophy in the place of first philosophy; it is metaphilosophy-first.17 The first Critique and its project of a propaedeutic philosophy in fact only concerns one part of Kant’s metaphilosophy, while others can be found elsewhere.18 When talking about Kant’s “general metaphilosophy”, we should have his full conception of philosophy in mind, i.e., his concept of “cosmic philosophy”19 [Weltbegriff der Philosophie] (KrV, A838/B866), which conceives of philosophy as “the science of the 17

Thus, although I focus on epistemological readings of Kant’s philosophy, I do not think that his project should be characterized as an epistemology-first project. Its primary objective is not to provide a theory of knowledge and justification, but to show how metaphysics is possible as science. Certainly, determining whether theoretical philosophy can count as science (and if so, under which conditions) requires a theory of knowledge and justification, but it also includes a theory of representation and of object constitution. 18 There is certainly also a lot to be said about the transition from the A to the B Edition. Kant shifts from a focus on the question of why our representations have objective validity to a focus on the possibility of synthetic a priori judgments. Förster has offered convincing arguments to suggest that this shift also happens for metaphilosophical reasons (2012, p. 46f.). Kant notes in his now-famous letter to Kästner in August of 1790: “[T]he efforts I have heretofore made are in no way meant (as they may appear to be) to attack the Leibniz-Wolffian philosophy (for I find the latter neglected in recent times). My aim is rather to pursue the same track according to a rigorous procedure and, by means of it, to reach the same goal, but only via a detour that, it appears to me, those great men seem to have regarded as superfluous: the union of theoretical and practical philosophy. This intention of mine will become clearer when, if I live long enough, I complete the reconstruction of metaphysics in a coherent system.” (AA 11, 186 [emphasis added]). 19 See Ferrarin (2015) for a thorough study on Kant’s concept of “cosmic philosophy”.

20

Introduction

relation of all cognition to the essential ends of human reason” (KrV, A839/B867), including its concern with moral philosophy, “the entire vocation of human beings” (ibid.). His complete metaphilosophy includes not only thoughts on how metaphysics can become a science, but also concerns with morality, religion, and anthropology. What interests me in this work, then, is in fact Kant’s meta-metaphysics, and not his general metaphilosophy.20 For purposes of simplicity, however, I will keep referring to this project as Kant’s metaphilosophical project. As mentioned above, Kant’s conception of propaedeutic philosophy as a philosophical discipline, which must “transform the accepted procedure of metaphysics” into proper science, is inextricably connected to the project of finding the appropriate methodology for doing so. What’s more, he seems to hold that an execution of his propaedeutic project depends on his success at finding the right method of philosophising in the mode of metaphilosophy. As becomes evident in the B Preface, and elsewhere, Kant was very aware that in order establish metaphysics as science, he not only needed to determine its own conditions of possibility, but also a designated procedure through which such a self-investigation would be possible at all. The Critique is not only a work on philosophical method, but also a work of a particular philosophical method. This is where we get to argumentative core of this book. I shall argue that, from his quest to establish a philosophy that responds to the requirements of metaphilosophy-first, Kant gives rise to a family of research programmes which are explicitly concerned with finding a philosophical methodology that is appropriate to this task. Looking at the B Preface again, it is quite easy to see that, firstly, Kant does offer a particular methodological approach by virtue of which he aims to execute the propaedeutic project, and secondly, that this methodological approach—in part explicitly, in part implicitly—contains the specific resources which are required for executing the type of investigation that is pursued in metaphilosophy which has philosophy as its 20 Note that this must not imply that I am committing to a metaphysical reading of Kant, as we find for example in Heidegger (1990). In fact, I will interpret Kant more along the lines of so-called epistemological readings (e.g., Cohen (1885)), since this is how he was interpreted by Maimon and Schelling, and this is the conception that their programmes are responding to.

Scientific tools for metaphilosophy: finding the right procedure

21

goal. I contend that Kant thereby initiated a specific research program in metaphilosophy which specifically connects the problem of scientificity in theoretical philosophy with the problem of metaphilosophical methodology. Recognising this metaphilosophical programme offers a new interpretative approach to many strands of Early Post-Kantian philosophy. Beginning with Reinhold’s method of reflection, Maimon’s method of fictions, through to Jacobi, Fichte, and Schelling’s methods of philosophical construction, and maybe ending with Hegel’s development of the speculative method, the philosophical discussion of this period is rife with methodological proposals of this nature. Under my hypothesis, these conceptions of philosophical method should be classified under the proposed description of metaphilosophy-first, which sets out to determine the possibility of theoretical (and practical) philosophy, thus providing a rationale for their continuous appearance in Early Post-Kantianism.21

0.3 Scientific tools for metaphilosophy: finding the right procedure In this book, I include three in-depth studies of such methodological programmes: Kant’s propaedeutic method, Solomon Maimon’s method of fictions, and Friedrich Schelling’s method of nature- construction. I shall argue that Kant’s idea of a propaedeutic method is adopted and advanced by Maimon and Schelling, whose metaphilosophies can thus be seen as versions of the former’s research programme; they can be seen as ways of conceiving of philosophical procedures that validate their own legitimacy through the course of their application, all while investigating the possibility of that activity which they are themselves instances of, i.e., theoretical philosophy. Regarding the previous discussion of the particular problems which metaphilosophical projects of this type will have to face, Kant, Maimon, and Schelling’s procedures reflect different 21 There exists a variety of studies on individual authors and their thoughts on philosophical method, but almost no systematic work on the metaphilosophical debate throughout German idealism. Exceptions are Ende (1973) and Taureck (1979), on the method of construction in German idealism, as well as more recently Franks (2005), on the notion of “philosophical systems”.

22

Introduction

strategies for how to respond to the charges of partiality and circularity. It is part and parcel of their metaphilosophical procedures that these adhere to a scientific standard themselves. To understand what links these three metaphilosophies more specifically, our theoretical lens must transcend the purely philosophical context (if there is such a thing). Michael Friedman (1992) famously argued that Kant’s philosophical project was deeply committed to finding a philosophy adequate to the sciences of his day. He interprets the first Critique as a crucial part of Kant’s attempt to articulate the metaphysical foundations of all sciences. I shall argue that this concern with the mathematical and natural sciences of his day not only plays an important role when it comes to Kant’s metaphilosophical project, but also in Maimon and Schelling’s. What unites their methodological approaches, beside their specific metaphilosophical goals and tasks, is that each of these philosophers saw his method as standing in continuity with some, or many, methods of the natural sciences. While Kant’s conception of propaedeutic method links to experimentation in the natural sciences, and in particular to experimental chemistry, Maimon’s method of fictions draws on theories of differential calculus as they were applied to experimental physics and other areas of science at the time. Schelling, finally, derives much of his conception of nature-construction from these two contexts but also, and more importantly, from the contexts of new experimental sciences such as magnetism, galvanism, and electricity. For the remainder of the introduction, I want to elucidate some general scientific concepts or procedures which had an important influence on the development and progress of those sciences which influenced Kant, Maimon and Schelling’s methodological conceptions. One of the major motivations for Kant’s first Critique was to articulate a conceptual framework that makes explicit the concepts and principles that were used by the science of his time to describe, predict, and explain nature. As indicated in earlier sections, Kant believed that, by accounting for the possibility of the a priori cogntion of objects in general, his proposal would not only explain the possibility of metaphysics, but also the possibility of all other sciences. The achievement of this goal was deeply intertwined with Kant’s conceptualisation of a historical event which is usually referred to as “the mathematisation of

Scientific tools for metaphilosophy: finding the right procedure

23

nature”. An important reason for the rapid development and progress of the natural sciences, as well as the spread of new sciences around the 17th and 18th century, was the successful application of mathematics to nature, and the practices of experimentation and idealisation that came with it.22 One influential expression of this “mathematisation thesis” can, for example, be found in Galileo’s statement that “[the book of nature] is written in the language of mathematics, and its characters are triangles, circles, and other geometrical figures, without which it is humanly impossible to comprehend a single word of it” (Galileo, 1957, p. 3). Taking up the intellectual heritage of Archimedes, scientists after Galileo began to “mathematise” physics. That is, they committed to the assumption that the relationships between physical magnitudes can be expressed as mathematical relations. Mathematics was thereby elevated from the status of an “abstract science exploring the relation of numbers”, as the guiding assumption had by then become that “in these relations lies a model of physical reality” (Hall, 1960, p. 80). Scientists such as Galileo, and later Newton, held that mathematical equations can be used to express the necessary and universal connections between physical objects and properties, e.g., to describe the motion of bodies. Depending on its formulation, the mathematisation thesis includes one or two separate assumptions. First, the assumption just discussed, namely that mathematical objects represent physical phenomena. Second, it could moreover include the additional assumption that mathematical representations are appropriate means for the representation of nature because physical objects are mathematical objects. Many of the natural philosophers who influenced Kant’s views on science argued for realist accounts that explained nature in terms of mathematics, thus proposing that nature is inherently mathematical (e.g., Galileo, Newton, Leibniz).23 This viewpoint goes against the Aristotelian framework of science, which took mathematical properties to be accidental in nature, and the study of planetary motion to be concerned with celestial, i.e., non-physical, phenomena. The new astronomers, with Kepler lead22 For an overview of the mathematical revolution in natural philosophy, see Mahoney (1998, pp. 702-55). 23 For an overview of these developments, see the classics Burtt (1925) and Hall (1963, pp. 36-103), or more recently Gaukroger (2010).

24

Introduction

ing the way, advocated for an alternative approach that treats planetary motions as terrestrial, physical phenomena. According to the latter’s Platonism, the Copernican hypothesis24 was re-interpreted as an argument for the inherent mathematical character of nature, and fuelled the mathematical realism of Early Modern scientists. Thus, the programme of mathematising nature would not only take as an assumption the epistemic thesis that mathematical reasoning and equations are useful to make nature’s phenomena computationally tractable and predict their course (e.g., Copernicus 1973 [1543]25 ), but would sometimes also adopt the ontological thesis that nature’s structure is in fact inherently mathematical, and that it is by virtue of its ontology that nature is representable and describable by virtue of mathematics. In the spirit of Newton’s dictum, Kant, too, committed to the view that “with the help of philosophical geometers and geometrical philosophers, instead of the conjectures and probabilities that are blazoned about everywhere, we shall finally achieve a science of nature supported by the highest evidence” (Newton, 1984). Kant aimed at the construction of a philosophical theory that would not only ground metaphysics as science, but also the mathematised natural sciences as such. In doing so, he assumed a specifically modern conception of science, whose goal it is to capture the totality of all natural phenomena by establishing the general and necessary laws that govern them. He took these laws to be attained (i) through the language of mathematics, and geometry in particular, (ii) through scientific experimentation that allows this language to make physical arguments, and (iii) through employing deliberate idealisations in scientific descriptions and explanations, which allowed natural phenomena to be amenable to mathematical repres24 Copernicus still adhered to an Aristotelian picture of science, looking at the celestial and the terrestrial as two different domains of being that are governed by distinct laws; see Martens (2000, pp. 26-30). 25 With regard to the ‘mixed sciences’, i.e., sciences that both included mathematical computation and empirical physics, such as astronomy and optics, Copernicus defended the view that since these “cannot in any way attain to the true causes, [they] will adopt whatever suppositions enable the motions to be computed correctly from the principles of geometry . . . these hypotheses need not be true or even probable.” (1978/1543, p. xvi)

Scientific tools for metaphilosophy: finding the right procedure

25

entation.26 His propaedeutic was then to propose an account of the mind-world-relation that coheres with this scientific framework, in the sense that its facticity informs the way in which Kant models the exact relation between different kinds of cognition and their objects, such as mathematical cognition or empirical cognition. Now, many studies on Kant’s philosophical methodology have focused on his philosophy of mathematics, and especially on his understanding of Euclidean geometry as a language in which one could both state laws of nature and make physical arguments through diagrammatic constructions.27 However, as I will explain in the first chapter, correctly understood, Kant’s conception of Euclidean construction does not play any important role in the development of his metaphilosophical procedure in the Critique.28 In the Doctrine of Method, he argues that philosophical cognition is disanalogous to mathematical 26 This programme thus integrates so-called ‘mathematical’, as well as ‘experimental’ traditions. See Kuhn (1977, pp. 31–65) and Hacking (1975), the latter of whom contrasts the ‘high’ (i.e., mathematical) sciences with the ‘low’ (i.e., probabilistic) sciences such as medicine and alchemy, which reason probabilistically rather than conclusively. 27 For Kant, Euclidean geometry presents a methodological ideal of how a priori cognition of objects is possible, and hence how scientific cognition is possible. On his account, the geometrical method serves as a model for a type of scientific cognition which can define its concepts in a synthetic manner and thus can not only infer but demonstrate their reality. Crucially, he thinks mathematics is capable of something all other sciences lack, namely, of constructing its concepts in pure intuition. On his view, “to construct a concept means to exhibit a priori the intuition corresponding to it” (KrV, A713/B741). In contrast to physicists, geometers can “construct a triangle by exhibiting an object corresponding to this concept, either through mere imagination, in pure intuition, or on paper, in empirical intuition, but in both cases completely a priori, without having had to borrow the pattern for it from any experience” (KrV, A713-4/B741-2). What characterises mathematical construction, in a nutshell, is that it has a way of proving its propositions through the immediate construction of the sensible objects which they refer to, generating a priori singular instances, i.e. intuitions, of the concepts constructed, which at the same time exemplifies their universal properties. Mathematical argumentation can proceed progressively, “through a chain of inferences that is always guided by intuition” (KrV, B745), thus ensuring the objective validity of and non-triviality of its propositions. 28 Rather, these reflections play an important role in his analysis of the nature of synthetic a priori judgments.

26

Introduction

cognition, and hence to geometrical construction, because it cannot exhibit its concepts in pure intuition and hence cannot be constructive (KrV, A713/B741).29 Moreover, in the B Preface, Kant introduces his propaedeutic method as “imitated from the method of those who study nature” (KrV, Bxviii), characterising this method as consisting in an “experiment of pure reason” that “has much in common with what the chemists sometimes call the experiment of reduction, or more generally the synthetic procedure” (KrV, Bxxi). In his attempt to explain the method of that philosophy, which investigates the conditions of possibility of philosophical cognition, Kant turns to sciences that are also only in the process of forming, and to sciences that he himself deems impure because they could not be mathematised. I thus propose that the scientific concepts and procedures relevant to Kant’s project of developing a procedure for metaphilosophy do not concern geometry and its specific methods of demonstration. Instead, Kant looks for alternative methods to enable (i), (ii), and (iii), which he finds in the concepts and practices of experimentation and idealisation that were being employed by the natural sciences of his time, and in experimental chemistry in particular. Maimon, on the other hand, will disagree with Kant’s focus on synthetic geometry, and argue for the primarcy of analytic geometry, i.e., the use of differential calculus as a primary language to describe, predict and explain nature, whose idealisation practices he analyses and makes use of for the development of his method of fictions. Moreover, we will see that he disagrees with Kant’s assessment that geometrical demonstration succeeds at making physical arguments, assigning that role only to the empirical and experimental sciences. On the contrary, however, Schelling will emphasise the importance of different layers of conceptual frameworks that derive from many “pure and impure sciences”, e.g., physics, chemistry, galvanism, magnetism, early biology, etc., which are systematically connected to make nature intelligible to that degree where it can become the content of an intuition at all. On his view, experimentation will become the dominant paradigm of physical argumentation, and the way in which nature is made amenable to conceptualisation. 29

For a thorough discussion of these issues, see Posy (1992) and Shabel (2003).

Scientific tools for metaphilosophy: finding the right procedure

27

In order to have a rough idea of what general scientific conceptions stand behind the more specific ones which inform Kant, Maimon, and Schelling’s methodologies, let us take a brief look at two of them: experimentation and idealisation. 0.3.1 Baconian and Newtonian experimentalism Bacon’s Novum Organon (1620) bore a new philosophical conception of what science ought to do.30 In opposition to the mechanistically oriented natural philosophy of his day, Bacon, much like Kant, advances a revolutionary programme to reform the sciences of his time.31 Stressing the significance of method in science, he introduces a new scientific method to be recommended for adoption. He claims that the sciences can only legitimately call themselves such if they become experimental. By institutionalising sets of rules, norms, and techniques, which mainly regulate the way in which scientific knowledge is to be produced, the Royal Society of London, and later also the Paris Academy, helped experimentalism grow into a new scientific paradigm, which was intimately connected with the development of an explicit standard for science. As the source of many of its main ideas, Bacon postulates the superiority of the senses over reason. According to his diagnosis, nature should not be studied using armchair methods such as the metaphysician’s rational reflection, but should be accessed empirically, through the senses. In order to serve the purposes of science, however, the senses must be assisted, as their unmediated grasp of nature is often distorted, or just false. This is where experimentation enters the picture. By assistance, Bacon does not refer 30 It is important to be aware here that, although Bacon’s experimentalism is often viewed as paradigmatic of experimental science, and is presented in such a manner by Kant and the German idealists, experimenting scientists such as Galileo and Newton learned only little from Bacon and his philosophical writings. For an extensive study on Bacon’s influence on science over his lifetime, and how his methodological reflections begin to shape a research programme, see Gaukroger (2001). 31 In fact, Kant dedicated the Second Edition of the Critique of Pure Reason to “Bacon of Verulam” (see (KrV, Bii)). What’s more, he also cites some remarks Bacon makes about his project of a “great instauration” of the sciences.

28

Introduction

to the employment of instruments such as a magnifying glass or ruler, but to the employment of scientific experiments: “the senses judge of the experiment, [and] the experiment judges of the thing” (Bacon, 2000, p. 19). In contrast to earlier science, Bacon stressed the importance of relying on evidence produced through actual experience instead of thought experiments and imaginative exercises: “the true Philosophy… is to begin with the Hands and Eyes, […] to be continued by Reason; […] [and] to come about to the Hands and Eyes again” (Hooke, 1664, sig b2). That is to say, generalisations that arrive at explanatory hypotheses are only licensed if they are generalisations over sets of facts that were actually established as part of data collections.32 Nature cannot be approached by virtue of a priori hypotheses or rational reflection, but rather must be observed as phenomena; natural facts must be collected prior to explanation, and explanation must be based on facts about the phenomenal world.33 As a scientific paradigm, Baconian experimentalism propagated a practical turn in the sciences. This firstly included a general highlighting of the practical use of scientific knowledge. Viewing science in its practical function, Bacon emphasised the usefulness of knowledge, and he defined the concept of truth in virtue of successful application, i.e., experimental generation of the phenomenon under question.34 More importantly, one of the main characteristics of his “new method” was that he understood experiments to constitute active interventions into nature.35 For Bacon, experiments do not consist in passive observations 32 Both Boyle and Hooke come up with their own accounts of what the construction of natural history should look like (see Boyle (2008)). 33 See Gaukroger (2010, pp. 157-163). 34 “It may be that there are some on whose ears my frequent and honorable mention of practical activities make a harsh and unpleasing sound because they are wholly given over in love and reverence to contemplation. Let them bethink themselves that they are the enemies of their own desires. For in nature practical results are not only the means to improve well-being but the guarantee of truth. The rule of religion that a man show his faith by his works holds good in natural philosophy too. Science also must be known by works. It is by the witness of works, rather than by logic or even observation, that truth is revealed and established. Whence it follows that the improvement of man’s mind and the improvement of his lot are one and the same thing.” (Bacon 1964, p. 93 [emphasis added]) 35 On the experiment as intervention, see Tiles (1993).

Scientific tools for metaphilosophy: finding the right procedure

29

of natural phenomena, or in thought experiments that imagine or infer possible outcomes. Instead, they constitute actual interventions into the course of nature, “for we are not to imagine or suppose, but to discover, what nature does or may be made to do” (Bacon, 1960, p. 130). Through experiments, scientists create artificial settings in which they can isolate, manipulate, and control different conditions and factors that contribute to the production of some natural phenomenon. Experiments as interventions enable scientists to actively “reconstruct” nature, since under the right experimental set-up, “nature takes orders from man and works under his authority” (Bacon, 1905, p. 403).36 The data of natural histories are “works [emphasis added]” (Bacon, 1964, p. 93), objects and phenomena meshed with, and generated through, human experimental activity. In the hands of British experimentalist philosophers such as Boyle and Hooke, Bacon’s “New Method” became increasingly authoritative for different fields of knowledge production.37 As members of the Royal Society, Boyle and Hooke greatly contributed to the development and vindication of the new experimental method, which eventually took great influence and became one of the dominant conceptions of scientific explanation.38 In opposition to the axiomatico-deductive model of science, they envisioned a model for inquiry that was, firstly, based on the systematic and rule-governed collection of (natural) facts through observations, testimony, and experiments. These are then used to construct so-called ‘natural histories’.39 Only on the basis of such data collections, Bacon writes, should scientists form hypotheses for induction.40 Further, among these manifold instances of experiments, 36 Much later, Hacking (1983) would defend a version of this view and, once again, turn it into a very influential subject for the philosophy of science and experiment. 37 For a broad overview, see Anstey (2002). 38 In their seminal work, Shapin & Schaffer (1983) identify the experimental culture around the British Royal Society as it was developing from the 1660s onward by describing the material, literary, and social techniques that characterised it. 39 See Anstey (2002) and Hunter (2007). 40 Note that Bacon envisioned a specific type of induction, i.e., eliminative induction. “True induction is founded on exclusion, but it is not completed until

30

Introduction

Bacon identified a particular kind of experiment in terms of the work it does within a particular research programme: the crucial experiment. Crucial experiments present supreme instances of evidence that often (but not necessarily) lead to the definite confirmation (or falsification) of a hypothesis, thereby settling a dispute about a particular issue (Bacon, 2000, pp. 137-39). Finally, Bacon’s interventionism not only affects the objects of scientific study, nor just the experimental and observational data studied in natural histories, but also the behaviour of scientists, both as individuals and within communities. Bacon suggested that research culture can be changed not only by a different experimental practice, but also by shaping norms for the behaviour of scientists.41 Since scientists naturally fall prey not only to “illusions of the mind” (i.e., false doctrines) but also innate errors of reasoning and representation, their minds must be trained and guided by behavioural norms in order to be less susceptible to, or avoid altogether, these naturally occurring errors (Bacon, 2000, p. 19).42 In a way, Bacon already makes a case for a scientific self-critique of reason long before Kant’s Critique of Pure Reason, in proclaiming that the “human intellect is the source of its own problems” (KrV, Avii), and in offering a methodological solution to this problem. Additionally, Bacon and the members of the Royal Society pushed for a public scientific discourse that documents its experimental set-ups and results

it reaches affirmation” (Bacon, 2000, p. 130). Compared to a naïve understanding of induction, eliminative induction does not simply infer from a small set of observations to a general law but isolates and probes different factors possibly contributing to an effect, and by virtue of exclusion determines whether they play a causal (and therefore necessary) role in producing the said effect, or not. 41 See Gaukroger (2001, p. 10ff.). Bacon writes, “[b]ut certain it is . . . that as the most excellent of metals, gold, is of all other the most pliant and most enduring to be wrought; so of all living and breathing substances, the perfectest (Man) is the most susceptible of help, improvement, impression, and alteration. And not only in his body, but in his mind and spirit. And there again not only in his appetite and affection, but in his power of wit and reason” (1905, vii.99). 42 In an interesting parallel to Kant, Bacon argues for an epistemological position that identifies these natural errors under the name of ‘idols of the tribe’, as stemming from reason, and not from the senses.

Scientific tools for metaphilosophy: finding the right procedure

31

in great detail, and thus allows for reconstruction and re-evaluation of all experiments admitted into natural history.43 Although Bacon’s experimentalism is often viewed as paradigmatic for experimental science, experimenting scientists such as Galileo and Newton learned only little from Bacon and his philosophical writings.44 Instead, they were inspired by prior attempts at applications of mathematical method to physical phenomena, e.g., Archimedes’ statics, or later Copernicus’ and Kepler’s mechanics of celestial bodies.45 It is certainly true that, since the Bacon-Boyle-Hooke method was invoked in contexts which still resisted mathematisation, i.e., chemistry, electricity, magnetism, and early biology, and forms a substantial part of the metatheoretical decisions leading up to the ‘chemical revolution’ in the late 18th century, that Kant, Maimon, and Schelling would refer to it and note its importance.46 However, their conceptions also clearly depart from this traditional picture of experimentation. In his article on Kant’s philosophy of experiment, Alberto Vanzo makes an important argument to that point. He observes that, while Kant might have drawn from Bacon’s ideas about the active character of scientific experimentation, the former’s ideas about the use of hypotheses clearly diverge

43 See for example Shapin & Schaffer (1983) or Bensaude-Vincent & Blondel (2016) on the notion of the scientific spectacle and the idea of experiments as public demonstrations. 44 Shapiro even calls “[t]he assumed continuity between Newton’s usage [of experimental philosophy] and that of the early Royal Society […] largely an illusion.” (2004, p. 185) He argues that Newton mainly associates his work with Boyle and Bacon to make a point against natural philosophers like Descartes and Leibniz, while, originally, Newton developed his method of analysis and synthesis after the example of ancient Greek mathematics in order to transfer it to the discovery of laws to explain natural phenomena (see, e.g., ibid., pp. 196-97, 207-216). 45 See Hall (1969, pp. 103-109). 46 For a detailed discussion of the Kuhnian account of the chemical revolution, see Hoyningen-Huene (2008).

32

Introduction

from Bacon’s programme and instead rather admit of what I will call a “Galilean-Newtonian view”.47 Vanzo argues that, although Kant had neither any practical knowledge about experiment, nor think much about the experimental logic advanced by Bacon, he presented interesting reflections about the nature and use of scientific assumptions or hypotheses.48 Against the “BaconBoyle-Hooke conception of experiment”, Kant held that assumptions and expectations are in fact not detrimental to scientific research but instead play an indispensable and fundamental role within the process of knowledge production. Still, “it is necessary to test and assess them in order to either reject them as false, or else to transform them from mere opinions to certain truths” (Vanzo, 2012, p. 92). As Gloy and Vanzo have pointed out in their studies on Kant’s views on experimentation, these views do not reflect the Bacon-Boyle-Hooke conception of experiment, but rather the views of Galileo and Newton (Vanzo, 2012, p. 79; Gloy, 1996, pp. 72-73). In the style of Vanzo’s distinction, I decided to call this view on experimentalism the “Galileo-Newton view”. Let me explain in some detail where these two conceptions collide in order to mark the specific turn taken by Kant’s views, as well as Maimon’s, and Schelling’s after him. The Bacon-Boyle-Hook view of experimentation is committed to the rejection of hypotheses. I explained above that, according to this view, the main function of experimentation was neither to test theories against experimental evidence, nor to determine the laws of nature. Rather, experimentation served to collect as much observational data, 47 The same general direction of argumentation can be found in Falkenburg (2018), and Olson (2018). Although I agree with Olson that Kant’s gesture to Copernicus really is a gesture to the latter’s non-speculative use of hypotheses and his procedure of inferences to the best explanation, I think he goes wrong in suggesting that all scientific revolutions are essentially characterised by the way in which they conceive of their object (2018, p. 109). Instead, I argue that what these revolutions share is their focus on scientific procedure and the conception of the mind-world relation that is implicit in such a method. 48 Note that while Kant had no experience with the practice of experimentation (see, e.g., Vanzo 2012, p. 76), both Maimon—who was a pharmacist by training—and Schelling—who treated laboratory experience as part of the natural-philosophical profession—in fact did.

Scientific tools for metaphilosophy: finding the right procedure

33

i.e., “factual information on the properties and behaviour of bodies in determinate circumstances”, as possible (Vanzo, 2012, p. 77). More precisely, Bacon-Boyle-Hooke- styled experimentalism was committed to a two-stage process of experimentation.49 In a first step, large sets of data would be gathered and organised into natural histories. The second step would then consist in the generation of inductive generalisations, and the development of theories based on these data. Since the second step could only take place upon the completion of the first step, the construction of hypotheses and theories was seen as a project of the future.50 It should thus be acknowledged that, while Kant does cite Bacon to emphasise a particular image of science (i.e., as activity of observation and experimentation), he does not do so to learn from his views about the logic and art of experimentation.51 For these matters, he turns to another family of views about experimentation that is better explained in relation to figures such as Galileo and Newton. On the Galileo-Newton view, scientific inquiry does not set out to merely collect facts, but to arrive at mathematical and nomological explanations of phenomena. Kant takes up this model and reconceives its understanding of experimentation. On his conception, the main function of experiments is to test, that is, to verify or falsify hypotheses.52 In the B Preface, and in the Doctrine of Method, he argues that experimentation can neither begin in the void, nor consist in the mere observation of nature.53 Rather, when experimenting, “we must always first presuppose something here (begin with a hypothesis) from which to begin our course of investigation”, since to just “venture forth blindly” is simply “bad advice for 49 See Vanzo (2012, pp. 77-79). See also Hooke (1705, p. 7) and Boyle (1662; 1666). 50 See also Parker (1666, p. 45-46). 51 According to Galileo’s biographers, Galileo was a student of Archimedes’ work his whole life, and even called the latter “my master” (Bethune, 1829, p. 101). His work on the science of motion, for instance, takes inspiration from Archimedes’ On the Equilibrium of Planes, see Meli (2016, p. 644). Also see Hall (1963, p. 74, 80ff.). 52 See Gloy (1996) or Vanzo (2012). 53 See (KrV, Bxviii, A769-783/B 797-811). For more detailed accounts of Kant’s discussion of assumptions and hypotheses, see Chance (2015) and Butts (1961).

34

Introduction

inquiry” (AA 07: 117–333, as quoted in Vanzo, 2012, p.79). Moreover, experiments must not only begin with hypotheses, but go on to intervene into nature “like an appointed judge” (KrV, Bxiii). In this respect, Kant’s conception of experimentation does pick up on the Bacon-BoyleHooke view of experiments as active intervention into nature, but it uses these interventions to establish the laws of nature through testing hypotheses for their correctness. It is this conception of experimentation that all three philosophers (Kant, Maimon, and Schelling) will rely on in developing their metaphilosophical methods. Compared to Kant, Maimon certainly does pay attention to the logic of Baconian experimentation and acknowledges its use in scientific inquiry. Still, his metaphilosophical method makes extensive use of hypotheses, and assumes that philosophy as such must begin on one specific assumption. Schelling, too, bases his method of nature-construction on the legitimacy of constructing and testing hypotheses. Unlike the other two, his conception breaks down the boundary between philosophical and empirical experimentation, envisioning science as a whole, i.e., as grounded by Naturphilosophie, to constitute a methodological process that is experimental to the core. 0.3.2 Galilean idealisations The second important scientific practice appropriated by Kant, Maimon, and Schelling has to do with the way in which the new natural philosophy reconceived the nature and purpose of scientific representations, as well as their relation to real-world phenomena. Even within early modern natural philosophy, there existed a growing awareness of the fact that not all of the theoretical representations successfully used in the sciences constituted accurate representations of the world “as it really is”. Various thinkers drew attention to the fact that, often, scientific prediction and explanation depended on the employment of explicit, deliberate misrepresentations, which would present distorted and idealised representations of their target systems. They doubted that procedures which were based on an employment of “untruths” could truly serve to arrive at true statements about the world. Confronted with such accusations, many natural philosophers began to defend their

Scientific tools for metaphilosophy: finding the right procedure

35

methodology, which meant that they provided explanations and justifications of such employments of deliberate misrepresentations. Figures as diverse as Hume, Leibniz, and Galileo acknowledged the epistemic usefulness of employing illusions and fictions for getting at the truth of nature. Not only did this rethinking have a profound influence on the scientific practice of Kant’s day, but also on the predominant philosophy of science. One common type of idealisation that, according to McMullin (1985), characterises much of the rise of early modern natural science, and continued to impress the philosophers of the 18th century,54 was what he calls a “Galilean idealisation”. Essentially, Galilean idealisations differ from other types insofar as they simplify theories, i.e., create simplified models of their targets, “in order to make [those theories] computationally tractable” and come with an expectation “of future de-idealization and more accurate representation”.55 Galilean idealisations are employed for the purposes of both theoretical and experimental practice. McMullin distinguishes (i) distortions in theoretical structures, where idealisation consists in a simplification of the conceptual structure which represents an object, and (ii) distortions of real-world situations, in which the simplification concerns an actual physical situation or object. The first type of idealisation (i.e., “construct idealisation”) can be differentiated into formal and material idealisation. While formal idealisations simplify or omit certain features (even when they might be relevant to the explanation) in order to make the derivation of theoretical laws easier, material idealisations leave certain features unspecified because these features are deemed irrelevant to an explanation (McMullin, 54 Note that there are much earlier instances of admitted use of idealisation. In fact, the Preface to Copernicus’ De Revolutionibus, written by Osiander, tries to moderate the former’s claims by stating that this work “does not think up [hypotheses] in order to persuade anyone of their truth, but only in order that they may provide a correct basis for calculation.” (Quoted in Donahue (2016, p. 568)). 55 In more recent discussions on the nature and role of idealisations, this type is demarcated as one amongst other types of idealisations. In an authoritative survey, Weisberg distinguishes three types of idealisation, one of which is Galilean idealisation, and the other two are minimalist and multiple-models idealisation. (2007, pp. 640-41, 45)

36

Introduction

1985, p.258). The second type of idealisation (i.e., “causal idealisation”) directly concerns real-world situations, since it consists in an elimination of causal factors in order to reduce the complexity of nature to a level on which the effect under investigation becomes tractable (ibid., pp. 265-268).56 One central use of these idealisations, which shall serve as example here, is in mathematics.57 Along with the project of the mathematisation of nature (“to make geometry the language of physics” (ibid.) came the difficulty of bringing the orderly realm of ideal mathematical objects together with the messy realm of actual physical objects or natural phenomena. As part of his fictional discussions in Two Chief World Systems (1967/1632), Galileo refers to this problem, in challenging the legitimacy of his own theory, by raising the following objection: how can it be that representations of mathematical objects, e.g., simplified geometric diagrams, match some of the properties of actual physical objects and systems? He then asserts that geometry and its constructions do not, and also cannot, describe the actual, qualitative properties of physical objects or natural phenomena. In order to make geometrical constructions applicable to nature, different operations of idealisation are necessary (1967/1632, p. 203). In response to this issue, Galileo both recognises and defends the use of idealisations. He observes that it is only by simplifying target systems, and replacing them with model-systems that provide simplified representations of the original problems, that these problems can be solved in a satisfactory manner. Since “[t]he errors lie, then, not in the abstractness or concreteness, not in the geometry or physics as such, but in a calculator who does not know how to keep proper accounts” 56 McMullin dubs this type of idealisation “the most distinctively ‘Galilean’ in origin” (ibid, p.265). 57 McMullin discusses three types of idealisations, namely mathematical, construct, and causal idealisations. While the first “is a matter of imposing a mathematical formalism on a physical situation, in the hope that the essentials of that situation (from the point of view of the science one is pursuing) will lend themselves to mathematical representation” (ibid., p. 254), the latter two are concerned with simplifications of either a conceptual representation (in the case of construct idealisation), or a real-world phenomenon (in the case of causal idealisation).

Scientific tools for metaphilosophy: finding the right procedure

37

(quoted in McMullin (1985, p. 251)), it is on his view possible (at least in principle) to calculate the effects of such distortions precisely enough to enable verification of the theory or concept at issue. From this, Galileo draws the conclusion that idealisation can deal with the “practical difficulty in realizing the simple relations of the mathematical system within the complexity of the material order” (McMullin, 1985, p.251). On his assessment, natural philosophers in fact rely on the practices of distorting representations because these idealisations and models are necessary in order to make true predictions about experimental outcomes, to arrive at the correct inferences about the course of events, to determine the properties of natural objects, and eventually to construct a unified system of natural laws.58 Further, idealisation comes in different degrees. On the one hand, some idealisations only constitute small (and it that sense negligible) departures from reality that (at least in theory) could be corrected. Such small-scale idealisations make their target phenomenon computationally tractable, where it would otherwise outrun comprehension due to its complexity. On the other hand, there are also idealisations which contain larger distortions, the result of which can still “be modified in order to make allowances for the ‘departures of truth’ that the original idealization required” (McMullin, 1985, p. 257). Although Galilean idealisations are employed for pragmatic reasons, namely, to make theoretical and physical situations computationally tractable, they are ultimately used to achieve the long-term goal of de-idealisation. Kant, Maimon, and Schelling all characterise and adopt operations of idealisation, each emphasising different aspects of what was just discussed. Kant is the first to describe the necessary constituents of our conceptual apparatus as idealisations, and also to engage in a deliber58 According to McMullin’s influential paper, Galilean idealisations “signify a deliberate simplifying of something complicated (a situation, a concept, etc.) with a view to achieving at least a partial understanding of that thing. It may involve a distortion of the original or it can simply mean a leaving aside of some components in a complex in order to focus the better on the remaining ones. The point of idealization […] is not simply to escape from the intractable irregularity of the real world into the intelligible order of Form, but to make use of this order in an attempt to grasp the real world from which the idealization takes its origin.” (1985, p. 248)

38

Introduction

ate form of modelling practice to determine the possibility of a priori cognition. Maimon’s philosophical method, too, explicitly relies on the idealisation practices of natural philosophy; its core idea being that, since philosophy cannot be grounded in constitutive principles, it must work with pragmatically justified construct idealisations or fictions. Schelling turns this point on its head by arguing that it is exactly in virtue of construct and causal idealisation that philosophy can formulate constitutive principles.

0.4 Overview Throughout the next three chapters, we will learn that Kant, Maimon, and Schelling’s metaphilosophies all belong to the same family of research programmes because their views on metaphilosophical method all result from an intensive engagement with the scientific theory and practice of their time. Indeed, it is by transforming and appropriating the methodological paradigm of the mathematised and non-mathematised natural sciences into that of metaphilosophy that each philosopher manages to develop a method appropriate to the tasks of the philosophy of philosophy. This paradigm contains a new brand of experimental philosophy, designed to serve the purposes of metaphilosophy, with the ultimate goal of achieving the scientification of philosophy as such. In chapter one, I look at Kant’s conception of a propaedeutic philosophy in the first Critique and, more importantly, its revolutionary methodology that consists in the performance of “experiments of pure reason”. He initiates the metaphilosophy-first programme, calling for a methodological solution to the problem of transforming metaphysics into a proper science. His methodological solution consists in his development of the propaedeutic philosophical discipline, whose task it is to investigate philosophy for its conditions of possibility, and by introducing a new procedure to philosophy that is specifically suited to the requirements of metaphilosophy-first. In the footsteps of scientific practices such as Baconian and Newtonian experimentation, as well as more specific reflections about experimental chemistry, Kant devises a new type of philosophical experimentalism. Through experiments of pure reason, propaedeutic philosophy is equipped with the

Overview

39

odological resources to investigate the fundamental cognitive elements which condition the possibility of scientific, philosophical cognition. To explain his methodological conception, he employs an analogy to experimental philosophy, as well as to the idealising and experimental practices of experimental chemistry. By taking this analogy to be authoritative regarding Kant’s metaphilosophy-first project, the first chapter offers a new reading of the latter’s philosophical procedure in the first Critique, and lays the groundwork for interpreting Maimon and Schelling’s metaphilosophical projects. In chapter two, I discuss Maimon’s method of fictions. In the same vein as Kant, Maimon also designs a new philosophical method specifically suited to the task of investigating the possibility of scientific philosophical cognition. Unlike Kant, he believes that this commitment to metaphilosophy-first and the methodological issues that come with the task of investigating the very activity it is engaged in must lead to a philosophical methodology that employs the use of fictions. As an alternative to Kant’s propaedeutic method, Maimon delivers a methodological conception that relies on an analogy to the hypothesising and idealising practices that come with theories and procedures of differential calculus in various sciences. From there, he derives a procedure for producing “philosophical fictions” by virtue of which his philosophy of philosophy proceeds to determine the conditions of possibility of philosophical science. On his view, philosophy must always remain a hypothetical endeavour, since, as metaphilosophy, it must remain a science without solid foundations. In connecting Maimon’s method of fictions to his commitments to rational dogmatism, as well viewing it through the theoretical lens of the metaphilosophy-first program, I present a novel reading of the latter’s methodology and put it in direct relation to Kant’s propaedeutic method and Schelling’s Naturphilosophie. In chapter three, I turn to Schelling’s method of nature- construction as it is applied in his early Naturphilosophie. His philosophical project, too, sets out to develop an appropriate propaedeutic philosophy that is fit to ground philosophical science as the science of science. In contrast to Kant and Maimon, however, he is committed to finding a propaedeutic method which is not only able to grasp necessary conditions or hypothetically derived conditions, but moreover the unconditioned

40

Introduction

principle of all science. It is for this reason that Schelling conceives of propaedeutic philosophy as Naturphilosophie, and believes it can only succeed if proceeding by virtue of constructing nature through experiment. Relying on the familiar conceptions of experimentation and idealisation, Schelling suggests that the required philosophy of philosophy must construct nature itself by first postulating an absolute hypothesis, by virtue of which the conceptual structure that is realised in natural phenomena is continually discovered and tested. On his conception, philosophical experimentation is not limited to the domain of philosophy anymore; the method of nature - construction becomes the experimental method that is science. By focusing on Schelling’s new conception of scientific experiments as the real creation of phenomena, I propose a revised reading of the method of nature-construction, and what we should understand by the philosophical presentation [Darstellung] of the infinite self-production of nature.

1 Kant’s propaedeutic method

Kant wrote The Critique of Pure Reason in an effort to “promise […] to metaphysics the secure course of a science” (KrV, Bxviii-ixi). In his own words, this promise could only be fulfilled if the Critique attempts “to transform the accepted procedure of metaphysics” (KrV, Bxxii [emphasis added]). Although it has been discussed for centuries that, through his critical philosophy, Kant did advance a novel conception of philosophical method, scholarship has not paid enough attention to this method’s relation to its specifically metaphilosophical purposes. In the introduction, I have claimed that, in order to understand the nature of Kant’s philosophical method in the first Critique, we need to consider that it is tailored to the purposes and tasks of metaphilosophy-first. The ultimate goal of the first Critique is to establish what scientific philosophical cognition consists in, to show whether, and how, metaphysics can be transformed into a science. On my reading, Kant’s view requires that theoretical philosophy begins with propaedeutic philosophy, which studies the nature of metaphysical and other kinds of cognition. Due to its specific goals, propaedeutic philosophy can only get off the ground if it also presents a new method of philosophising. This method enables the kind of investigation that must be characteristic of the metaphilosophy-first approach. In this chapter, I investigate Kant’s explicit metaphilosophical views about the nature of this method, which he presented in the B Preface to the first Critique. In the introduction, I claimed that one important aspect which sets the metaphilosophy-first programmes of Kant, Maimon, and Schelling apart from those of other contemporaries is their specific relation to the theories and practices of the sciences of their time. I took this aspect to account for their specific family resemblance.

42

Kant’s propaedeutic method

This chapter makes evident that Kant’s propaedeutic method clearly testifies to this thesis. It argues that Kant is drawing on the theories and practices of several experimental sciences in order to explicate the nature of the metaphilosophical method, which is to investigate the possibility of a scientific metaphysics. Kant characterises this method as a specific type of experimentation, namely “experiments of pure reason”, whose argumentative procedure is to be explained by virtue of an analogy to chemical experimentation (KrV, Bxxi). Following the pioneering work of Brigitte Falkenstein (2018), Ian Proops (2021), and Jan Sticker (2017), I shall claim that, in describing this method, Kant employs a two-layered analogy to the natural sciences. One the one hand, he drew on a specific conception of experimentalism, which he arrives at through his analyses and theoretical development of the scientific works of Bacon, Galileo, Newton, and others. On the other hand, he combines his theory of experimentation with some further reflections on the procedures of chemical analysis and synthesis, at work in the emerging science of experimental chemistry propagated by figures such as Stahl, Erxleben, Karsten, and others.1 Only by understanding these analogies can it become clear how Kant’s methodological solution to the problem of the scientification of metaphysics works out. Interpreting Kant’s analogy to the natural sciences, the following chapter will provide a new perspective on Kant’s project and method in the first Critique. In light of his descriptions in the B Preface, propaedeutic philosophy must be understood as employing an experimental method which establishes the possibility of scientific cognition in metaphysics. This experimental method can be put to use by critical philosophers, who thereby attain the necessary means to engage in an investigation of the human cognitive faculties. Since this investigation constitutes one in which the thing that is being investigated is identical with the thing that is executing this same investigation, it can be described as a self-investigation of reason (cf. KrV, A754-5/B782-3).2 1 See McNulty (2014b, pp. 13-39) for an overview of the chemists that mattered for Kant’s own work. 2 Proops notes that Kant identifies the critical experiment taking place in the Dialectic as an experiment which philosophers must undertake for themselves. He says that “Kant assumes, with some plausibility, that insofar as each of us

Kant’s propaedeutic method

43

On my view, Kant takes the propaedeutic method to carry out a selfexperiment on human cognition. This experiment proceeds under the assumption that the possibility of scientific metaphysics depends on the correctness of a specific theory of cognition. This is what the experiment of reason intends to prove. I shall argue in more detail that experiments of pure reason are designed to investigate inner cognitive structures, whose properties explain the epistemic qualities of our cognition, in the same way that chemistry investigates the inner structures of substances whose properties explain the qualities of these compounds. Chemical analysis and synthesis provide the specific methods of discovery and proof which critical philosophy, in its metaphilosophical functions, requires in order to arrive at non-trivial and valid conclusions. Kant hypothesises that those cognitive structures consist in heterogeneous kinds of a priori representations (this is the “Heterogeneity” thesis), and that these a priori representations must cooperate in any cognition for it to qualify as such (this is the “Cooperation” requirement). Its focus on establishing and investigating the elements of cognition, i.e., the a priori representations which must be involved in cognition in general, suggest a distinct interpretation of the goal and outcome of the Dialectic. If the analytic and synthetic procedures of his experiment of reason are understood as constituting operations of chemical decomposition and recombination, then it becomes clear how philosophical experimentation can prove that the products of metaphysical speculation are (or are not) of the right representational composition to count as scientific cognitions of objects. Consequently, Kant’s description of experiments of pure reason reveals that their conclusions concerning transcendental idealism constitute meta-epistemological, and not metaphysical, statements. This chapter proceeds in five steps. In a first step, I give a brief sketch of Kant’s propaedeutic philosophy as a metaphilosophy-first programme, as well as some relevant concepts and terminology. The second step contains a short outline of what we should understand instantiates ‘common human reason’ we are thereby in possession of reason’s incipient dialectic and so able to develop it for ourselves (compare 8: 226 note *)” (2021, p. 215).

44

Kant’s propaedeutic method

by Kant’s two-layered analogy. Steps three and four will each contain detailed analyses of the analogy between metaphilosophical method and Baconian and Newtonian experimentalism on the one hand, and chemical analysis and synthesis on the other. By connecting these insights to interpret Kant’s methodological remarks in the B Preface, the fifth step presents a novel description of Kant’s conception of what metaphilosophical procedure should consist in, and how this meshes with the methodological requirements of metaphilosophy-first.

1.1 Propaedeutic philosophy is the study of philosophy There is little doubt that Kant, in many ways, has changed the fate of metaphysics. One major source of his contributions in this field are his metaphilosophical studies on the nature of metaphysics and his concern for the scientification of this philosophical discipline. As stated previously, I take it that Kant argued for a metaphilosophy-first approach which holds that the scientification of one philosophical discipline can only be achieved through the inception of another philosophical discipline: metaphilosophy. I interpreted Kant’s views as expressing a metaphilosophy-first approach; metaphysics can be transformed into a science if its nature is first studied and determined through metaphilosophy. In this section, I show why this metaphilosophy must be identified with that “special science, which can be called critique of pure reason” (KrV, A11/B24), and why we should take seriously Kant’s statement that the first Critique has the character of a “propaedeutic” (KrV, A841/B869, cf. KrV, Bxviii-xix).3 To articulate the specific purpose and tasks of this propaedeutic, it will be necessary to introduce some key concepts from Kant’s theory of cognition and metaphysics. In the Prefaces and Introductions in particular, Kant clearly presents his metaphilosophical project as a condition for the transformation of metaphysics into a proper science.4 Although it is true that the 3 It is important to keep in mind that, for reasons that I gave in the introduction, I will only be concerned with Kant’s theoretical philosophy, and not his practical philosophy. 4 Similar takes can be found in Serck-Hanssen (2017), Heidegger (1990),

Propaedeutic philosophy is the study of philosophy

45

answer to the metaphilosophical question (i.e., ‘What makes philosophy scientific?’) is connected to the idea that speculative philosophy as first science must be some form of epistemology or new metaphysics, it is still this metaphilosophical question that is prior in the order of investigation. Hence, I do not contend that Kant’s project can be characterised in either epistemological or metaphysical terms. Its primary objective is not to provide a theory of knowledge and justification, or a new form of general metaphysics, but to show that, and how, metaphysics is possible as science. Certainly, an account of what theoretical philosophy must be like in order to count as science includes a theory of knowledge and justification, as well as a general metaphysics, a theory of representation, and of object constitution. Yet, all of this is done in service of reforming philosophy, of metaphilosophy-first. Kant’s metaphilosophy-first program leads him to develop a special philosophical discipline, which studies and determines the nature of metaphysics. He repeatedly characterises this discipline as “propaedeutic” (KrV, A841/B869; cf. Bxviii-xix). The task of this propaedeutic is to determine the concepts and principles in virtue of which human cognisers can know of objects a priori. As “propaedeutic (preparation)”, this philosophical procedure has to “investigate […] the faculty of reason in regard to all pure a priori cognition”, to establish which cognitive conditions are constitutive of the scientific cognition of objects (KrV, A841/B869; cf. Bxviii-xix). By conceiving of its goals in the terms just introduced, propaedeutic philosophy transforms its initial question (Is metaphysics possible as science?) into a more specific one: how is scientific metaphysical cognition possible? Kant’s metaphilosophy-first project makes an important theoretical assumption: that an answer to the metaphilosophical challenge must include the development of a theory of cognition. To determine the possibility of scientific metaphysics, he believes, the propaedeutic procedure must be employed to establish the conditions of the possibility of representing objects. In that sense,

lison (2004), as well as in Tonelli & Chandler (1994), which interpret the first Critique as a methodological work in meta-epistemology, or metaphysics, respectively.

46

Kant’s propaedeutic method

Kant’s metaphilosophical project does take a decisively epistemological (or meta-epistemological) turn. Now, if we believe Kant’s words in the B Preface, he arrived at this new assumption through surveying the development of several sciences, e.g., logic, mathematics, and physics. In doing so, he was looking for potential causes of their transformation into proper sciences. His analysis leads him, first, to recognise a general “change in our ways of thinking” (KrV, Bxvi), which he believes to make for a shared property of all scientific revolutions. He conceives of this change as constituting a fundamental shift in the ways in which we theorise the relation between mind and world. This reconfiguration of the relation between mind and world is what came to be known as Kant’s “Copernican revolution”:5 Up to now it has been assumed that all our cognition must conform to the objects; but all attempts to find out something about them a priori through concepts that would extend our cognition have, on this presupposition, come to nothing. Hence let us once try whether we do not get farther with the problems of metaphysics by assuming that the objects must conform to our cognition, which would agree better with the requested possibility of an a priori cognition of them, which is to establish something about objects before they are given to us. This would be just like the first thoughts of Copernicus, who, when he did not make good progress in the explanation of the celestial motions if he assumed that the entire celestial host revolves around the observer, tried to see if he might not have greater success if he made the observer revolve and left the stars at rest. (KrV, Bxvi [emphasis added])

Without getting into the nitty-gritty details of Kant’s analogy with Copernicus’ account in De revolutionibus orbium coelestium (1543), three points remain important for our purposes. First, on Kant’s view, the Copernican revolution implies a reconfiguration of metaphysical theories and their fundamental concepts. In fact, when proposing to address “the problem of metaphysics” through the hypothesis “that the objects [of a priori cognition] must conform to our cognition”, Kant opts for an epistemological theoretical solution (KrV, Bxvi). Further, he 5 It seems quite obvious that, by his assumption that the observer revolves around the sun, Kant did not mean to say that we, the epistemic subjects, revolve around the sun, but primarily the planet we are located at. Otherwise, the comparison is not analogical in content (see Schönecker, Schulting, & Strobach (2011)).

Propaedeutic philosophy is the study of philosophy

47

believes that, through this reconfiguration of the conceptual apparatus used, we will not only be able to show the possibility of metaphysics, but also explain why science can in fact ‘carve nature at its joints’. The second salient point to note about Kant’s analogy is that this reconfiguration of the relation between cognition and its objects also has important implications for conceptualising the relation between scientific theories and their objects. According to the Copernican thesis, sciences become successful once they have begun treating their objects of study as things “produce[d] according to [reason’s] own design” (KrV, Bxiii). The third feature to note (as pointed out by Olson (2018, p. 95)) is that Kant’s analogy is not only one of content but also one of form.6 Kant simultaneously suggests that philosophy should adopt Copernicus’ idea of a change in perspective, and that it should take an example from his method of hypothesis-construction, which will be central for my argumentation in chapter one. So, Kant argues that any epistemic practice which has been transformed into a proper scientific method must have first undergone this paradigmatic shift.7 That is, both mathematics and physics have fun6

“If one attempts to give some determinate content to the formal similarity of Copernicus’s and Kant’s changes in perspectives, the analogy is consumed by a rather striking irony. Under greater scrutiny, Kant’s Copernican revolution appears considerably more Ptolemaic than Copernican. If Copernicus’s great contribution lies in his heliocentric reorganization of Ptolemaic astronomy, the broader consequences of his hypothesis urge a reconsideration of the priority given to the human, terrestrial perspective and its supposed centrality in the universe as a whole. Kant’s proposed revolution in metaphysics, on the other hand, argues for the irreducibility of the active contribution of the thinking subject to cognition, and so argues in favor of just that centrality of the human perspective that Copernican astronomy rejects.” This apparent contradiction can be resolved if we interpret the analogy to be based on “the human perspective” as “where human thinkers are located, i.e., planet earth”, whose movement and thus contribution gives the necessary ingredient to correctly calculate the planetary trajectories (Olson 2018, p. 95). 7 According to Kuhn, concepts or theories are ultimately embedded within a scientific paradigm: a set of theoretical assumption guiding the direction and methods of inquiry, the standards of truth which define “normal science” within scientific discipline (see Kuhn, 1962, p. 102). His main point in describing science according to “paradigms” is to be able to appropriately describe scientific change and progress: the history of science is presented as a succession

48

Kant’s propaedeutic method

damentally (but implicitly) reconceived the constellation of concepts by virtue of which the relation between concept and object is explained. He contends that, by implicitly assuming that “reason has insight only into what it itself produces according to its own design”, these sciences discover new methods of inference and evidence-production (KrV, Bxiii).8 Mathematicians, for example, have revolutionised their method by understanding that in order to know something about a geometrical concept, they “had to ascribe to the thing nothing except what followed necessarily from what [they] had put into it” (KrV, Bxii [emphasis added]). Constructing a triangle means producing a triangle in accordance with its concept. Mathematicians derive a non-trivial proposition that is not analytically entailed by the concept, and know of its truth, because they generate a triangle in pure intuition with such and such properties. For Kant, these examples of mathematical cognition reveal the fundamental insight: that we can only know a priori what comes from us—we can know it because we generated it. The same goes for the empirical sciences, which, “since the suggestion of the ingenious Francis Bacon partly occasioned this discovery and partly further stimulated it” (Bxii), “owes the advantageous revolution in its way of thinking to the inspiration that what reason would not be able to know of itself and has to learn from nature, it has to seek in the latter (though not merely ascribe to it) in accordance with what reason itself puts into nature” (Bxiii-iv). Experiments, as described and classified by Bacon, were conceived of as active interventions into the course of nature, which led to the conception of experimentation which treated nature as something to be controlled and manipulated through experiments. This conception likewise, and even more importantly when speaking of Copernicus’ hypotheses about the central laws of motion, prompted Newton to postulate another theory of the same laws, unified by the law of universal gravitation (KrV, Bxxii). In these examples, a theory precedes its objects. Hypotheses are constructed before observation and data collection, and even more, hypotheses and of paradigms, which are interrupted through scientific revolutions, leading to the exchange of an old paradigm for a new one. 8 This is also reminiscent of Vico’s “verum et factum convertuntur” [the true and the made are convertible] (see DA 131/45).

Propaedeutic philosophy is the study of philosophy

49

theories inform and control the way in which their object of study will be investigated; they enable insight into what they have produced ‘according their own design’. Consequently, Kant suggests that metaphysics should also be reconceived according to this model. The possibility of a scientific metaphysics, too, must be considered from an epistemological perspective. Propaedeutic philosophy must engage in the study of the nature of metaphysics, with the goal of establishing the conditions of possibility of metaphysical cognition. Kant’s metaphilosophical programme assumes that the intended self-investigation of human cognition must consist in a determination of those a priori representations, which make possible the various kinds of a priori cognition through which metaphysics and the other sciences proceed. Since Kant’s Critique will be concerned with many kinds of cognition, and not just the metaphysical one, the propaedeutic simultaneously takes up the role of that procedure which establishes the meta-epistemological framework that underlies any scientific practice of description, prediction, or explanation. Now, two assumptions stand at the beginning of this investigation into the a priori concepts and principles of human cognition. First, the assumption that “all our cognition begins with experience” (KrV, B1). Kant starts from the epistemological assumption that human cognition is finite cognition, from which he infers that finite cognition arises from experience. Unlike infinite or divine cognition, which creates its objects of cognition, human cognition as finite cognition is conditioned, in the sense that its objects must be given to it from without, through experience. And this, Kant stresses, “is possible only if [they] affect the mind in a certain way” (KrV, A19/B33 [emphasis added]).9 In contrast to infinite cognition, finite cognition must be discursive cognition. Discursive cognisers must be in possession of two faculties of cognition: one that passively receives sensory information about particular objects, and one that spontaneously orders and categorises the received sensible data under general concepts (KrV, A50–1/B74–5). Second, “although 9 Kant says that intuitions depend on their objects, because they stand in an immediate relation to their objects. To have an intuition is to know of the immediate presence of an object (B32, 72).

50

Kant’s propaedeutic method

all our cognition commences with experience, […] it does not on that account all arise from experience” (KrV, B1). As discussed above, Kant assumes that our cognition of objects is conditioned by a determinable set of a priori representations. From these two assumptions, Kant begins to explicate his theory of which elements constitute the cognition of objects, and in what way. This theory is grounded in two theoretical commitments: the Heterogeneity thesis and the Cooperation requirement.10 (Both commitments will also play a crucial role in the interpretations of Maimon and Schelling that I will discuss in subsequent chapters.) Very broadly, the Heterogeneity Thesis states that the cognition of objects can only be explained if we posit that it is constituted through the contribution of two separate sources of cognition, sensibility and understanding.11 Kant assumes that the a priori forms of cognition, which are the conditions under which any object is presented to human cognition, are heterogenous in nature. Kant’s rationale for assuming this theses is roughly the following: only if there is some element in human cognition that explains how it can relate to objects that are not of its own creation—a receptive faculty through which objects “are given”, and a spontaneous faculty through which objects “are thought” (KrV, A19/B33), can we make intelligible how human cognition can determine something about an object in general.12 10

See, e.g., Land (2012, p. 1). I leave open the question how to specify the explananda: whether “things” are, say, entities or events. On dualism, compare Guyer (2000), who argues that the dualism of concept and intuition is fundamental. I agree that the dualities of understanding and sensibility, reality and appearance, depend on the duality of concept and intuition. But I find the duality of finite and infinite intelligibility to be still more fundamental. As Guyer documents, Kant took numerous steps towards the concept/intuition duality throughout his precritical work, but the final, critical step was taken only with the notion of a form of intuition in the first Critique. On my account, that crucial, innovative notion is Kant’s response to the realisation that finite intelligibility is neither reducible to infinite intelligibility, nor supervenient upon infinite intelligibility, nor fully self-standing, nor dispensable. For some reflections on this line of thought, see Franks (2017). 12 “In whatever way and through whatever means a cognition may relate to objects, that through which it relates immediately to them, and at which all 11

Propaedeutic philosophy is the study of philosophy

51

Moreover, Kant does not only determine the conditions of possibility for representing objects through the postulation of a priori cognitive elements. By arguing for two heterogenous kinds of cognitive elements, one sensible, one conceptual, he can explain the possibility of thought, and hence of cognition, as well as the possibility of that which is thought or cognised. From this flows Kant’s Cooperation Requirement: that “[t]houghts without content are empty” and that “intuitions without concepts are blind” (KrV, A51/B76). For cognition of objects to be possible, he argues, both conditions must be employed in cognition. Kant claims that “[w]ith us understanding and sensibility can only determine an object only in combination” (KrV, A258/B314).13 This will play an important role when it comes to the evaluation of Kant’s propaedeutic project, which will make the claim that certain types of metaphysical arguments do not count as cognition of objects. From the Heterogeneity thesis and the Cooperation requirement, Kant derives another important theoretical consequence which specifically concerns the nature of metaphysical cognition and its objects: the thesis that transcendental idealism is true (KrV, A369).14 For Maimon and Schelling, transcendental idealism expressed a metaphysical thesis which distinguishes between two classes of objects: appearances and things in themselves. Objects of the first class, i.e., things in themselves, exist independently of human cognisers to perceive them, and so do their properties. Objects of the second class, i.e., appearances, only exist in relation to human cognisers—their properties depend on human cognisers—which is why they have often been characterised as mental entities or mental representations. Readings of this type have thought as a means is directed as an end, is intuition.” (KrV, A19/B33) Kant will specify these conditions as the pure forms of sensibility (i.e., space and time) and the pure forms of the understanding (i.e., the categories). Heidegger, for one, raises this point (see Heidegger (1997, pp. 15-18, 30, 50-51)), as well as many contemporary interpretations after him, e.g., Allison (2004, pp. 13-14) and Engström (2006, pp. 9-13). 13 See also (KrV, B296ff/A237). 14 My reading here agrees with Allison’s (2004), insofar as I interpret Kant’s theses about transcendental idealism to be consequences of his epistemological theses (and not the other way around). This is the view, as we will see in sections III and IV, which Kant’s chemical analogy suggests.

52

Kant’s propaedeutic method

later been categorised under the notion of the “two-objects view”.15 In anticipation of the new interpretive perspective that my reading of Kant’s two-layered analogy provides, however, I want to draw attention to a second possible reading, which distinguishes itself from the former view by proposing a “two-aspects view”, and more particularly, an epistemological interpretation of this reading. According to Allison’s influential interpretation, Kant’s transcendental idealism should be understood in epistemological terms, that is, as I suggested above, transcendental idealism should be taken to be a consequence of Kant’s theory of cognition.16 This “epistemologically based understanding of transcendental idealism”, then, requires that the transcendental distinction between appearances and things in themselves be understood as holding between two ways of considering things (as they appear and as they are in themselves) rather than as, on the more traditional reading, between two ontologically distinct sets of entities (appearances and things in themselves). In this regard, it may be characterized as a “two-aspect” reading. (Allison, 2004, p. 32)

Against “metaphysical interpretations”, Allison assumes that the distinction between appearances and things in themselves gets its meaning in virtue of the contents of the Heterogeneity thesis (and the Cooperation requirement). According to the Heterogeneity thesis, the representation of objects from a human standpoint is conditioned by two classes of cognitive elements: a priori forms of sensibility and a priori forms of understanding. If we consider objects as objects of cognition, we must consider them as falling under their relevant “epistemic conditions”, otherwise we “would not be cognizing but misrepresenting them” Stang (2022, section 4.2).17 So, to consider objects as objects of our cognition is to consider them as objects which fall under the specific epistemic conditions of our spatiotemporal discursive cognition. If we consider them 15 For an updated presentation and discussion of these and other readings of Kant’s transcendental idealism, see Stang (2022). 16 In fact, this type of reading was first advanced by Graham Bird (1962) and Gerold Prauss (1974). 17 “Epistemic condition” is Allison’s term for those a priori representations which must be applied to objects in order for them to be constituted as objects of cognition. As Stang puts it, “[i]f E is an epistemic condition then necessarily if we know an object O, in knowing it we represent it using E.”; see Stang (2022).

Propaedeutic philosophy is the study of philosophy

53

as things in themselves, on the other hand, we do not represent these objects according to the epistemic conditions which affords cognition of objects, but only according to the epistemic conditions of discursive cognition. In this case, we represent objects as satisfying the conditions of the understanding, but not the conditions of sensibility. And this way of considering objects will not count as representation of objects of cognition, but only of “objects [as they] are merely thought” (KrV, Bxviii; cf. Bxxvi). So, via the transcendental idealism thesis, but ultimately because of his theory of cognition, Kant derives the following results regarding his initial, metaphilosophical research question: [f]rom this deduction of our faculty of cognizing a priori in the first part of metaphysics, there emerges a very strange result, and one that appears very disadvantageous to the whole purpose with which the second part of metaphysics concerns itself, namely that with this faculty we can never get beyond the boundaries of possible experience, which is nevertheless precisely the most essential occupation of this science. (KrV, Bxix)

Human finite cognition, since it depends on two stems of cognition, one of which is receptive in nature, cannot cognise things in themselves, but only things as they appear to us (KrV, Bxx). Propaedeutic philosophy establishes the a priori conditions of possibility for the cognition of objects, and these objects are such that they are represented according to the a priori forms of sensibility. Since, for Kant, the “second part of metaphysics” (metaphysica specialis)18 is the science of things as they are in themselves, it cannot be pursued in a scientific manner. This is what is argued in the third part of the Doctrine of Elements, the Dialectic, which is also structured according to the three objects of special metaphysics (the soul, the world, and God). We will return to these matters in section 4.2. It is certainly important to keep in mind that propaedeutic philosophy not only determines whether, and under what conditions, metaphysical cognition is scientific, but also does so for mathematical and empirical cognition. As Kant states in the B Preface, “with this altera18 It has been argued that Kant hereby alludes to the traditional distinction between general and special metaphysics—see for example De Boer (2015). For a discussion of the history of this distinction, see Vollrath (1962, pp. 258-284).

54

Kant’s propaedeutic method

tion of our way of thinking we can very well explain the possibility of a cognition a priori, and […], we can provide the satisfactory proofs of the laws that are the a priori ground of nature, as the sum total of objects of experience” (KrV, Bxix). According to the B Preface, then, propaedeutic philosophy is concerned with the representational structures that constitute the a priori cognition of all objects within the domain of objects of possible experience. In fact, all propaedeutic philosophy can establish is the representational structure that conditions objects of cognition in general. It must determine how, and in virtue of what, any scientific (or non-scientific) cognition of objects is possible. An anticipation of the domains of objects we can access in this way, and hence what kinds of knowledge claims (scientific and nonscientific) will be possible for us to make, is part of this determination. Propaedeutic philosophy determines how knowledge claims can have objective validity. For the purposes of illustration, consider a physicist who is trying to determine the laws of motion of bodies through experiments and observation. For her to be able to track an object, this object needs to have a spatiotemporal location, by virtue of which it can be identified and re-identified. Similarly, the physicist assumes that it can enter into causal relations which explain the changes in state it undergoes. For Kant, these are all features according to which physical objects must be structured in order for physical knowledge claims to have objective validity, for statements referring to them to be true or false. In this way, Kant’s new philosophy sets out to determine the metaphysical foundations of the natural or “exact” sciences.19 This general outlook on Kant’s metaphilosophy can be summarised 19 Friedman (1992; 2013). Kant does not fully achieve this foundation of the sciences in the Critique of Pure Reason. While the first Critique demonstrates why and how it is that the objects of our senses afford mathematisation, he has not yet shown why possible experience must necessarily be structured according to the a priori laws of physics. In his thorough study, Friedman shows that it is in the Metaphysical Foundations of Natural Science (originally published in 1786) that Kant specifies the concepts and principles of the understanding with the given concept of matter to derive the mechanical laws and other a priori principles of physics. On his interpretation, Kant demonstrates that an application of the concepts and principles of understanding and intuition to the empirical concept of motion yields the basic principles of Newtonian physics.

Propaedeutic philosophy is the study of philosophy

55

by means of the following theses: Firstly, Kant asks whether a scientfication of metaphysics is possible. Secondly, he argues that an answer to this question can only consist in a self-investigation into our cognition, requiring metaphysics to be preceded by a metaphilosophical propaedeutic. This is why Kant calls his Critique “a treatise on the method and not a system of the science itself” (KrV, Bxxii), formulating the procedure to transform metaphysics into a proper science. Yet what does this procedure on which the fate of all future metaphysics depends consist in? Interpreting the Critique as metaphilosophy-first means that we have to look for Kant’s explicitly metaphilosophical views. This again means that the focus of this study will not lie in Kant’s views about the nature of science, or scientific cognition in general. The Critique enters metaphilosophical terrain by theorising about the scientification of philosophy, and, more specifically, the nature and methods of metaphysical cognition. As a metaphilosophy-first project, it provides explicit reflections about what kind of procedure is needed for its metaphilosophical purposes, to establish its own legitimacy without running into a regress or admitting circularity. This then leads me to my main question: in virtue of what method does propaedeutic philosophy proceed; how can it reach its purpose? On my interpretation, metaphysics must undergo a methodological revolution in order to become a science, not only because it must determine its own conditions of possibility, but also, and more importantly, because Kant thought that this propaedeutic needed to come up with a procedure in virtue of which philosophy can investigate its own nature and method. The Critique is not only a work on philosophical method, but also a work of a particular philosophical method. There are at least three options which Kant could have chosen in dealing with this metaphilosophical circle. The first option would entail that the Critique itself was produced by a different method than the one it arrives at. Then, however, Kant has to show the legitimacy of his first methodological choice, and is facing the same charge again. On the second option, Kant could have admitted that the Critique pursues its goals in a circular manner by declaring that the Critique already applies the same principles it is arguing for. A third, and less attractive, option regarding the Critique’s own method has to be reckoned with as well: that it was produced in

56

Kant’s propaedeutic method

accordance with no particular method at all. I shall argue that Kant’s propaedeutic combines the first and the second option. Propaedeutic philosophy “under the name of a discipline erects, as it were, a system of caution and self-examination out of the nature of reason and the objects of its pure use” (KrV, A711/B739) that only ever prepares reason for its scientific use in metaphysics. For this purpose, it adopts its own philosophical procedure which—as we will see—consists in the performance of so-called “experiments of pure reason” (KrV, Bxxi). Moreover, this specific methodological procedure reflects an awareness of the specific problems that metaphilosophy as philosophical discipline must face. It offers a way, so Kant believes, to arrive at an impartial decision about the scientificity of metaphysics. On my view, what justifies his views are the specific features that characterise his experimentalist methodology. His metaphilosophical procedure employs methodological paradigms from the experimental sciences, specifically from the chemical sciences. On this conception, the propaedeutic method turns out to be impartial—because it is experimental—and circular in a nonvicious way—because it is experimental in the chemical sense.

1.2 A two-layered analogy: Baconian and Newtonian experimentalism, and chemical analysis and synthesis So far, we have seen that a possible response to Kant’s guiding question requires a specific conception of metaphilosophy as propaedeutic, and the development of a philosophical method suited for these purposes. In what follows, I want to motivate my claim that a correct analysis of the nature of this method must take seriously Kant’s remarks that his methodology imitates the methods of certain experimental sciences. It is rarely noted that, in the B Preface to the first Critique, Kant introduces a two-layered analogy, with the explicit goal of explaining the specific procedure that constitutes his propaedeutic philosophy. He states that his own metaphilosophical procedure successfully imitates the experimental methods of the natural sciences, and more specifically, those of experimental chemists. He takes this to indicate that there exists a relation of similarity between two relations between dissimilar things:

A two-layered analogy

57

Kant draws an analogy between metaphilosophical and some naturalscientific procedures and their respective methods of proof: both are experimental. Karin Gloy (1996) and Brigitte Falkenburg (2018) have argued that in these passages Kant introduces what should be labelled a methodological analogy.20 I agree on this general point, which emphasises two of the major claims of this book, namely that Kant did conceive of his metaphilosophical method by comparing it to the theories and practices of the sciences of his time, and that he saw a continuity between the methods of experimental philosophy and propaedeutic philosophy. Yet, while Gloy and Falkenburg argue that this methodological analogy is “between scientific [naturwissenschaftlich] and metaphysical cognition” (Falkenburg, 2018, p. 655), I will argue that Kant’s methodological analogy supports inferences about the nature of metaphilosophical cognition, about the kind of cognition which the Critique of Pure Reason is a treatise of. Before introducing the analogy itself, let me briefly explain what I take analogical reasoning—as exemplified in Kant’s case—to consist in. For an argument to constitute an analogy, it must assume that the target system, which it aims to explain, resembles the base system in some relevant respects. Very broadly, analogical arguments then proceed by stating one or several accepted similarities between two objects, or systems of objects, to justify the claim that there exist further similarities between the two. Thus, analogical arguments need to invest the assumption that the base system and the target system share certain properties, and that because the same properties give rise to further properties in the base system, this must also be the case for the target system.21 As for Kant, we can specify this description in the following way. He holds that an analogy “surely does not signify, as the word is usually taken, an imperfect similarity between two things, but rather a perfect similarity between two relations in wholly dissimilar things” (AA 04, 20 What’s more, Fulkerson-Smith (2013) connects Kant’s methodology in the Critique to a particular conception of experimentalism, i.e., that of Bacon and his notion of a crucial experiment. However, for reasons I will discuss in III.1, I warn against overemphasising Kant’s debt to Bacon. 21 Bartha (2019) makes a similar point.

58

Kant’s propaedeutic method

357). That is, on his understanding, analogies exhibit the structure of a double ratio, in which A stands to B just as C stands to D (“A : B = C : D”) (Falkenburg, 2018, p. 653). One example he uses to illustrate his point concerns the relation that a watchmaker has to their watch and the relation God has to the world (Falkenburg, 2018, p. 653; AA 04, 357). So, what relations is Kant’s B Preface analogy concerned with?22 Kant says that [t]his method [i.e., the propaedeutic method], imitated from the method of those who study nature, thus consists of this: to seek the elements of pure reason in that which admits of being confirmed or refuted through an experiment. (KrV, Bxviii [emphasis added])

The first layer of the analogy, as it concerns his philosophical method, works between the experimental method of the natural sciences as base system and the experimental method of propaedeutic philosophy as target system. As I will explain in section III.1, the analogy obtains between the relation between empirical cognition and experimental validation, and philosophical cognition and consistency.23 Kant’s analogy is grounded in a relatively new theoretical conception of how experimentation in the natural sciences works. Before going into the details of this analogical layer, let me also indicate the second layer of this same analogy. Some paragraphs later, Kant further specifies the nature of philosophical experiments by introducing a second layer of the analogy to the natural sciences. He points out that [t]his experiment of pure reason [i.e., the philosophical experiment that takes place in the Critique] has much in common with what the chemists sometimes 22 Another positive analogy that I will not discuss here is the analogy to juristic practice and the legal notion of ‘deduction’ (see Henrich (1969), Proops (2003), and Møller (2013, 2020)). Møller shows that although Kant “compares different aspects of the critique to different methods and procedures at courts of justice”, the claim that one legal practice, e.g. the ‘Deduktionsschriften’, “should be a ‘methodological paradigm’ for the entirety of the Critique of Pure Reason, as Henrich claims, […] [is] ill-founded.” (2013, p. 318) What is more, on this particular analogy, it does not become clear in what sense philosophical argumentation can be distinguished into analysis and synthesis. 23 Here, I follow Falkenburg’s suggestion (2018, p. 656).

From empirical experimentation to a priori experimentation

59

call the experiment of reduction, or more generally the synthetic procedure. (KrV, Bxxi [emphasis added])

More precisely, [t]he analysis of the metaphysician separated [schied] pure a priori knowledge into two very heterogeneous elements, namely those of the things as appearances and the things in themselves. The dialectic once again combines them, in unison with the necessary rational idea of the unconditioned, and finds that the unison will never come about except through that distinction, which is therefore the true one. (KrV, Bxxi)

In this second layer of his methodological analogy, Kant proposes that the specific procedures which constitute philosophical experiments do not only stand in analogy to the experimental procedures of the natural sciences in general—which could include physics, as well as chemistry or mechanics—but that they stand in specific analogy to a combination of procedures used in experimental chemistry, and more precisely, the procedures commonly known as chemical analysis and synthesis. Kant submits that the relation between chemical analysis and synthesis is analogical to the relation between the philosophical procedures in the Analytic on the one hand, and the philosophical procedures employed in the Dialectic on the other.

1.3 From empirical experimentation to a priori experimentation Before turning to Kant’s notion of the philosophical experiment, it is crucial to understand his general theory of experiments. As we will see, this conception arises from his engagement with two distinct conceptions of empirical experimentalism. By connecting the two, Kant introduces a revolutionary conception of experimentation which, although it was certainly already used in the scientific practice of his day, he was the first to theorise about. In Kant’s time, philosophical experimentalism was certainly not something entirely new. Empiricist philosophers such as John Locke and David Hume had already suggested that philosophers should introduce the experimentalist methods of the early natural sciences into their own disciplines. Their philosophical projects were committed to the

60

Kant’s propaedeutic method

experimentalist programs of Bacon and Newton. For them, it was clear that natural philosophy, and philosophy more generally, must become experimental, and abandon the speculative explanations of 17th century mechanical philosophy. Against armchair methods of inquiry such as speculative demonstration from first principles, the new experimentalists promoted the use of observation and experiment in natural science and beyond.24 As discussed in the introduction, two of the most central doctrines of this movement are perhaps Bacon, Boyle, and Hook’s conception of natural histories and Newton’s thoughts on the proper use of hypotheses. On the former view, experimental philosophy—before it can erect any hypotheses—must dedicate itself to the observational and experimental collection of large sets of natural facts, which are then to be organised into natural histories. In a similar vein, Newton had argued that experimental philosophy must adopt a proper use of hypotheses in order to distinguish itself from “speculative” 17th century philosophy. He states that it is never legitimate for scientists “to feign hypotheses, for whatever is not deduced from the phenomena is to be called a hypothesis; and hypotheses […] have no place in experimental philosophy” (Newton, 1999, p. 493). Building on these methodological programs, Locke and Hume ventured to restructure other philosophical disciplines, most important among them the study of human understanding and moral philosophy, to conform to the methodological ideals of experimental philosophy. In the introduction to his Essay Concerning Human Understanding (1975/1690), for example, Locke claims to have adopted an experimental method to investigate the nature of human understanding. Reflecting on the “Design” of his philosophical investigations, he notes that [i]t shall suffice to my present Purpose, to consider the discerning Faculties of a Man, as they are employ’d about the Objects, which they have to do with: and I shall imagine that I have not wholly misimploy’d my self in the Thoughts I shall have on this Occasion, if in this Historical, Plain Method, I can give any Account of the Ways, whereby our Understanding comes to attain those Notions of Things, and can set down any Measure of the Certainty of our 24

For an overview, see Anstey & Vanzo (2012).

From empirical experimentation to a priori experimentation

61

Knowledge, or the Grounds of those Perswasions, which are to be found among Men. (I.1.1.2; N, 44 [emphasis added])

Here, Locke makes what looks like an obvious reference to the BaconBoyle-Hooke programme of constructing natural histories. If this is indeed what he meant to convey, and many have believed so, then this passage provides us with evidence that Locke thought of his science of the human understanding as a form of experimental philosophy. If we take this reference seriously, then this would mean that through his “plain, historical method”, Locke was conducting research in the style of collection of observations and experiential evidence for the purposes of philosophical processing. And, indeed, it has been argued that Locke, in the spirit of Bacon’s natural histories and Newton’s hypotheses non fingo, promoted his “science of man” as a science which begin its inquiries on the solid ground of experiential observations and evidence.25 On this view, philosophy about human nature—like natural philosophy—must proceed by experimentally generating different phenomena and types of cases to construct and organise natural histories of man. Only such organised collection can deliver the necessary material to generalise over, to be able to draw inductive inferences from, in order to eventually arrive at law-like propositions. A few decades after Locke, Hume, too, started advocating for the replacement of traditional speculative philosophy with a philosophy that adheres to the methodological principles of the experimentalist program. Most notably, he took his Treatise on Human Nature (2011/1739-40) to be “an Attempt to Introduce the Experimental Method of Reasoning into Moral Subjects”. Following Locke’s impulse to introduce the experimentalist paradigm into different philosophical domains, Hume “promises to draw no conclusions but where he is authorized by experience” and “talks with contempt of hypotheses” (T, 403-405, 407). He even asserts that “such of our countrymen as have banished them [hypotheses] from moral philosophy, have done a more signal service to the world, than my Lord Bacon, whom he considers as the father of experimental physicks” (T, 407). Amongst the other philosophers mentioned in this respect, we also find John Locke’s name. Hume was not only 25

See Wood (1990, pp. 96-97).

62

Kant’s propaedeutic method

loyal to the experimentalist program, but also delivered a specific conception of how experimentation was to proceed in moral philosophy. In a recent paper, Demeter showed that, using the idea of an experimentum crucis, Hume developed a “method of experimental reasoning” (2012, pp. 584-589).26 He argues that, for Hume, the adoption of this method did not just entail an employment of empirical methods, i.e., argumentation based on experiential evidence, but moreover the use of an experimental method at the level of reasoning.27 Methodologically, Hume’s moral philosophy then presents an experimental approach insofar as it aims at the theoretical processing of empirical data. By “us[ing] both history and observation as sources of experimenta crucis”, Demeter explains, Hume aimed to show “the explanatory strength and plausibility of [a philosophical] theory by comparing and contrasting phenomena to assess the truth or falsehood of alternative explanations” (2012, p. 582).28 It is plausible, thus, to describe the latter’s method as 26

An important starting point for this reading is found in Hume’s reflections on the experimental use of natural history in the Enquiry Concerning Human Understanding ([EHU]) (Hume 1975). There, Hume asserts that the chief use of natural history is “to discover the constant and universal principles of human nature, by showing men in all varieties of circumstances and situations, and furnishing us with materials from which we may form our observations and become acquainted with the regular springs of human action and behaviour. These records of wars, intrigues, factions, and revolutions, are so many collections of experiments, by which the politician or moral philosopher fixes the principles of his science, in the same manner as the physician or natural philosopher becomes acquainted with the nature of plants, minerals, and other external objects, by the experiments which he forms concerning them. Nor are the earth, water, and other elements, examined by Aristotle, and Hippocrates, more like to those which at present lie under our observation than the men described by Polybius and Tacitus are to those who now govern the world” (EHU, 8.7). 27 It is important to note that experimental philosophy was not just about a reliance on empirical evidence in the sense of common knowledge, i.e., beliefs about what happens most of the time, or thought experiments about what will probably happen, but also (and more importantly) about the artificial production of natural facts. In contrast to the use of common knowledge, it would rely on the experimental creation of phenomena that would otherwise not occur in nature, thereby generating singular instances of events which happened at a specific time and place. These instances could serve as evidence for the truth of a general claim; see Anstey & Vanzo (2012, p. 90). 28 Hume states in the Enquiry Concerning the Principles of Morals ([EPM]) (in

From empirical experimentation to a priori experimentation

63

an experimental one, which relies on empirical data in assessing the validity of different explanations, and hence grounds its theory-construction in experimentally-produced, “telling” instances. Now, what about Kant’s experimental philosophy? Not only is the Second Edition of his Critique dedicated to Bacon, but he also imported the latter’s programme of an Instauratio Magna into metaphysics (KrV, Bii).29 In contrast to some of the rationalist metaphysics he otherwise clearly cares for, Kant does not “imitate” logical or mathematical methods, for example, but rather the methods of “those who study nature” (KrV, Bxviii). On my view, in arguing so, Kant meant to express his commitment to at least some core elements of the experimentalist program. Like some experimental philosophers before him, Kant held that, in order to change “the fate of metaphysics”, philosophy must take a stand against the unrestricted speculative use of reason within its domain (KrV, Bxiv). In continuity with Locke and Hume’s experimentalist philosophies, Kant also believed that this goal could only be achieved by introducing an experimental method into his critical philosophy. In the course of doing so, Kant both offers his reasons for adopting an experimental method and provides an explicit account of philosophical experimentation in the Critique. As we have seen, Kant used the B Preface to clarify his views about the project, goals, and methods of critical philosophy. One of the central notions he employed in this context was that of an “experiment of pure reason” (KrV, Bxxi). Before we can turn to an analysis and explication of this method of experimenting with pure reason, we must turn to Kant’s Hume 1975) that “we can only expect success, by following the experimental method, and deducing general maxims from a comparison of particular instances” (EPM, 1.10), insisting that the natural history of man ought to be used for the production of particular cases, which can then be tested against different theories, to evaluate their strength. In his Abstract, he specifies this method of reasoning as follows: “if, in examining several phænomena, we find that they resolve themselves into one common principle, and can trace this principle into another, we shall at last arrive at those few principles, on which all the rest depend. And tho’ we can never arrive at the ultimate principles, ‘tis a satisfaction to go as far as our faculties will allow us. (T, Abstract.1) 29 On the similarities and differences between Kant and Bacon’s philosophies, see Kim (2008).

64

Kant’s propaedeutic method

general conception of experimentation—before explaining the nature of his philosophical experiments, he makes some important remarks about the nature of experiments as they are used in the natural sciences. The conception emerging thereof exhibits at least two features that make it continuous with earlier experimentalist conceptions. Their combination, however, also led to the development of a new, distinct theory of what experimentation consists in. Regarding earlier experimentalist philosophies, Kant’s conception draws on some ideals which can be assigned to “Baconian” experimentalism, and some which rather meet the requirements of “Newtonian” experimentalism.30 A good place to look for an exemplification of both ideals is in Kant’s famous description of how the natural sciences came to revolutionise their methods: Reason, in order to be taught by nature, must approach nature with its principles in one hand, […], and, in the other hand, the experiments thought out in accordance with these principles—yet in order to be instructed by nature not as a pupil, who has recited to him whatever the teacher wants to say, but like an appointed judge who compels witnesses to answer the questions he puts to them. (KrV, Bxiii)

According to Kant’s description, experimentation cannot simply consist in the mere observation of nature, that is, of nature in its usual course. Experimentation, on his model, must entail a process of active intervention into natural processes. In this respect he follows the Baconian dictum that natural science must proceed “through the deed” (quoted in Falkenburg 2018, p. 643. Also cf. 641-644). In Kant’s remarks, experimenters are likened to judges who “compel” nature to respond to their specific questions, to produce the evidence that is needed to support their hypotheses and theories. Since the production of experimental evidence requires human activities and interventions, these pieces of evidence are often described as artificial facts of nature. At the same time, Kant makes it very clear that his theory of experimentation is not just the Baconian view. An important difference comes with his rejection of the view that the main goal of experimentation should lie in the construction of natural histories. Instead, Kant holds 30

See Falkenburg (2018), Gloy (1996), and Vanzo (2012).

Kant’s conception of the philosophical experiment

65

that experimentation, correctly understood, must be guided through the correct use of hypotheses.31 He explicitly states that experiments must be “thought out in accordance with […] principles”—principles which scientists induce or deduce from previously established phenomena. It is only in this way, he thinks, that scientists can hope to arrive at nomological (or mathematical) explanations of nature. As we have seen earlier, this model of experimentation is closer to Newtonian experimentalism, and stands in contradiction with the Bacon-Boyle-Hook view, which advocates for a philosophy that mainly collects data in order to construct systematic natural histories. As a consequence, Kant’s conception of experimentalism responds to Bacon’s call for an “interventionist” experimentalism, while also taking up the Newtonian push for empirical but theory-driven experimentation. In that sense, his views reflect points of both Locke’s and Hume’s experimentalism. In virtue of integrating both demands into one conception, however, Kant becomes the first “actual theoretician” of a new experimental program.32

1.4 Kant’s conception of the philosophical experiment How does this general conception of experimentation influence Kant’s conception of the philosophical experiment? It seems reasonable to assume that when Kant describes his philosophical method in the Critique as a procedure which “seek[s] the elements of pure reason in that which admits of being confirmed or refuted through an experiment”, he is drawing an analogy to a similar conception of experimentation (KrV, Bxviii). Taking seriously this analogy would imply that critical philosophers, such as natural scientists, ought to engage in a specific practice of experimental reasoning, which constructs hypotheses, engages in deduction or induction from existing collections of phenomena, and tests those hypotheses through some kind of active interventions into 31 Vanzo’s investigations, on the other hand, show that Kant—in many of his works—argues for an experimentalism that takes the construction and investment of hypotheses to be a legitimate means for arriving at the truth, focusing on the development of a theory about the proper use of hypotheses in science, see Vanzo (2012, p. 72). 32 See Gloy (1996, p. 72).

66

Kant’s propaedeutic method

the mind. Looking at the B Preface, we find that this is exactly the kind of procedure that Kant describes. As noted earlier, Kant aims at an appropriate description of the procedure which establishes whether philosophical cognition as scientific cognition is possible. In the B Preface, he suggests that this procedure might finally be brought to success if it adopts the implicit epistemological assumption underlying the transformation of mathematics and physics into proper sciences: that “we can cognize of things a priori only what we ourselves have put into them” (KrV, Bxviii). Kant proposes to “try and see whether we do not get further with the problem of metaphysics by assuming that the objects must conform to our cognition” (KrV, Bxiii). It is this theoretical assumption which is to be tested in the announced philosophical experiment. Moreover, Kant qualifies the Dialectic as the experiment “that will provide a splendid touchstone of what we assume as the altered method of our way of thinking” (KrV, Bxviii). The Dialectic therefore assumes a dual function: to confirm Kant’s theory of cognition through an application of this same hypothesis to the objects of special metaphysics. All in all, then, this philosophical experiment provides the methodological means to test the hypothesis that Kant inferred from his “history of science”. In section I of this chapter, I showed that Kant articulates the general hypothesis from the B Preface through his Heterogeneity and Cooperation theses. The first thesis, which can also be allocated the status of a hypothesis, assumes that the objects of our cognition must satisfy two different sets of conditions. Each set depends on the contribution of one of two separate sources of cognition, each of which generates representations that are distinct in kind. The Cooperation thesis then adds the claim that we can only cognise objects which satisfy these two sets of conditions. Then, toward the end of the Analytic, these two hypotheses lead to the formulation of another hypothesis, which is derived from the former: the theory of transcendental idealism.33 As I said above, this theory postulates a distinction between “two ways of 33 Thus, I disagree with Fulkerson-Smith’s approach (2013), which takes this second hypothesis to be independent and to be the central hypothesis of Kant’s experiment.

Kant’s conception of the philosophical experiment

67

considering things”, namely things as they appear and things as they are in themselves. On this distinction, objects can be represented in (at least) one of two ways. They can be represented (i) qua the conditions of sensibility, as things as they appear to us, that are structured in the right way to be subsumed under the conditions of the understanding to be cognised. Alternatively, objects can be represented (ii) qua the conditions of the understanding as things in themselves, which results in the representation of objects as they are thought independently of the epistemic conditions that arise from the receptivity of our mode of cognition. In this second case, however, the epistemic conditions used to represent an object do not satisfy the conditions for cognitions of objects. Given his Heterogeneity hypothesis, Kant constructed a theory of cognition that postulated different kinds of a priori representations as its theoretical concepts. From his understanding of these concepts, and their specific ways of relating in acts of cognition, he infers to a theory which explains the possibility of representing things such that they can be cognised. From this same theory it now also follows that we should distinguish between at least two ways of considering things as the objects of our representation. As stated above, we can distinguish between considering things as appearances and as things in themselves, namely, in virtue of positing a difference in representational structure. The Cooperation thesis, if correct, further leads to the conclusion that only one of these ways of considering or representing things will amount to the cognition of objects. If these are the hypotheses Kant constructs, what is he testing them against? Or in other words, what are the objects of Kantian philosophical? Experimentation in physics or chemistry intervenes into the course of nature to compel it to answer its questions; both sciences test their hypotheses against the objects of nature. Things lie differently with philosophy, however. Just as in empirical experimentation, Kant considers the conclusions he draws from the Critique to result from an experimental procedure of reasoning. However, and here lies the important difference from the experimental philosophies introduced earlier, Kant stressed that “the propositions of pure reason, [...] admit of no test by experiment with their object (as in natural science)” (KrV,

68

Kant’s propaedeutic method

Bxviii; [emphasis added]). Instead, “to experiment will be feasible only with concepts and principles that we assume a priori” (KrV, Bxviii). The objects of philosophical experimentation à la Kant are to be identified with a priori concepts and principles, namely those which condition the cognition of natural objects. It is for this reason that philosophical experiments can be described as active intervention into the mind. They constitute active interventions into “theoretical reason”, the human capacity to think, cognise and know things. Against this background, we can now also make sense of Kant calling those experiments “experiments of pure reason” (KrV, Bxxi). Experiments of pure reason offer a specific procedure for the manipulation and artificial preparation of the a priori structures conditioning our minds’ operations. At this point, Kant’s philosophical experimentalism parts ways with earlier experimental philosophies. Kant’s innovation consists in conceiving of a non-empirical experimental philosophy, or in Kantian terms, of a priori experimentation. For philosophers such as Locke and Hume, to apply the experimental method to metaphysics or moral philosophy meant that philosophy—in continuity with the natural sciences—would, in one way or another, proceed empirically. Kant’s philosophical experimentalism, on the contrary, tests hypotheses to “seek the elements of pure reason” (KrV, Bxviii). On his novel conception, an experimental inquiry into the nature of finite cognition must neither appeal to collections of discrete observations, nor to other evidence given through experience. Experiments of pure reason do not operate on human cognition to infer its causes through an experimental-inductive method of reasoning. Instead, they operate on objects that are a priori in origin, and namely those a priori representations which condition the possibility of cognition in general. On Kant’s revolutionary view, philosophical investigations, as they are undertaken in the propaedeutic, are both abstractions from, and reflections on, a priori concepts and principles, and are still experimental in the proper sense. Therefore, the type of experimentalism he pursues should be qualified as a priori experimentation. For Kant, philosophical experimentation in the Critique entails the performance of a priori experiments on reason itself. Note that, even at this stage, we can see how the propaedeutic method provides a novel

Kant’s conception of the philosophical experiment

69

model of philosophical argumentation that is specifically suited to answer the special needs of a metaphilosophy-first programme. Since successful experiments always transform the epistemic status of the propositions that they test, this experimentalist conception of philosophising neither merely presupposes the validity of its initial assumptions, nor simply assumes contingent facts about human cognition. The inherent circularity of philosophical experiments proves a valid methodological tool to literally “transform” metaphysics into a science, without facing charges of partiality or vicious circularity, because the experiment validates what is first only assumed hypothetically. One important question remains. If experiments of pure reason are performed with the goal of testing at least two hypotheses, i.e., the Heterogeneity thesis and the Transcendental Idealism thesis, how can a priori experiments gather proper evidence to confirm or reject their guiding hypotheses? In her paper on Kant’s experiments of pure reason, Falkenburg has shown that, analogously to natural-scientific experiments, philosophical experiments confirm or reject hypotheses not in virtue of whether they are in accordance with natural processes or objects, but by checking whether they are in accordance with reason itself (2018, p. 645). Consequently, experiments of pure reason certainly offer the means to refute hypotheses, namely by showing that their acceptance leads reason into “an unavoidable conflict”, i.e., produce contradictions (Bxix). Her proposal is that Kant’s analogy works in virtue of comparing empirical cognition to philosophical cognition, precisely by pointing to a similarity in how both types of cognition arrive at experimental validation. While what supports the truth of experimental hypotheses in empirical cognition is their agreement with nature, what supports the truth of philosophical hypotheses is their agreement with reason. Hence, philosophical cognition and consistency stand in the same relation as empirical cognition and experimental validation (Falkenburg 2018, p. 656). This, however, does not yet clarify why we should describe Kant’s procedure as an experimental method of reasoning. If philosophical hypotheses are tested through an a priori procedure whose only criterion is that these hypotheses should be consistent or contradiction-free, it remains at best unclear why such a method of reasoning should qualify as

70

Kant’s propaedeutic method

experimental in the discussed sense, and even less clear how it can claim to confirm any hypotheses about the elements of pure reason. A first solution to this interpretive difficulty is to suggest that Kant understood philosophical experiments as instances of thought experimentation.34 While natural scientists forcefully join conditions or create states of nature which would otherwise not occur together, experiments of pure reason forcefully combine concepts and principles that otherwise would not appear together in cognition, were it not for the active intervention of the philosopher.35 By testing whether several theoretical terms can be postulated jointly, without running into contradictions, experiments of pure reason can certainly produce one of their desired outcomes: they can falsify theories, namely those which are inconsistent. And indeed, I believe it makes sense to assume that the part of Kant’s Critique which consists in the articulation of criticisms of other metaphysical theories— which renders it “critical” philosophy—in fact proceeds analogously to the process of falsification. In the B Preface specifically, Kant refers to his argumentation in the Dialectic as one which reveals the inconsistency of holding the view that Transcendental Realism is true (KrV, Bxvii).36 However, for his argumentation to qualify as providing experimental proof, and as licensing the confirmation of hypotheses, more is required. This interpretive tension only intensifies when considering Kant’s affirmations that, indeed, to experiment will be feasible only with concepts and principles that we assume a priori by arranging the latter so that the same objects can be considered from two different sides, […]. If we now find that there is agreement with the principle of reason when things are considered from this twofold standpoint, but that an unavoidable conflict of reason with itself arises with a single standpoint, then the experiment decides for the correctness of this distinction. (KrV, Bxviii) 34 See, for example, Buzzoni (2011; 2019), Fehige (2012), Fehige & Stuart (2014, p. 143), Kalin (1972, pp. 322-323), and Falkenburg (2018, p. 645). 35 Lichtenberg is first to transfer this idea to philosophical method; see Fehige & Stuart (2014, pp. 184-185). 36 I here understand transcendental realism as the view that spatiotemporal objects exist independently of our experience of them. Kant states that “[i]f intuition has to conform to the constitution of the objects, then I do not see how we can know anything of them a priori; but if the object (as an object of the senses) conforms to the constitution of my faculty of intuition, then I can very well represent this possibility to myself.” (KrV, Bxvii)

Chemical analysis and synthesis

71

Certainly, Kant knew that an agreement between his theory of cognition and the principle of reason would only show that the theory thereby tested is logically possible, namely insofar as it is contradiction-free.37 Yet, how mere consistency can possibly decide for the correctness of a distinction, or joint postulation of theoretical terms, remains to be explained. In the next section, I shall argue that, in order to understand why, on Kant’s account, experiments of pure reason can provide valid evidence to reject and confirm hypotheses, we must move away from the debate on thought experiments and rather consider his analogy to chemical analysis and synthesis in the B Preface. This will clarify the methods of discovery and demonstration by which these experiments can establish the truth of their hypotheses.

1.5 Chemical analysis and synthesis At the beginning of section 2, I showed that Kant is comparing his philosophical method not only to empirical experimentation in general, but that he further specifies the content of his methodological analogy to one particular experimental discipline: [the] experiment of pure reason has much in common with what the chemists sometimes call the experiment of reduction, or more generally the synthetic procedure. (KrV, Bxxi [emphasis added])

On my view, Kant not only considers his brand of experimentalism to resemble natural science in general but, very specifically, he takes it to

37 The next time Kant mentions a principle of reason is in the Dialectic, where he understands “the principle of reason” as consisting in the principle that “when the conditioned is given, then so is the whole series of conditions subordinated one to the other, which is itself unconditioned, also given (i.e., contained in the object and its connection)” (KrV, A307-8/BB364). It is this principle which his experiment will identify as postulating a new type of concept (i.e., concepts of reason), which I take to be intended here. The resulting hypothesis consists in this: that it is only possible to postulate the principle of reason as part of our theory if this theory also postulates a specific difference in the nature of a priori representations.

72

Kant’s propaedeutic method

resemble the experimentalism of chemistry in his day.38 He specifically refers to “the experiment of reduction, or more generally the synthetic procedure”. In the passage following the just cited remark, he makes it quite explicit, however, that his procedure not only bears comparison with chemical reduction but, more broadly, the set of chemical procedures which were widely known as chemical analysis and re-synthesis. To arrive a complete description of Kant’s philosophical experimentation, then, we must first take a closer look at his understanding of these chemical practices to then be able to show how these are implemented through experiments of pure reason.39 The notion of chemical analysis and synthesis arose from non- scientific contexts of laboratory practice, such as pharmacy and metallurgy, in combination with the chemical notion of “composition”.40 Chemists brought the procedures of analysis and synthesis into scientific contexts to learn about certain chemical substances by treating substance compounds as composites of more fundamental chemical constituents or principles.41 One famous example of this method, and thereby of its 38 In this and the following paragraphs, I draw on materials from an unpublished manuscript on “The Experiment of Pure Reason: A chemical reading”, Schmid (2022). 39 In what follows, I will not be concerned with the details of Kant’s philosophy of chemistry. For information on this topic, see Carrier (1990), Friedman (1992, chapt. 5), and McNulty (2014a; 2014b; 2015). 40 Klein shows that the concept of the chemical compound was at the centre of the emerging scientific culture of chemistry. Roughly, she defines this conceptual system as having “pure chemical substances” as its basic chemical entities. As it was based on the procedures of analysis and synthesis to bring about and study chemical change, which again assume “the conservation of substance-related physical parts”, this system “implies the existence of empirically determinable lawful relations between chemical substances, relations that explain chemical syntheses and analyses”, in Klein (1994a, p. 173). 41 “[C]hemistry […] is the art of resolving mixed, compound, or aggregate (aggregata) bodies into their principia; and of composing such bodies again from those principiis.” Stahl (1720, p. 1); “The effects produced should rather consist in a separation of the parts, such as they were orginally, than be changed by the operation: for as bodies, by this art, are resolved either into their natural parts, or part acquired in the operation, it is plain that the first alone should be here to be obtained; so that by re-adjusting the same parts, we may be certain of producing the original subject again.” Boerhaave (1753, p. 115); “[S]eparating

Chemical analysis and synthesis

73

implicit conception of composition, is mentioned in the B Preface. As part of his history of the experimental sciences, Kant notes that “Stahl changed metals into calx and then changed the latter back into metal by first removing something and then putting it back again” (KrV, Bxiii). Although this claims to prove what we now know to be a false theory, this experiment still exemplifies the basic idea of the relation between analytic and synthetic operations, and chemical composition. On this conception, chemical experimentation can determine the nature of substances in virtue of the possibility to produce “artificial resolutions and combinations” of substances, which, as reversible operations, determine the simple principles or elements that fundamentally make up matter (Stahl, 1720, p. 12). German chemist Georg Ernst Stahl (1659-1734) became known for developing one variant of the theory of phlogiston to explain the processes of calcination and combustion.42 This theory posits phlogiston as that substance which is released and absorbed by the air when something (i.e., organic or inorganic substances) is burned,43 in a process Stahl called “dephlogistication”. Experimentally, dephlogistication was demonstrated through the oxidation of a metal oxide (e.g., metallic lead), which was processed into lead calx. Stahl explains this transformation as a calcination, arguing that metal lead, when heated, reacts with air, resulting in the “release” of the phlogiston it contains, which in the different substances and rejoining them to make the mist reappear with all its properties” Macquer (1749, p. 1). 42 Stahl’s experiment and theory of phlogiston were inspired by previous work by J.J. Becher, who had already identified a principle for the explanation of the flammability of substances. He had argued that, for example, limestone changes to quicklime—when heated—because of the presence of the principle terra pinguis. Through the process of adding heat, he reasoned, terra pinguis is expelled from the limestone, which is thereby transformed into another distinct substance, i.e., quicklime, with chemical properties different from limestone itself. Stahl renamed terra pinguis to “phlogiston”, and argued that this same element is also involved in processes of combustion. See McNulty (2014b, p. 24). 43 He states that it is not the sulphur itself, “but probably in the sulfur, [that we find] just the same burning basic being, which also in these metals, indeed in all combustible things, constitutes the true actual and specifically combustible main being” (Stahl, 1718, p. 36; my translation).

74

Kant’s propaedeutic method

turn compounds with phlogiston-poor air. This reaction produces a sample of lead calx (and phlogisticated air) (Stahl, 1718, p. 119). Further, he inferred that, if all processes of calcination and combustion occur according to this same principle, then it should also be possible to chemically reverse this reaction. And indeed, a second type of experiment showed that the relevant metal oxide could be regenerated from its calx by heating it up in combination with a different source that is also ‘rich in phlogiston’, e.g., charcoal—due to the reaction between the lead calx and the phlogiston, this would yield a once more “phlogisticated” compound. This second part of the experiment is what he identified as the “reduction”. For Stahl, these experiments proved that it was always the same element, namely phlogiston, which brought about these reactions and chemical transformations.44 He therefore espoused a view that conceives of chemical experimentation as an analytic-synthetic procedure which determines the fundamental entities of chemistry through reversible reactions. Kant followed Stahl’s assessment and, in addition, used it in his account of an experimental philosophical method. Broadly construed, chemical analysis embodies the analytic operations of separation and isolation (in one word: decomposition), while chemical synthesis is the operation of combination. Chemical combination can consist in the reversal of prior decomposition, in which case chemical synthesis takes up the role of confirming the success of previous analyses. What renders chemical analysis and synthesis as elements of a distinctive method is the requirement of removing or adding of a third thing in order to initiate the reaction that either leads to decomposition or (re-)combination (Sticker, 2017, pp. 44-45). To assess the value of this analogy for understanding the nature of metaphilosophical argumentation, we must first clarify how and why experimental procedures of decomposition and combination (analysis and synthesis) allow the chemist to determine the properties of substances. This will then make it clear how philosophical analysis and synthesis—in an analogous 44 As he summarises in the Zymotechnia fundamentalis, “I could still show with various experiments how the basic combustible substance from the fats, coals, etc. itself passes very quickly into the metals, and how it presents the same from the burnt limes again in their fusible state, stretchable under the hammer and convenient for amalgamation” (Stahl, 1734, p. 184; my translation).

Chemical analysis and synthesis

75

manner—are supposed to be establish the properties of elements of pure reason. Some additional remarks on the analogy to experimental chemistry should help to consolidate both my claim that this description of chemical argumentation constitutes a fair description of Kant’s own views, as well as give further insight into the exact nature of his conception of chemical analysis and synthesis. In a first step, on decomposition Kant writes that the chemist is reducing a compound to the simple substances that constitute it insofar as these substances are “separated” stating: “that chemical influence whose effect is to isolate two matters dissolved into one another is decomposition” (MNS, AA 04, 530). Chemical analysis, thus, is understood as a type of decomposition which consists in operations of separation and isolation, allowing the chemist to study the pure substances which make up this compound, as they are independent of their connection to other substances. Kant conceives of chemical synthesis, on the other hand, as an operation through which two substances are made to (re)unite. It is important to note that his view tracks the idea of “chemical affinity”, which was first described by Isaac Newton (1643-1727) and later popularised by Étienne F. Geoffroy (1672-1731).45 This concept explains chemical “laws” as expressing the propensities or “elective attractions” of substances to combine with other substances. In order to investigate the affinity that one substance has to another, a compound chemical substance either had to be brought to react with another substance, or with a third substance. In the first case, the chemist would obtain what is called a “symmetrical reaction”, through which both substances would be made to compound together.46 In the second case, a substance compound would be made to react with a third substance to create a “displacement reaction”: if the third substance had a stronger affinity to one of the substances of the compound, it would decompose the compound and take up the latter’s place, creating a new material compound.47 As Sticker shows in 45 For an excellent overview of how this concept formed and evolved throughout the discourse of 18th century chemistry, see Carrier (1986). 46 The phlogiston example is an instance of symmetrical synthesis. 47 See Klein (1994a, pp. 172-173). For example, “an analyst adds alkali to a solution of calcareous earth in hydrochloric acid” to get the desired reaction of

76

Kant’s propaedeutic method

his paper on the chemical analogies in the Critique of Practical Reason (2002/1788), Kant was aware of this specific understanding of chemical synthesis and also employed it to explain practical judgment.48 Another important feature of this emerging conception, which previous theoretical conceptions lacked, is that it did not conceive chemical compounds as natural bodies, or “mixts” (compounds of different materials found in nature). Rather, it conceptualised them as artificial products of laboratory practice, as idealised artefacts of previous refinement procedures, that as such could only be analysed and synthesised as compounds of heterogenous substances.49 This new chemistry was grounded on the assumption that, in their purified state, substances can be made up from compounds and broken down into pure substances again.50 Kant mentions chemical practices of purification at least twice, once prominently in the first Critique by asserting that one of the problems of earlier metaphysics—as opposed to chemistry—exactly consists in its failure “to present [the pure elements of cognition] in a manner sufficiently purified of everything foreign to [them]” (KrV, A842/B870 [emphasis added]).51 On this view, chemistry deals with idealised entit“the hydrochloric acid at once releas[ing] the lime and uniting with the alkali, [whereby] the lime is precipitated” (KpV, AA 05, 92.27-93.10). 48 Sticker shows that Kant was familiar with the chemical theory of affinities that emerged in the 18th century, particularly with Geoffroy’s theory (1718), which abandoned all speculation about atoms or corpuscles and constructed a theory that was based only on reversible reactions (2017, p. 45). See also Klein (1994b; 1995) and Chalmers (2012). In both the Anthropology and the Critique of Practical Reason, Kant defines chemical laws in virtue of “elective affinities”. He also expresses the view that affinities can only be investigated by means of experiments that employ decompositional and recompositional operations (MNS, AA 04, 534; 23, 284). Thus, the assumption that Kant understood chemical synthesis in this specific sense seems justified. 49 Stahl (1720, p. 76). Against the Paracelsian tradition and its theory of principles, Stahl argues that natural bodies (or “mixts”) are not homogenous bodies whose properties ought to be explained in virtue of a small number of principles, and indeed accused his predecessors of an ignorance concerning the possibilities of “artificial resolutions and combinations” (ibid., p.12). For Stahl, a mixt was “defined as a compound in which the simple principles preserved in the state of bodily compounds or atoms”. See Klein (1994b, p. 161). 50 See Klein (1994a, pp. 170-172). 51 Kant gives his clearest description of this particular analogy with chemical

Chemical analysis and synthesis

77

ies which only can be determined as distinct from each other by virtue of this idealisation. Kant asserts that although it seems “hard to find pure earth, pure water, pure air, etc.”, or even more, that “their complete purity” has its “origin only in reason”, “concepts of them are required” nevertheless, because it is only through determining “the share that each of these natural causes have in appearance” that the chemist can determine the chemical properties of a compound (KrV, A646/B674). Chemistry can demonstrate that pure substances have stable properties. Those properties consist in their affinities to combine with other substances, through chemical combinations and reversible decompositions. Since these processes could be reproduced under controlled circumstances in the laboratory, chemists began to assign pure substances an explanatorily basic role in chemistry, despite their idealised nature.52 Before moving on, I should address one lingering worry. Around the time of the publication of the Critique of Pure Reason, chemistry was not considered a proper science yet. Although chemical change was being studied in various scientific contexts, and had actually been used in technological and practical contexts,53 chemists had not yet been able to postulate axioms and first principles which would ground and unify their subject matter (e.g., laws of chemical change) as Newton’s universal theory of gravitation had done for physics.54 Due to the uncertain experimentation in a passage in the Critique of Practical Reason. While, he still equates the nature of mathematical and chemical idealisation practices in the first Critique, he articulates them as clearly distinct in his second Critique. There, he notes that philosophical experimentation resembles chemical experimentation because it analyses our cognitions of objects “into their elementary concepts, and in default of mathematics, adopt a procedure similar to that of chemistry—the decomposition, by repeated experiments on common human understanding, of the empirical from the rational might be found in them—and come to know both of them [i.e. the elements of practical reason] pure and what each can accomplish of itself” (KpV, AA 04, 163.15-26 [emphasis added]). 52 See Klein (1994a, pp. 173-174). 53 Klein (1994a; 1994b) and Klein & Lefèvre (2007) show how the notion of a chemical compound, and the procedures of analysis and synthesis by virtue of which it is defined, arose from the technological-practical context of metallurgy and pharmacy (instead of from a purely scientific context). 54 See McNulty (2015, p. 89-93). Joseph Black (1728-1799) gives a good statement of the general sentiment: “[C]hemistry is not yet a science. We are

78

Kant’s propaedeutic method

status of their research discipline, chemists of the 17th and early 18th century worked with a combination of theoretical principles, sometimes adopted from other sciences (e.g., corpuscular theory), and a broad range of experimental and technical procedures. In line with this general tendency, Kant described chemistry as “systematic art or experimental doctrine”55 (AA 04, 471). Here’s the worry: why would he consider the Critique’s procedure analogously to chemistry, if chemistry was not yet a science yet, but metaphysics was in need of a method that would transform it into science? Considering that the Critique, in its function as propaedeutic, also needed to take shape in the absence of first axioms or principles, the analogy should appear less odd. Kant does not propose to analogise between chemistry and philosophy as systems of empirical and contingent laws, but between chemical and philosophical methods, as both of those are (to some degree) nascent, experimental doctrines. As ‘sciences in formation’, both philosophy and chemistry require experimental methods to establish concepts and principles, in order to eventually establish solid foundations. Using this sketch of Kant’s understanding of chemical analysis and synthesis, I now turn to explicate my full argument. My claim is that the methods of discovery and proof which are applied in the Critique, through the experiment of pure reason, are best explained as proceeding by virtue of “chemical” operations of decomposition and recombination.

very far from the knowledge of first principles. We should avoid every thing that has the pretensions of a full system” (as quoted in Donovan (1975, p. 133)). 55 Prima facie, this could lead us to believe that Kant viewed chemistry as a purely practical science. However, his comments about the nature and procedures of chemistry suggest that, by the term “art”, he was referring to chemistry’s reliance on “artificial experiences” (AA 29, 102). “Artificial experiences” demarcate a class of experiences that are distinct from “mere perceptions” insofar as they are produced through an enhancement of the senses, or through experimental intervention.

Philosophical procedure in the Aesthetic/Analytic and the Dialectic

79

1.6 Philosophical procedure in the Aesthetic/Analytic and the Dialectic

In what follows, I offer a novel interpretation of Kant’s proposition that the Critique contains an “experiment of pure reason”, which “seek[s] the elements of pure reason” and “has much […] in common with what the chemists sometimes call […] the synthetic procedure” (KrV, Bxxi). As promised above, I shall argue that Kant’s chemical analogy holds between the chemical procedures of analysis and synthesis, on the one hand, and the philosophical procedures as applied through the Aesthetic/Analytic and the Dialectic on the other. I claim that, interpreted in this way, this analogy explains why experiments of reason can reject or confirm metaphilosophical theories by establishing unanimity or consistency between the different elements postulated by these theories. If the combination of the philosophical procedures applied in the Critique are understood to yield a method that is analogous to the chemical one, then metaphilosophical experimentation can discover and prove theories through the decomposition and recombination of pure elements of cognition. First of all, Kant’s philosophical absorption of chemical concepts leads to a new understanding of the proper objects of philosophical experiments. By the pure elements of reason, Kant understood the specific sets of a priori representations which are jointly necessary and sufficient to explain how human thinkers can represent, cognise, and finally know objects. I believe that his analogy provides an illuminating suggestion for how we should conceive of the relation between a priori cognition of objects and the pure elements of cognition. Cognition of objects is to its pure elements as water is to H2 O. And, as with H2 O and water, the elements of our cognition do not offer a conceptual analysis of the concept of a priori cognition of objects. Instead, Kant’s analogy with chemistry suggests the view that the human cognition of objects has “a hidden, inner structure that is explanatory of its more evident, surface properties, and which may bring with it some surprising consequences”

80

Kant’s propaedeutic method

(Currie, 1995, p. 160).56 Just as knowledge about H2 O can help to explain the surface properties of water, e.g., its behaviour under certain circumstances, the cognitive structures which Kant dubs “elements of cognition” can serve to explain the epistemic properties of human cognition, and some consequences arising from this, i.e., that we cannot cognise things in themselves.57 Furthermore, unlike other scientific experiments, chemical experiments work under the hypothesis that chemical compounds—by virtue of their heterogeneous composition—can be analysed and synthesised in virtue of the pure principles or elements that constitute them. On my reading, Kant, like Stahl and others, conceives human cognition of objects as analysable and synthesisable in virtue of the pure elements that constitute it.58 In accordance with the Heterogeneity thesis, he hypothesises that the elements involved in cognition are heterogeneous to each other. And, further, given the Cooperation thesis, Kant assumes that, for an act of representation to qualify as an act of cognition, it must involve a representational structure which is composed of both kinds of elements. Now, the chemical terms of analysis and synthesis refer to the set of experimental operations that determine the “laws” according to which complex chemical compounds can be decomposed into “simple” substances and recombined into mixed substances again. Following Kant’s analogy, this means that philosophical analysis and synthesis must refer to the experimental operations that establish the laws according to which human a priori cognition of objects can be decomposed into its constitutive representational elements (analysis), and the laws according to which they can be combined with other ele56 Perhaps surprisingly, the extent of Kant’s analogy became clear to me when reading Currie’s quite unrelated paper. 57 I do not believe that the experiment of reason shows that some things, e.g., objects of the senses, “are combinable with things in themselves” if they are taken “as not being identical with them (as Transcendental Realism holds), but rather only as being the appearances of them (as Transcendental Idealism holds)”, as suggested by Proops (2021, p. 15). I take it that what is being decomposed and recombined are the elements of cognition, i.e., a priori representations of different kinds, whose properties can explain why one way of considering things qualifies as cognition of objects, while another does not. 58 See Stahl (1720, p. 76) and Klein (1994a, p. 169).

Philosophical procedure in the Aesthetic/Analytic and the Dialectic

81

ments and still yield representations of possible objects of cognition (synthesis). It will turn out that, taking this reading for granted, it makes sense to accept an epistemological, instead of metaphysical, reading of the Dialectic. Finally, chemical analysis and synthesis operate on the assumption that chemical compounds are compositions of idealised entities. Before chemists can discern the fundamental constituents of the substances under investigation, in virtue of their properties, they must implement specific procedures of idealisation. As discussed above, Kant was aware of the chemical practices of purification that had to be applied to mixts or complex substances in order to achieve their decomposition and the isolation of their constitutive elements. So, when Kant compares what “chemists do in analysing materials” to the philosopher’s task, i.e., to “isolate the cognitions that differ from one another” and “carefully avoid mixing them together with others which they are usually connected in their use” to be able “to present […] in a manner sufficiently purified of everything foreign to it”, I take him to be referring to this shared feature of the two procedures (KrV, A842/B870).59 Kant’s method operates within the territory of those sciences which must first prepare and idealise their proper objects before being able to study them and determine their nature. And, as quoted above, he takes these pure elements to have an objective reality which cannot be empirically proven, but must nevertheless be assumed for the sake of scientific explanation (KrV, A646/B674). The analogy to chemical experimentation also provides us with the relevant information which allows us to arrive at a proper understanding of how Kant theorises about the type of philosophical analysis and synthesis at work in the Critique, as we shall now see. As I explained in the previous section, chemical analysis and synthesis were usually understood as two procedures: (i) an analytic procedure decomposes and isolates the components of a “mixt”, and (ii) a synthetic procedure recombines these components into the original compound. On this understanding, chemical experimentation can discern substances by virtue of these “artificial resolutions and combinations”, which, as reversible 59

See also (KrV, Bxxiv).

82

Kant’s propaedeutic method

operations, determine the simple principles or elements that fundamentally make up matter. As demonstrated through the discussion of Stahl’s experiments on phlogiston, one or both of these procedures can include the removal or addition of a third element, e.g., phlogiston, which initiates the process of decomposition or recombination. Further, while the analysis serves to separate and isolate the fundamental constituents of the compound that is being investigated, the synthesis not only serves to recombine these elements into the same compound again, but also to confirm the validity of the hypothesis constructed through analysis. On Kant’s analogy, philosophical analysis moves from the representational whole that is involved in the constitution of the objects of our cognition to its idealised ingredients, the pure elements of cognition. Philosophical synthesis then serves to confirm Kant’s theory of cognition and the theoretical concepts it postulates, which are nothing other than the elements of cognition, understood as heterogeneous kinds of a priori representations. In what follows, I will show that Kant connected his procedure of experimental analysis and synthesis to the discussed procedures of chemical analysis and synthesis. In a footnote in the B Preface, Kant explains his experimental method as follows: [t]he analysis of the metaphysician separated [schied] pure a priori knowledge into two very heterogeneous elements, namely those of the things as appearances and the things in themselves. The dialectic once again combines them, in unison with the necessary rational idea of the unconditioned, and finds that the unison will never come about except through that distinction, which is therefore the true one. (KrV, Bxxi)

If Kant’s experiment works analogously to a chemical reduction, then it must begin with the “taking away” of one element which will later be added again. Following his quote, this should be the representational element involved in the metaphysical cognition of objects: the “necessary rational idea of the unconditioned” [notwendige Vernunftidee des Unbedingten].60 It is this same element of cognition which is added back in the synthetic part of the experiment, to provide “a checkup 60 By “Vernunftideen des Unbedingten” Kant seems to mean “ideas” or “concepts of reason” and not, as Guyer and Wood suggest, “rational ideas” (KrV, Bxxi).

Philosophical procedure in the Aesthetic/Analytic and the Dialectic

83

[Gegenprobe] on the truth of the result of that first assessment of our rational cognition a priori” (KrV, Bxx). Analogously to the chemical procedures of analysis and synthesis, the Aesthetic/Analytic and the Dialectic carry out procedures of decomposition and (re)combination of these representational compounds and their simple elements. It is the specific method, through which the combined procedures of analysis and synthesis construct and test hypotheses, which will explain how experiments of pure reason can do the same, despite their reliance on the principle of non-contradiction or consistency. In its style of a “chemical reduction”, the experiment of reason proves the correctness of its analyses through applying them to further consequences, a procedure which requires the forceful combination of the analysans with a further element. Just as with chemical reactions, what provides the necessary evidence for validating these theories is the fact that the postulated elements do have affinity, which, in the case of cognitive elements, means that they do combine into a consistent theoretical framework. As far as I am aware, Kant’s analogy to experimental chemistry has only been noted by three scholars: Brigitte Falkenburg, Ian Proops and Martin Sticker.61 They all agree that we should interpret Kant’s experiment of pure reason as consisting of two procedures, which are analogous to some form of chemical analysis and synthesis. Since Sticker’s arguments mainly concern Kant’s practical philosophy, I will not be concerned with his argumentation here.62 Against Falkenburg’s63 and 61

Falkenburg (2018), Proops (2021), and Sticker (2017). At the end of the first Critique, Kant hints at the possibility of another experiment in practical philosophy: he says that “[y]et another experiment remains open to us: namely, whether pure reason is also to be found in practical use, […], and thus whether from the point of view of its practical interest reason may not be able to guarantee that which in regard to its speculative interest it entirely refuses to us” (KrV, A804/B832). Kant therefore suggests that there could be an experiment of pure practical reason which in some way does prove the possibility of special metaphysics. I am indebted to Sticker’s reading, insofar as it draws special attention to the concept of elective attraction, and the role that this plays in chemical analysis and synthesis. 63 Falkenburg (2018) argues that Kant’s analogy to chemistry should be understood as comparing “Analytic” and “Dialectic” to chemical analysis and synthesis (p. 656). In contrast to Proops’ reading, Falkenburg characterises the specific philosophical procedures which take the analogous places of chemical analysis 62

84

Kant’s propaedeutic method

Proops’64 interpretation, who both argue that the experiment of pure reason simply works to reject and (partially) confirm metaphysical theories, I argue that Kant is in fact testing two hypotheses: that his theory of cognition is correct, and that his transcendental idealism is correct. I hold that Kant’s experiment serves to confirm the whole theory of cognition, which is thus proved through application to its further consequences, namely those of transcendental idealism, leading him to a conclusive answer to his initial metaphilosophical question. For lack of space, I will not consider all of the details of how Kant carries out either procedure, but I will rather point out some key passages which support my reading of the latter’s methodological selfunderstanding. First, I consider the “analysis of the metaphysician which separated a priori knowledge into two very heterogeneous elements, namely those of the things as appearances and the things in themselves” (KrV, Bxxi). I argue that this part of Kant’s argumentation (i.e., the Aesthetic and the Analytic) should be interpreted in analogy to chemical analysis, and, more precisely, to the procedures of separation and synthesis through reference to the Newtonian analytic-synthetic method. According to the Newtonian model, analysis has as its goal an inference to best explanation, and must precede synthesis. The function of the latter lies in confirming the adequacy of the established explanation by showing its validity for several known consequences. Although this touches on an important part of the explication of Kant’s analogy, her interpretation lacks a clear argument for why Kant would specifically choose the analogy with chemical analysis and synthesis. Under Falkenburg’s argumentation, the method in question could just as well have been another instantiation of this argumentative procedure, for example, Newtonian-style analysis/synthesis as it was employed in experimental physics itself. But I will show that Kant in fact did have his reasons for choosing chemical analysis and synthesis to explain how experiments of reason arrive at valid conclusions. 64 For Proops, the important chemical concept is that of a “fire assay” [Feuerprobe]—a test used by metallurgists to analyse a sample for its precious metalcontent. On his interpretation, Kant uses chemical experimentation as a base system for his analogy because, on this model, we can understand how a philosophical experiment can reject or confirm hypotheses partially. Just as a metallurgical assay can establish the precious metal content of some sample, the Critique can establish the valuable parts of different metaphysical theories without rejecting or confirming them as a whole. Although Proops refers to the operations of analysis and synthesis, he gives no further elaboration of how exactly either of these procedures work (see Proops (2021, pp. 8-12, 15)).

Philosophical procedure in the Aesthetic/Analytic and the Dialectic

85

and isolation as these are applied in chemical oxidations of the type described through the example of Stahl’s dephlogistication. By separating the a priori representation specifically involved in metaphysical cognition, Kant aims to establish the heterogeneous elements which are required for the representation of the objects of our spatiotemporal discursive cognition.65 As a first step, he formulates his initial hypothesis in the form of the Heterogeneity thesis: to say that the objects conform to our cognition is to say that they must conform to two sets of epistemic conditions, namely those arising from our faculty of intuition and those from the constitution of our faculty of thought (KrV, Bxvii). Through the Cooperation thesis, he adds that something can only count as an object of our cognition if it satisfies the epistemic conditions of both faculties, since otherwise it would not count as an object of our specifically human cognition.66 Each cognitive faculty contributes a priori representations that are heterogeneous in kind, the forms of sensibility and understanding, which together ground the possibility of cognition of objects. Other than mathematicians, who—in order to determine their essential properties—can treat their objects as homogenous composites (e.g., spatial magnitudes are “the mere syntheses of the homogenous manifold” (KrV, A720/B748)), propaedeutic philosophers must instead treat their objects of inquiry as composites of heterogeneous elements. In both “Metaphysical Expositions” of the “Transcendental Aesthetic”, Kant aims to establish a first kind of a priori representations, namely those representations which constitute things insofar as they are possible objects of the senses. By analogy to chemical analysis, Kant describes philosophical analysis as proceeding through the operations of separation and isolation. The philosopher “isolate[s] sensibility by separ65 In this and the following paragraphs, I draw on materials from an unpublished manuscript on “The Experiment of Pure Reason: A chemical reading”, Schmid (2022). 66 My reading here agrees with Allison’s, insofar as I interpret Kant’s theses about transcendental idealism to be consequences of his epistemological theses (and not the other way around) (see Allison (2004)). This is the view which Kant’s chemical analogy suggests, since experiments of pure reason are consistently concerned with representations and their components (and not with things).

86

Kant’s propaedeutic method

ating off everything that the understanding thinks through its concepts, so that nothing but empirical cognition remains”, and then “detach[es] from the latter everything that belongs to sensation, so that nothing remains except pure intuition and the mere form of appearances” (KrV, A22/B36).67 By separating everything that is not of sensible origin (concepts of the understanding) or not pure (sensations), Kant describes the pure forms of sensibility as space and time (KrV, A22/B35). In the “Analytic of Concepts”, Kant repeats the same operations in order to determine the pure forms of the second source of human cognition; he undertakes “an analysis of the faculty of understanding itself” (KrV, A65/B90). He claims that “[i]n the transcendental logic we isolate the understanding (as we did above with sensibility in the transcendental aesthetic)” by “separating it […] from all sensibility” (KrV, A65/B89), to “expound the elements of the pure cognition of the understanding and the principles without which no object can be thought at all” (KrV, A62/B87 [emphasis added]). Again, philosophical analysis serves to determine the pure forms of our cognition, this time separating and isolating a second kind of a priori representations, pure concepts and principles of the understanding. Thus, philosophical analysis operates on human cognition of objects under the assumption that what constitutes this species of cognition are two distinct sets of epistemic conditions (i.e., specific kinds of a priori representations) which must be applied to objects in order to cognise them. Through the procedures of separation and isolation, Kant’s analysis establishes those representational elements which “we ourselves have put into them [i.e., into the objects of our cognition]” (KrV, Bviii). Kant’s analysis not only establishes the cognitive structures with which human thinkers can represent objects, but, through keeping them in isolation, his analysis also specifies their specific properties as heterogeneous kinds of a priori representations, namely, in terms of logical quantity and structure, as well as in terms of reference. For example, the decomposition of the elements involved in the constitution of objects 67 It is important to keep in mind that all pure elements are separated from the investigated compound and determined in isolation, i.e., by virtue of the properties they have independently of their relation to the pure elements of another cognitive source.

Philosophical procedure in the Aesthetic/Analytic and the Dialectic

87

of the senses in the Aesthetic reveals that these consist in the a priori representations of space and time. As products of the receptive faculty of cognition, these representations relate immediately to their objects, and so assume the epistemic role of providing our cognition with objective content. Kant further qualifies this kind of a priori representation as intuitions. Intuitions provide singular and immediate representations (KrV, A19/B33; A68/B93), which exhibit the logical structure of whole-part relations, in which each part is constituted as a limitation of an infinite whole that is prior to its parts.68 A similar analysis takes place with the other set of a priori representations, namely the categories, which must be applied to objects of the senses in order to cognise them. Based on his analytic experiment, Kant claims to have arrived at an explication of the a priori elements which compose the representational structure that must be applied to objects to cognise them. Combining his Heterogeneity thesis with the Cooperation requirement, his analysis not only discerns the elements his theory of cognition must postulate, but also their modes of interaction. It is in virtue of their special properties as representations of one kind or another that Kant describes the ways in which they can be combined. One result of Kant’s analysis is that the representational elements belonging to the cognitive structure can only be applied to representational elements that have the properties arising from the structure of sensibility. In other words, universal and mediate concepts can apply to objects as represented by virtue of singular and immediate intuitions. Cognition requires that its objects fall under both sets of epistemic conditions, meaning that the inner structures which partially make up the objects of our cognition are composed of a combination of sensible and conceptual elements. In the Third Chapter of the “Analytic of Principles”, Kant further argued that the result of this analytic procedure allows an inference to the following hypothesis: that we can only have cognition of “things as appearances”, but not of “things in themselves”. On this distinction, objects can be represented in (at least) one of two ways. They can be represented (i) qua the conditions of sensibility as things as they appear to us, which are structured in the right way to be subsumed under the 68

See McLear (2015).

88

Kant’s propaedeutic method

conditions of the understanding to be cognised. Alternatively, objects can be represented (ii) qua the conditions of the understanding as things in themselves, which results in the representation of objects as they are thought, independently of the epistemic conditions that arise from the receptivity of our mode of cognition. In this second case, however, the epistemic conditions used to represent an object do not satisfy the conditions for cognitions of objects. From his theory of cognition, it follows that only things as appearances, which are represented using a priori representations of space and time, can become objects of our cognition. If the epistemic conditions used in representing a thing do not exhibit the same composition of elements as the one established through the analysis, then—by hypothesis—this act of representation will not qualify as finite cognition of objects. Kant inferred that for the case of the objects of special metaphysics, this would mean that unconditioned things such as the soul, the worldwhole, and God can never become objects of our cognition. This is so because, as human thinkers, we cannot cognise things that we cannot represent in space and time (KrV, A239). Although we can represent them according to one set of epistemic conditions, whatever the resulting representation consists in will not yield cognition of the objects of special metaphysics (KrV, Bxx, A254). Kant believed that this second part of the experiment can verify the hypotheses of the first (analytic) part if, contrary to competing theories, it can also show how metaphysical cognition fits into the picture. If successful, his theory articulates the conditions of the possibility of object cognition, including those of metaphysical cognition, and does so by accepting the same hypotheses that were constructed in the analytic part. On this experimental procedure, the legitimacy of the hypotheses is established by showing their consistent application to further consequences. Following the chemical analogy, it makes sense to assume that the synthetic part of Kant’s metaphilosophical experiment will consist in some operation of (re)combination. Indeed, Kant’s wording invites such a reading. He refers to the Dialectic as that part of the Critique which “once again combines them [i.e., the elements that constitute appearances and things in themselves], in unison with the necessary rational idea of the unconditioned” (KrV, Bxxi). On this chemical analogy,

Philosophical procedure in the Aesthetic/Analytic and the Dialectic

89

we should draw the following conception: the Dialectic “adds back” the ideas of the unconditioned, and so adds another kind of a priori representations which had been separated off at the beginning of the “chemical reduction”. Kant adds these a priori representations, once as conceived through former metaphysical theories, and once as conceived through his theory of cognition and its postulate of transcendental idealism, which is proven to be the correct one. In this sense, the conclusions reached by the experiment of reason do not have the status of merely possible explanations, but of best explanations. For example, in the Antinomies, we get explanations which satisfy this standard through the successful syntheses of antinomy-resolution and through the failing syntheses presented through antinomy-construction. In what follows, I will use the example of the Antinomies to explain Kant’s experimental procedure of synthesis. In the B Preface footnote, Kant makes the claim that the Dialectic proves the hypothesis that transcendental idealism is true. Now if we find that on the assumption that our cognition from experience conforms to the objects as things in themselves, the unconditioned cannot be thought at all without contradiction, but that on the contrary, if we assume that our representation of things as they are given to us does not conform to these things as they are in themselves but rather that these objects as appearances conform to our way of representing, then the contradiction disappears; and consequently that the unconditioned must not be present in things insofar as we are acquainted with them (insofar as they are given to us), but rather in things insofar as we are not acquainted with them, as things in themselves: then this would show that what we initially assumed only as an experiment is well grounded. (KrV, Bxx-xi)

The Antinomies provide such a proof through two steps (which is why this proof might be labelled indirect) (KrV, A506/B534). First, antinomy construction shows that another hypothesis, namely the hypothesis that “transcendental realism is true” is in fact false. As mentioned above, Kant describes this part of the Dialectic as follows: Kant falsified the opposing hypothesis, identified as “transcendental realism” (KrV, A369; A491/B519), through antinomy-construction, which roughly consists in bringing together, or constructing a pair of, sound arguments in order to make evident their apparently contradictory conclusions. Kant took his experiment to show that a theoretical position

90

Kant’s propaedeutic method

committed to transcendental realism cannot satisfy the standard of unanimity, because the forceful combination of its theoretical postulate must lead to “an unavoidable conflict of reason with itself arises” (KrV, Bxix). What Kant described as experimental synthesis in the B Preface, however, does not primarily refer to his procedure of antinomy construction, but rather his subsequent procedure of antinomy resolution (KrV, A491/B519; A506ff./B534ff.). His resolution of the antinomies of reason provides a proof of his whole theory of cognition (including the thesis of transcendental idealism). And from this it follows that the original metaphilosophical question, which initially motivated Kant’s performance of an experiment of reason, must receive a partially positive and a partially negative answer. According to the chemical analogy, this synthetic part of the experiment must be conceived as proceeding through the forceful recombination of the results of prior analyses. However, Kant argues, this recombination will not only include the elements of cognition postulated in the Aesthetic and Analytic, but also add another “third” element: the Vernunftideen of the unconditioned. On his view, prior to the experimental analysis of the composite of a priori representations that are required for human cognition of objects, he had separated from it that element of cognition which is specifically involved in metaphysical speculation and its representation of objects. Like Stahl’s phlogiston experiment—“by first removing something and then putting it back again”—Kant’s experiment of reason in the Aesthetic and Analytic decomposes the pure elements of cognition by “removing” the idea of the unconditioned in order to study the properties of the other constitutive elements of cognition in isolation. By “putting it back again”, the Dialectic investigates whether the ideas of the unconditioned can compound with the other representational elements to grant their unanimity within one theory of cognition, and what properties ideas of reason as a priori representations must have (KrV, Bxiii). In style of a “symmetrical reaction”, Kant’s experiment confirms that we can only be said to cognise things which we have applied certain a priori representations to. As predicted by the Heterogeneity thesis and the Cooperation thesis, the experiment shows that we do arrive at a consistent theory of human cognition if we accept the heterogeneous

Philosophical procedure in the Aesthetic/Analytic and the Dialectic

91

nature of the a priori representations involved, and further that we only get to represent objects in the required way if the structures involved are a compound of sensible and conceptual elements. Under a chemical description, the synthetic part of Kant’s experiment shows that what we can cognise about objects (which is equivalent to “what we have put into them”) is grounded in heterogeneous cognitive elements which bear certain affinities to combine. The synthesis shows that conceptual elements can only be applied to representational compounds which already contain sensible elements. All Kant’s experiment can show at this point is that it is possible to arrive at a consistent theory of cognition if his two hypotheses are accepted. Through recombination, the experiment can demonstrate that the postulated elements of cognition do have the propensity to combine to represent the objects of our cognition. Moreover, in the style of a “displacement reaction”, the experiment also shows that the newly introduced representational element cannot assume the place of the categories of the understanding, since it cannot apply to objects that are represented according to space and time. It turns out that this representational element can only combine with representations of the second type, the categories. Ultimately, Vernunftideen are concepts of concepts, and only show affinity to other conceptual elements, and so they can only help represent things as they are thought, but not as they are cognised (KrV, Bxii). This is how Kant conceives of his experimental validation of the hypotheses which guided the analytic experiment. By adding them to a further consequence, his experiment shows that the Heterogeneity thesis and the Cooperation requirement can explain both why human cognisers engage in metaphysical speculation and why these acts of representation do not amount to cognition of objects: the Vernunftideen—in virtue of their representational nature— bear no affinity to the a priori representations which must be in every cognition of objects, namely space and time.69 As a consequence (Kant claims), transcendental idealism is shown to be correct (KrV, A369). Unlike Falkenburg’s and Proops’ interpretations, my interpretation thus 69 Kant further investigates the properties of this special type of conceptual representation at the beginning of the Dialectic (KrV, A310-38/B367-96) and returns to make further analyses in the Appendix (KrV, A642-68/B670-96).

92

Kant’s propaedeutic method

suggests that we should understand Kant’s experimental conclusions as meta-epistemological, and not as metaphysical. Let me explain. When read through the lens of Kant’s analogy to experimental chemistry, the experimental conclusions of the Dialectic now appear as statements about the nature of our cognition, and not as statements about what there is.70 On my reconstruction, Kant derived the transcendental idealism thesis as a theoretical consequence of the Heterogeneity thesis and the Cooperation requirement. This reading is in line with what has been called the “two-aspect view”, most prominently argued for by Allison, Bird, and Prauss.71 On this “epistemologically based understanding of transcendental idealism”, “the transcendental distinction between appearances and things in themselves [is] understood as holding between two ways of considering things (as they appear and as they are in themselves) rather than as, on the more traditional reading, between two ontologically distinct sets of entities (appearances and things in themselves)” (Allison, 2004, p. 32). Against “metaphysical interpretations”, this view holds that the distinction between appearances and things in themselves gets its meaning in virtue of the contents of the Heterogeneity thesis (and the Cooperation requirement). According to the Heterogeneity thesis, the representation of objects from a human standpoint is conditioned by two classes of cognitive elements: the a priori forms of sensibility and the a priori forms of understanding. If we consider objects as objects of cognition, we must consider them as falling under their relevant epistemic conditions, otherwise we “would not be cognizing but misrepresenting them” (Stang, 2022, sect. 4.2). So, to consider objects as objects of our cognition is to consider them as objects which fall under the specific epistemic conditions of our spatiotemporal discursive cognition. If we consider them as things in themselves, on the other hand, we do not represent these objects according to the epistemic conditions which afford cognition of objects, but only according to the epistemic conditions of discursive cognition. In this case, we represent 70 I thereby by no means intend to claim that this shows the invalidity of metaphysical readings of Kant. I only claim that if we read the Critique through Kant’s remarks on his conception of experiments of pure reason, then we must arrive at this consequence. 71 Allison (2004), Bird (2017), and Prauss (1974).

Philosophical procedure in the Aesthetic/Analytic and the Dialectic

93

objects as satisfying the conditions of the understanding but not the conditions of sensibility. And this way of considering objects will not count as representation of objects of cognition, but only of “objects [as they] are merely thought” (KrV, Bxviii; cf. Bxxvi). As my argument shows, Kant’s description of his method in the Dialectic reveals that its proofs of transcendental idealism must be taken in this sense. Falkenburg and Proops, on the contrary, both presuppose that transcendental idealism expresses a metaphysical thesis which distinguishes between two classes of entities: appearances and things in themselves. Objects of the first class, i.e., things in themselves, do exist independently of there being any human cognisers around to perceive them, and so do their properties. Objects of the second class, i.e., appearances, only exist in relation to human cognisers—their properties depend on human cognisers—which is why they have often been characterised as mental entities or mental representations. Readings of this type have later been summarised from the perspective of the “two-objects view”.72 I propose that if we only pay attention to Kant’s metaphilosophical remarks we find reasons to doubt this interpretation of the latter’s distinction between appearances and things in themselves. From Kant’s conception of how experiments of pure reason proceed, it follows that their conclusions concern the nature of our cognition and the representations involved, and not metaphysical statements. Kant’s experiment of pure reason “decide[s] for the correctness of this distinction”, since its synthetic procedure confirms that his theory of cognition can work as an explanans for more than one domain of objects. He confirms its correctness through what we might call an “inference to the best explanation”.73 The philosophical principles 72 For an updated presentation and discussion of these and other readings of Kant’s transcendental idealism, see Stang (2022). 73 “Inference to the Best Explanation can be seen as an extension of the idea of ‘self-evidencing’ explanations, where the phenomenon that is explained in turn provides an essential part of the reason for believing the explanation is correct. […] Self-evidencing explanations exhibit a curious circularity, but this circularity is benign. […] According to Inference to the Best Explanation, this is a common situation in science: hypotheses are supported by the very observations they are supposed to explain. Moreover, on this model, the observations support the hypothesis precisely because it would explain them. Inference to the Best

94

Kant’s propaedeutic method

which he hypothesises have “the special property that [they] first make […] possible [their] ground of proof, namely experience, and must always be presupposed in this” (KrV, A737/B765). Kant’s experiment moves from one kind of cognition, e.g., empirical cognition, to the hypothetical conditions which, if correct would best explain the epistemic qualities of such cognition, i.e., the elements of pure reason. He further confirms this theory by showing its adequacy for another kind of cognition, namely metaphysical cognition. His synthesis reveals that metaphysical cognition, too, is best explained if these same hypotheses are accepted. Finally, Kant’s hypothesis also explains why other theories which have not adopted a transcendental idealist approach will run into contradictions. These theories postulate a false theory of metaphysical speculation because they do not presuppose Kant’s original Copernican hypothesis for thinking about the nature of these concepts in the first place. Kant amends this in the Dialectic and its Appendix by discussing the specific nature of the representational element responsible for thinking the unconditioned—the Vernunftideen, or concepts of reason—as regulative or problematic concepts.

1.7 Metaphilosophy as experimentalist practice Let me conclude with what I take to be Kant’s conception of propaedeutic method. Instead of just presenting a new theoretical framework to take the place of earlier metaphysical systems and invest it with the label of “scientific metaphysics”, Kant’s Critique of Pure Reason provides a special procedure which is to investigate the conditions under which metaphysics as science is possible. We saw that this methodological solution is coupled with his development of the philosophical discipline of propaedeutic. As a propaedeutic to the system of metaphysics, the Critique is assigned the metaphilosophical task of investigating the conditions of metaphysical science by virtue of reconceiving the relation between cognition and its objects. By focusing on some remarkable but often ignored passages in the B Preface, this chapter then Explanation thus partially inverts an otherwise natural view of the relationship between inference and explanation.” Lipton (2000, p. 185).

Metaphilosophy as experimentalist practice

95

presented an argument for why we can only understand the nature of Kant’s methodological solution if we analyse his metaphilosophical method as grounded in a two-layered analogy to the experimental sciences of his day. The propaedeutic method draws on Baconian and Newtonian conceptions of experimentation, as well as specific reflections about experimental chemistry and its use of idealisation. His experiments of pure reason are specifically designed to resolve the problem of a scientification of metaphysics, and the problem of providing a valid method for the philosophical investigation of human cognition. Through idealisation, they make possible an investigation of the fundamental structures that condition human cognition and explain its manifest epistemic qualities. Through the design of an experimental method of hypothesis-testing, Kant’s metaphilosophical method can establish justified theoretical claims without producing partial or circular results. The analogy to chemistry, moreover, illuminates how metaphilosophical hypothesis-testing works without producing trivial results. The propaedeutic method is described in a way that renders it a legitimate exercise of self-investigation, insofar as Kant proves the validity and non-triviality of philosophical propositions through the experimental decomposition and recombination of an idealised modelsystem of cognition. On my view, philosophical experimentation is concerned with establishing the composition of the representational structure that is required for the cognition of objects. Under the premise that objects conform to our forms of cognition, philosophical experimentation sets out to investigate the heterogeneous representational compounds that condition their representation and cognition. In analogy to the chemical procedures of analysis and synthesis, the Aesthetic/Analytic and the Dialectic carry out procedures of decomposition and (re)combination of these representational compounds and their simple elements. It is this specific method, through which the combined procedures of analysis and synthesis construct and test hypotheses, which will explain how experiments of pure reason can do the same despite their reliance on the principle of non-contradiction or consistency. In the style of a “chemical reduction”, the experiment of reason proves the correctness of its analyses through applying them to further consequences, a

96

Kant’s propaedeutic method

cedure which requires the forceful combination of the analysans with a further element. Just as with chemical reactions, what provides the necessary evidence for validating the theories is the fact that the postulated elements do have an affinity, and in the case of cognitive elements this means that they do combine into a consistent theoretical framework. In achieving this, Kant devises a new type of philosophical experimentalism specifically suited to solve the metaphilosophical problems introduced in the introduction. There, I claimed that to understand the nature of Kant’s philosophical method in the Critique we need to consider that it is tailored to the purposes and tasks of metaphilosophyfirst. The ultimate goal of the Critique is to establish what scientific philosophical cognition ought to consist in, and thereby whether, and in virtue of which conditions, metaphysics can be transformed into a science. Through the propaedeutic method, Kant proposes that this metaphilosophy-first can only be realised through a specific methodological programme suited to proceeding in the absence of certain propositions, and to testing all assumptions it relies on. His propaedeutic method provides a novel model of philosophical argumentation that is specifically suited to answering the special needs of a metaphilosophyfirst program by proposing a metaphilosophical method that is, in key respects, continuous with those of earlier experimental philosophies. Since successful experiments always transform the epistemic status of the propositions that they test, this experimentalist conception of philosophising can proceed without merely presupposing the validity of its initial assumptions, and without simply assuming contingent facts about human cognition. The inherent circularity of philosophical experiments therefore proves a valid methodological tool for literally “transforming” metaphysics into a science, without facing charges of partiality or vicious circularity, because the experiment validates what is first only assumed hypothetically. Furthermore, it departs from the traditional paradigm of either establishing logical, metaphysical, or psychological propositions as fundamental principles grounding philosophical science. Since he identifies these principles with cognitive elements and their modes of interaction, Kant also holds that metaphilosophical experimentation must engage with the nature of a priori representations as conditions for object

Metaphilosophy as experimentalist practice

97

nition, treating them as idealised structures which are not immediately accessible through experience but which must be artificially prepared, modelled, and manipulated. Finally, to engage in metaphilosophyfirst means to engage in abductive reasoning—reasoning which aims at best explanations—which is at the same time experimental insofar as it proceeds through reversible decompositions—analogous to forceful combinations—of those mental entities which condition the possibility of scientific representation.

2 Maimon’s method of fictions

In a letter accompanying the copy of his Essay on Transcendental Philosophy (1790), Maimon addresses Kant as “a man who has reformed philosophy and thereby all other sciences as well” (Letter [365], April 7, 1789; AA 11, 14-17). For Maimon, Kant had asked the right question to improve the situation of metaphysics, namely whether metaphysics is possible as a science. From the beginning, Maimon supported the latter’s metaphilosophical programme: to provide for metaphysics the secure course of a science, it must become metaphilosophy-first. For Kant, this possibility depended on a methodological issue. Entrusted with the tasks of metaphilosophy-first, propaedeutic philosophy needed to yield a procedure for establishing the conditions of possibility of philosophical cognition, and thereby the conditions of scientific cognition as such. Metaphysics could therefore only be transformed into science through the performance of a so-called experiment of reason. Maimon, whom Kant called his “best critic,” (Letter [362], May 26, 1789; AA 11, 48-49) expands the latter’s research program, devoting his work to the same question. Moreover, he agrees with Kant that ultimately, philosophy1 can only become science by virtue of a methodological solution. On Maimon’s view, however, this propaedeutic philosophy must embody a “science of the limits of appearances (Ideas)” (Maimon, 2010, p. 239). He argues that, owing to its special object of 1 Maimon will often refer to the target of his metaphilosophy simply as “philosophy”, except in a few passages of an article in the Berlinisches Journal für Aufklärung, in which he directly compares his research programme to Kant’s, and identifies its target as lying in the explanation of the science of “metaphysics” (GW II, IX, 52, 73). For the remainder of this chapter, I will use these terms interchangeably.

Maimon’s method of fictions

99

study, the possibility of metaphysics cannot, pace Kant, be investigated through a propaedeutic experiment, but must turn instead to imitating another method. This method provides another testable solution to the research programme of metaphilosophy-first, and is derived from the realm of the mathematised natural sciences: “the method of fictions” (Strf, GW IV, 39). While Maimon accepts Kant’s general hypothesis, according to which philosophy can only become science if it—via a theory about the relation between mind and world—investigates the conditions of possibility of philosophical cognition, he rejects the latter’s articulation of this hypothesis through the Heterogeneity thesis. Maimon criticises Kant’s theory of cognition for accepting, rather than proving, the factum of science, that is, for presupposing that scientific cognition exists and that one class of scientific judgments can be characterised as causal judgments. Furthermore, Maimon also contends that Kant cannot show why we are justified in combining one set of cognitive elements with another, that is, in virtue of what an application of the forms of the understanding (i.e., of the categories) to objects falling under the conditions of sensibility (i.e., given sensible objects) is justified. Thus, Maimon aims at showing that, pace Kant, propaedeutic philosophy fails on its own terms because it has neither the factual evidence, nor the justificatory power, to prove its claims (i.e., the Heterogeneity thesis, the Cooperation thesis, and the Transcendental Idealism thesis), and that its experiments must produce contradictions. On his alternative account, only a framework of infinite intelligibility can guarantee the possibility of genuine explanation. Since this kind of explanation demands overstepping the boundaries of experience and spatiotemporal cognition, the “new metaphysics” must be conceived as “the science of the limits of appearances”. Maimon argues that his philosophy must also determine the cognitive conditions under which metaphysical cognition can fulfil the criteria of scientific cognition. Unlike Kant, however, his notions of scientific cognition are bound to his deep commitments to explanatory rationalism and the infinite intelligibly of objects of cognition. From this it follows that his metaphilosophy will succeed at its task if it presents a model of human cognition of objects that are intelligible without limits, and

100

Maimon’s method of fictions

thus completely explicable, at least in principle. Since the elements which must be postulated as part of the desired theory of cognition are “impossible as objects of our [finite] cognition” (Maimon, 2010, p. 249), his metaphilosophy requires a procedure that can deal with objects of this kind: the method of fictions. As a consequence of his methodological solution, Maimon’s metaphilosophy ultimately cannot end up with verified theories of cognition. Instead, it can only hope to construct consistent but hypothetical models of human cognition, the assumption of which would entitle science to its statements. By turning metaphilosophy into a method of fictions, it can establish the conditions under which metaphysics would be a science without arbitrariness or circularity. In this chapter, I offer a new interpretation of Maimon’s ‘philosophical science’2 as a metaphilosophy-first project. I show that Maimon— continuing Kant’s project—develops a methodological solution to the metaphilosophical issue of transforming metaphysics into a proper science. This solution is informed by his philosophy of science, and more precisely by his thoughts on differential calculus and the use of fictions in science.3 With regard to the vast amount of work that Maimon devoted to discussing metaphilosophical issues, as well as their relation to 2 When talking about Maimon’s metaphilosophy and the philosophical science he envisions, I will just refer to his theoretical philosophy as “philosophy” (as opposed to “new metaphysics” or similar). I thereby follow Maimon’s own practice: although he calls for a re-definition of the term “metaphysics”, in later writings he resorts to using the term ‘philosophy’ whenever he discusses its scientific status and method. Although he does not give his reasons for this choice, we can assume that this was a measure of precaution against being taken to advocate for “metaphysics as science of things in themselves” (as Kant had defined it). 3 Maimon expresses this view perhaps most clearly in a letter to Fichte, in which he reacts to the latter’s proposal for a deductive-axiomatic science: “I am expecting with joy the time when, as you say, ‘philosophy should be a systematic science’. On my part, too, I will not fail to contribute towards this goal as much as is in my weak forces. We will meet on the very same way, even though it seems that we will travel it in opposite directions. You wish to travel it from top to bottom (from the concept of a science as such to the concrete sciences), but I want to travel it from bottom to top” (Maimon to Fichte, Berlin, October 16, 1794; GW VI, 449-450), translated in Freudenthal (2019, pp. 65-66).

Maimon’s method of fictions

101

the history, concepts, and methods of science4 , it comes as quite of a surprise that much of his efforts have so far gone largely unnoticed.5 In response to this lacuna in scholarship, the following chapter develops an account of Maimon’s method of fictions as an answer to the metaphilosophical challenge posed by Kant. This method functions as a consequence of his thoughts on mathematical modelling and its exemplary function in a rationalist foundation of science in general. In order to address the challenges of this metaphilosophical research programme, Maimon looks for solutions in the scientific methods of calculus and comes to argue that metaphilosophy must appropriate the idealising and modelling procedures of the mathematical sciences in order to determine the principles of a priori cognition of objects. As systematic fictions, he contends, philosophical models do not claim objective validity, butdo provide methods for how to treat objects, namely as though they were intelligible without limits. I begin this chapter by showing how Maimon takes up the metaphilosophical challenge to determine whether and how metaphysics can become a science. This will include an initial exposition of his alternative conception of metaphysics as the science of the “limits of appearances”. In the second section, I then turn to Maimon’s famous criticism of Kant’s dualist system. This criticism both doubts the fact (also known as the argument quid facti?) and the legitimacy (argument quid juris?) of an application of elements of the understanding to elements of sensibility, thus attacking Kant’s explanatory strategy of positing the Het4 See for example Ankündigung und Aufforderung zu einer allgemeinen Revision der Wissenschaften (1792), Streifereien im Gebiete der Philosophie (1793), Vorrede zu Bacons von Verulam Neues Organon (1793), Anfangsgründe der Newtonischen Philosophie, Vorrede (1793), Ueber den Gebrauch der Philosophie zur Erweiterung der Erkenntniß (1795), and Pragmatische Geschichte des Begriffs von Philosophie, und Beurtheilung der neuern Methode zu philosophiren (1797). 5 Some scholars devote short chapters and remarks to the method of fictions (e.g. Cassirer (1974, pp. 97-125); Atlas (1967, pp. 34-37); Freudenthal (2010, pp. 97-103)), or touch upon the subject in the context of Maimon’s philosophy of mathematics (e.g., Bergman (1967, pp. 140-155); or Pringé (2018, p. 40)). A notable exception is Breazeale (2003; 2018), although he does not approach the topic from within Maimon’s systematic philosophy, but rather focuses on the method of fictions in relation to other historical figures.

102

Maimon’s method of fictions

erogeneity thesis, showing that it therefore becomes impossible to argue for the Cooperation requirement. Section three then investigates the basic epistemological presuppositions of Maimon’s alternative account, which is presented as a combination of rationalism and scepticism. I show that, for Maimon, the latter position is a result of the former: it is because of his rationalist commitments—to a standard of explanatory rationalism and infinite intelligibility—that he must also adopt the position of a sceptic with regard to the realisation of this standard in the sciences. In section four I explain how Maimon conceived of his alternative account of cognition, which arises from his engagement with the concepts and theories of calculus. This part of the latter’s overall programme is illuminated through an in-depth discussion about the philosophy of mathematics, which includes his criticism of Kant’s philosophy of mathematics and Maimon’s vindication of analytic geometry and the method of calculus. Using these reflections, it then will be possible to give an adequate interpretation of Maimon’s account of human cognition of objects, which rests on his “Homogeneity thesis” (i.e., that the form and content of cognition are homogeneous) as well as his conception of the relation between divine and human cognition. Section five introduces Maimon’s engagement with the method of fictions, which I take to result from his analysis into the scientific tools which led to the successful application of calculus within a number of scientific disciplines. On his analysis, the theories and practice of calculus are grounded in the use of a specific type of scientific representation, i.e., scientific fictions, which he categorises into several classes in order to describe their individual properties and shared functions. After this discussion of Maimon’s description of scientific fictions, section six will offer a novel interpretation of the latter’s conception of philosophical fictions and their role within scientific inquiry in general, as well as their role within philosophy itself. It will be argued that philosophical fictions should be conceived of as a consequence of Maimon’s rationalist commitments, whose standards of explanatory rationalism and infinite intelligibility can only ever be articulated through the employment of fictions of systematicity. Consequently, in the final part, I will argue that Maimon’s

Philosophy is the science of the form of all sciences

103

ulation of the metaphilosophy-first program should be conceived as advancing a particular conception of metaphilosophy as model-based science.

2.1 Philosophy is the science of the form of all sciences To consider how Maimon provides a new foundation for metaphysics, let us briefly consider how he entered the debate on Kant’s critical philosophy toward the end of the 18th century. Salomon Maimon was born as Shlomo ben Joshua in a small Jewish village in Lithuania, and although, in his Autobiography and letters, he laments his years in Lithuania because he was “deprived of knowledge” (AA 09, 15-17, Letter No. 352 [330]), it is worth noting that Maimon, whose father was a rabbinic scholar, led quite a “bookish life”, studying the Torah, Talmud, Kabbalah, the works of Maimonides, whose name he later adopted as his own (in 1784), as well as mathematics and natural sciences (Maimon, 2019, pp. 1-192). Around 1778, Maimon left his wife and children to pursue his study of the sciences in Berlin. Due to his financial and professional situation, as well as his political status as a Jew and his free appreciation for Spinozism, it took him several attempts to finally take up permanent residence in Berlin. When he finally did so in 1785, he began studying Kant’s Critique of Pure Reason (Maimon, 2010, pp. 230ff.). This reading prompted him to draft an unusual script, half commentary, half philosophical system, in which he tries to combine Kant’s philosophy with his readings of Hume, Leibniz, and Spinoza. It was through their mutual friend Marcus Herz that this manuscript, titled Essay on Transcendental Philosophy, finally reached Kant. Upon reading it, Kant would note that “none of his opponents has understood him better than Mr Maimon” (Maimon, 2010, p. xvii), and later, more deferential still, Gottlob Fichte would remark that “his respect for Maimon’s talent knows no bounds” since “he has turned upside down the Kantian philosophy as it has been generally understood” (Maimon, 2019, p. xiii). Despite the philosophical interest he inspired in some of the most important philosophers of the period, Maimon received little to no support in getting his works published, or in being hired, and was subject to much discrimination and anti-semitic sentiment. The importance of

104

Maimon’s method of fictions

his writings was only rediscovered in its own right in the early 20th century,6 and it was not until quite recently that more scholarship would engage with Maimon’s thought, and not just with the role it played in the development of German idealism.7 Maimon first entered the domain of philosophical debate around Kant’s philosophy with a text that is half commentary, half independent study. His Essay on Transcendental Philosophy (2010/1790) comprises both a tripartite overview and discussion of the three main parts of Kant’s Doctrine of Elements in the Critique of Pure Reason and a collection of several independent works on metaphysics, ontology, and philosophy of language.8 Hotly debated amongst his contemporaries,9 the Essay only marks the beginning of Maimon’s continuous engagement with metaphilosophy, and in particular with the question of how metaphysics (i.e., theoretical philosophy) is possible (Tr, GW II, 3, 335). In 1790, shortly after finishing the Essay, Maimon also published a short commentary (“Kant und Baco”) about the parallels and differences between Kant’s and Bacon’s scientific programmes. Regarding their shared project of a “reformation of philosophy” (KB, GW II, 503), Maimon notes that both Bacon and Kant realise that philosophy consists neither in logic (which concerns objects of thought but not of experience), nor in an uncritical metaphysics (which presupposes objective reference without having demonstrated the possibility thereof) (KB, GW II, 504). Instead, they both suggest a procedure for philosophy, in virtue of which it can determine the relationship between the forms of our cognition and the forms of real objects. Hence, in a different vocab6

See, e.g., Kuntze (1912). Important groundwork was laid by authors such as Atlas (1964) and Bergman (1967). 8 Ehrensberger (2004, XIV-XVII). 9 It has to be noted at this point that Maimon’s philosophical work received nowhere near the consideration it deserved, and this was no accident. Throughout his short philosophical career, the Jewish Maimon faced a great deal of anti-semitism, not least from Kant himself. In a letter to Reinhold, for example, he writes that “what a Maimon wants with his correction of critical philosophy (which Jews often like to try, in order to give themselves a sense of importance at the expense of others), [he] was never really able to grasp and whose rebuke [he has] to leave to others” (see AA 11: 476)). 7

Philosophy is the science of the form of all sciences

105

ulary, they both recognise that philosophy must be reformed through self-investigation, and that this procedure must begin with a hypothesis about the nature of the relationship between mind and world. Maimon argues that, while Bacon accepts the general possibility of applying logical forms on to the objects of nature in general, without requiring further argumentation, he only seems to be concerned with understanding their application in particular cases. Kant, on other hand, is portrayed as asking the more general question of whether human cognition entails an application of logical rules to objects of nature at all, or whether cognition of objects contains elements of the understanding as well as of sensibility. Maimon reads Kant’s propaedeutic project as trying to show no less than this: that all possible objects of cognition must fall under the general forms of the understanding and sensibility, and that only on these grounds can it be explained why universal and necessary connections obtain between the objects of nature. Bacon’s programme, on the other hand, departs decisively from the Kantian reformation programme, because his method of self-investigation is not only experimental but experimental in the empirical sense. It is for that reason that the goal of his programme takes a descriptive turn. On Maimon’s analysis, Bacon’s restorative philosophy adopts an empirical stance to examination and determination of how logical form has been applied to particular natural phenomena or objects. Maimon himself, however, sides with Kant’s research programme, taking the core of this program not to consist in empirical, descriptive explanations of the cognition of objects, but to consists in justificatory explanations, i.e., explanations providing reasons for why applications of certain forms of cognition to its objects are warranted. In agreement with Kant, Maimon believes that metaphysics must take a step back and engage in a methodological investigation of its own nature, to figure out whether, and in virtue of which conditions, it can be pursued as a science. Further, he agrees with Kant that a sufficient answer to this question consists in demonstrating “how […] philosophy as a pure a priori cognition [is] possible”.10 This view results from his 10 “So, the question is: how is philosophy as a pure a priori cognition possible? The great Kant posed this question in his Critique of Pure Reason, and answered

106

Maimon’s method of fictions

adoption of the latter’s paradigm of a Copernican revolution, conceiving objects as conforming to our forms of cognition, since we can “only know of objects what we ourselves have put into them”. Furthermore, he also follows Kant in committing to an explanation of specifically human cognition, which to him also means discursive spatiotemporal cognition. Maimon concedes that finite cognition only arises through its object being “given to it from somewhere else” (Tr, GW II, 3). As discussed in the previous chapter, Kant assumed that our a priori cognition of objects should be explained by virtue of the epistemic conditions (that is, the a priori representations) which we must apply to objects in order for them to be cognisable by us. Like Kant, Maimon assumes that these conditions must include the forms of space and time, which conditions the specifically human cognitive capacity to individuate and differentiate objects spatiotemporally: any “intrinsic diversity […] shows up in two forms: topical or positional discreteness, and sequential discreteness” (Lachterman, 1992, p. 502). At the same time, he identifies the essential function of the categories of thought as generating “unity in manifold” (Tr, GW II, 16), giving an account of cognition which must include both the forms of thought and the forms of sensibility in order to represent and cognise objects of our cognition.11 it as well, by showing that philosophy must be transcendental if it is to be of any use; that is, it must be able to relate a priori to objects in general, and is then called transcendental philosophy. This is therefore a science that relates to objects determined through a priori conditions, and not a posteriori through particular conditions of experience: this distinguishes transcendental philosophy as much from logic (which relates to an undetermined object in general) as from the doctrine of nature, which refers to objects determined through experience” (Tr, GW II, 3). 11 Consider Maimon’s own words: “So if A and B are completely identical, there is in this case no manifold. There is therefore no comparison, and consequently no consciousness (and also no consciousness of identity). But if they are completely different, then there is no unity and, once again, no comparison, and consequently no consciousness, not even consciousness of this difference, since, considered objectively, difference is just a lack of identity (even though, considered subjectively, it is a unity, or relation of objects to one another). As a result, it cannot have objective validity. So space and time are these special form by means of which unity in the manifold of sensible objects is possible, and hence by means of which these objects themselves are possible as objects of consciousness.” (Tr, GW II, 16)

Philosophy is the science of the form of all sciences

107

Now, although Maimon agrees with Kant that metaphysics must depart from the standpoint of finite cognition, his new philosophy proposes quite a different route to understanding the metaphilosophical challenges thus raised. Against Kant’s proposal of establishing the propaedeutic procedure, Maimon offers the following reflections: Concerning the final question, namely: How is metaphysics possible? we must first determine what is meant by ‘metaphysics’. I think I am in agreement with Kant as to the definition of metaphysics, namely, metaphysics is the science of things in themselves. I differ from Kant only in this: according to him things in themselves are the substrata of their appearances in us, and are quite heterogeneous with these appearances so that this question must remain unresolved insofar as we have no available means of cognition of things in themselves abstracted from the way they affect us. According to me, on the other hand, cognition of things in themselves is nothing other than the complete cognition of appearances. Metaphysics is thus not a science of something outside appearance, but merely of the limits [Gränze] (ideas) of appearances themselves, or of the final members of their series. Now these are indeed impossible objects of our cognition, but they are so closely connected to the objects that without them no complete cognition of objects themselves is possible. We approach ever closer to cognition of them according to the degree of completeness of our cognition of appearances. (Maimon, 2010, pp. 248-49 [emphasis added in italics])

First, Maimon here obviously misreads Kant’s metaphilosophical position, since Kant explicitly distinguishes between two kinds of metaphysics (general and special) and only declares the latter as unamenable to scientific inquiry. Kant makes clear that his experiments of reason not only lead to both positive and negative results (KrV, Bxviii). Second, Maimon seems to believe that adopting the Heterogeneity thesis entails that one also has to adopt a Heterogeneity thesis with regard to the metaphysical entities which Kant’s theory postulates, i.e., Maimon defends a two-objects view, taking Kant’s transcendental idealism hypothesis to postulate the existence of two kinds of objects: real and mental entities. He infers from this that this view would again create the need for another account of cognition which would state the epistemic conditions under which the cognition of this class of objects is possible. But this will ultimately create a contradiction, since these objects, i.e., things in themselves, are per definitionem objects whose cognition cannot be conditioned through epistemic conditions. Yet—and this is his key divergence from Kant—Maimon does not

108

Maimon’s method of fictions

believe that metaphysics must therefore abandon its aspirations, but merely reconceive of its actual objects of investigation. Attacking the premises of the two object-view, Maimon invokes one of Kant’s key premises, that if “reason has insight only into what it itself produces according to its own design” (KrV, Bxiii), we cannot say anything about objects which do not fall under the epistemic conditions that are characteristic of human cognition, and so we cannot posit anything intelligible “outside appearance[s]”. Consequently, metaphysics—according to Maimon’s redescription—must determine how a priori cognition of things as appearances is possible, that is, it must determine “the limits [Gränze] (ideas) of appearances themselves”, which are equivalent to “the final members of their series”. Maimon’s metaphilosophical position adheres to the traditional view of metaphysics as “a science of things in themselves”, yet it reconceives the notion of things in themselves from a perspective that is thoroughly epistemological: as the conditions of the a priori cognition of objects, “they are so closely connected to these objects” that the complete cognition of objects themselves entails the cognition of their conditions of possibility. In fact, it will turn out that only the end of science itself can act as proof of the propositions of metaphilosophy, through a complete cognition of all objects. Even in this initial quotation we can see that Maimon holds that assertions about the truth of metaphysical judgments can only be justified through a complete cognition of all objects or appearances.12 We will see in sections III and IV that what underlies this conception of metaphysics is in fact Maimon’s commitment to rationalism, which for him includes a demand for explanatory rationalism and infinite intelligibility. What should be noted here is that, 12 This is, however, something which finite cognition can only do in principle but not in reality: “The complete consciousness of all parts of synthesis and consequently also of the synthesis itself, is not a representation, but rather a presentation of the (understanding’s) thing itself. It is, however, to be noted that both the primitive consciousness of a component (of a synthesis) not related to something, as well as the consciousness of the complete synthesis are mere ideas, that is, they are the two limiting concepts [Gränzbegriffe] of a synthesis, in that no consciousness is possible without synthesis, while the consciousness of the complete synthesis grasps the infinite itself, and consequently is impossible for a limited faculty of cognition” (Tr, GW II, 349 [emphasis added]).

Philosophy is the science of the form of all sciences

109

contrary to Kant, who believes that metaphysics can become a science just by virtue of an experiment of reason, Maimon thinks that a true metaphysical theory is what stands at the end of science and its progress as such. A true metaphysics is only reached through the completion of all sciences and their cognitions of objects. At the same time, metaphysics as metaphilosophy is also the first science, insofar as it provides the methodological hypotheses which are needed to begin that whole process which we call science. It is also the last science, insofar as it is the science of the limits of all appearances, which will only be revealed once the scientific process as a whole reaches its conclusion. On Maimon’s view, then, metaphilosophy determines objects only with regard to their form, since it is only the form of the objects of our cognition that can be known a priori and thus with absolute certainty. Consequently, Maimon’s metaphilosophy can be classified as a “formal” science, since it does not employ or make reference to any a posteriori evidence. As a formal science, this philosophy has a special purpose: it unites all sciences under one form, no matter their individual objects or methods. Since there exist different scientific methods (e.g., mathematical methods, inductive and experimental methods in the natural sciences, etc.), which produce different varieties of cognition that do not, for Maimon, cohere with each other, it only makes sense that it should be part of the scientific enterprise to look for a method fit to equip all sorts of knowledge claims from the various individual sciences with scientific form.13 The different sciences must be united under one criterion that makes them what they are: sciences. And indeed, Philosophy is the science of all sciences, through which they [i.e., all other sciences] only ever acquire their status as sciences. […] If the objects of nature are ordered philosophically under principles, are brought into a system, they become a natural science proper. (Strf, GW IV, 34 [emphasis added])

13 This, for Maimon, results from the fact that some of the methods of the mathematical sciences apply a different standard of truth and inquiry than the experimental natural sciences. I will turn to this point in section III, as well as in chapter 4.

110

Maimon’s method of fictions

Maimon holds that the proper task of philosophy is to be metaphilosophy-first, which entails not only the determination of those conditions which determine the possibility of scientific cognition in philosophy, but at the same time it aims at the discovery of the scientific form of all sciences, the forms of whose cognitions must be the same as in philosophical cognition. Thus, although each science proceeds according to a different method, they are unified under one form, the scientific form (as such). In this way, Maimon’s metaphilosophy assumes a normative role with regard to the individual sciences, as they can only claim their status as science once the epistemic conditions that ground the instances of cognition from which they are constituted have been established.14 With metaphilosophy’s role and goal thus identified, its object must consequently be determined as “the form of science in general, or of a whole [ein Ganzes] of cognition” (Strf, GW IV, 132). Now, using these starting premises, this chapter will argue that Maimon is also dedicated to developing a “testable” version of the metaphilosophy -first research programme. Following Kant, Maimon uses his knowledge of the sciences of his time to develop a method that is specifically suited to deal with the difficulties of philosophical self-investigation, which in his case demands a method which can determine objects that are beyond our cognition insofar as they can only be proven through, and cognised by, an infinite intellect. Unlike Kant’s, Maimon’s version of an experiment of reason uses another method for theory-construction and testing, and is grounded in different epistemological presuppositions. We will see that these are the exact reasons that Maimon took to show that metaphysics can only be transformed into a proper science by virtue of an employment of scientific fictions. Before turning to the motivation and investigation of the philosophical procedure which enables the execution of Maimon’s metaphilosophical programme, a better understanding of Maimon’s criticism of Kant’s cognitive theory (section II) and his own alternative proposal is required (sections III and IV).

14 “[Philosophy is] a science that relates to objects determined through a priori conditions.” (Tr, GW II, 3)

Against propaedeutic philosophy: quid facti? and quid juris?

111

2.2 Against propaedeutic philosophy: quid facti? and quid juris? As I mentioned above, Maimon entered into the philosophical debate surrounding Kant’s work with a study titled Essay on Transcendental Philosophy. In this work, Maimon raises two famous criticisms to Kant that have come to be known under the titles of two questions: quid facti? and quid juris?. While that quid facti doubts that we are in possession of a fact 15 that proves the actual application of the pure forms of understanding to particular objects given to the senses, the quid juris demands an argument to determine whether we are justified in applying the pure forms of understanding to some given cognitive matter. As discussed in chapter 1, Kant assumes that human cognition involves the contribution of two faculties of cognition, sensibility and understanding. On Kant’s cognitive dualism, if you will, sensibility contributes the content of cognition, which is structured according to the pure forms of space and time. The understanding is then responsible for further determining and cognising this content via application of a priori concepts and principles. Kant maintains that the connection between elements of both faculties is possible in virtue of their pure forms, that is, he believes that the concepts and principles of the understanding can be applied to all sensible objects via their temporal schematisation, insofar as all appearances are structured with respect to the pure form of time. Two things are important to remember looking forward into this chapter: firstly, Maimon will argue that Kant’s account of object cognition can neither show the actuality or possibility of an application of the categories, nor that it can show how we could justify such applications. Secondly, while it is impossible for any account of cognition to resolve the question of facts (quid facti?), there is an account, i.e., Maimon’s own account, that provides a sufficient answer to the question with

15 For example: “Its [i.e., quid facti] meaning is this: how do we know from our perception that b succeeds a that this succession is necessary, whereas the succession of the very same b upon c (which is equally possible) is accidental?” (Tr, GW II, 71).

112

Maimon’s method of fictions

of justification (quid juris?). In what follows, I will first explain the argument quid facti and then the argument quid juris.16 2.2.1 Quid facti? Maimon’s first criticism, namely his argument quid facti, stands in close connection to his revival of Humean scepticism, which he defends against a response which he ascribes to Kant.17 Humean scepticism amounts to scepticism about whether we should assume that necessary and universal connections obtain between sensible appearances, e.g., that the concept of cause applies to the objects we perceive through the senses. One of the results of Kant’s propaedeutic is that appearances do stand in necessary and universal connection, as the cognition of objects requires that appearances be subsumed under the categories of the understanding. One of the a priori representations that must apply to objects in order to cognise them is the category of cause and effect (KrV, A189-211/B232-256). Now, on Maimon’s reading, Kant’s argument for positing connections of this kind is grounded in the assumption of an undeniable fact: that there are actual judgments of experience. Maimon’s criticism is that neither scientific nor ordinary experience can produce judgments of this sort, i.e., judgments that are expressive of ne16 It is important to keep in mind that neither Maimon himself nor my interpretation claims that this engagement with Kant’s philosophy aims at a correct exegesis of the latter’s philosophy. Rather, Maimon tries to elaborate on what he takes to be “the most important truths of this science”, and he comments and criticises where he sees these truths as insufficiently developed or realised (Ehrensberger 2004, pp. VIII-IX). In a similar spirit, my interest here lies in unfolding Maimon’s strategy for revising philosophy and its role in the scientific enterprise, and not in evaluating whether his criticism is fair to Kant or not. 17 Maimon himself connects quid facti to Hume’s scepticism, e.g. “the question quid facti?, to which Hume’s objection appears to be irrefutable” (Tr, GW II, 9), also see (Tr, GW II, 120, 261). On the quid facti? argument in general, see Franks (2003, pp. 215-30; 2007) and Senderowicz (2003). Like many contemporary scholars, Maimon takes Kant’s philosophy to be a direct response to Humean scepticism, although I think Franks is right to point out that Kant thought Hume to have asked the right question (2014, p. 20). Kant’s analysis intends to show the basis on which scientific explanations can be subject to sceptical doubt. His constitutive account of cognition distinguished between scientific explanations which can be subject to scepticism and those which cannot.

Against propaedeutic philosophy: quid facti? and quid juris?

113

cessary and universal connections amongst appearances. That is, there are no scientific nor ordinary causal judgments that would serve as proof of an actual application of the categories to particular objects given to the senses.18 Therefore, Maimon concludes, the Critique presupposes what it is supposed to prove.19 With regard to the causal judgments of ordinary experience, Maimon uses Kant’s own example (KrV, A190-92/B235-37). In this example, the latter compares the succession in the representation of a moving object to the succession in the representation of a still object. More concretely, we are asked to imagine a ship moving down a river and a house on the riverbank (II, 188). Kant points out that, while we represent the house as a whole without regard for the order of successive perceptions, the order in which we represent the moving object matters. With regard to the ship moving downstream, causal judgments constrain the order of successive perceptions as irreversible.20 In order to perceive successive events in the world, we need a representation of succession as irreversible, or in Kant’s preferred vocabulary, we need the category of causality that represents the flow of events as rule-governed. Maimon, however, submits that “these two kinds of succession are not in fact distinct from one another and so when someone asserts that the ship actually moves down the stream, he does not in fact know what he means by the word ‘actually’ ” (Tr, GW II, 188-9). By judging a succession of perceptions as causal events we already apply or rather impose the category of causality on to the succession, without being able to substantiate a criterion that explicates the grounds on which we do so. All that one’s actual experience effectively admits of is that one does perceive regularities and subjective compulsions which serve as the grounds for judgments of subjective or comparative necessity (Tr, GW II, 173-176).21 Hence, we have no reason to prefer an explanation that 18

See Franks (2003, p. 219; 2014, pp. 40-41). Maimon argues that “on Kant’s assumption, experience is possible because it is actual, and this is why these concepts have objective reality” (Tr, GW II, 185 [emphasis added]). 20 See Thielke (2001b, p. 444). 21 In Maimon’s words: “But if we want to consider the matter more precisely, we will find that the expression, ‘objective necessity’ is actually meaningless, 19

114

Maimon’s method of fictions

posits the objective application of categories over one that posits the mere subjective application of categories. The fact that the judgments of experience we make in our everyday life exhibit this specific form (i.e., taking appearances to be causally structured) does not force us to posit necessary and universal conditions of experience that are thus a priori.22 In fact, Maimon suggest that these judgments are just as well explained by virtue of habituation and psychological laws (II, 72, 129).23 Rather than postulating a transcendental framework to explain the imposition of a causal structure on to reality, we should explain the content of our perception judgments by virtue of principles that are continuous with those of the natural sciences.24 Furthermore, Maimon also investigates the possibility of whether the because necessity always signifies a subjective compulsion to accept something as true” (Tr, GW II, 173-176). Likewise: “I doubt the reality of experience itself so that for me the categories are not, as they are for Kant, conditions of experience (objective perception); rather they are conditions of perception in general, which no one can doubt” (Tr, GW II, 261). “Kant merely presupposes the fact, but he does not prove it. So these principles remain merely probable, but not necessary” (Tr, GW II, 342). 22 “But to this David Hume would reply: it is not true in this case that I perceive a necessary succession; I certainly use the same expression than others use on this occasion, but I understand by it only the often perceived succession of the warming of the stone upon the presence of the fire and not the necessity of this succession. It is merely an association of perceptions, not a judgment of the understanding” (Tr, GW II, 72-73 [emphasis added in italics]). 23 In fact, Maimon argues, we even have more reason to adopt a naturalistic explanation, since we thereby avoid violating the principle of parsimony. As Franks points out, Maimon refers us to Newton’s methodological principle that “one should assume no new principle for the explanation of a phenomenon, which may be explained from other, long since known principles” (Strf, GW IV, 239n, as cited in Franks (2014, p.43)). 24 Sometimes Maimon also alludes to the power of judgement, which must be cultivated in order to pick out the right connections; see for instance the following quotations: “Its [QF’s] meaning is this: how do we know from our perception that b succeeds a that this succession is necessary, whereas the succession of the very same b upon c (which is equally possible) is accidental? Kant indeed notes (and rightly) that the answer to this question depends only on the power of judgment” (Tr, GW II, 71). Or “whether the perception is itself correct amounts to the answer to the question quid facti? It is based only on the power of judgment (Beurteilungskraft) and no further rules can be given for this” (Tr, GW II, 129).

Against propaedeutic philosophy: quid facti? and quid juris?

115

desired fact can be extracted from the practice of science. His analysis of the nature of scientific judgments, however, yields a negative result.25 For one, the kind of mathematical judgements made by Newtonian scientists do not actually employ a concept of cause and effect. Rather, these judgments are grounded in the concepts of natural forces (BNO, IV, 374) and the principle of continuity [Stätigkeit] (BNO, GW IV, 366, 380).26 Secondly, he notes that the experimental Baconian sciences most certainly do not assume necessary and universal connections amongst sensible appearances. These sciences self-consciously produce judgments that only express “comparative necessity and universality”; they are concerned with “empirical laws that determine real relations between particular objects” (Strf, GW IV, 35). Thus, contra Kant’s inaccurate description of scientific judgments and the kind of constitutive principles they assume, Maimon argues that the judgmental practice of the sciences can be both be explained according to a priori or a posteriori principles, none of which include the category of causality. In addition to this analysis of ordinary and scientific judgments, Maimon puts forward a second point concerning the arguments quid facti. He says that: “either the fact itself […] is false, and the cited examples [e.g., of empirical judgments] are based upon a deception of the imagination, and […] the categories have no use at all; or it is true in itself, and then it has no knowable ground, and the categories remain, even after their laborious deduction and schematism, as before, mere forms that can determine no object.” (Logik, GW V, 250 [emphasis added])

Even if there should be some a priori forms that determine the way in which we can gain empirical knowledge, these principles themselves 25 Thus, I disagree with Senderowicz’s claim that Maimon’s argument is not based on a “scientific refutation” of the fact (2003, pp. 176-77). Pace Senderowicz, Maimon does not only refer to examples of ordinary experience, but already in the Essay (Tr, GW II, 140), and even more so in his writings on mathematical and natural science analyses, various examples of scientific experience. See especially the Preface to the translation of Bacon’s New Organon (BNO, GW IV, 359ff.). 26 Of special interest in this respect is Maimon’s modified concept of causality at work in scientific judgments, e.g., (Tr, GW II, 80); see Bergmann (1967, pp. 127-137) on this matter.

116

Maimon’s method of fictions

must remain indeterminate or unknowable to us. That is to say, we could accept that Kant’s account of cognition shows that we have some perceptual experience, and that in order for this experience to be possible at all, all its objects must be constituted as undergoing a series of events that each have some indeterminate cause. Yet, this account still does not give us a rule—a “knowable ground”—on which it becomes transparent how a determinate cause is assigned to a particular event (Franks, 2003, pp. 224-29). “From this principle it only follows that objects of experience in general must be thought of as causally related to one another, but in no way that it must be just these objects that stand in this relation” (Logik, GW V, 489).27 A fact could justify the claim of an application of categorical forms to particular objects given to the senses only if it could explicate the rule according to which determinate causes are assigned, through which it would truly be shown that scientific experience is possible. 2.2.2 Quid juris? As well as criticising the strategy of assuming an indubitable fact, Maimon raises a second worry under the label of the question quid juris?. This question and its corresponding argument directly attack Kant’s strategy of explaining our conceptual apparatus as constituted by two heterogeneous kinds of representations. Rather than asking for a fact that can serve as evidence for the actual application of our a priori forms of cognition to particular sensuously given objects, the quid juris demands a description that articulates how it possible to apply the categories to particular sensible objects, and thus why we would be justified in doing so. Maimon contends that such justification cannot be given 27 Consider the entire quotation: “How, from the principle that everything that appears does so according to the law of causality, can I derive, through the given objects of determinate propositions, that the sun’s rays necessarily melt the ice? From this principle it only follows that objects of experience in general must be thought of as causally related to one another, but in no way that it must be just these objects that stand in this relation. [Kant’s] answer to this question then fails, according to me: we know synthetic judgments merely in relation to an object of possible experience in general, but nothing of synthetic judgments that relate to determinate objects of real experience.” (Logik, GW V, 489–90)

Against propaedeutic philosophy: quid facti? and quid juris?

117

through a dualist account because it cannot make intelligible how our cognitive forms can determine the matter of cognition, since form and matter of cognition are heterogeneous to one another. Franks (2005) argues, and I think rightly so, that Maimon reads Kant’s dualism as expressing a real distinction between exercises of faculties, i.e., that each exercise is intelligible independently of the other (pp. 54-61). For Kant’s dualist account, this means that these exercises must stand in an external relation to each other. In order for both exercises to be intelligible independently, it must thus be intelligible how the receptive faculty is affected by something outside of itself which is neither itself nor the other spontaneous faculty. This is exactly the point that Maimon presses: how can we make sense of this external relation, that is, can we show how spontaneous and receptive capacities cooperate in cognition if their relation is an external one? His worry consists in this: the dualistic distinction between receptivity and spontaneity “makes unintelligible the relation between what is distinguished” and thereby fails its explanatory task.28 Why? In the previous chapter, I showed that, on Kant’s view, the possibility of scientific explanations can only be accounted for if we assert that human cognition depends on its objects being given to it from somewhere else. But this does not necessarily mean that what is given to sensible intuition is prior to thought. Rather, Kant reasons that given intuitions must be thought in order for there to be objects for us. And if anything must be thought in order to become an object for us, then thought is prior to sensible content and indeed must determine it according to its forms. Yet Maimon reads this dualism as entailing neither a modal nor rational distinction, which would leave some room for explaining how one form applies to the other.29 Rather, he thinks Kant is positing a real distinction that conceives the forms of thought and that which is given as completely heterogeneous. And, for Maimon, Kant fails to answer the quid juris because he 28

See Brandom (2002, p. 263). Examples are Franks (2003, 2005), who interprets Kant’s dualism as a modal distinction, and Conant (2016), whose “transformative” reading implies a rational distinction. 29

118

Maimon’s method of fictions

thinks it is constitutively impossible for a dualistic account to adequately ground the possibility of scientific explanations: Is it conceivable that a priori forms should agree with things given a posteriori? […] how can we conceive of matter arising, as something merely given and not thought, by assuming an intelligence, since they are so heterogeneous? […] That is to say, how can the understanding subject something (the given object) to its power (to its rules) that is not in its power? According to the Kantian system, namely where sensibility and understanding are two entirely different sources of our cognition, this question is insoluble as I have shown”. (Tr, GW II, 63-4 [emphasis added])

Read in this way, the explanatory role of Kant’s cognitive dualism becomes unintelligible, as it becomes impossible to show that and how the a-priori-representations conditions of the understanding apply to sensible manifold ordered through the forms of sensibility, since both cognitive elements, as well as the ways in which they order content, are argued to be completely heterogeneous to each other. If it is assumed that the elements of cognition are in fact—as Kant claims—heterogeneous to each, then it also becomes incomprehensible how the receptive and spontaneous capacities are supposed to cooperate, since it is not intelligible how something can be computed according to rules which it is not susceptible to. In other words, Maimon contends that if Kant accepts the heterogeneity thesis he cannot show why and how the Cooperation thesis is true. This charge has the same target as what has been described by Conant (2016) under the label of the “layer-cake model of cognition”. If we conceive of receptive and spontaneous capacities as self-enclosed, independent layers of a cake, it becomes inexplicable how these layers are supposed to mix or mesh in an orderly, i.e., non-arbitrary and comprehensible, fashion. Maimon’s charge is this: if our cognitive access to the world is constituted such that it is grounded in a real dualism, then it is by virtue of this cognitive constitution impossible to explain how universal categories (i.e., concepts) can intelligibly determine particular objects given to the senses (i.e., intuitions). Maimon’s argument quid juris puts forward a version of the argument that would later become known as the “myth of the given”.30 This argument claims that it is already immanent to the notion of given 30

Franks (2014, p. 54) and Thielke (2003, p. 111, in reference to McDowell)

Against propaedeutic philosophy: quid facti? and quid juris?

119

perceptual content that it cannot play an appropriate justificatory role, because a so-conceived given manifold “does not provide any a priori criterion, whereby one could know whether [it] can be thought in a unity of form in general, still less any criterion by which one could know in which unity” (Logik, GW V, 476). 31 Which is to say that Maimon claims that if intuitions do not admit of some conceptual foundation which is at least potentially accessible to conceptual rules, then the criteria according to which we apply these universal rules to particular institutions become inconceivable.32 Even if all intuitions share formal, i.e., spatiotemporal features, this can only explain how categories apply to intuitions in general, but never how they discriminate between particular intuitions and their contents. From this argument, we see the consequence, that it remains unintelligible how we could have a principled way for applying our conceptual scheme onto the world if that which it is to be applied to does not already comprise conceptual structure, and thus offer a potential principle or criterion for its ‘processing’. Additionally, Maimon contends that Kant’s real distinction between the exercise of faculties commits him to an incoherent metaphysical view that assumes the existence of things themselves, since in order for both exercises to be intelligible independently, it must thus be intelligible how the receptive faculty is

make this connection. The myth of the given was originally described by Sellars (1956). 31 In the Philosophical Dictionary (1791), Maimon seems to go even further by stating this to be the problem or “antinomy of thinking itself” (PhWb, GW III, 162). The thought is roughly this: thinking is constituted in applying form to matter; form needs matter as the condition for it to have reality, for us to be able to have consciousness of it; and matter needs form, in order for it to be anything for us. Hence, each is constitutive of the other, leading to a problem of what grounds what. Franks labels this transcendental antinomy scepticism, (2014, p. 49), the question of whether experience is possible at all, as thinking itself always runs into an antinomy of form and matter, which are both necessary conditions of the possibility of experience. 32 Thielke stresses this point as being Maimon’s most potent criticism (2001a, p. 112). See also Atlas (1967, p. 69), Franks (2014), and Engstler (1990, pp. 71-190).

120

Maimon’s method of fictions

being affected by something outside of itself, that is neither itself, nor the other spontaneous faculty.33 Yet, this is not Maimon’s last word. As part of his criticism of Kant’s work, he begins to sketch out an alternative account for philosophy that escapes the consequences of the quid juris argument. Although philosophy must be sceptical in the way discussed above, it has more options than just accepting a Hume-inflected naturalism. We have seen in the first section that Maimon in fact sides with Kant’s project to deliver a metaphilosophy which shows how metaphysics as science is possible. His criticisms make clear why this philosophy cannot adopt the Kantian strategy specifically, because of Kant’s commitment to the Heterogeneity thesis. Maimon’s Essay shows why Kant’s experiment of pure reason cannot show what it promises: it cannot deliver a consistent model of cognition because the hypotheses it tests must run into contradiction, e.g., if the Heterogeneity thesis is true, then the Cooperation requirement cannot be shown to be true.34 Hence, the theory it advances does not even qualify as a possible explanation, and much less a best explanation of the conditions which make possible our cognition and representation of objects. Maimon therefore advances another theory of cognition which opposes the Heterogeneity thesis, and rather adopts what I will refer to as the Homogeneity thesis. He believes that it is only by adopting this hypothesis, that a priori representations that are epistemic conditions of human cognition are homogenous in nature, can we arrive at a theory that is consistent with the Coopera33 This bears some similarity to the criticisms raised by Jacobi (1787) and Schulze (1996/1792), which were taken up again by McDowell (1994, pp. 41-3). 34 Franks (2003) argues that Maimon’s criticisms must be seen as motivated by the latter’s rationalist commitments to infinite intelligibility. Whilst I agree with his account of how Maimon’s rationalism and skepticism are connected, I am not convinced that this is also how we should understand Maimon’s criticisms of Kant’s dualism. Maimon seems to believe that Kant’s Heterogeneity is not just incapable of giving a rationalist explanation for why the Cooperation requirement should be consistent with it, but he believes that Kant’s theoretical apparatus is simply incapable of giving any explanation, because these two hypotheses are inconsistent with each other. On my tentative view, Maimon’s rationalism only comes into play when we are concerned with his alternative account. I am indebted to Marialena Karampatsou for discussing this point with me.

A coalition system

121

tion requirement and, unlike Kant’s model, still manages to explain the possibility of metaphysics as a science of things in themselves. To understand Maimon’s alternative solution to the Kantian metaphilosophical question, we first need to consider how this solution proceeds from fundamentally different presuppositions:35 those of infinite intelligibility and explanatory rationalism.

2.3 A coalition system: Maimon’s rational dogmatism and empirical scepticism In order to understand the presuppositions guiding Maimon’s own philosophy in its attempt to determine the formal conditions of all scientific cognition, one must shift focus from his criticisms to the alternative epistemological position that he proposes. In characterising the nature of his own philosophy as an alternative to Kant’s propaedeutic, Maimon employs the notion of a ‘coalition system’ (LG, GW I, 557).36 This system37 combines two epistemological positions regarding the possibility of scientific explanation: empirical scepticism and rational dogmatism. In contrast to his own position, Maimon ascribes to Kant the position of an empirical dogmatist and a rational sceptic. Kant is an “empirical dogmatist” because he is not sceptical about our empirical cognition, nor our judgments that concern the empirical world (Tr, GW II, 435). And he is a “rational sceptic” because he denies (certain types of) metaphysical cognition, that is, cognitions which falsely take something besides appearances as their objects (ibid.). So, first, what makes Maimon a rationalist? He identifies his position as “rational dogmatism” because he is committed to an unlimited 35

See Franks (2003, p. 201). “I had already appropriated the systems of Spinoza, Hume and Leibniz in the same way, it was natural that I sought to find a system agreeing with all these systems—a coalition-system. And in fact I found it and established it in the form of comments and explanations of the Critique of Pure Reason which finally gave birth to my Essay on Transcendental Philosophy” (LG, GW I, 557 [emphasis added]). 37 It will turn out that this proposed system is really a “non-system” (II, 443), since it reduces philosophy to a purely methodological science. 36

122

Maimon’s method of fictions

version of the principle of sufficient reason as explanatory standard. On this standard, to provide a genuine explanation of a fact, cause, or thing is to know its sufficient reason; for Maimon this means to provide a complete explanation. The chain of reasons explaining a thing “neither goes on forever, nor turns in a circle, nor terminates arbitrarily; instead, the series of reasons ends with an absolute reason that is self-explanatory, or wholly beyond the need for explanation” (Franks, 2003, p. 202). Maimon’s metaphilosophical programme is grounded in a different presupposition than Kant’s because, for the former, science must commit to a rationalist standard of inquiry, and due to this dogmatic commitment it must also commit to the view that only complete explanation can count as genuine (and thus scientific) explanation. It is this commitment to rationalism which Maimon presupposes dogmatically, taking the stance of a rational dogmatist. It should be noted here that even though Maimon gives no explicit reason for this commitment to infinite intelligibility, it would be wrong to look at it as just some arbitrary preference that happens to inform his method of philosophising.38 His point is not that philosophy should regress to the state and method of metaphysics before the Copernican revolution. On the contrary, his conclusions about the nature and methods of philosophy and science more generally lead Maimon to a metaphilosophical view that adopts and develops the Kantian project of an experiment of reason. For Maimon, it is this same commitment which instructs him in his choice of theory when it comes to the explanation of cognition, that is, by virtue of what epistemic conditions we can cognise objects. He says that rational dogmatists assert that both the forms and the objects of our cognition themselves are in us a priori, and that this faculty [Vermögen] does not consist merely in recognizing objects given to us by means of forms of thought by us, but also in producing the objects themselves by means of those forms. (Tr, GW II, 436) 38 Nisenbaum explores whether Maimon offers such an argument and comes to a negative conclusion (2018, p. 60). She seems to be right that there is no real argument for the Principle of Sufficient Reason (henceforward “PSR”) in Maimon’s philosophy, since everything he argues, including his arguments for the possibility of a priori cognition within a framework of infinite intelligibility, is already driven by a commitment to an unlimited PSR.

A coalition system

123

It seems quite uncontroversial to argue that it is his commitment to a standard of infinite intelligibility and explanatory completeness that makes him choose a model of cognition that accommodates these two demands. As we will see in the next section, Maimon’s rationalism motivates and instructs his hypothesis-construction regarding the nature of human cognition. His rationalism leads him to propose a variety of cognitive elements through which the representation of objects that are intelligible occurs without limits, allowing them to become objects of complete, and thus genuine, explanations. More precisely, it will be shown that Maimon’s account of cognition explains the form and content of cognition as homogeneous, and thus raises the possibility of understanding the application of a priori forms to cognitive matter as rule-governed application. Using this solution to the quid juris problem, he provides a model on which a priori cognition can be warranted in issuing necessary and universal judgments, since on this model we can explain how our forms of cognition apply to its content. Now, scholars have long been debating whether Maimon’s coalition system is more rational than sceptical, or the other way around, while still others—following Maimon’s self-description—have argued for the more plausible view that he is both: an empirical sceptic and a rational dogmatist.39 The interpretation corresponding to the one defended here aligns with what Paul Franks, Peter Thielke, and others have argued, that it “is only because he is a rational dogmatist that [Maimon] can be or perhaps, can only be an empirical skeptic” (Franks, 2003, p. 201).40 39 Franks (2003, p. 200) lists the main proponents of each view, as well as those who have argued that Maimon follows a so-called “middle path” (e.g., Beiser (1987, pp. 303-309) and Cassirer (1974, p. 103). He himself joins Engstler (1990) in arguing for the view which I endorse in this chapter, that Maimon’s view exactly consists in holding both positions at the same time. 40 Franks (2003) defends this point in stating that “Maimon’s rational dogmatism drives his empirical scepticism” because “that scepticism presupposes an commitment to infinite intelligibility as alone genuine” (p. 205), as does Thielke (2008) with his interpretation of Maimon, defending what he calls an “apostate rationalism”: “reason makes unavoidable demands on us that nonetheless cannot be met, while our natural propensities lead us to adopt beliefs that cannot be rationally defended” (2008b, p. 616, cf. Thielke & Melamed (2008)). The latter also explicitly connects Maimon’s thoughts to a particular

124

Maimon’s method of fictions

As finite cognisers, we are never in possession of the complete concepts or representations of objects, and are therefore also never in possession of the explanans necessary to produce a complete explanation. This is how Maimon’s rationalist commitment leads to the position of an “empirical sceptic”: it is by committing to a rationalist standard of what can count as scientific explanation, namely only a complete and unified explanation, that we will find our cognitive and scientific resources to be limited to a degree that makes such explanation impossible, thus forcing us to take up the position of the sceptic. For Maimon, this same commitment has consequences for his account of cognition: as opposed to the rational dogmatist, the empirical sceptic doubts the fact itself, i.e., that reason possesses or uses these forms; they will only hear of a single form, namely that of identity and contradiction, to which they attribute objective reality. By contrast, they are certain that a merely subjective reality should be attributed to the other forms, but because of their universality in relation to us, these forms serve in exactly the same way as if they had objective reality, so that the interest of reason is not narrowed in any way. (Tr, GW II, 436-7)

Since Maimon holds a standard of explanatory completeness, there can be no proof of the objective validity of any category that goes beyond their merely logical use. This is so because Maimon’s explanatory standard settles the criteria which a theory of cognition would have to satisfy if it were to satisfy this standard. Considering our standpoint as humans, and thus finite cognisers, however, this is a standard that our explanations—in the discipline of metaphilosophy—cannot meet. Hence, taking up the epistemological position of a rational dogmatist entails two things with regard to the development of a theory of cognition. First, it provides clear criteria for how the objects of cognition must be conceived: as objects that are intelligible without limits, which—at least in principle—can be represented under the epistemic conditions of human cognition. If such an account can be given, it will, in turn, provide an answer to the quid juris question. Second, this epistemological position has the consequence that no philosophical argument reading of Hume’s sceptical naturalism that explains this as a result of the latter’s underlying explanatory rationalism. See also Nisenbaum (2018).

The analogy to calculus

125

can provide sufficient proof, that is, an explanation of the theory of cognition that would satisfy this standard of justification. If this is true, then Maimon’s alternative theory will provide no answer to the quid facti question. Now, in his remark on rational dogmatism, Maimon hints at the major change that his hypotheses include in opposition to Kant’s. He seems to be suggesting that, with regard to the nature of our epistemic conditions, metaphilosophy cannot adopt a Heterogeneity thesis, but rather a Homogeneity thesis. In order to determine the structure of nature via determining the nature of our cognitive structures, the a priori representations that we establish must not only structure the form of our objects of cognition but also their content. Otherwise, it would not be intelligible how the content of cognition can play the justificatory role in judgements that it is supposed to. On Kant’s model, Maimon argues, we have no available explanation for why the application of some rules of the understanding to the content of a given intuition is justified, i.e., rule-governed (Tr, GW II, 61-64).41 Maimon’s alternative account therefore demands that “we must have some insight into how cognitive content arises” (Thielke, 2014, 684). And this is where the mathematical method of calculus comes in.

2.4 The analogy to calculus 2.4.1 Maimon’s philosophy of mathematics For both Kant and Maimon, mathematics, and the way it cognises its objects, were of crucial importance in understanding the way in which philosophical cognition works and should be justified. At the beginning of this chapter, I suggested that Maimon intends to develop a method for theoretical philosophy after the model of “the method of fictions”, and that it was the mathematised natural sciences, employing calculus for this purpose, which lead him to the key to conceiving of a cognition whose objects are intelligible without limits, and thus “impossible as objects of cognition”. Section II made it clear that, for Maimon, Kant’s model of 41

See McDowell (1994, p. 8).

126

Maimon’s method of fictions

cognition fails because it explains the possibility of a priori cognition as constituted through two heterogeneous kinds of representation, as well as for grounding its proof in presupposed facts of experience. I shall argue that Maimon proposes a different model of cognition in response, which uses specific insights from how the sort of cognition at work in differential calculus deals with its objects. It is in doing so that he comes up with a theoretical solution which explains our cognition as rational and thus scientific, drawing on a model that explains how cognitive form and content can arise without reference to intuitions or facts of perception. In the previous chapter, I briefly mentioned Kant’s philosophy of mathematics and his fascination with the geometrical method, especially the geometers’ ability to construct their objects of inquiry. On his view, geometry defines its concept by constructing in intuition those objects which simultaneously exemplify a correct application of its principles, rules, or theorems. By “exhibiting a priori the intuition corresponding to a concept” (KrV, A714/B741)—through intuiting the concept— geometers can reflect on its properties. Geometrical objects are intelligible in terms of their generation (Entstehung) because we know how to produce them: we can make explicit the rules according to which mathematical objects are generated, and it is through this process of spatial visualisation that we can learn about their defining properties. Maimon also turns to mathematics as paradigmatic science when it comes to finding a model to explain the possibility of a priori cognition of objects. In stark contrast to Kant, however, Maimon argues that it is not geometrical construction in intuition, but a different family of mathematical theories and practices, which can really make intelligible how mathematical objects come about: differential calculus.42 Consider the following example of a mathematical definition: “a straight line is the shortest distance between two points”. On Kant’s account, the mathematician arrives at this definition by constructing the concept of a straight line in intuition. She determines the properties 42 This is so because, “[i]n the differential calculus, space is considered as concept abstracted from all quantity, but nevertheless considered [as] determined through different kinds of quality in its intuition” (Tr, GW II, 22-23).

The analogy to calculus

127

of a straight line by reflecting on this same activity of construction; constructing a straight line between two points she learns that it has the property of demarcating the shortest distance between any two points.43 According to Maimon, the problem with this conception is that it understands mathematicians to gain their insights by directly “reading them off” intuitions. Intuitions are sensible images or representations of objects, which as diagrams show that the shortest distance between two points is a straight line. But this is exactly the problem: all one can learn from this exercise is the fact that a straight line is the shortest line between two points, but not why this is so:44 “even if I already see the meaning of the proposition that a straight line is the shortest between two points (by constructing a straight line), I still do not know how I arrived at this proposition” (Tr, GW, II, 43).45 From an explanatory standpoint, construction in pure intuition does not make available any clues as to why a mathematical concept applies to an intuition, all it does is to state that it just seems to be so. While the geometer actively constructs a figure in space, she only passively perceives the properties that this figure as sensible representation exhibits.46 A cautionary note here: Kant and Maimon have two different concep43 “[T]he straight line between two points is the shortest is a synthetic proposition, for my concept of the straight contains nothing of quantity, but only of quality. The concept of the shortest is therefore entirely additional to it, and cannot be extracted out of the concept of the straight line by any analysis. Help must here be gotten from intuition, by means of which alone the synthesis is possible.” (KrV, B16) 44 Both Bransen (1991, p. 73) and Pringé (2018, p. 37) emphasise this point. 45 Or, alternatively: “All that cannot be constructed otherwise cannot be cognized otherwise in a construction. Such an acclaimed principle of the possibility of a construction reduces itself to the barren identical proposition” (KUmG, GW VII, 399, as translated in Pringé (2018)); similar passages can be found in (Tr, GW II, 59) and (Logik, GW V, 472-3). 46 “The Understanding prescribes the productive imagination a rule to produce a space enclosed by three lines. The imagination obeys and constructs the triangle, but lo and behold! three angles, which the understanding did not at all demand, impose themselves. Now the understanding suddenly becomes clever since it learned the connection between three sides and three angles hitherto unknown to it, but the reason of which remains unknown. Hence it makes a virtue of necessity, puts on an imperious expression and says: A triangle must

128

Maimon’s method of fictions

tions of mathematical construction. Kant’s point is that the cognitive process do not consist in “read[ing] off, as it were, from the properties of the figure” (KrV, Bxii), but that construction entails a form of practical, self-reflexive knowledge. For Maimon, however, such diagrammatic reasoning is bound to intuition and is therefore not completely intelligible,47 as we cannot give sufficient reasons for the constructions of diagrams, e.g., why a straight line must be constructed thus and not otherwise, and why this endows it with the property of being the shortest line between two points. If this is all that construction in pure intuition can show, Maimon concludes, then such mathematical judgments can only express comparative necessity and universality (Tr, GW II, 72).48 Just as with any object or state of affairs encountered in experience, the repeated perception of a necessary connection does not amount to more than subjective necessity; the judgment is only highly probable.49 Only if the production of mathematical facts can be understood according to a framework that shows how (and not just that) intuitions arise from rules—how the production of mathematical facts can be understood according to their rules of generation—can they be said to exhibit the right kind of necessity. For such a framework, Maimon points us to the method of analytic geometry, insofar as it uses calculus. Unlike synthetic geometry, this

have three angles!—as if it were here the legislator whereas in fact it must obey an unknown legislator” (PhWb, GW III, 185-201; see also BNO, GW IV, 449-450). 47 Freudenthal argues that Maimon ultimately thinks so because geometrical truths rely on diagrammatic, i.e., sensible information, and sensible truth can never be necessary truth because we could have had other forms of sensibility (2019, p. 55). 48 However, I think Freudenthal (2006; 2010) is wrong when he argues this to be Maimon’s final judgement about mathematical construction. On my account, this fate only applies to mathematical construction within the Kantian framework. 49 “I note, however, that even if such propositions express necessity, this does not establish that they contain (objective) necessity. For example, my judgment that a straight line is the shortest between two points can derive from my having always perceived it thus so that for me subjectively it has become necessary. The proposition has a high degree of probability, but no objective necessity.” (Tr, GW II, 173 [emphasis added])

The analogy to calculus

129

method considers space as a concept (Tr, GW II, 22-3).50 In the example of a straight line, this means the following: the concept of the shortest line between two points gives rise to and thus is prior to the sensible image of a straight line.51 In order to be able to explain this, Maimon needs a nondualist account of space and time as forms of finite cognition of sensible objects. And this is just what he proposes. Maimon defines space and time as formal features of cognition and, since there can be no account of cognition that includes heterogeneous elements, he defines these formal features of cognition as concepts (and not intuitions). As conceptual representations, space and time are the conditions of possibility for a finite intellect to be able to represent and cognise unity in the manifold, i.e., to have cognition of something (whether objective or subjective).52 For Maimon, the forms of space and time enable us to perceive identity in difference; they are the conditions of differentiation—space as concept determines the simultaneous “being-apart of objects”, which at the same time is a cancellation of “being in the same place”—and individuation—time as concept of the “preceding and succeeding of objects with respect to each other”, which entails a cancellation of simultaneity (Tr, GW II, 16).53 On top of that, space and time, as intuitions, 50 On the connection between quid juris and differential calculus, see Altas (1964, pp. 109-123), Cassirer (2000, pp. 93-100), Duffy (2014), Engstler (1990, pp. 45-70, 124-164), Guérolt (1929, pp. 59-86), Kauferstein (2006 pp. 309348), Kroner (1921, pp. 353-356), and Thielke (2003, p. 110ff.). 51 “As soon as the understanding prescribes the rule for drawing a line between two points (that is, that it should be the shortest), the imagination draws a straight line to satisfy this demand” (Tr, GW II, 19). 52 Space and time “are these special forms by means of which unity in the manifold of sensible objects is possible, and hence by means of which these objects themselves are possible as objects of our consciousness” (Tr, GW II, 16). 53 Nisenbaum explains that “if I wish to view two things as different from each other, I must view them ‘simultaneously, that is, in one and the same point in time’ (Tr, GW II, 16). Stated otherwise, if I wish to view two things as different or outside of each other, I must, as it were, cancel their trajectory in time and focus only on their position in space. Yet if I wish to compare a variety of different things, I must, as it were, cancel their separate positions in space and focus successively on the aspects and relations that I am comparing. Space and time are therefore, together, the conditions for differentiation and individuation, and as concepts or rules they specify what we must do in order to be able to think of anything at all” (2018, p. 79).

130

Maimon’s method of fictions

are required to represent one particular object as individuated and differentiated in relation to other objects.54 The straight line drawn out by the mathematician is then nothing other than a sensible image of the concept “the shortest distance between two points”. Consequently, constructions in pure intuition explicate the rule for how to sensibly represent55 concepts of difference: “the sensible representation of difference is therefore a schema of the concept of difference, i.e., space as intuition” (Tr, GW II, 346).56 Still, that which functions to define a determination of space is a conceptual rule, and it is not the construction of the sensible image, but the definition of the conceptual rule as an analytical judgment, which proves its validity.57 54 E.g. “Space and time are as much concepts as intuitions, and the latter presuppose the former” (Tr, GW II, 18 [emphasis added]). For further elaboration this point, see Thielke (2003), Duffy (2004, p. 230) and Pringé (2018, p. 38). 55 Viewed in this way, Maimon’s re-interpretation bears a striking resemblance to Shabel’s reading of mathematical construction in Kant (see, e.g., 2003, pp. 109-133). As Pringé illuminates: “Maimon explains how a concept of difference (minimal distance) relates to its sensible schema (straight line), thereby showing in a concrete case how a certain determination of space as a concept produces its corresponding intuitive representation, i.e., how an intuition arises according to a certain rule. […] In contrast [ to Kant], it is rather the case that we impose such a property as the rule according to which the straight line can arise at all. In this case, we comprehend the way in which the straight line arises without intuiting it as already arisen.” (Pringé 2018, p. 38) Note that this is not an adequate understanding of Kant’s concept of construction, as I have shown in the first chapter. 56 As Maimon never defines it, it seems appropriate to use Kant’s definition: “[T]f I place five points in a row this is an image of the number five. On the contrary, if I only think a number in general, which could be five or a hundred, this thinking is more the representation of a method for representing a multitude (e.g., a thousand) in accordance with a certain concept than the image itself […]. Now this representation of a general procedure of the imagination for providing a concept with its image is what I call the schema for this concept” (KrV, A140/B179). 57 See for example (Tr, GW II, 42, 61, 65ff.). See Freudenthal (2006, pp. 8-39) for Maimon’s conception of mathematical judgments as analytic judgments. Although Maimon does not believe that he ultimately succeeded in proving the analyticity of geometrical judgements, he held this to be possible: “By contrast [to Kant], I pose the question in the following way. Since all a priori cognition must be analytic, and can be derived from the principle of contradiction, how can we make those propositions that are synthetic due to a lack in our cognition into analytic ones? […] I do not want to take on the task of developing all such

The analogy to calculus

131

This has interesting consequences for Maimon’s conception of geometrical construction, or rather his explanation of how diagrammatic reasoning works. Although the mathematician can explicate the rules that generate the sensible image of concept in order to demonstrate the properties that define it, the procedure for doing so must remain infinite: “there is no other means to construct a concept absolutely a priori than progressus in infinitum, as has already been shown” (BNO, GW IV, 447). While the mathematician can show the formal completeness of a concept because she can list the totality of all conditions by which a circle has to be thought, i.e., its rules of production, the complete material proof of a geometrical concept depends on an infinite procedure. For example, the concept of a circle is defined as “a plane figure bounded by one line, and such that all right lines drawn from a certain point within it to the bounding line, are equal” (Euclid & Heath 1956, p. 4). For a construction to reach material completeness and actually prove the adequacy of a definition by virtue of fact, it would have to produce an infinite number of equal lines from one centre point. The same goes for any straight line: to prove its material completeness the mathematician would have to infinitely divide the line to show that each segment thereof has the same angle as all of the others. Still, the definition is formally complete, as any further construction does not add to the conditions or rules of production that define the concept. Mathematical proofs as generated by geometry “are complete according to their rule, but in their presentation they are always incomplete” (BNO, IV, 447). Thus, compared to Kant’s model of construction, which he criticised as only showing that a mathematical fact obtains—instead of demonstrating why it does (i.e., we can construct the circle but cannot make explicit the rule that determines the production of this sensible image)—Maimon’s revisionary model suggests the reverse: geometry that makes use of calculus can show how mathematical objects are produced. While it can make explicit the rules for generating these objects, propositions in this way in order to make them satisfy my requirement; it is enough that I hold it not to be impossible” (Tr, GW II, 178–79 [emphasis added]; cf. II, 323).

132

Maimon’s method of fictions

calculus cannot actually produce these objects as sensible images which would prove the material completeness of the definition; “in spite of their material incompleteness, these concepts, or rather ideas of the understanding, are nevertheless correct because their rules can be made comprehensible by means of what is always given in intuition. For their material completeness they require only a continual repetition of this very rule. But since this repetition, in accordance with the concepts’ conditions, must be infinite, they remain mere ideas, and their application has the same degree of correctness as the degree of their material completeness” (Tr, GW II, 79ff.).58 Thus, mathematical judgments are necessary and universal because mathematical cognition can determine space as a concept; it defines the conceptual rules from which sensible representations of mathematical objects are generated (e.g., through the methodus indivisibilium (Tr, GW II, 274)).59 The method of indivisibles works on the basis of decomposing geometrical magnitudes, such as lines, planes, or figures, into infinitely small elements or building blocks that are of the same dimension as those magnitudes themselves. In Maimon’s own words, the method of indivisibles “treats a continuous (and hence infinitely divisible) magnitude as if it was composed of indivisible parts (a line as composed of points, a plane as composed of lines, a figure as composed of planes)”, and uses the known ratio between the indivisible parts to determine the ratio of the magnitudes they compose (Strf, GW IV, 51). We determine “something pertaining to a real object” by treating it both as infinitely divisible and made up of smallest indivisible parts, by treating the relation obtaining between the divisible parts as equivalent to the relation between the indivisible parts. According to this method, mathematical objects are determined to be generated from the 58 For Maimon, construction in pure intuition can also never be a priori (nor can we explain how we can gain explanatory insight from intuitions in general), because construction presupposes movement (e.g., movement of a point), which is an empirical concept (BNO, GW IV, 444). 59 “This is why I also hold that geometrical constructions can be demonstrated far more powerfully through the methodus indivisibilium, or the differential calculus, than in the usual way” (Tr, GW II, 274). Or, even stronger: “So, there is no other means to construct a concept absolutely a priori than progressus in infinitum, as has already been shown” (BNO, GW IV, 447).

The analogy to calculus

133

rules which govern the ratios between their differentials. Though these rules constitute the necessary conditions for the spatial representation of mathematical objects (e.g., representing a circle), these rules—as the determining values—cannot be represented. This might be explained better by way of an example.60 Imagine a body moving in space. Now, the trajectory of this moving body can be taken as a sensible representation of its instantaneous velocity because the instantaneous velocity is the rule from which the trajectory arises. Pringé explains this point well: Now, the velocity at each instant is a real object (a determinate intensive magnitude), a quantum of determinate quantity. But we cannot have any cognition of this determinate quantity through the velocity in itself, but only through its effect, namely through the space that a body with this velocity would traverse (if the velocity remained constant); but neither the duration of the movement nor the space traversed in this time are part of the essence of the velocity. So we must think the latter abstracted from the former, i.e. we must reduce them to an infinitely small space and an infinitely small time; but this does not make them any less real. (Pringé, 2017, pp. 39-40)

As rules which determine the generation of these trajectories, instantaneous velocities are referred to as “limiting concepts”. Limiting concepts are concepts which we can incrementally approach, but at the same time can never reach. According to Maimon, “[t]hey arise through a continuous regress or endless diminution of the consciousness of an intuition” (Maimon, 2010, p. 19). Thus, although the instantaneous velocity of a moving body cannot itself become an object of observation, or in Maimon’s terms, an object of intuition, it may be represented by virtue of its sensible image, which is nothing other than its trajectory. “[T]he instantaneous velocity (the rule of generation) is the derivative of the trajectory (the generated)” (Pringe, 2018, p. 40). Thus, differential and integral calculus shows us a way in which we can determine the rules of the generation of spatiotemporal representations, although what appears to us as determinate is only the generated, which, if thought of as consisting of infinitely small changes in x and y rates, yields its gener-

60 I am here following Pringé’s excellent reconstruction of this point; see Pringé (2018).

134

Maimon’s method of fictions

ating laws, although these laws cannot directly become objects of our experience. We will see in the next section that Maimon uses this reduction to an infinitely small space and an infinitely small time to explain the elements that constitute our intuitions, i.e., the sensible manifold of intuition, which is what he identifies as the content of cognition. It is this theory of differentials which drives Maimon’s explanation of a priori cognition. This explanation works by way of claiming a merely gradual difference between human and divine cognition, thus guaranteeing accordance with his standards of explanatory rationalism and infinite intelligibility. This again raises the issue of how to explain the relation between these two modes of cognition. Maimon then employs the theoretical conception of calculus to explain the relations within the fully determined relational whole that is God’s thought, which in turn provides the generative laws that give rise to what appears to us as sensible manifold or content of cognition. So, the relation between God’s thought and our thought is such that finite cognizers must represent the conceptual relations thought by God through sensible images, to cognize them as determinate (sensible) objects. 2.4.2 An immanent account of cognition Just as Kant made use of the concepts and practices of the experimental chemistry of his time, Maimon employs concepts from the mathematical sciences in order to conceive of his objects of study, i.e., the elements of cognition. Unlike Kant, however, Maimon suggests that we should not treat these elements like chemical elements, whose inner structures explain the epistemological properties of the instances of cognition of which they are ingredients. Rather, he proposes to adopt the paradigm of calculus which helps him to conceive of the relation between the forms and the content of cognition as analogous to the relation between functions and their differentials. Maimon’s idea is that the sensible content of cognition is the product of a conceptual synthesis, just as a differentiable function is the result of an integration of its differential. He takes this to entail what I shall call a “Homogeneity thesis”, instead of Kant’s Heterogeneity thesis, since those things which affect the mind

The analogy to calculus

135

must not be conceived as external metaphysical entities. Rather, he argues for an immanent model of cognition, and this means that he argues for an account which explains the form and content of sensible objects as homogenous in some sense. More precisely, he explains cognition —as John McDowell would say—as “conceptual all the way down”. In a first step, Maimon argues that we have no good reason to infer from the passivity of perceptive content to effects from external objects. Although Maimon does agree that sensible content can be described as that which is passive in cognition, he denies that we should understand this passivity in the referential sense.61 He states that a case in which something appears as “given” to our cognition “signifies only this: a representation that arises in us in an unknown [unbekannt] way” (Tr, GW II, 203). So, the given manifold of sensation which is said to constitute the content of our cognition should not be explained as the effect of something external and outside of us, but rather is simply a modification of consciousness, which arises according to a rule, but one that we are not aware of. As content of cognition, these modifications which arise in an unknown way form the objects of our cognition. Thus, Maimon explains that finite cognition must be explained as consisting in the application of the rules of understanding to these “infinitely small elements of every sensible intuition, which provides the matter to explain the way that objects arise” (Tr, GW II, 82). Differentials are the qualities that make up the matter of cognition; they are “the infinitely small of every sensible intuition and its forms, which provides the matter to explain the way that objects arise” (Tr, GW II, 82 [emphasis added]). Yet, it would be absurd if these elements of cognition were also produced by the finite cognising mind. Indeed, Maimon explicitly states that “these forms [the forms of human understanding] serve only to connect objects but not to produce them” (Tr, GW II, 212). This is where the conception of an infinite intellect comes in. In a second step, Maimon turns to the model of an infinite understanding, which he employs in order to explain how apparently passive sensations of external things in fact arise from intellectual, conceptual activity. He proposes that “we assume an infinite understanding (at 61

See Thielke (2003, p. 113).

136

Maimon’s method of fictions

least as idea), for which the forms are at the same time the objects of thought, or that produces out of itself all possible kinds of connections and relations of things (the ideas). Our understanding is just the same, only in a limited way” (Tr, GW II, 64-65). Thus, our apparent passivity is explained through the subconscious activity of an infinite understanding [unendlicher Verstand], for which “everything is in itself fully determined because it thinks all possible real relations between the ideas as their principles” (ETP 86n.1.). Maimon proposes that “what belongs to sensation must be ordered in relations if it is to be perceived, even if I cannot directly perceive these relations” (Tr, GW II, 205). If we posit that this infinite understanding, whose purely relational activity generates the rules from which the manifold of sensations arises, differs from our understanding only in degree, we no longer have to assume a second source of cognition, but rather sensibility and understanding are explained as flowing from the same cognitive source. Consider the following discussion, in which Maimon imagines a model of divine cognition: God, as an infinite power of representation [Vorstellungskraft], from all eternity, thinks all possible essences, i.e., he thinks himself as limited in every possible way. He does not think discursively as we do, rather, his thoughts are simultaneously presentations [Darstellungen]. Should someone object that we do not have a concept of such a way of thinking, my answer is: we do in fact have a concept of it, since we ourselves possess it [i.e., this way on thinking] in part. (Strf, GW IV, 42 [emphasis added])

The divine intellect doesn’t need a theory of representation because, for him, the act of cognising an object is identical with the act of generating it.62 Human cognition, however, is dependent on its objects being given to it. As we have seen, these objects are somehow given to it as determinate values, and must be represented according to a system that is relativised to the finite cogniser and her limited resources. Moreover, Maimon believes that “indeed, God thinks real objects, not only according to the principle of contradiction praised so highly in our 62 This idea of an intellect that “creates all things in knowing them, for the definition or thought of an object is at the same time the law of its generation”, stems from Maimon’s engagement with Maimonides and the latter’s Guide of the Perplexed (see Nisenbaum (2018, p. 99)).

The analogy to calculus

137

philosophy, but as we do (although in less perfect manner) when we think the objects of mathematics, i.e., through thinking these objects, he simultaneously brings them about” (Strf, GW IV, 42). Since the divine intellect’s cognition does not simply consist in purely logical thinking that relates to all possible essences, but is creative through and through, Maimon also argues that “our cognition relates to objects of mathematics”, like infinite cognition “to all objects of nature as such” (Strf, GW IV, 629).63 The infinite intellect thinks all possible things (or: “worlds” (Strf, GW IV, 58)), and by thinking all things, determines the laws that govern their relations.64 This thinking qualifies as real thinking because it simultaneously generates its objects. More precisely, it is in virtue of this act of thinking all of the relations between all possible objects that the infinite intellect simultaneously generates all objects as completely determined.65 The mathematical analogue for a pure thinking of relations that goes hand in hand with a generation of its object is “genuine algebra”. Maimon defines “[t]he object of pure arithmetic [as] number, whose form is pure time as a concept” (Tr, GW II, 22). Further, he defines numbers as pure ratios (Tr, GW II, 69). Any number (where n > 1) is defined and generated through the ratio n:1—for example, the number 666 is defined and generated through the ratio 666:1.66 Numbers, as the objects of pure algebra, are thereby generated from relations between them as sequentially-ordered magnitudes. Finite understanding can 63 He also states that “God generates the objects of nature in the same way than we generate the objects of mathematics through real thinking, i.e., through construction” (Strf, GW IV, 58). It is important to remember at this point that Maimon explains construction via analytic geometry. 64 Engstler (1990, pp. 152-156) argues that we can understand Maimon’s conception of infinite mind as analogous to that of Leibniz and Spinoza. Maimon explicitly gives his definition of infinite intellect with reference to “Leibniz’ system” and “if the Leibnizian does not want to admit it, it shall be called Spinoza’s system” (Strf, GW IV, 58). 65 “The infinite mind comprehends the real object as it is, as something thought and nothing more, as something identical with the concept, just as the intelligible is identical with the intelligens. In this sense Maimon, as we will see further on, uses the term ‘real’ (reell).” (Bergman (1967, p. 32). 66 See Lachterman (1992, pp. 503-511) for further illuminations on Maimon’s model of algebra.

138

Maimon’s method of fictions

generate these mathematical objects according to its pure forms of difference (i.e., time) without having to refer to individual representations, i.e., “[e]numerated sets of (sensible) objects”, since with this domain of mathematical objects, “the pure relation n:1 suffices to define, and, indeed to generate any [of these objects], i.e., number” (Lachterman, 1992, p. 503).67 To “assume an infinite understanding (at least as idea)” then means to assume the idea of an understanding “for which the forms are at the same time objects of thought, or that produces out of itself all possible kinds of connections and relations of things (the ideas)” (Tr, GW II, 64-65). Divine cognition, which “lets it [i.e., the given] arise [lässt entstehen] in accordance with these rules [i.e., the rules of understanding]”, embodies a kind of cognition that simultaneously defines and generates its objects through purely conceptual relations. Maimon contends that this model of cognition is the “only way” to explain a priori cognition in such a way that it “answer[s] the question quid juris? in a wholly satisfactory way)” (Tr, GW II, 81).68 As a consequence, Maimon takes mathematical knowledge to be a proof of the fact that “we are similar to God” (ibid.). Thus, our cognition must not be seen as categorically, but only gradually, different from that of an infinite intellect.69 “[T]he understanding can and must be considered in two opposed ways[:] 1) [a]s an absolute understanding (unlimited by sensibility and its laws)[; and] 2) [a]s our understanding, in accordance with its limitation” (Tr, GW II, 226-7), and it is only by entertaining the perspective of an infinite intellect that we can genuinely 67 This is also reminiscent of Spinoza’s account of intuitive knowledge. According to Nassar (2013b), intuitive knowledge is exemplified through the mathematician’s knowledge, who “sees the idea that underlies and determines the numbers and their relations. The idea, therefore, is not an abstraction, but is immanently realized in numerical relations; the relations are singular manifestations of the idea” (2013b, p. 245). 68 Other passages include: “[T]he quid juris? is easily explained according to my theory because the elements of appearances […] are not themselves appearances” (Tr, GW II, 193). And: “The metaphysically infinitely small is real because quality can certainly be considered in itself abstracted from all quantity. This way of considering it is also useful for resolving the question, quid juris?” (Tr, GW II, 355). 69 Maimon states that “our understanding is just the same, only in a limited way” (Tr, GW II, 65).

The analogy to calculus

139

explain the possibility of infinite intelligibility and object cognition.70 Maimon defines infinite intellect as an understanding that can “produce objects out of itself according to its self-prescribed rules or conditions without needing to be given something from elsewhere” (Tr, GW II, 63).71 Maimon argues that “this is not the case” for our finite intellect, since “the objects that are subject to its rules and conditions must be given from elsewhere” (ibid.), as our understanding is limited. Humans are rational, but finite, cognisers; we are not in possession of the complete concepts or representations of objects, and so these objects do not appear as intelligibly generated according to rules, but as (to some degree at least) mind-independent objects that are given to a ‘passive’ mind. It is for this reason that the quid juris arises for finite minds72 at all. However, if philosophers want to explain how there can be infinite intelligibility and complete explanation, they can only do so by positing the idea of an infinite intellect (Tr, GW II, 85).73 This is so because it is only if we treat the object of science as if we were in the position to provide the 70

In this reconstruction, I follow Engstler’s account (1990, pp. 143-165), which also appears in Bergman (1967), although I will ultimately disagree with this reading of Maimon’s idea of an infinite intellect, as he does not take into account Maimon’s method of fictions and its consequences for his conception of mathematics, and hence also misrepresents his conception of differentials and the way in which the constructive activity of the infinite intellect should be conceived. 71 As he puts it, “the objective order (of an unlimited cognitive faculty) is: 1) ideas of the understanding (in this case there is no sensibility and no intuition, but only the representations of all possible things; 2) concepts of the understanding (these connect these ideas in a unity of apperception); 3) ideas of reason (the representation of this cognitive faculty itself, as absolute substance, highest cause, etc.” (Tr, GW II, 376). 72 Maimon’s statements regarding the isomorphism between finite and infinite understanding should not be understood as comparing finite and infinite rationality, but rather the same rationality in a finite—and thus ‘constrained’—being and in an infinite being (II, 65). That is, although our understanding, in its capacity as rational, has the potential to produce its own objects, it cannot actualise this potentiality because it is constrained. Mind is cognising and producing, and it is in its productiveness that it is similar to God (Strf, GW IV, 42, 58; IV, 629). 73 For a discussion on whether we should interpret Maimon’s infinite intellect as a regulative idea or as a metaphysical reality, see Socher (2006, p. 96ff.)

140

Maimon’s method of fictions

complete chain of its conditions that we claim to be truly seeking its explanation.74 And the only perspective from which this possibility is truly intelligible is from that of an infinite intellect. But how can this model explain the representation and recognition of spatiotemporal objects from the human cognitive perspective? Maimon explains that: [s]ensibility […] provides the differentials to a determined consciousness; out of them, the imagination produces a finite (determined) object of intuition; out of the relations of these different differentials, which are its objects, the understanding produces the relation of the sensible objects arising from them. The differentials of objects are the so-called noumena; but the objects themselves arising from them are the phenomena. (Tr, GW II, 31-32)

Like calculus, finite cognition treats the smallest elements of cognition like differentials. Maimon proposes that these “infinitely small” elements are given to finite cognition from the infinite intellect that produces them through purely relational activity. In saying that finite cognition can determine objects in an a priori manner, Maimon assumes that our forms of cognition are applied to the differentials as the elements of intuitions, which are at the same time the real objects determined by the infinite understanding. The objects of intuition (but not of cognition per se), however, are not produced through the understanding, but through the activity of the power of our imagination. The imagination generates sensible representations, i.e., sensible images, through ordering the manifold given to it according to the (conceptual) forms of space and time.75 Following Maimon, the “spatiotemporal determination of appearances [which arise from this process] cannot serve as schemata of the categories because they do not represent objective relations” (Engstler, 1990, p. 164). Instead, the understanding applies 74 “But my question is: how is it comprehensible? (quid juris? for me means the same as quid rationis? because what is justified [rechtmässig] is what is legitimate [gesetzmässig], and with respect to thought, something is justified if it conforms to the laws of thought or reason). […] I merely ask: what sort of hypothesis must I adopt for it to be comprehensible?” (Tr, GW II, 363 [emphasis added in bold]) 75 “Consciousness arises through an activity of the faculty of thought. But in the reception of individual sensible representations this faculty is merely passive.” (Tr, GW II, 29 [emphasis added]).

The analogy to calculus

141

its forms to the “relation of these different differentials”, and not to the relation of the objects of intuition. This can be illustrated through Maimon’s example of the triangle. Through this, he aims to show that if we reduce any extensive magnitude such as a triangle to its differentials, then we can still think of it as existing because of the relation that remains the same, no matter how small the magnitude. Maimon says that if we think of a triangle, one of whose sides moves in the direction of the angle lying opposite to itself, in such a manner that the side constantly remains parallel to itself, and do so until the triangle becomes smaller and smaller ad infinitum (differential), we find that the extensive magnitude has ceased … but the relations of the sides of the triangle always remain the same … In this way, the intensive magnitude (the quality of the quantity) becomes the differential of the extensive, and the extensive of the integral of the intensive. Quality abstracted from all extensive quantity can, nevertheless, be thought in a quantitative relation … This relation does not exist among the lines insofar as they are measured but insofar as they are determined qualitatively. (As quoted in Bergman, 1967, p. 60)

Maimon imagines a triangle, whose extensive magnitude is infinitely reduced while its intensive magnitude, i.e., its quality, remains the same. His point in this comparison is that even if we constantly move the one side in the direction of the angle, and thereby infinitely minimise the size of the triangular shape, we must admit that the relation between the sides, as well as the angles they enclose, remain the same. While continuing this exercise to infinity, Maimon argues, we begin to realise that this relation between magnitudes, which is the triangle, does not exist “insofar as it is measured but insofar as it is determined qualitatively”. Hence his point that we can think of “intensive magnitudes within extensive magnitude relations”. We thus get a model of how we can determine objects by virtue of their qualities (which are determined in virtue of their a priori relations to all other qualities), even though we encounter these objects as quantities. In the same way, the elements or differentials of cognition, as purely conceptual relations, only appear to us through the objects of intuition which the imagination accordingly produces.76 76 “[S]ensibility provides the differentials to a determined consciousness; the imagination produces from these finite (determinate) objects of intuition; the

142

Maimon’s method of fictions

This is what Maimon means in saying that “taken in its strictest sense, absolutely a priori cognition [Erkenntnis a priori im engsten Verstand und absolut betrachtet] is the cognition of a relation between objects that is prior to the cognition of the objects themselves between which this relation is found” (Tr, GW II, 169 [emphasis added]). Differential and integral calculus provides a model for how the gap between extensive and intensive magnitudes can be bridged without illegitimately going beyond the level of extended magnitudes (“quality abstracted from all extensive quantity can be thought in a quantitative relation”). Whereas the analogy to differential calculus shows us how to move from the sensible to the supersensible, integral calculus teaches us how to make the inverse transition, that is, how to move from the supersensible to the sensible. The first method is analogous to the way in which we arrive at our empirical concepts; the second method aims to explain how our mind produces spatiotemporal intuitions, or appearances. Maimon’s views on the role of imagination can be better understood by turning to some of his remarks about the general nature of fictions in his Philosophisches Wörterbuch. There, he explains that, because of its finite nature, human cognition must in fact always involve fictionalisation (PhWb, GW III, 60). As finite cognisers, we only possess abstract and universal concepts, by which we can determine the objects of our cognition partially, but not completely. For an object to be presented to our consciousness, however, it must be a determinate object. This is why empirical objects have to be determined further through the forms of space and time, by virtue of which they are individuated as particular objects. On Maimon’s model, finite cognition can only individuate and differentiate objects through an operation of the imagination. Imagination employs the forms of space and time to produce particular and sensible images of the conceptual differences that are thought as determinate by the infinite understanding. Thus, by employing spatiotemporal representation, finite cognisers map incomplete conceptual differences on to another order of differentiation in order to enable absolute deunderstanding produces from the relations of these different differentials, which are its objects, the relation from which arise sensible objects” (Tr, GW II, 32).

The analogy to calculus

143

termination and individuation according to spatiotemporal location. In doing so, it becomes possible, for example, to differentiate between two identical raindrops, namely as ‘being-apart’, although we do not know of their conceptual and thus real difference. Since imagination is responsible for the operations by which we, as limited cognisers, attribute certain properties to objects which in themselves these objects don’t possess, Maimon describes the faculty of imagination as the “faculty of fictions” (Tr, GW II, 19). We will return to this point when discussing the nature of philosophical fictions. So, throughout the Essay, Maimon repeatedly states that quid juris can only be answered through endorsing the method of differentials and the idea of an infinite mind to model an account of cognition. With this endorsement, the possibility of a priori cognition becomes explicable, from the perspective of human cognition, as continuous to infinite cognition. He posits a first principle (i.e., the infinite intellect as pure generative relationality) and uses it to explain the possibility of a priori cognition (i.e., finite cognition treats these ultimate qualities as differentials). Rational dogmatists assert that things are not just thus and so, but that they are thus and so for a reason that can (at least in principle) be made explicit. On Maimon’s view, philosophy must approach science with its own standards and principles. That is to say, if metaphilosophy wants to argue for a particular conception of scientific explanation, and a model of cognition which accounts for it, it must take a normative position. Philosophy can suggest that if we want to conceive of the possibility of a scientific explanation that assumes necessary and universal connections between things, then we must assume infinite intelligibility.77 And to articulate the conditions of possibility of this standard means nothing other than to articulate the corresponding model of cognition, according to which cognition which achieves this standard of genuine explanation is shown to be possible. And the only model which can accommodate the possibility of explanatory rationalism, Maimon argues, 77 In fact, I take this to (at least partially) explain Maimon’s underlying motivation for publishing texts such as the Ankündigung und Aufforderung zu einer allgemeinen Revision der Wissenschaften (1792).

144

Maimon’s method of fictions

is one in which not only the form, but also the content, of cognition is determined a priori. On his alternative model, the a priori cognition of objects must be explained according to the Homogeneity thesis, which, on Maimon’s view, is the only thesis that can simultaneously license us in positing the Cooperation requirement. Although he accepts Kant’s demand to articulate the epistemic conditions of specifically human cognition, he contends that these cognitive elements must be homogenous for us to be justified in claiming their cooperation in a priori cognition. Section IV established that Maimon’s theory of differentials provides the right resources to explain the possibility of a priori cognition. However, this description also comes at a cost, namely that of empirical scepticism concerning the experience of any particular spatiotemporal objects of science. The theory of differentials, and its postulation of an infinite intellect, yield a description under which we are entitled to explanations that meet rationalist standards; yet, it cannot ground this description on immanent principles, nor prove its actuality through the facts of cognition, or of science. Maimon is committed to being an empirical sceptic concerning the actual application of his account to particular objects because these can only be given to consciousness as determinate spatiotemporal magnitudes, and their rules of generation can only be derived as relatively probable laws with regard to their relations to other determinate magnitudes. Although he understands this order of intelligibility as mapped on to the order of infinite intelligibility, its lack of completeness makes judgements of a necessary and universal manner impossible. Since the standard of complete explanation within a framework of infinite intelligibility calls for an explanation that is grounded in an absolute reason, and not a probable one, the new philosophy must depart from both a rational dogmatist position—insofar as it maintains a rationalist standard of intelligibility and explanation—and an empirical sceptical position—insofar as it maintains an empirical standard towards the intelligibility and explicability of any particular object. Following from this, Maimon can only explain the possibility of a priori cognition if he can point to a method by virtue of which philosophy can act as governed by both positions at once.

The method of fictions

145

Now, this is the point from which it can finally be explained what this mathematical method has to do with the possibility of metaphysics and the choice of an appropriate method for this new philosophy. Maimon cares not only about the theory of differential calculus and the ways in which it conceives of qualities as smallest quantities, but moreover he is interested in the methodological tools which allow mathematicians to do. Remember that his metaphilosophy is only possible as “science of the limits of appearances”; it must establish those a priori representations in virtue of which “a complete cognition of appearances” is possible, while respecting that it can never prove their objective validity through a complete explanation. Nevertheless, metaphilosophy can show which elements of cognition, i.e., epistemic conditions, are required in order for us to represent and cognise objects, given that these objects must be conceived as intelligible without limits and susceptible to complete explanation. On Maimon’s analysis, theories and practice of calculus are grounded in the use of a specific scientific tool, which we encountered in the introduction under the umbrella term of Galilean idealisations. For Maimon, it is because of “the mathematical method of fictions” that calculus can determine its objects as discussed (Strf, GW IV, 51). It is thus this same methodological procedure which enables the possibility of this science. And, as indicated earlier, it will be this procedure with which Maimon also identifies his metaphilosophy. His key metaphilosophical claim will be this: metaphysics is only possible as proper science if it makes use of fictions.

2.5 The method of fictions Through his study of calculus, Maimon notes that the very thing which instructs us how to treat spatiotemporal shapes, namely infinitesimals, involves a contradiction—after all, infinitesimals are at once indivisible (as an infinitely small point of difference) and ‘building blocks’ of a spatial magnitude that is infinitely divisible. But it is exactly by treating geometrical figures as if they were made up by these contradictory entities that we that we can define and calculate geometrical shapes through the relations between infinitely small changes of x-rates and

146

Maimon’s method of fictions

y-rates. Through this observation, it becomes clear to Maimon that mathematics, and also other sciences, depend on the use of fictions. And it is from this insight into the concepts and methods of mathematical practice that Maimon begins to develop a methodological solution to the task which his new metaphysic faces. Maimon claims that the definition of the conceptual rules that determine the generation of spatiotemporal representations must involve the use of a fiction. This fiction consists in the articulation of a conceptual representation which cannot be true, simply because it predicates of something that it is both A and not-A, e.g., magnitudes that are at the same time divisible and indivisible. According to the concept of the infinitesimal we should treat extensive magnitudes as if they really were composed of infinitely small quantities which ultimately resolve into intensive magnitudes, even though we also hold extensive magnitudes to be infinitely divisible. In a letter to Reinhold, Maimon voices the opinion that all “principles of higher mathematics […] are mere fictions” (Strf, GW IV, 224). What makes these principles successful tools for scientific inquiry is exactly that they are treated as mere fictions or methods, instead of true representations of the world. Maimon stresses that “the method of indivisibles, the infinite series, the differential calculus and such like necessarily lead to contradiction if they are considered to be more than mere methods. […] [R]eason […] declares them to be what they really are: mere fictions” (Logik, GW V, 263-64). Now, if the analogy to calculus serves to determine an adequate model of human cognition, it only makes sense to assume that the procedure of metaphilosophy can be explained via a methodological analogy to calculus. Indeed, this is exactly what Maimon suggests: “[T]here is another method, whose reality in philosophy I will assume only problematically […]. This method I will call the method of fictions, of which the mathematician has made use with the greatest success” (Strf, GW IV, 39). Like Kant, Maimon therefore develops his methodological solution in order to transform metaphysics into a science through an in-depth engagement with the methods and practice of the mathematical and natural sciences of his day. Maimon is by no means the first philosopher to discover this use of fictions in science: on the contrary, he joins the ranks of a whole group of rationalist philosophers, perhaps

The method of fictions

147

most famously Leibniz and Spinoza, who recognise the existence of such fictions.78 Unlike his predecessors, however, he delivers a detailed analysis and explanation of the nature and function of fictions in science, which culminate in his appropriation of the method of fictions for the purposes of philosophy itself. In order to understand how this method makes possible his concomitant commitment to rational dogmatism and empirical scepticism, let us briefly explore how Maimon defines scientific fictions and their role within scientific inquiry. By scientific fictions, Maimon identifies a specific class of representations which are used in the mathematical and natural sciences of his time. As can be inferred from its name, the method of fictions operates by means of representations (e.g., theories, models, or other theoretical instruments), which somehow deviate from or deliberately distort reality. In contrast to the “science fiction” of faster-than-light travel and endless interstellar war, these fictions constitute what Maimon characterises as “useful fictions [nützliche Fikzionen]” (Strf, GW IV, 78). He distinguishes this type of fiction from others by virtue of (i) the specific ways in which they misrepresent (ii) their epistemic or cognitive role. By noting that mathematics, as well as its application to the natural sciences, makes use of idealisations, Maimon focuses on what I discussed earlier in my discussion of Galilean idealisations.79 The main goal of 78 Maimon on multiple occasions (maybe most explicitly in Strf, GW IV, 5154) connects his theory of fictions to Leibniz and the latter’s treatment of infinitesimals as fictions. Leibniz sometimes, but not always (see Sherry & Katz 2013), ascribes fictional status to infinitesimals, e.g., “For I consider infinitesimal quantities to be useful fictions” (PE, 230); in other places he also calls them “well-founded fictions” (GM IV, 110) or “mental fictions” (G 2, 305). Such fictions are non-real, that is, they are ideal or imaginary entities “abbreviating thought and aiding in discovery, and they cannot lead to error” (GP 2, 305). For discussions of Leibniz’ fictionalism concerning calculus, see Levey (2008) and Jesseph (2008). Besides mathematical fictions, Leibniz expresses a similar attitude to settling the dispute between a heliocentric and geocentric hypothesis: “with respect to the very things we need to deal with, one hypothesis might be more intelligible than another and more appropriate for a given purpose”, “for a hypothesis to be true is just for it to be used purposely” (PE, 91). This is much closer to Maimon’s conception of useful fictions, yet it still gives no firm ground for Maimon’s substantive claim that the Leibnizian system of monads is to be given a fictionalist reading (cf. e.g., Strf, GW IV, 53). 79 Interestingly, there exists a related debate in contemporary philosophy of

148

Maimon’s method of fictions

Galilean idealisations is to create simplified models of their targets in order to make them computationally controllable. Furthermore, they are set up with the expectation of a future process of de-idealisation which approximates these systems in order to accurately represent the ‘real-world’ target-system.80 Maimon’s analysis of fictions in science focuses on a subtype of fiction that McMullin characterises as ‘construct’ fictions, i.e., fictions which consist in a simplification or distortion of the conceptual structure used to represent some target-system (instead of causal idealisations that ‘distort’ real-world situations). Maimon further analyses these practices of idealisation by differentiating them into two subtypes: the self-contradictory fictions of mathematics and the consistent fictions of the mathematised natural sciences. Mathematical idealisations fall under the first category because they involve contradictions. To use a familiar example, Maimon explains that the method of indivisibles “treats a continuous (and hence infinitely divisible) magnitude as if it was composed of indivisible parts (a line as composed of points, a plane as composed of lines, a figure as composed of planes)”, and uses the known ratio between the indivisible parts to determine the ratio of the magnitudes they compose (Strf, GW IV, 51 [emphasis added]). By decomposing geometrical magnitudes into infinitely small elements that are of the same dimension as the magnitudes themselves, the mathematician treats a surface “just as if [it] were composed of lines” (ibid.), thereby gaining a method for determining the ratio between different magnitudes. As we saw, Maimon takes this representation to involve a contradiction in itself since infinitesimals science, which has only recently been reanimated (see e.g., Fine’s (1993) article), and which can be traced back to Kant and Maimon. For a critical collection of essays on the topic of fictions in science, see Suárez (2008). This debate was sparked by Fine’s essay about fictions as they were discussed in Vaihinger’s Philosophy of As If (1935). The latter’s extensive account of the genesis and function of fictions (see e.g., Suárez 2008, p. 3), traces the origin of this theoretical tool back to Kant’s regulative ideas and Maimon’s theory of fictions. 80 This is somewhat equivalent to Vaihinger’s distinction between real fictions and semi-fictions. While semi-fictions are fictions that can be assimilated into reality through processes of approximation or de-idealisation, real fictions, by virtue of their contradictory nature, make such an endeavour impossible (1935, p. 16).

The method of fictions

149

constitute the indivisible constituents of an infinitely divisible extensive magnitude. Now, although the method of indivisibles is based on a contradictory concept, it can serve to determine mathematical facts such as the surface of a geometrical figure, or the continuous rate of change in a curve (which in turn helps in predicting, for example, the actual growth rate of some bacteria cultures).81 Maimon also introduces other types of useful fictions, which I summarise under the category of “consistent fictions”. These are applied in physics and other natural sciences. One of his examples concerns the “method of interpolation”, used by astronomers to determine, on the basis of several discrete observations of some celestial body, its position during periods in which it was not (and indeed could not have been) observed.82 He compares this method of interpolating with Leibniz’ doctrine of petites perceptions (Strf, GW IV, 52). Maimon says that “it must not mean: the soul still has representations during sleep, but: one must, according to the law of continuity, interpolate the representations, i.e., assume that if man could make observations about himself during sleep, he would find these representations in himself; and the same may be the case with innate ideas. More or less just as one can determine backwards the position of the planets before the origin of the world” (Strf, GW IV, 52-53). Again, Maimon explains useful fictions as a type of misrepresentation, this time not because it would in itself be contradictory, but because it is impossible to determine its truth value from empirical observation alone. Now, let me move on to the second feature which characterises scientific fictions as such, namely their epistemic role. According to Maimon’s view, it is this role which renders scientific fictions “scientific”. Contrary to other fictions, scientific fictions are useful or expedient 81 On Maimon’s engagement with differential calculus and its use in geometry and physics, see Duffy (2014) and Pringé (2018). 82 Consider another example: “An astronomer cannot, as such, be expected to determine whether the run revolves around the earth or the earth revolves around the sun; and if he assumes the latter, he does so not because of the objective truth of this hypothesis, considered in itself, but because only on this assumption is a system of the universe possible.” (Logik, GW V, 35-36 [emphasis added]).

150

Maimon’s method of fictions

to the scientific enterprise. They are instruments or means “for the discovery of new truths” (Strf, GW IV, 53) because they can “serve to determine something pertaining to a real object” (Strf, GW IV, 39).83 Maimon contents that, as purposeful misrepresentations, their function is not to “mirror” the world, that is, to be true (or in the bad case: false) representations of the world. Rather, scientific fictions serve as idealisations within other types of model-system, from whose construction and manipulation we can still learn about the world. In that sense, scientific fictions only really misrepresent when “they are considered to be more than mere methods”, that is, when their function is mistaken to consist in accurate representation (Logik, GW V, 263). What does it mean that scientific fictions serve as model-systems, and thereby help to discover new truths? Roughly, the idea is that: first, a representational system is modelled in order to determine X in accordance with some fictional characterisation Y, e.g., celestial bodies at unobservable positions are treated as if they were governed by the same laws as celestial bodies in observable positions. Through constructing a simplified model-system, it becomes possible to investigate other Yrelated features that X would instantiate if it really did instantiate Y, e.g., the unobservable positions of celestial bodies. Second, this hypothetical model and its properties and behaviours are then projected on to the target-system, which must bear some sort of similarity or resemblance to the model-system, e.g., the laws of movement of celestial bodies (cf. Camp, 2020, pp. 313-317). Ideally, then, by learning about the modelsystem, we also learn about the target-system. Scientific fictions must enter the process of scientific knowledge production as conscious misrepresentations of the world. In order for the fictional method to be successful, scientists must be aware that their theoretical tools employ fictions, and not mistake them for true representations of the world. There are several ways to know or become aware that one’s theoretical tools make use of fictions; Maimon seems to gesture to two in particular: (i) the fiction is in itself contradictory and therefore cannot be objectively real, e.g. infinitesimals, or (ii) there 83 Breazeale (2003; 2018) was the first to note Maimon’s specific conception of scientific representations as fictions.

The method of fictions

151

is no way of proving the objective reality of these theoretical concepts, e.g., interpolation. In that sense, scientific fictions are not simply misrepresentations, errors or falsities that are “opposed to the truth” and should be discovered and resolved (PhWb, GW III, 73). Rather, scientific fictions mark a class of representations that do not directly aim at truth. Although the function of scientific fictions is not to represent the world, they are treated and maintained as if that were the case, and are therefore used as means to arriving at the truth about the world.84 Owing to their nature, at least on most views,85 scientific fictions cannot in themselves be explanatory. The instances through which Maimon illustrates scientific fictions include models, theories, and other representations, which, although they are instrumental in arriving at true statements about the world, are not themselves veridical statements. Consider a sixteenth century realist about the Copernican hypothesis who uses the Ptolemaic hypothesis to predict the apparent motions of the planets. While he believes that the Ptolemaic hypothesis can “save the phenomena” through (more or less) adequate predictions, it cannot explain them, since it does not adequately represent the world. Explanations require an actual understanding of how the world really is. In order to explain a target system (explanandum) one needs a set of judgments that are true of the world (explanans).86 Scientific fictions, however, only serve an instrumental role within the process of scientific explanation; they function as theoretical tools on the supposition of which one arrives at the desired explanans. Now, this seems to be at tension with Maimon’s proposal to make 84 This is what distinguishes fictions from hypotheses: while hypotheses are tested for their appropriateness to represent target systems, scientific fictions do not “demand verification” because their function is not to represent the world (Vaihinger (1935, p. xlii). Even though this might point one toward pragmatist readings of Maimon’s philosophy (see e.g., Kuntze (1912, pp. 37677) or Vaihinger (ibid.)), we will see that Maimon endorses a form of explanatory rationalism and scientific realism that directly stands at odds with such readings. 85 See Fine (1993) and Vaihinger (1935, p. xv). 86 This is of course exactly what the use of idealisations and other fictions in science puts in question. See Cartwright (1983) for an argument on the ideality of the physical laws employed in physical explanations, or, more recently, Bokulich (2009) on the explanatory power of scientific fictions in general.

152

Maimon’s method of fictions

the method of fictions the new scientific procedure of theoretical philosophy. If philosophy can only become a proper science by virtue of a metaphilosophy which employs fictions, then, at least prima facie, it seems to be the case that philosophy cannot be an explanatory science. This, however, goes against the grain of what philosophy had traditionally been taken to be, namely an explanatory science that tries to answer ‘why-questions’, if anything at all. What’s more, Maimon tasks philosophy with the role of determining the possibility of a priori cognition, which he takes to entail showing that finite cognition of objects can, at least in principle, answer all why-questions. After all, philosophy as the science of the limits of appearances must provide the a priori principles of the generation of these limits in order to explain the possibility of the cognition of objects that are intelligible without limits. At the same time, this philosophy lacks the resources to prove an actual application of these principles to any particular object that appears to us. How do scientific fictions provide a methodological solution for this difficult situation? The answer to this question leads us back to Maimon’s dual position as a rational dogmatist and an empirical sceptic.

2.6 Philosophical fictions Before turning to the nature of philosophical fictions, it is important to recall that, for Maimon, the task of metaphilosophy has a clear definition: it must determine “the form of science as such” (Strf, GW IV, 35). More precisely, he explains that his “philosophy is the science of all sciences, through which they only ever acquire their status as sciences”: only “if the objects of nature are ordered philosophically under principles, are brought into a system, do they become a natural science proper” (Strf, GW IV, 34). Thus, what philosophical fictions seem to be useful for is the determination of “the form of science as such”. Yet, what exactly does it mean to say that fictions can order the objects of nature systematically under principles? In light of our previous discussion, we can assume that philosophical fictions will involve deliberate misrepresentations that are nevertheless expedient or instrumental for scientific inquiry. According to Maimon,

Philosophical fictions

153

“[philosophical fictions] have by no means the utility that fictions have in mathematics. In mathematics, differential calculus serves for the discovery of new truths; in contrast, one can, at most, presuppose that the Monadology provides an explanatory ground of natural appearances, without employing itself for that purpose.” (Strf, GW IV, 53) If fictions do not contain any contradictions, they can be used as principles of reason for the explanation and systematic ordering of cognition [Begründung und systematischer Ordnung der Erkenntnis]. (Logik, GW V, LIV, Annotation “u”)

With regard to the first quote, Maimon’s point is not to say that philosophical fictions have no utility compared to, say, mathematical fictions. Rather, he thinks that their epistemic role is different from that of useful fictions in other sciences. In these other sciences, fictions serve as methods by which we can generate new truths about the world, even though they cannot themselves be assigned truth values. Philosophical fictions, on the other hand, are not instrumental in arriving at new truths. Rather, they are instrumental in explaining why different kinds of objects or phenomena are of the right structure or form to be explicable by the sciences. This is what I take Maimon to be saying when he states that philosophical fictions provide “an explanatory ground of natural appearances”: philosophical fictions serve to explain why the objects of our cognition are such that they can be known and studied by the sciences. In that sense, philosophical fictions operate on a second-order, or transcendental, level of explanation. For Maimon, philosophy constructs models in order to determine whether the cognitions that ground our explanations on the first-order level of explanation, i.e. those of experience and phenomena, can be warranted, or not (Thielke 2014, 225-227). In other words, therefore, philosophical fictions state the epistemic conditions under which our judgments would be justified, and thereby take up the explanatory cognitive role that metaphysical laws, which governed the general form of natural appearances, used to occupy. Note that Maimon insists that his new philosophy is “as little concerned with the (metaphysical) truth of the principles from which it proceeds, as with the results to which it finally arrives, but merely with the suitability

154

Maimon’s method of fictions

of the principles as principles or the preservation of the highest possible unity of reason [Vernunfteinheit]” (Logik, GW V, XXV). Furthermore, unlike many mathematical fictions, philosophical fictions are not contradictory, but rather must be consistent. What distinguishes them from truth-apt representations is not that they cannot represent the world for logical reasons, i.e. because they are contradictory, but because their truth could only be proved by virtue of an infinite operation of representation. Although we require them as “principles of reason for the explanation and systematic ordering of cognition”, we cannot determine their truth values because they are the conditions of the possibility of the explanans, e.g., empirical judgments, which are required for scientific explanations. Philosophical fictions do not refer to possible objects of experience, and could only be proven through a determination of all objective facts. Since this goal constitutes an infinite task, Maimon defines philosophical fictions as hypothetical models which present the conceptual conditions under which any given science can achieve the highest systematic unity. More precisely, he says that a genuine philosopher cannot be expected to determine whether our cognition possesses a real ground outside our faculty of cognition, a ground that can be derived from this faculty (as the Critical philosophers maintain); and if he nevertheless adopts such a position, he does so only because this will produce the highest possible degree of systematic unity in our cognition, by means of which everything therein will be explicable in the most exact interconnection. The assertion of the dogmatists concerning things in themselves are, in contrast, quite superfluous, since nothing within our cognition can be explained thereby. (Logik, GW V, 35-36).

Thus, Maimon commits to the metaphilosophical view that the “explanatory ground of natural appearances” can only be posited as a theoretical tool, a “necessary assumption” that instructs us on how to treat natural appearances in general, namely as exhibiting some sort of necessary connections. Kant’s philosophy, for example, suggests treating appearances as standing in necessary causal relations. Although explanations in the natural sciences rely on particular observed phenomena, they have to take for granted the universal principles that ground any explanation of natural phenomena (BNO, GW IV, 533). That is, any scientific treatment of empirical phenomena implicitly relies on a theory of how phenomena in general are structured, for example as standing in

Philosophical fictions

155

tions of causal dependence.87 Whether certain forms of explanation, for instance causal explanations, are valid depends on their being grounded in a specific order of intelligibility, and in turn a specific order of grounding, e.g., physical grounding.88 While Kant subscribes to the view that the natural sciences depend on philosophy because the latter establishes those principles that are constitutive for the possibility of experience, and hence the structure of empirical objects, Maimon thinks that natural sciences depend on philosophy exactly in the absence of this constitutive relationship. We have seen above how Maimon tries to show that critical philosophy, as it stands, can neither answer the question quid juris? nor the question quid facti?. For one, critical philosophy can neither demonstrate by what right we apply the concept of cause, nor whether we do in fact apply the concept of cause.89 The reason for this defect of critical philosophy is grounded in Kant’s positing of the Heterogeneity thesis. However, after criticising the latter’s theoretical choices, Maimon also offers a charitable interpretation of the Critique by trying to conceive of it in terms of a fictional or hypothetical model.90 On his reading, transcendental philosophy should aim at determining the conceptual structure that is necessary in order to explain how we can predicate of objects of perception in general as conforming to certain categories, e.g., how we can judge them to be objects that enter 87 “All alterations occur in accordance with the law of the connection of cause and effect” (KrV, B232). 88 For Kant, empirical or special laws of nature have their status as laws in virtue of being grounded in some kind of first-order laws, which are the principles that structure anything that can ever become an object of experience. On the distinction between first- and second-order laws in nature, see Kitcher (1986). 89 “My skepticism thus is grounded in this two-horned dilemma. Either the factum is in itself (that we use the form of hypothetical judgments for empirical objects) false, and the cited examples rest on an illusion produced by imagination, as I have demonstrated multiple times, the categories then have no use; or it is in itself true [i.e., the factum], and then it does not have an understandable reason, and the categories after their arduous deduction and schematism, remain mere forms that cannot determine objects.” (Logik, GW V, 250) 90 Sometimes, it seems that Maimon pushes a more radical agenda that presents critical philosophy as not having presented any explanations. At the same time, however, he calls Kant’s system the “completest” to date (Strf, GW IV, 73).

156

Maimon’s method of fictions

causal relations. With regard to determining particular objects, however, the categories can only have a regulative function, since no matter how many singular instances of a conjunction of natural events we encounter, it will never ‘add up’ to an experience that has universal character. Therefore, Maimon argues that “the concept of cause is not a category, but an idea, which can be approximated through ever more complete induction but never reached” (BNO, GW IV, 442 [emphasis added]). Since an explanation of regularities in perception can just as well posit a priori structures as it can posit empirical laws of association, categories must be treated as ideals that singular judgments about empirical regularities instantiate. Each time we judge that event B was caused by event A, we not only approach the verification of some special law of nature, but we also simultaneously approach the verification of the more general, metaphysical ‘laws’ that are involved in making judgments of this kind. Thus understood, the continued process of scientific inquiry and explanation can only infinitely increase the probability of complexes of empirical judgments, and thereby infinitely increase the probability of their being actual expressions of an objective order of nature. On this view, scientific reasoning, then, does not culminate in fictions in order to stop the regress of grounds, but rather begins with fictional principles. Maimon’s own model is grounded in the metaphilosophical assumption that philosophical fictions should not explicate the models that are already implicit in scientific practice, but instead that philosophical science should take a normative stance on what kind of explanation should qualify as genuine. Consequently, he notes, his project calls for a revision of the established epistemic practices of science, not merely a new, adequate theoretical grounding. As a rational dogmatist, Maimon is committed to infinite intelligibility and complete explanation: the possibility of a priori cognition will have been demonstrated once it can be shown that our cognition can determine its objects in virtue of their sufficient reasons. This and nothing less can count as the standard for scientific reasoning and explanation. At the same time, Maimon is an empirical sceptic, who insists that no science dealing with sensible objects (i.e., phenomena) can give factual proof of its application of the forms of understanding (and thus of the objective necessity of its laws and theories). Reason cannot satisfy its

Philosophical fictions

157

demand for infinite intelligibility and complete explanation, in either the exact nor experimental sciences, as long as they are concerned with the realm of sensible objects.91 The only proper scientific procedure for philosophy must therefore consist in positing a fiction that conceives of the world in such a way that it becomes explicable while asserting the fictional status of this posit, and which would therefore accommodate both rational dogmatism and empirical scepticism. Maimon’s method of fictions is driven by his rationalism, which can only be articulated through philosophical fictions, which only lead to empirical scepticism in a second step. I shall propose that we can only come to understand all conditions of Maimon’s philosophical method if we adopt this interpretive strategy. I thereby counter dominant interpretations which take Maimon’s employment of philosophical fictions to be a consequence of his empirical scepticism.92 In contrast to this line of argumentation, I argue that we should understand philosophical fictions as both a result and an integral part of Maimon’s rationalism, as well as his idea of a rational science. I take it those previous misconceptions and confusions have stemmed from not adequately separating Maimon’s general theory of fictions from his account of scientific fictions, which embody two separate kinds of fiction: (i) fictions of the imagination and (ii) fictions of reason. Views such as those found in Breazeale’s influential studies garner some plausibility from close readings of what Maimon says about the general nature of fictions in his Philosophisches Wörterbuch. As explained in section III, Maimon argues for the view that human cognition always involves fictionalisation because of the role that imagination plays in the constitution of representation, i.e., in generating spatiotemporal images of conceptual differences. Since imagination is responsible for the operations by which we, as limited cognisers, attribute certain properties to objects which they do not really possess, Maimon describes the faculty of imagination as the “faculty of fictions” (Tr, GW II, 19). 91 The only exceptions are the science of logic and the science of pure numbers, i.e., algebra. 92 For example, see Zubersky (1925) and Breazeale (2002; 2003; 2018).

158

Maimon’s method of fictions

Unlike scientific fictions, Maimon seems to believe that fictions of the imagination constitute representations that aim at being true, and, if taken to represent things as they are in themselves, do constitute false representations. Yet, it is my contention that the kind of fiction that is expounded here is exactly not what Maimon is referring to when ascribing to philosophy the task of producing fictions. To get a better grip on the distinction between the two types of fiction at issue, it is helpful to consider Maimon’s reaction to criticism of his method of fictions. In a letter from 1794, he distinguishes between two kinds of fictions, the first of which are indeed products of the imagination, the second of which, however, he identifies as “the work of reason” (Logik, GW V, 35). As discussed above, fictions of reason do not pretend to represent what they are not a representation of; rather, they constitute a different kind of representation which also serves a different cognitive role. Unlike the imagination, which presents as objective what is in fact just subjective, reason “declares [these representations] to be what they really are: mere fictions” (Logik, GW V, 263-4). Using the example of an infinite series, Maimon explains that the problem with its representation does not consist in a general human inability to represent infinity. In fact, we can and do represent infinity without running into problems, for example in mathematics, where we can represent this series through a definite conceptual rule that progresses infinitely. Such representations only assume the character of false representations if they represent an infinite series and its last member as given. Thus, fictions of reason turn into fictions of the imagination only once they are “considered to be more than mere methods” (Logik, GW V, 263- 4), namely if imagination produces a sensible image of this infinity. In summary, Maimon makes clear that philosophical fictions constitute members of the class of fictions of reason. To wrap up, philosophical fictions of reason are useful fictions insofar as they are instrumental in determining the conditions of possibility under which rational explanation, and hence rational science, is possible. In that sense, they serve as conditions of possibility of taking the stance of a rational dogmatist within the human cognitive enterprise. Only if we treat the objects of science as if we were in a position to provide

Philosophical fictions

159

the complete chain of their conditions can we claim to be truly seeking their explanation. And the only perspective from which this possibility is truly intelligible is the provided by the useful fiction of an infinite intellect and its generation of the objects of the understanding: “the question is not, how far we can get in this endeavour but just from which perspective we should treat the object (of inquiry), in order to be able to judge about it in appropriate manner” (PhW III, 222). Since taking this perspective presupposes the a priori determination of objects that are “impossible” as objects of finite cognition, it demands adopting specific practices of idealisation. Yet the fictions employed are not products of the imagination because they are not taken to represent real objects but mere methods. They enable us to arrive at a model which specifies what it means for scientists to treat every object, event, or fact as if it were completely explicable, thus, under what conditions scientific judgments are justifiable. While theoretical philosophy is not, and cannot be, in possession of a fact that would prove the legitimate application of a priori forms to cognitive content, it is in possession of a model of cognition that shows the conditions under which we would be justified in applying a priori forms to cognitive content. To connect this to what I said about his criticisms of Kant at the beginning of this chapter: Maimon’s philosophical fictions provide a possible answer to the quid juris. While Kant subscribes to the view that the natural sciences depend on philosophy because the latter’s principles constitute the possibility of scientific experience, Maimon thinks that natural sciences depend on philosophy exactly in the absence of such a constitutive relationship. In delivering consistent conceptions of objectivity, philosophy delivers systems of principles that provide the conditions under which we would be justified in assuming that, e.g., necessary and universal connections obtain between sensible appearances. In doing so, philosophical model-systems yield possible descriptions of how it is that science can explain one thing in terms of another. And Maimon claims that it is only his system that can deliver a description that satisfies a rationalist standard of explanation. As opposed to Kantian regulative ideas, to draw an illustrative comparison, Maimon’s philosophical fictions do not emerge “from the bottom-up” to stop the infinite regress of grounds. Rather, they work from the “top-down”, to

160

Maimon’s method of fictions

show that it is possible to arrive at a model of finite cognition which can show rational inquiry to be legitimate. As a consequence, one crucial element—which could easily lead to an interpretation of Maimon’s philosophy as broadly Leibnizian or Spinozistic—of the latter’s theory of cognition must be reconceived: Maimon’s conception of God, or the divine intellect. For Maimon, to conceive of the possibility of a priori cognition meant to conceive cognition from a particular perspective, namely that of divine cognition. It now turns out that positing such a perspective entails positing a fictional standpoint. Although Maimon never calls the infinite intellect a fiction, he does call it an idea: “we assume an infinite understanding (at least as idea), for which the forms are at the same time objects of thought, or that produces out of himself all possible kinds of connections and relations of things (the ideas)” (Tr, GW II, 64-5).93 Although it is safe to assume that Maimon would never have called God a scientific fiction, he does employ the idea of infinite mind in a functionally equivalent way. The point is not that ‘God is a fiction’, but that positing an infinite intellect is a method of proceeding as if it the world could be cognised from a divine perspective, that is, as completely explicable.94 Postulating an infinite intellect does not explain the world itself, but it provides a method for how scientific practice is to treat its objects if it wants arrive at scientific explanations— it must treat them as infinitely intelligible and therefore that any explanation will point to further conditions or reasons that must be sought out until it arrives at an ultimate self-explanatory ground. What’s more, imagining the possibility of this absolute standpoint demands that we conceive a model of cognition that explains objects from their manner of origination, thus explaining them as produced (and not given), since this is the only way,

93

For other examples, see (Tr, GW II, 265) or Maimon (2000, pp. 53, 81). Another interpretation that also makes this connection is Kuntze’s opum magnum on Maimon (1912, p. 84). Although loosely connecting the latter’s concept of an infinite mind with the theory of fictions, he fails to elaborate on the distinctive grounding role that the infinite mind plays as scientific fiction, as well as the methodological reflections on fictions which stand in the background here. 94

Philosophical fictions

161

according to Maimon, in which an insight into how cognitive content arises is conceivable. Since Maimon’s philosophy is the science of the form of all sciences, its method of fictions has broader consequences for the image of science as such: his philosophy can only ever serve “for the determination of an object of nature in general but not for the determination of any particular object” (Strf, GW IV, 362). Although mathematics is capable of real thought insofar as it remains pure algebra, Maimon’s attempts to extend mathematisation to the realm of natural phenomena have failed. Construction, whether conceived from a Euclidean perspective or from the perspective of calculus, cannot serve to mathematise natural phenomena, since it will have to make use either of fictions in the one sense (i.e., intuitions), or in the other (i.e., contradictory fictions). Mathematics can determine particular objects of nature only insofar as they are determined as quantities. That is, mathematical equations determine only “the ratios between particular objects, which are generated through induction and spring from universalized empirical laws” (Strf, GW IV, 364-6). Yet, the empirical laws which mathematics “views as quantas” (Strf, GW IV, 366) are only discovered through “experiences, observations, and experiments”, or in short, “through experimental art” (IV, 365). Natural science, as “the science of the real ratios […] between empirical objects” (IV, 365), is concerned with “the particular causes of natural phenomena” (Strf, GW IV, 374), and can only produce laws and theories of comparative universality and necessity (Strf, GW IV, 373). On Maimon’s analysis, theoretical philosophy determines the form of all sciences in that it determines the form of objects of experience in general, and thus cannot establish with absolute certainty the form of the particular objects on which the natural sciences experiment. The comparatively universal and subjectively necessary principles articulate a systematic whole of the rules of generation of natural phenomena. A chain of reason that renders a phenomenon intelligible to the fullest degree, however, must be one that ends in a first unconditioned principle, from which all other principles, all ‘rules of production of the natural world’, as it were, are transparently generated. It is only by constructing a fictional model of divine cognition that human cognition can

162

Maimon’s method of fictions

be explained, and in turn that a completely (i.e., infinitely) intelligible production of all natural objects can be explained.95 When Maimon asks “what sort of hypothesis [he] must adopt for [object cognition] to be comprehensible?” (Tr, GW II, 363 [emphasis added]), he uses the term ‘hypothesis’ in a similar way to Kant, namely as the positing of a theoretical structure serving as best explanation, and not a proposition to be tested by empirical experiment.96 The infinite intellect is not a hypothesis that is tested for its truth; rather, it should be understood as a condition for hypothesis-testing in general. The hypothesis of an infinite intellect cannot be falsified, even if science produces a whole range of failing experiments—these can never show the inadequacy of the hypothesis, nor its falsity, but only these experiments’ inappropriateness for verifying this irrevocable hypothesis.97 The infinite intellect as a (fictional) first principle is a requisite for explaining the possibility of a priori cognition, given Maimon’s commitment to infinite intelligibility and complete explanation; from the perspective of Maimon’s overall picture of science, it is also a necessary requisite for the possibility of testing (i.e., verifying or falsifying) hypotheses within science as a systematic whole.98 In order for his account to not be potentially self-refuting, it must rely on an untestable hypothesis, the scientific fiction of an infinite intellect, to construct a hypothetical model under which his conception of philosophy and science can be described as possible. This allows him to both endorse a rationalist position while being 95 “Here, the question is not, how far we can get in this endeavour but just from which perspective we should to treat the object (of inquiry), in order to be able to judge about it in appropriate manner” (BNO, GW IV, 443). 96 Thus, Nisenbaum goes wrong in just reading ‘hypothesis’ in the simple sense of ‘being one possibility’ (2018, p. 101). 97 Here, we see why Maimon’s image of science is very far away from what is discussed in post-Popper and post-Kuhn philosophy of science. Maimon nowhere takes up the idea that scientific progress is built on failure and error in theories and experiment, nor that it undergoes radical paradigm changes. This is interesting, considering that some of his examples for scientific fictions exactly refer to failed theoretical models that still served some calculational function despite their inadequacy as representations (e.g., Strf, GW IV, 51). 98 We will see below and in the next chapter that Maimon therefore posits a precursor to Schelling’s “absolute hypothesis”.

Philosophical fictions

163

aware of its dogmatism, which he realises through his methodological conception of philosophy as producing scientific fictions.99 Finally, treating the phenomenal world as completely explicable also brings with it the idea of infinite progress in science: “reason thus demands an infinite progress by which the rationalization of the object is constantly increased” (PhWb, GW III, 193). As the infinite can only be thought of in terms of an infinite approximation in time, science can be described as an infinite, self-correcting progression of knowledgeproduction that approaches, but never realises, its perfection (Strf, GW IV, 55, BNO 443ff.).100 Science is conceived as a constant process of improvement within the bounds of a set of methods, or, to put it another way, within the bounds of a methodological paradigm. From this perspective, Maimon’s philosophy advocates for a research programme which, although (according to its own standards) it cannot prove the objective validity of its governing principles (quid facti), scientists should commit to in order to secure the possibility of scientific explanation at all: the idea of a unified science based on rational explanation (quid juris), “the acceptance (of which) involves a commitment to confront any future phenomena by means of the conceptual resources of his theory. It determines the terms in which we shall seek explanations” (Van Fraassen, 1980, p. 12). Even so, these terms are only fixed in a methodological sense: an explanation must be rational and complete, and must strive to produce intelligible facts. Postulating an infinite intellect does not explain the world itself, but it provides a method for how scientific practice is to 99 Breazeale (2002; 2018) fails to acknowledge the fact that, for Maimon, there is one particular philosophical fiction that is superior to all other philosophical fictions, for one because it does not just unify or systematise already existing cognitions (and thus qualify as philosophical fiction), but also because it is the only philosophical fiction that can serve as an explanatory ground for all natural appearances if one accepts a standard of complete explicability: the fiction of an infinite intellect generates the objects of cognition. 100 E.g. “The more complete [vollkommener] an intellect is, the more complete its presentations as compared to all other possible ones will become. An infinite intellect herein reaches the maximum of completeness. All others can approximate this completeness to a higher or lesser degree” (Strf, GW IV, 57 [emphasis added]).

164

Maimon’s method of fictions

treat its objects if it wants arrive at scientific explanations, namely as ultimately having rules of productions that can be articulated to completion. In contrast with Kant, Maimon does not think we should limit reason’s rational demands, but rather that it is only by fully articulating their interest that we can gain a secure foundation for explanation.101 It is the demand for genuine explanation, grounded in an order of infinite intelligibility, that renders the employment of fictions necessary, and it is through these fictions that this demand can be coherently articulated and implemented into the process of scientific inquiry. From here, we can also understand why philosophical fictions are not explanatory in the same sense: they deliver methods for describing the conditions under which a priori cognition becomes possible without claiming the apodictic necessity of these conditions. Maimon’s coalition system ultimately delivers a novel method for philosophy, through which it can generate methods for the sciences, which in turn explain how their method can be described in such a way that it is consistent with a framework of infinite intelligibility, i.e., a framework within which science is possible—that is, his metaphilosophy. In conclusion, it is because Maimon is a rational dogmatist, and not because he is an empirical sceptic, that he comes to advocate a metaphilosophy that proceeds on the basis of useful fictions. For Maimon, to answer the question of whether metaphysics is possible as a science is to determine whether science as such is possible, and, in turn, to determine whether science as such is possible is to determine whether philosophy can construct a model of cognition which shows rational inquiry to be possible. Philosophical fictions are the methodological means by which theoretical philosophy can produce hypothetical models which investigate the conditions under which particular explanations would be justified, i.e., under which complete explanation is possible.102 101 For these reasons, I do not agree with readings such as Bransen’s (1991), which solely emphasise Maimon’s scepticism and attribute a merely aporetic function to his philosophy. Maimon’s method of fictions offers a positive conception of how to do science as well as problematising its premises. 102 Nisenbaum argues for a similar view, stating that Maimon thinks that the “principal task of the philosopher is not to answer the question whether we are entitled to form empirical judgments making claims to universality and necessity,

Maimon’s metaphilosophy: philosophy as modelling practice

165

2.7 Maimon’s metaphilosophy: philosophy as modelling practice In the sloganeering of contemporary analytic philosophy, we might conclude that Maimon’s methodological solution for metaphilosophy-first advocates a kind of ‘modelling’ practice. Unlike Kant’s experiment of reason, his procedure establishes the conditions of philosophical reasoning not by positing a cognitive theory that is then tested and verified, but through the employment of scientific fictions which keep their hypothetical status while nevertheless guiding scientific inquiry in all of the sciences. To transform metaphysics into a proper science, metaphilosophy must adopt a method by which it can show the possibility of the science of the limits of appearances. A theory explaining this possibility will, according to Maimon, be a theory of cognition which is committed to a standard of rational explanation, and thus to the view that the objects of cognition are intelligible without limits. Maimon’s theory of cognition meets these demands by employing philosophical fictions. He contends that, in order to satisfy the rationalist commitment, his metaphilosophy must construct a cognitive model according to which complete explanation is possible, while at the same time postulating theoretical entities that must nevertheless be understood as fictional, i.e., whose function is not to represent, but rather to enable science to treat its objects in a rational manner. Philosophical fictions, per Maimon, constitute a type of modelsystem that are manipulated and investigated in order to make clear the possibilities and limits of genuine scientific explanation.103 Instead of constitutive principles, metaphilosophy can only offer model-systems to but to describe the overall framework in which we are entitled to do so, and this descriptive method pertains not to just this particular question, but to all philosophical problems” (2018, p. 101). 103 In that sense, I take Maimon to broadly agree with the metaphilosophical stance taken up by contemporary philosophers such as Godfrey-Smith (2006), Paul (2012), and Williamson (2017), in expressing the view that “[m]etaphysical system-building is model-building” (Godfrey-Smith, 2006, p. 6). On their shared view, some instances of philosophy should be understood as methodologically similar to the sciences insofar as they, too, operate as model-based science (see Godfrey-Smith (2006) and Weisberg (2007b)).

166

Maimon’s method of fictions

investigate which fundamental structures and appearances must figure as parts of genuine, i.e., rational, explanations. In delivering consistent conceptions of objectivity, philosophical fictions deliver systems of principles that provide the conditions given which we would be justified in assuming that necessary and universal connections obtain between sensible appearances. In doing so, these model-systems provide descriptions of how it is that science can explain one thing in terms of another. As I will now explain, the stance of Maimon’s metaphilosophy is therefore more radical than Kant’s proposals in some ways but less radical in others. First, it is more radical in the sense that it assigns to philosophy (correctly understood) the role of choosing a standard of intelligibility and explanation which it ought to impose on all other sciences as the standard for scientific explanation—to be scientific is to adhere to this norm. In contrast to Kant, Maimon does not believe that the sciences already prove the fact of an actual application of forms of understanding. Likewise, for Maimon, just because the sciences do assume one order of intelligibility (e.g., that of causal explanations), this does not mean that they rightly do so. In fact, Maimon is committed to the opposite: the sciences can only do right if they area also committed to the rationalist standard of infinite intelligibility which is entailed by his own position. Consequently, his metaphilosophy-first not only explains the possibility of philosophical science, but also sets the standard for what should count as scientific explanation more generally. Given his settled view on the possibility of proving the correctness of his theory, however, Maimon’s metaphilosophy is less radical than Kant’s insofar as it denies that metaphilosophy in particular has the capacity to do so. This is so because, in order to prove the correctness of his theoretical propositions concerning the limits of the appearances, Maimon would have to provide a complete explanation of all appearances. Since finite human cognition can only ever approach such an explanation, he has to maintain the problematic status of these premises (in their capacity as models) and the theory that interconnects them. Like Kant, Maimon considers more than one possible theory on the way to his conclusion; unlike the former, however, he does not claim that his experiment arrives at the only possible and true model of

Maimon’s metaphilosophy: philosophy as modelling practice

167

tion. Rather, Maimon’s experiment of reason tests different fictions that function as methods104 for determining the object of science in relation to explanatory completeness (e.g., those of Kant, Hume, Leibniz, etc.). Although he grants superiority to his own model in terms of its compatibility with a rationalist mode of explanation, he seems to leave open the possibility of other philosophical fictions (such as Leibniz’s system, “correctly understood”) that could take its place without causing any loss. Since Maimon not only uses his metaphilosophical conception in developing his own position but also in evaluating others, he also offers his own “history” of philosophical systems (Strf, GW IV, 3-56).105 Of these systems, he identifies a few which he takes to have already made implicit use of scientific fictions, e.g., Leibniz, Kant, and arguably also Spinoza (e.g., Tr, GW, II, 73, 338, 437, 563; Strf, GW IV, 51-54). What distinguishes these from Maimon’s own fiction is merely that their authors did not consider their systems to contain scientific fictions, or at did not explicitly label them as such.106 But Maimon argues that what unites all these fictions (i.e., implicit and explicit ones) is their systematic form. More precisely, systematicity is the formal feature which characterises scientific philosophical fictions.107 The purpose of any philosophical fiction, should it satisfy the standards of Maimon’s science, is to “subsume […] the highest possible manifold under the highest unity of principles by establishing the most complete systematic 104 Freudenthal seems to vaguely agree with my assessment: “But what if his philosophy is to be understood as a methodus indivisibilium, i.e., as a method and not dogmatic ontology?” (2010, p. 98). 105 He does so especially in his later Pragmatische Geschichte des Begriffs von Philosophie, und Beurtheilung der neuern Methode zu philosophiren (1797). 106 Closest is probably Kant’s treatment of regulative ideas. I treat this problem elsewhere (Schmid 2021a; 2021b). 107 As Maimon states in the Announcement (1792) for a “general revision of science”, “[t]he most universal form of any science is the form by virtue of which it is a science at all, namely the highest possible degree of unity of principles, and a systematic order of the subsumed manifold […]. This form is a demand of reason with regard to any object to be treated as object of science in general” (AARW, GW III, 340-1). Or, “[t]his form [i.e., systematic form] is a demand of reason that concerns each and every object if it is to be treated as object of science” (1792, p. 43).

168

Maimon’s method of fictions

unity” (Strf, GW IV, 63).108 Thus, the systematising function of philosophical fictions brings the greatest aggregate of phenomena under the smallest number of principles, turning the aggregate into a system of intelligible facts.109 In more abstract terms, metaphilosophy, through positing fictional concepts and principles, methodologically grounds a relational whole of why-questions and their answers. Philosophical fictions establish the ‘form of all science’ through positing theoretical structures, through which “the objects of nature are ordered philosophically under principles, are brought into a system” (Strf, GW IV, 349). In short, the main function of philosophical fictions, in virtue of which they qualify as scientific, is to establish a systematic form according to which scientific explanation is possible.110 This is also why Maimon describes the products of philosophical science as “formal inventions” (VI, 16), or, more precisely, as “forms, systems, methods” (Strf, GW IV, 28; cf. 16). Philosophical fictions as fictions of systematicity enable scientists to conceive of different sorts of judgments as standing in explanatory relationships by projecting the form of an organised, interconnected body of knowledge. Only by treating the object of science as an ultimately explicable and coherent whole of intelligible determinations can our scientific explanations be understood as on the road to becoming a true representation of the world. These fictions guide science to produce a unified system that justifies its knowledge claims by virtue of their integration into the systematic whole, and which continually approaches an objectification of systematicity. The unity of a systematic whole presents a means to ground scientific 108

On the notion of systematicity in Maimon, see Franks (2017, p. 128ff.). I also think Kant’s definition of a system is revealing here: “[S]ystematic unity is that which first makes ordinary cognition into science, i.e., makes a system out of a mere aggregate of it. I understand by a system, however, the unity of the manifold cognitions under one idea (KrV, A833/B 861)”. 110 In this way, we can make sense of his simultaneous praise of Kant’s critical philosophy as “systematic to the highest degree” (Strf, GW IV, 73), Newton’s “world system, wherein can be found the highest unity in the greatest manifold of appearances” (Strf, GW IV, 63), and the Leibnizian system (Strf, GW II, 437). It is only on grounds of philosophical fictions that incorporate the demand for complete explanation that we can show its objective purport. 109

Maimon’s metaphilosophy: philosophy as modelling practice

169

claims, since each is determined through, and grounded in, its relation to the whole. This suggests an answer to question of what it is that the property of systematicity adds to a body of knowledge to make it a scientific body of knowledge: fictions of systematicity account for the possibility of scientific explanation without having to posit first, constitutive principles. Systematic form can figure as a normative goal that continuously informs the ways in which scientists treat and investigate their objects of study, construct hypotheses, experimental procedures, and explanations in order to construct a systematically ordered nature.111 On this account, reason’s demand for systematicity, then, does not emerge ‘from the bottom-up’,112 but rather from the ‘topdown’, as a demand for rational (that is, complete) explanation, which can only be accommodated by virtue of fictions of systematicity. This top-down demand for complete explanation, which is realised through a philosophical construction, will also become relevant to Schelling’s reflections on philosophical method. Like Maimon, he argues for a philosophy that employs useful fictions, which provides a model for understanding the possibility of science unified under one paradigm of rationality. However, Schelling realises that this possibility demands a complete revolution of theoretical philosophy as it stands: philosophy must be transformed into Naturphilosophie. As such, it entirely abandons the standpoint of human cognition, but not the standpoint of a divine intellect. Rather, it uses methodological fictions to explain everything to do with the forces of nature, and tests the constructions of Naturphilosophie through positing an absolute hypothesis.

111 Still, we should keep in mind that Maimon does not think that science can ever achieve this goal. For example, with regard to explaining some natural event, he states that “[i]f I notice this again and again, so that these two appearances are ever more strongly connected in me, then at last (through complete induction) this subjective connection reaches its highest degree, and becomes equivalent to the objective” (Tr, GW II, 72). However, this does not mean that we can actually carry out a complete induction, but only that, from an explanatory standpoint, scientists must understand their work accordingly. 112 I am thankful to Peter Thielke for discussion of this point.

3 Schelling’s method of nature-construction

In one of his earliest writings, the young Schelling notes that he learned from the latest writings by Salomon Maimon—a work worthy of closer examination that [Schelling has] so far been able to accord to it—that the need for a complete solution of the entire problem, which had been so far a barrier to all attempts at shaping a universally valid philosophy, is beginning to be felt more generally than has been the case until now. (ÜMFP, HKA I/1, 267)

This “entire problem” illuminated by Maimon’s philosophy is nothing other than the challenge which has occupied us in previous chapters: “the entire problem of the possibility of philosophy as such” (ÜMFP, HKA I/1, 266). Unlike Kant (and much more like Maimon), Schelling does not focus on the isolated problem of transforming metaphysics into a science; rather, he sees this as a problem for philosophy as such. Schelling is led to the first works of Post-Kantian philosophy1 “through the study of the Critique of Pure Reason, in which nothing seemed more obscure and harder to understand […] than the attempt to lay the foundation for a form for all philosophy” (ÜMFP, HKA I/1, 265 [emphasis added]). Schelling interprets the reactions of both Kant and his predecessors as revolving around one particular question: how is philosophy as science possible? From the very beginning, his philosophical project is committed to these metaphilosophical concerns. In this chapter I shall argue that it is with his early system of Naturphilosophie that Schelling delivers his most original methodological solution to the problem: in order to become science, the philosophy of philosophy must construct nature. 1 Schelling explicitly mentions Schulze’s Aenesidemus (1792), Reinhold’s Essay on a New Theory of the Human Capacity for Representation (1789), and Fichte’s Concerning the Conception of the Science of Knowledge Generally (1794).

Schelling’s method of nature-construction

171

Schelling praises Maimon for his New Essay on Logic (1794) and its exposition of the problems that a philosophical science must face. As we saw in the last chapter, the latter had become a firm defender of the method of fictions in philosophy by this time. Maimon also begins the Logic with a restatement of the basic premises of the Essay. He reiterates his account of how the different sciences are differentiated, individuated, and how they relate to each other within a purposefully organised whole (Logik, GW V, 13-17).2 Maimon re-emphasises the interdependence between his metaphilosophy and the sciences. Only in the right combination can they begin to generate a true philosophical system. He states that “[w]ithout philosophy no science at all is possible, because it determines a priori the form of a science in general. Without positing any other science, philosophy has no meaning for us” (Logik, GW V, 19). Maimon’s method of fictions only provided the resources to model an account of a priori cognition, detailing the conditions according to which rational inquiry would be possible. It cannot, however, produce the facts themselves which would confirm its philosophical hypotheses. Schelling agrees with Maimon’s assessment that philosophy can only determine the form of science, through determining the possibility of a priori cognition, if it is committed to explanatory rationalism and infinite intelligibility. Through his engagement with Fichte’s subjective philosophy, moreover, he develops an account that grounds the possibility of a priori cognition and rational explanation in a pre-subjective principle. This proposal attains maturity in his (1799) system of Naturphilosophie, which argues that philosophy can only legitimise itself as science if it posits an unconditional principle in nature itself, and thereby constructs consciousness from nature and not vice versa. To not only prove its legitimacy, but also its reality, philosophy must put its theoretical constructions to the test. On my interpretation, Schelling holds that the rational construction of nature is presented through scientific experimentation and is therefore empirically verified. 2 Maimon describes the relations between the different sciences and their connection into a systematic whole, or “unity within the manifold”, with the help of functional metaphors, i.e., the metaphors of a tree and a machine (V, 14). Each science is defined and individuated in relation to all other sciences, as well as in relation to the systematic whole which they constitute.

172

Schelling’s method of nature-construction

Since it would be untenable on grounds of the textual evidence3 to claim that Maimon directly influenced Schelling in the development of his metaphilosophical programme, my argument is more modest: I show that Schelling continues Kant and Maimon’s metaphilosophical research programme by offering a new and testable version of it.4 I argue that Schelling’s Naturphilosophie presents another metaphilosophy-first project which grounds the possibility of science in the possibility of philosophy. In continuity with both Kant and Maimon, Schelling develops a methodological solution to the metaphilosophical problem of determining this possibility. And, just like the former two, his methodological proposal is informed by his philosophy of science. By virtue of reconceiving of the experimental methods of the natural sciences through his account of the natural-philosophical experiment, Schelling develops the method of nature-construction. While there exists an interesting array of scholarship on his philosophy of science, only few have tried to connect it to his metaphilosophical views. By investigating this connection exactly, I show how we can understand Schelling’s method of nature-construction through his specific conception of the 3

Schelling only mentions Maimon in his early writings, i.e., (ÜMFP, HKA I/1, 267) and (VIPP, HKA I/2,151). It is maybe for this reason that (at least to my knowledge) there exist only two papers focusing on the connection between Maimon and Schelling (see Fincham (2016) and Franks (2020, p. 81)). 4 It is certain, though, that Maimon influenced the young Schelling through the philosophy of Fichte. Fichte was not shy about pointing out the importance of Maimon for philosophers working on the critical project, and was one of the most important philosophical authorities for the young Schelling. He explicitly embraced Maimon’s criticism of Kantian dualism and sought to develop a philosophy fit to deal with this challenge (see Beiser (2003)). I agree with most of Beiser’s assessment, although he seems to overestimate the similarity between their respective approaches: although Fichte follows Maimon in “curing skepticism through the practical instead of self-evident first principles”, Maimon’s pragmatism stems from his reflections on the nature of science and, more precisely, scientific fictions. Fichte, on the other hand, ultimately cures this problem through practical self-determination in an intersubjective world with other self-determining agents (e.g., Fichte 1794; 1796). While Fichte charts new territory (i.e., the debate about rational foundations and intersubjective recognition), Maimon is still part of a theoretical paradigm. We will see that Schelling reprises Maimon’s critique of dualism, and also takes this as a necessary starting point of his metaphilosophical theorising.

Schelling’s method of nature-construction

173

unconditional principle of nature as “absolute hypothesis”, as well as his conception of scientific experimentation as the presentation of “the finite in the infinite”. In doing so, I will show how he thereby arrives at a new model of propaedeutic philosophy that—by taking science to be grounded in a methodological principle and scientific experiments to be the presentational media of its construction—shows how philosophy as first science can be conceived in terms of Naturphilosophie. Since the underlying thought is that Schelling’s metaphilosophical programme is already present in his early writings, and that his first original proposal is a result of his early Naturphilosophie,5 I begin in the first section with an analysis of his early philosophy, specifically with his version of the metaphilosophical conception that philosophy must be preceded by a propaedeutic science in order to establish the possibility of philosophy. We will see that, in agreement with Maimon, Schelling argues that philosophy cannot explain this possibility if it posits a cognitive dualism. Instead, philosophy must find one principle—a fundamental axiom or Grundsatz—that grounds all science as its unconditioned first principle, and from which everything can be explained as infinitely intelligible. The second section will show that in contrast to his predecessors, Schelling maintains that there is yet another dualism which must be overcome if philosophy is to be possible as science: the dualism between the subject of consciousness and the object it is the consciousness of. The problem of dualism between epistemic subject and object, on his view, can be resolved by assuming a pre-subjective principle, which alone can serve as unconditioned first principle of all science. The third section elaborates on why Schelling’s search for an unconditioned first principle lead him to conceive a metaphilosophy that explains everything using natural forces. For philosophy to be possible, it must be grounded in one pre-subjective principle, and this can only be provided through a system of Naturphilosophie. In a fourth section, I explain Schelling’s mature Naturphilosophie as grounded in a 5 His earlier writings in Naturphilosophie, i.e., Timaeus (1794), Ideas for a Philosophy of Nature 1797), and On the World-Soul (1798), still conceptualise nature (in one way or another, e.g., following Kant or Fichte) as a structural moment of consciousness. It is only from 1799 onwards that Schelling develops Naturphilosophie as a first philosophy.

174

Schelling’s method of nature-construction

methodological conception. Naturphilosophie can play the role of first philosophy because it is grounded in a pre-subjective principle that is posited as belonging to nature as an “absolute hypothesis”. By treating nature in accordance with this methodological principle, Naturphilosophie can proceed to construct nature as an infinite productivity, which inhibits itself in an infinite series of natural products. Sections five and six are then dedicated to a detailed elaboration of what Schelling’s special conception of scientific experimentation consists in, and why it allows him to conceive of experiments as mediums for the presentation of nature’s infinite activity. Schelling’s methodological solution yields a conception of Naturphilosophie as philosophical experiment which—under the functional posit of an absolute hypothesis—tests and gradually verifies a particular construction of nature.

3.1 Philosophy is the science of the unconditioned Let me begin with an investigation of how Schelling’s original solution to the challenges of devising the right metaphilosophy is prepared in his early philosophy. To comprehend how Schelling arrives at the standpoint of Naturphilosophie, we must trace how he develops this particular conception of science through an ongoing preoccupation with the concept of “the unconditioned of science”. In some of his earliest texts, namely On the Possibility of an Absolute Form of Philosophy (1980/1794) and Of the I as the Principle of Philosophy, or on the Unconditional in Human Knowledge (1980/1795), Schelling tries to conceive the scientific nature of philosophy, as well as its relation to other sciences through an analysis of the concept of an ultimate axiom [Grundsatz]6

6 In the English translation of Schelling’s early works, the German word Grundsatz is consistently translated with the English ‘axiom’. Since this could be a misleading translation (as ‘axiom’ is here understood in a different sense than, for example, in Euclid), I will stick to the German Grundsatz.

Philosophy is the science of the unconditioned

175

for science.7 Metaphilosophy, for Schelling, begins as the science of the nature of the unconditioned. Philosophy, he contends, can only be sure to stand on firm ground if it begins as a metaphilosophical investigation. That is to say, philosophy must begin by studying its own nature, method and role, and become philosophy of philosophy. Looking back to Kant’s First Critique, Schelling states that the “aim of philosophy is no mere reform of its discipline but a complete reversal of its principles, that is, a revolution which one can view as the second possible revolution in its field” (I/2, 77). As for the first revolution, Schelling identifies Kant’s proposal of the Copernican framework as a paradigmatic change in how to think the possibility of a priori cognition of objects, namely to conceive of objects as conforming to our forms of cognition. The second revolution, however, although maintaining the same direction of explanation, is only “made possible by a complete reversal of the principles” (ibid.). Instead of adopting Kant’s heterogeneous account of cognition, Schelling follows Maimon8 in proposing an account that grounds the possibility

7 Of his four early essays, these two engage with the unconditional and its meaning for theoretical philosophy, while the other two (Philosophical Letters on Dogmatism and Criticism (1795) and New Deduction of Natural Right (1796)) focus on its role in practical philosophy and the moral realm. 8 In his Treatises (1797/8), Schelling explicitly criticises Kant’s dualism in an analogous fashion to Maimon, thereby establishing the ground for his own monist conception (see Fincham (2016, pp. 1049-1055)). Understanding the former’s dualism as primarily a dualism between form and matter, his criticism is that, although for Kantians the world is “primordially alien to our spirit and the world bears no affinity to the spirit other than that of an accidental affect”, this very same world is “govern[ed] with laws that—they neither know how nor whence—have been implanted in their understanding” (AEIW, HKA I/4, 78-79) [emphasis added]). Or: “[E]ven if we understand the origin of a world external to ourselves, we still do not understand how the representations of this world could have entered into our consciousness. In the last effort, then, it had to be explained not how external things could have originated independently of ourselves—(for of these we cannot have any knowledge, since they themselves are the ultimate substratum for any explanation of external phenomena)—but how a representation of these [things] could have originated within us” (as quoted in Fincham 2016, p. 1053, Schelling (AEIW, HKA I/4, 83).

176

Schelling’s method of nature-construction

of a priori cognition in one, unconditioned principle.9 Before locating this principle within nature itself, however, he took his cues from Fichte’s philosophy of the absolute I, which posits this principle within the human mind.10 In On the Possibility (1980/1794), Schelling begins this investigation not by formulating the question, i.e., how is philosophy possible as science, but by stating the fact: “[p]hilosophy is a science, that is, it presents a specific content in a specific form” (I/1, 267). Thus, Schelling assumes that he has have found the specific content and the specific form that characterise philosophy as an individual science. He then asks whether it is possible to think of a reason that explains why philosophy has this specific form and this specific content. He then hypothesises that there could be “some common ground which would simultaneously furnish the form as well as the content” of philosophy, and he asks whether it could not be that “the very form of this science bring[s] along its content, or the content its form” (ibid.). Moreover, this interdependence of content and form has such an ascendancy over the mind that it must give rise to the thought that there might be a reason for it in man’s mind, but that philosophy has not yet found it. It looks as if that thought had guided philosophy to search for an absolute connection between a specific content and a specific form. This is an idea which philosophy could approach only step by step and which it could express only to a more or less limited degree as long as it could not find that reason, which is lodged in the human mind itself. This much is clear that, if the content of philosophy necessarily creates its form or the form its content, then there can only be one philosophy in line with the very idea of philosophy. (ÜMFP, HKA I/1, 268)

Schelling’s initial hypothesis states that what defines philosophy is its specific content and form, and that these depend on each other in such a way that one of them necessarily creates the other. Philosophical practice is depicted as an ongoing search for the specific content and form that this absolute connection entails, rendering the history of 9 Reinhold was the first to criticise Kant’s philosophy for not being grounded in one first principle (see Snow (1996, pp. 12-14)). 10 “Where is the principle on which the first form is based? Where is the principle from which the particular forms of thinking derive, those which Kant enumerates without relating them to a higher principle? These questions remain unanswered” (ÜMFP, HKA I/1, 289).

Philosophy is the science of the unconditioned

177

philosophy as a collection of more or less adequate expressions of this connection, which itself remains an idea. Of course, this search would terminate if philosophy were to find the reason or ground of this absolute connection, as it would identify one philosophy as the correct one. The reason which grounds this connection, as Schelling repeats twice, is to be found in the human mind itself. By arguing that to determine the necessary structure of the objects of science is to determine something about the structure of human minds, Schelling embraces a Copernican framework. Now, Schelling begins to investigate this hypothesis through an examination of the form and content of science in general. As we can infer from what was said above, each science is determined in virtue of a specific form and a specific content. No matter its specific content, Schelling contends, the general form of a science is “the form of unity” (ÜMFP, HKA I/1, 268). And the content of any science can be structured according to such form only if “all its parts are subjected to one premise, and when each part determines the other only insofar as the part itself is determined by that one premise” (ibid.). From a formal standpoint, therefore, each science consists of two elements, namely of “theorems” [Sätze] and “axioms” [Grundsätze].11 While theorems constitute the Sätze of a science, Grundsätze function as their unifying premises. Only if based on a Grundsatz can a body of sentences count as proper science.12 Thus, Schelling first identifies the “general form of all sciences” not as a systematic-teleological structure (as Kant and Maimon do) but as an axiomatic-deductive structure. Schelling distinguishes Grundsätze from theorems (or, simply, ‘sen11 F. Marti’s translation of the German Sätze as ‘theorems’ also strikes me as not quite correct in this context. I will therefore stick to the German here, too. 12 Schelling further determines that what makes any epistemic practice fall under the genus ‘science’ is its formal form, i.e., “the unity of a continuous connection of conditional theorems, of which the first, the axiom [Grundsatz], is not conditional” (ÜMFP, HKA I/1, 268). What distinguishes the sciences as particular sciences (i.e., their differentia) is that “each one of them depends on its specific content”. Whether this content is structured through the general form of science, or the general form is structured through the specific content of this science, the resulting form of science should be called its “material form” (ÜMFP, HKA I/1, 268).

178

Schelling’s method of nature-construction

tences’) by identifying the latter as conditional and the former as unconditional in nature (ÜMFP, HKA I/1, 270). He describes sciences as consisting of a body of theorems, which can all be derived from one and the same premise.13 As a Grundsatz, this first premise must be one that is accepted without further proof or justification because it is either self-evident or a presupposition. Although each science is grounded in one premise, in virtue of which its parts are unified, insofar as this premise provides the unconditional condition to all other, i.e., conditional parts, these Grundsätze only have the character of presuppositions. For example, Euclidean geometry consists in a set of problems and theorems that together determine the properties and relations of spatial magnitudes, and which are themselves grounded in and derived from a system of axioms, definitions, and postulates. Schelling argues that each science has only one premise or Grundsatz which truly unites all of its parts. On this view, then, it would for example be the constitution of space itself that qualifies as unconditioned of Euclidean geometry (and not Euclid’s own axioms, which are five in number).14 All of the theorems of geometry are conditional on this one assumption, namely that space is a determinable (with some formal and material features). Yet, if each individual science is grounded in an ultimate premise, how can these sciences possibly be connected with each other? Interestingly, Schelling does not say why it must be that all sciences connect to each other to form one system of science. Nevertheless, he asks what a Grundsatz would have to look like “to be a condition of the entire science” (ÜMFP, HKA I/1, 270). To explain this, we have to assume that Schelling shares Maimon’s commitment to infinite intelligibility and complete explanation. In contrast to Maimon, Schelling here sees this commitment as reflected in the form of individual sciences, 13 Schelling states that “[t]he axiom [Grundsatz] of each science cannot be conditioned by the science itself, but must be unconditional in regard to that science. For this very reason its axiom [Grundsatz] can be only one” (ÜMFP, HKA I/1, 270). 14 Thus, Schelling treats as theorems what Euclid considers to be differentiated into axioms, postulates, definitions, theorems, and problems. Furthermore, while Euclid’s axioms are multiple in numbers and, amongst other things, serve to define the three basic operations of construction in space, Schelling’s Grundsatz rather assumes the role of an absolute ground.

Philosophy is the science of the unconditioned

179

whose nature he uncritically analyses as axiomatic-deductive in nature. Schelling’s arguments about the nature of philosophy as propaedeutic science only follow on the basis of his prior assumption about the form of science as such, which at the same time is consistent with a commitment to infinite intelligibility. In particular, he argues that if the individual sciences were grounded in different Grundsätze that were not unconditioned, but rather afforded further conditions (e.g., the constitution of space in geometry, or gravitational forces in Newtonian physics), then explanation in these sciences could not arrive at a sufficient reason, but would instead have to stop at some arbitrary point. In this case, we would also not be able to explain how the different sciences relate to each other, and how the different standards and modes of explanation practised by each of them make sense with regard to the same natural world. It is because Schelling commits to infinite intelligibility, and thereby to the explanatory standard that all why-questions must be answerable until we arrive at an absolute ground, and because he defines sciences as necessarily having axiomatic-deductive structure, that he takes all sciences to be grounded in one ultimate Grundsatz. Philosophy is assigned the role of the science which determines that Grundsatz. In fact, Schelling contents that “philosophy—if it is to be a science at all—must be governed by a plainly absolute axiom [schlechthin unbedingter Grundsatz]” (ÜMFP, HKA I/1, 93). Schelling’s argument for this is a bit tricky, and again relies on his rationalist commitments. Remember his starting hypothesis concerning philosophy, i.e., that “philosophy […] is to be a science whose specific content is necessarily, not arbitrarily connected with a specific form” (ÜMFP, HKA I/1, 270). Now, if philosophy is to be a science, i.e., “an entirety governed by the form of unity”, and as a specific science is one in which the specific content necessarily relates to its form, then “it should not be conditioned by any other science” (ibid.).15 This is so because if philosophy’s specific content 15 Schelling correctly notes that this does not yet make philosophy the condition of all other sciences, since it could still turn out to be an independent, unconditional science that is not connected in any significant way to other sciences (ÜMFP, HKA I/1, 271).

180

Schelling’s method of nature-construction

is necessarily, not arbitrarily connected with a specific form, then its highest axiom [höchster Grundsatz] must establish not only the entire content and the entire form of the science, but it must have a content of its own which is connected with its own specific form inherently and not merely in an arbitrary way. (ibid. [emphasis added])

To show the possibility of philosophy as science, Schelling argues, the relation between content and form of philosophy should be defined not in an arbitrary way, but rather in a necessary way. Unlike Kant’s definition, which allows for brute facts (e.g., it just happens to be the case that our receptive faculty has the form of space and time, with no further explanation possible), Schelling thinks that the possibility of philosophical science can only be explained if it is completely explained. Thus, the reason for a necessary connection between its content and form must be grounded in the nature of this specific content and this specific form. And to show that their specific relation truly is necessary, philosophy must determine its entire content and form. Philosophy’s ultimate Grundsatz “must be a content which simply is, which unconditionally is”, and not in virtue of another science. On these grounds, it becomes clear why Schelling can argue that philosophy’s task is to determine the absolute Grundsatz which, as absolutely unconditioned, underlies all other Grundsätze of science. That is, if philosophy provides a content “which unconditionally is”, then it “furnishes the condition of every content in all sciences” (ÜMFP, HKA I/1, 271) and thus “contain[s] the ultimate conditions of all other sciences” (ÜMFP, HKA I/1, 272). Therefore, we could characterise philosophy as the science which articulates the structure of the unconditioned at its most fundamental level; it determines “the fusion of form and content of the ultimate axiom [die Verbindung der Form und des Innhalts des obersten Grundsatzes]” (ibid.). Schelling further clarifies that “an unconditional content can have only an unconditional form and vice versa since, if one were conditional, the other, even if it were unconditional, would have to be conditioned, owing to its fusion with something conditional” (ÜMFP, HKA I/1, 274).16 This Grundsatz 16 Here Schelling also seems to agree with Fichte, although the latter gives a much less detailed argument for why the ultimate axiom must have form and content. He only notes that “[n]o sentence is possible without content or

Philosophy is the science of the unconditioned

181

must consist in the fusion of an unconditional content with an unconditional form, which “are possible because each requires the other” (ibid.). This is so because there can neither be form without content nor content without form, and neither can only one be unconditional because it would therefore still depend on something conditional. Hence, both must simultaneously be posited as unconditional. As the science that provides an absolute Grundsatz of this nature, philosophy is assigned a specific role with regard to all other sciences, namely the role of the first science: [F]or that reason the question whether philosophy is possible at all places us within the domain of that first science, which could be called propaedeutic of philosophy (Philosophia prima), or better still, theory (science) of all science, archscience, or science κατ’ ἐξοχήν, since it is to condition all the other sciences. (ÜMFP, HKA I/1, 274 [emphasis added])

Just like Maimon before him, Schelling identifies the possibility of philosophical science with the possibility of philosophy as the science of all science. And, like Kant, he identifies this science as a propaedeutic philosophy, since it provides the ultimate condition under which all other sciences, including itself, become possible. The question ‘how is philosophy possible’ overlaps with the question ‘how is science possible’ in that, by articulating the conditions of the possibility of philosophy, one is simultaneously articulating the conditions of possibility of all science, insofar as it is connected by one unified system.17 Only if propaedeutic philosophy manages to determine this “plainly absolute axiom [schlechthin unbedingter Grundsatz]” does it become intelligible how science as a whole can treat its objects as infinitely intelligible, and how it can treat every thing, fact, and event as having a sufficient reason: the uncondiwithout form. There has to be something, of which one knows, and something, that one knows of it” (GA 1,2, 121). 17 With this characterisation of the task of philosophy as propaedeutic science, Schelling essentially takes Fichte’s remarks in On the Concept (1794) as read. There, Fichte characterises first philosophy as Wissenschaftslehre, which again is defined as “the science of science as such” (GA 1,2, 118). As the “science of science”, Wissenschaftslehre has as its task to ground the Grundsätze of all other sciences by showing how these are possible and certain (ibid., 120). Wissenschaftslehre arises as a science of knowledge, determining how anything can be certain, and thereby determines how we can make valid knowledge claims (ibid., 121).

182

Schelling’s method of nature-construction

tional Grundsatz of philosophy provides the absolute unconditioned, in which all chains of reasons or conditions are grounded. Therefore, philosophy is the science of the structure of the unconditioned per se. It is this conception of propaedeutic philosophy which Schelling will continue to develop throughout the subsequent years of his academic work, and which culminates in an original formulation in his Naturphilosophie in 1799. After emancipating his own view from Fichte’s philosophy of consciousness, Schelling joins Kant and Maimon in developing a methodological solution to the metaphilosophical challenge that is grounded in his philosophy of science, and in particular his philosophy of experiment. The reason for this philosophical innovation lies in Schelling’s altered analysis of the ultimate Grundsatz of philosophy. Through his engagement with Fichte’s philosophy of the absolute I, Schelling comes to realise that the reason for the necessary connection between the form and content of philosophy is in fact not “lodged in the human mind” (ÜMFP, HKA I/1, 268). As we will see, Schelling’s point is that the unconditioned cannot be grasped from the self-reflexive standpoint of consciousness. To explain the possibility of a priori cognition of objects, first philosophy must articulate a pre-subjective structure of the unconditioned which can explain the dualism between subject and object, rather than take it as primitive, and must therefore accept another real dualism. To understand how Schelling arrives at his own account, we need to first understand his attempt to reconstruct Fichte’s model. His reconstruction of the latter’s “original insight” leads him to observe a particular shortcoming of Fichte’s explanation of a priori cognition as grounded in the self-positing structure of the epistemic subject, from which he begins to develop his own position. To show this, I first lay out Schelling’s observation,18 that what characterises the unconditioned Grundsatz is not that it is ‘self-evident’ but that it is ‘self-causing’, to then explain Schelling’s point that philosophy’s unconditioned cannot be grasped from the standpoint of consciousness, but only from the stand18 As some might notice, this vocabulary is inspired by the omnipresence of Kimhi’s Thinking and Being (2018) at philosophy colloquiums during the years 2018 and 2019.

Fichte’s original insight, Schelling’s observation, and Schelling’s point

183

point of Naturphilosophie. Rational explanation that is committed to infinite intelligibility is only possible when grounded in a pre-subjective principle.

3.2 Fichte’s original insight, Schelling’s observation, and Schelling’s point So, what does this “plainly absolute Grundsatz” consist in? Schelling first excludes one kind of procedure, which we have already encountered in the previous chapters, namely the regressive or analytic procedure: Should we retrace our steps from axiom to axiom [von Grundsatz zu Grundsatz], from condition to condition, until we arrive at the ultimate, absolute, categorical axiom [categorischer Grundsatz]? In that kind of procedure we would necessarily have to begin with disjunctive propositions, that is, no axiom [Grundsatz]— inasmuch as it is determined neither by itself (for then it would be the ultimate) nor by one that is higher (for then we would already have the higher for which we are looking)—could serve as starting point for a regressive search. (ÜMFP, HKA I/1, 279 [emphasis added])

Philosophy cannot proceed in a regressive manner in order to discover the absolute Grundsatz because, in doing so, it would need to begin with something (e.g., a fact or event) that is by definition something conditioned. Thus, it is either something which is not determined by itself but by something that it is not itself, i.e., its condition, or it is determined by itself, in which case it could not serve as the starting point of an analysis because it is that which has no further condition, i.e., the unconditioned. Instead, Schelling suggests that the first criterion found in the concept of an absolutely unconditional proposition shows of itself the quite different way in which it must be sought. For such a proposition can be given only by its own criteria. But it has no other criterion than the criterion of absolute unconditionality. All other criteria which one might attribute to it would either contradict this criterion or be already contained in it. (ibid.)

To find this Grundsatz, one must not proceed in a regressive manner, but take seriously the only criterion that characterises the Grundsatz, namely absolute unconditionality, and proceed from it. Since for a content or a form to be absolutely unconditional is for it to not be

184

Schelling’s method of nature-construction

determined by anything but itself, only a procedure that determines something through itself, i.e., by generating its own criteria, instead of determining something in virtue of something else, can serve to determine the absolute Grundsatz. Thus, Schelling does not simply suggest that philosophy should proceed in a synthetic manner, and instead shows how to arrive at a principle from which synthetic procedure can depart: by constitution through self-relation. This is where Fichte makes his entrance. For those studying the early Schelling, Fichte is one of the most important influences to consider. And, importantly for our context, it is also through Fichte that Schelling is confronted with Maimon’s challenges to propaedeutic philosophy, as well as his new strategy for answering them.19 He agrees with the latter’s monist strategy, suggesting that: to end the “doubting an actual application of the category of reality” (GA 1,2, 261-2), philosophy must be shown to really determine not only the form but also the content of all science.20 Contra Maimon’s 19 In a letter to Reinhold in 1795, Fichte expresses his admiration: “My respect for Maimon’s talents knows no bounds. I firmly believe that he has completely overturned the entire Kantian philosophy as it has been understood by everyone until now, including you, and I am prepared to prove it. No one noticed what he had done; they had looked down on him from their heights. I believe that future centuries will mock us bitterly” (as quoted in Beiser 2003, p. 233). Sadly, the future centuries have not mocked Fichte bitterly. However, I hope this work will contribute its modest part to the mockery. 20 As always with Maimon, not much has been written about the connection between the two authors. The main contributions come from Beiser (2003) and Thielke (2001a). Both interpret Fichte’s epistemology as a response to Maimon’s scepticism. Beiser, in particular, identifies two lines of thought in Fichte’s philosophy that he sees as direct responses to Maimon’s challenge. According to him, Fichte’s first ‘line of defense’ addresses Maimon’s scepticism through his theory of the power of imagination as creating all objects of experience, thereby overcoming the problem of real dualism (Beiser 2003, pp. 239-243). The second, more promising, strategy can be found in Fichte’s postulation of a primacy of the practical, which ultimately constructs knowledge as a product of action, rather than a product of contemplation (GA 1,2, 368). Beiser’s closing statement, namely that Maimon himself had anticipated this line of argumentation, however, seems mistaken to me. Unlike Fichte, Maimon’s critical rationalism does not endorse a view of the empirical world as an imperfect construct that ought to be rationalised through human action, but rather as a construct that ought to be contemplated as if it were rational. In other

Fichte’s original insight, Schelling’s observation, and Schelling’s point

185

fiction of an infinite intellect, Fichte delivers a propaedeutic philosophy (Wissenschaftslehre) of the absolute I, which, as the first principle of all science, posits the form and content of all science. His view, i.e., the self-positing activity of the absolute I, which explains the form and content of science and thereby answers the quid juris, as well as the quid facti, was of great importance for the early Schelling’s own propaedeutic philosophy.21 And yet, it is wrong to assume that Schelling’s early philosophy is just a reprisal of Fichte’s doctrine. While Fichte presented this absolute I as a regulative ideal that could only be realised in the realm of the practical (e.g., GA 1/2, 429),22 Schelling employed it as a constitutive first principle of theoretical philosophy,23 since for Schelling the absolute I does not have the status of an idea, but of a words, although Maimon does attribute an important role to the pragmatic value of fictions, he argues not from standpoint of moral agents whose actions make nature conform to their purposes and ideals, but from the standpoint of scientists who contemplate the world according to pragmatic (but rationalist) fictions. 21 Fichte reacts to Maimon’s challenges in both his On the Concept (1794) and his Foundations of Natural Right (1796). From what is documented through letter exchanges, we can infer that, while Schelling was not familiar with the practical part of Fichte’s system (which might explain why he was prone to read Fichte in the manner he did), he did know On the Concept (see Nassar (2013a, pp. 161-2)), and had read it at a similar time to when he read Maimon’s New Essay on Logic (1794). 22 “An Realität überhaupt, sowohl die des Ich, als des Nicht-Ich findet lediglich ein Glaube statt” (GA, 1/2, 429). 23 See Beiser (2002, p. 217). I agree with Beiser insofar as his is an antifoundationalist reading, that as a form of ‘subjective idealism’ is opposed to Schelling’s ‘absolute idealism’ (for more details, see Rockmore (1996)). Although this reading is problematic from an exegetical perspective with respect to Fichte (see Neuhouser (1990) and Breazeale (1996) for substantial criticisms), it does mark out the telling difference that Schelling himself perceived to be at the core of his dispute with Fichte. Especially in the On the Concept (1794), which seems to have largely served as the foundation for Schelling’s early philosophy, Fichte clearly endorses the latter reading (GA 1/2, 150-152). Although the reason for Schelling’s misrepresentation of Fichte’s own philosophy might just be owed to the fact that he hadn’t read the third part of the Grundlage (Beiser 2002, p. 473), this point went on to be explored further, and indeed develops into the distinctive mark of Schelling’s mature Naturphilosophie. In particular, it is interesting to note that Schelling begins his philosophical work by adopting a misconceived foundationalism from Fichte, which then changed

186

Schelling’s method of nature-construction

real principle (VIPP, HKA I/2, 111-112, 130-133). Since Sandkaulen’s original interpretation,24 Schelling scholarship has shifted its tendency to read his early works with regard to their implicit criticism of Fichte’s work. In line with these interpretations, I show how Schelling’s divergent conception of the absolute I prepares his Naturphilosophie and its new method of construction. First, what was Fichte’s account? With regard to the question of how philosophy can become a science and deal with Maimon’s challenges, he developed the following idea, which is sometimes referred to as his “original insight”. 25 Philosophy as the science of science cannot begin with a principle that we arrive at by abstraction or reflection on a fact, but must begin with a synthetic principle, which can only be generated through an act.26 Fichte translates the proposal we already know from previous chapters, namely that we know something because we create it, into an act of consciousness. This synthetic act is the ground of yet another dualism, namely that between the subject and object of cognition. More precisely, he identifies this first synthetic principle as consisting in an activity that is at the same time a deed (i.e., a fact): a “Tathandlung” (GA 1,2, 46, 255ff.; 1,4, 219, 221). Now, Fichte’s original insight is this: this first principle is constituted through the particular relation that a consciousness has to itself as self-consciousness, into an anti-foundationalism in his Naturphilosophie from 1799, which has to begin from a standpoint that rules out foundationalism as a position. 24 Sandkaulen (1990) was (one of) the first to notice (and if not the first, the most thorough in noticing) that in his early philosophy, Schelling begins to develop an account of the absolute I that is distinct from the one that Fichte offers. Her idea was most thoroughly and productively developed by Nassar (2013a, 2016), who uses this thesis to argue that Schelling, in his early philosophy and through his dissenting interpretation of Fichte, is already working towards his later Naturphilosophie. 25 As is well-known, Henrich (1967) argues that Fichte’s theory can be seen in light of one “original insight”: that there can be no reflexive theory of selfconsciousness. I agree with Pippin (1988, pp. 80-84) and Neuhouser (1990, p. 69) that Henrich’s interpretation is too narrow, and that Fichte is more broadly concerned with a theory of consciousness and how it can explain the possibility of a priori cognition. Nevertheless, it does serve to crystallise the main point of disagreement between Fichte and Schelling. 26 As Fichte explicitly states, the first principle is “not a mere fact, but expresses an ‘act’ ” (GA 1/2, 46).

Fichte’s original insight, Schelling’s observation, and Schelling’s point

187

and this relation as on-going activity is the absolute I. However, this relation should not be understood as the reflexive act of a consciousness that turns back onto itself, and through that act is transformed into self-consciousness.27 Rather, consciousness, as it is found in reflection, is always already self-consciousness. In order to avoid circularity, self-consciousness must be taken as prior to consciousness. Further, in order to avoid regress,28 self-consciousness must be explained by virtue of a pre-reflexive activity, the Tathandlung. This activity, Fichte tells us, can be expressed through the proposition ‘I am I ’. He characterises this proposition as one whose form cannot be posited without its content, and vice versa: “I am posited, because I have posited myself. I am because I am.” (GA 1,2, 139-40 [emphasis added in bold]). The Tathandlung denotes an activity of self-positing, in which the content of positing is posited through the form of this positing, and in which the content of this positing posits the content. Or in other words, the self-positing activity is not distinguishable (at least not in any meaningful way) from that which it posits. This is because Fichte describes his Grundsatz from the perspective of an epistemic subject. The Tathandlung is an activity which presents itself to us as “the immediate consciousness that I am acting and [of] what I am enacting. It is that by means of which I know something because I do it” (GA I,4, 217). The absolute ground of knowledge is practical insofar as reason grounds itself through its own activity. Consequently, Fichte defines the Grundsatz as a sentence which “is certain, also when other sentences are not” (GA 1,2, 115). On his view, what distinguishes the Grundsatz from all other sentences is that we can 27 Franks likes to quote Novalis, who describes the problem posed by the phenomenon of self-consciousness in a particularly illuminating way: “What reflection finds, seems to [have] already [been] there” (2013, p. 2). 28 “The reason for this [regress] is easily grasped: If the self-awareness [i.e., self-consciousness] involved in my representation of X were itself a species of representation, then there would need to be another subject distinguished from and related to that representation of myself, and so on ad infinitum. An account of even the simplest state of consciousness would require the assumption of an infinite number of subjects and an infinite number of acts of relating [i.e., acts of reflection] between subject and representation” (Neuhouser, 1990, pp. 71-2).

188

Schelling’s method of nature-construction

and do know it with immediate certainty (GA 1,2, 112-115, 121). What characterises the Grundsatz as such is its self-evident nature, and so, for Fichte characterises unconditionality in terms of an epistemic quality, i.e., immediate certainty.29 The form of the unconditioned on Fichte’s analysis is self-evidence; it is defined in relation to an epistemic subject.30 The Grundsatz is unconditioned because it is evident independent of anything else, or (put simply) self-evident—it is evident through itself (GA 1,2, 114). This is why the Grundsatz can confer certainty onto other theorems. It is then by virtue of the Grundsatz that we can know other sentences, i.e., theorems, with certainty, namely when it can be shown that they are grounded in it and can be derived from it.31 Since everything is “for the I” (GA 1/2, 62), through adopting the Copernican framework that all objects, qua objects of consciousness, must conform to its structure, the Grundsatz can serve as the first principle of all science. Now, what is Schelling’s take on this description of the Grundsatz? In line with Fichte, he suggests that the only proposition which can serve as the ultimate Grundsatz is the proposition ‘I is I’ (I/1, 279-80). He thinks this must be so because “nothing can be posited absolutely except that which contains an absolutely original self and is posited not 29 It is important to note that Fichte does indeed mention the need for assuming the I as “absolute causality” (GA 1,2, 151). However, he attributes this absolute causality to the I which is opposed to the not-I, and which should be conceived of not as absolute causality but as “striving”. The I comes to know itself as ‘absolute causality’ only in the practical world, when it acts freely, thereby continuously realising freedom and absolute causality, which itself must remain a regulative ideal. (GA 2/3, 176). This stands is in direct opposition to Schelling’s view, which culminates in the statement that the “striving of the empirical I, and the consciousness stemming from it, would itself not be possible without the freedom of the absolute I, and absolute freedom is equally necessary as a condition for both imagination and action. For your empirical I would never strive to save its identity if the absolute I were not originally posited by itself, as pure identity, and out of its absolute power” (I/1, 181 [emphasis added]). 30 See Sandkaulen (1990, p. 25). 31 Fichte stresses that, in order to do so, a body of sentences must have systematic form. Only in this way can the Grundsatz confer certainty onto all other parts, making them theorems which can ultimately be derived from it (GA 1,2, 112-115).

Fichte’s original insight, Schelling’s observation, and Schelling’s point

189

because it is posited but because it is itself that which posits” (ÜMFP, HKA I/1, 280). Remember that for a content or a form to be absolutely unconditional is for it not to be determined by anything else, but only by itself. And to be self-determined in this special context, Schelling says, means that “[the absolute I] therefore posits itself (by absolute causality)” (ÜMFP, HKA I,1, 279). Self-determining as self-positing thus means simultaneously creating the very object of one’s determining activity by virtue of this activity; “the unconditional should realize itself, create itself through its own thought; the principle of its being and the principle of its thinking should coincide” (I/2, 87-88). The absolute I as the unconditioned principle “is posited not because it is posited (from without) but because it is itself the positing agent” (ÜMFP, HKA I,1, 279). At first sight, then, it looks as if Schelling—just like Fichte—identifies the absolute Grundsatz with “nothing other than the originally self-posited I, which is marked by all criteria enumerated” (ibid.). Upon closer examination, however, we can find two important clues that Schelling already disagrees with Fichte’s theory in these early writings. Firstly, and most importantly, he describes the positing I with the notion of “absolute causality”, and secondly, he expresses the absolute Grundsatz with the proposition “I is I” instead of Fichte’s proposition “I am I”. These are indicative of Schelling’s changed conception of the absolute Grundsatz, not as ‘that which we can know unconditionally’, i.e., as self-evident principle, but just as “that which is unconditionality’. We will see that, according to Schelling, the Grundsatz should not be interpreted as the self-reflexive activity of an epistemic subject, but rather as the self-determining activity of an ontological power. Schelling determines the unconditioned content as something “that posits itself (through absolute causality)” (ÜMFP, HKA I,1, 279). Unconditional content must be determined only through and by itself (ÜMFP, HKA I,1, 279). If the absolute I (as that which is absolutely posited) is defined not as this self-positing activity, but only as the immediate awareness of this self-positing activity, then it is conditioned by this positing activity of which it is only an immediate awareness. Schelling’s observation consists in this: into Fichte’s notion of immediate certainty and self-evidence is already built the standpoint of a knower, of an individual consciousness

190

Schelling’s method of nature-construction

that stands in a specific relation—of immediate certainty32 —to its object of knowledge.33 Consequently, the Grundsatz must not only express the self-evident knowledge of the absolute I’s self-positing activity,34 but also, and primarily, the structure of this self-determining power itself, by which it is transformed into the first constitutive principle of theoretical philosophy. Let us look at this observation in some more detail. In both his (1794) and his (1795) texts, Schelling makes a distinction between an inner and outer form of the Grundsatz (ÜMFP, HKA I/1, 274), which he sometimes equates with a distinction between a material and formal form (VIPP, HKA I/2, 217). For purposes of clarity, I will refer to these distinctions simply by virtue of the first formulation, i.e., as inner and outer form. Here are two exemplary passages: [T]he inner form of the content and the form of the axiom [des Grundsatzes] are each the form of being conditioned by itself, and only through this inner form does the outer form, the form of being posited unconditionally, become possible. (ÜMFP, HKA I/1, 274 [emphasis added]) The I posits itself absolutely, and posits all reality within itself. It posits everything as pure identity, that is, equal to itself. Thereby the material original form [Urform] of the I is the unity of its positing, inasmuch as it posits everything as equal to itself. The absolute I never steps outside itself. Through its original material form, however, a formal form, the form of positing in the I, is necessarily determined in the I as such. For the I is determined as the substrate of positability of all reality as such. Inasmuch as the I in its material form is the sum-total of all reality, it is also at the same time a formal condition of all positing, and thus I obtain a sheer form of the possibility of positing entities in the I at all. (ÜMFP, HKA I/1, 217 [emphasis added])

32 “Immediate certainty, Schelling argues, could not be the nature of the unconditioned because it is a result of the self reflecting on itself” (Nassar, 2013a, p. 242). 33 See Nassar (2016, pp. 124-127). 34 There are still passages in which it sounds as if Schelling does embrace the Fichtean image. For example: “However, if that ultimate itself is a condition of all knowledge, indeed a condition of its own being known, if it is the only immediacy in our knowledge, then we know precisely through it that we know; we have found the principle of which Spinoza could say that it is the light which illuminates itself and the darkness” (HKA 1/5, 155).

Fichte’s original insight, Schelling’s observation, and Schelling’s point

191

Here, Schelling notes that there is a difference between the act of positing and the formal structure of this act of positing.35 As discussed previously, the Grundsatz can be expressed through the content and form of a proposition. This proposition is not Fichte’s “I am I”, in which the I is that which knows (the form) and that which it knows about (the content). Rather, Schelling identifies this first principle with the proposition “I is I”, and “I is I is the material and the formal form” (ÜMFP, HKA I/1, 280). Now, Schelling’s distinction between an inner and an outer form of the Grundsatz intends to make us aware of the difference between “that which is known and known about” and “that which is determined and determined by itself”. The latter Schelling characterises as the inner and material form of the Grundsatz, which is more fundamental than the former, with the formal form depending on it. The material form is the form of “being conditioned by itself”; it is nothing other than the activity of positing itself [das Setzende] (ÜMFP, HKA I/1, 280).36 As content, is it the unity of this activity of positing, i.e., a pure activity that “never steps outside itself” (ÜMFP, HKA I, 1, 217); as a form, its activity is to determine itself. Or, its content is pure positing and its form is self-positing. On the other hand, the formal form “of [this] positing in the I” is the form of identity. That which is being posited unconditionally posits “everything as equal to itself”, and so the activity of self-positing takes a certain form, namely that of identity. Consequently, for Schelling, it is the material form of the Grundsatz which is fundamental, and which makes possible its formal form. This material form is the truly unconditioned principle, which must “precede all thinking and imagining” (VIPP, HKA 1/2, 90). It is the immediate awareness of this activity, which is expressed as formal structure of selfconsciousness. Thus, Nassar argues convincingly, that by distinguishing between the act of self-positing and its form, “Schelling distinguishes 35 See Nassar (2013a, p. 124): “[b]y indicating a difference between selfpositing, on the one hand, and the form of an unconditioned positing on the other, Schelling distinguishes between the I as absolute reality, and the I as a member of the formal structure of positing, a self-identical, self-reflexive consciousness” (ibid.). 36 See Sandkaulen (1990, pp. 24-26).

192

Schelling’s method of nature-construction

between the I as a self-causing cause, a self-determining reality, and the I as a member of the formal structure of positing, a self-identical, self-reflexive consciousness” (Nassar, 2016, p. 127). While, from the standpoint of self-consciousness, this activity appears to be the “the immediate consciousness that I am acting and [of] what I am enacting”, and “that by means of which I know something because I do it”, selfpositing as an activity must be explained in virtue of an absolute causality (GA I,4, 217). Careful readers of Fichte could take issue with this interpretation, and maybe rightly so. Was it not Fichte himself who recognised that the reflexive theory of self-consciousness faces an insoluble regress problem? Was it not Fichte’s original point to realise that consciousness requires self-consciousness, and that self-consciousness is not explicable as the product of reflection? Indeed, one of Fichte’s key points lies in the realisation that self-consciousness cannot be made intelligible through a consciousness that reflects back on itself. If self-consciousness is explained as a relation of identity between two relata, namely the subject-I that reflects and the object-I which it grasps as a result of reflection, then either the two relata are not identical—and hence the relation doesn’t hold—or they are identical, in which case the object-I cannot be a product of reflection, and hence the relation cannot be explained through reflection. In other words, I cannot know that what I have grasped through reflection is myself if I have not been acquainted with it before, or otherwise had a criterion to make sure that that what I have grasped is indeed myself and nothing else. A reflexive theory of self-consciousness has to presuppose the reflecting subject that is aware of itself as a thinker, but must also understand that the reflexive relation it enters into with itself is not constitutive of self-consciousness. Alternatively, it has to think of the relationship as constitutive of selfconsciousness. But in this case it remains unclear which criterion necessitates that the relation holds between the two relata (e.g., VIPP, HKA 1/2, 260).37 Fichte’s original point was to counter this problem through conceiving self-consciousness as a pre-reflective consciousness of one-

37

See Henrich (1967, pp. 35-39).

Fichte’s original insight, Schelling’s observation, and Schelling’s point

193

self as ‘knowing subjectivity’.38 Self-consciousness is self-positing, an immediate self-awareness of the knowing subject in its activity, which grounds any reflection and object-awareness.39 Now, while Fichte and Schelling agree that consciousness cannot be explained through self-consciousness as reflection, they differ in their positive theory. Schelling’s observation suggests that it would not suffice to propose a pre-reflective theory of self-consciousness, but his original point against Fichte is that a correct conception of the unconditioned must also be pre-subjective.40 In Fichte’s eyes, the system of science is grounded in the Grundsatz because it is known with absolute certainty; it “is self-evidently certain because its form and content mutually determine each other” (Nassar, 2013a, p. 170). Yet, this epistemological approach is still grounded in a dualism between act and fact; a dualism which Fichte cannot explain further. Schelling observes that “self-awareness implies the danger of losing the I” [Selbstbewusstsein setzt die Gefahr voraus, das Ich zu verlieren] (VIPP, HKA 1/2, 104). To have an “immediate consciousness that I am acting and [of] what I am enacting”—because the content of this act necessarily determines the form in which it is known, and because the form determines what is 38

“This primordial selfhood first allows a Self to work itself free from its connection to the world and to grasp itself explicitly as what it must have been previously, namely, knowledge that what it is, is knowing subjectivity. The possibility of reflection must be understood on the basis of this primordial essence of the Self” (1967, p. 40). 39 Henrich describes this relation between act of production and product with reference to a currency and its magnetic field, i.e., the act of production, in a way, is the product, and not its (temporal) cause (1967, p. 19). Still, he argues, the product must be distinguished from the act, in that the product is pure knowledge of the act and the act its pure ground. To him, I-consciousness is act-consciousness, and it is always already present in consciousness, and not just when made explicit through the summons. 40 This is therefore also where I disagree with Nassar’s reading (2013a; 2016). Nassar argues that Schelling challenges Fichte by showing that “because it precedes the self-reflective, self-determining structure of consciousness, the unconditioned cannot be grasped by it” (2016, p. 127 [emphasis added]). However, Schelling’s challenge really is to the necessarily subjective structure of consciousness, which he thinks is grounded in a neither subjective nor objective structure of self-determination.

194

Schelling’s method of nature-construction

known—still entails a moment of passivity that cannot be overcome.41 The Grundsatz still entails an original duplicity between the epistemic subject-as-subject and the epistemic subject-as-object. In other words, the Grundsatz is only immediately certain if it is still grounded in a dualism between fact and act. It cannot explain how we get from this absolute activity to the special structure through which it is known. It is this original dualism, which results in consciousness and knowledge, that Schelling’s inner and outer forms of the absolute Grundsatz wants to make explicit. For Schelling, there is a gap between a consciousness that is immediately aware of its own self-reflexive activity and this pure activity as absolute causality. Unlike Maimon’s proposal, Schelling does not radicalise Kant’s Copernican framework into a purely epistemological perspective (i.e., one on which metaphysics is the purely formal science of the limits of appearances), but claims that propaedeutic philosophy must take a pre-subjective, and thus ontological, standpoint. This is what I refer to as Schelling’s point. If we consider “that the I is no longer the pure, absolute I once it occurs in consciousness”, we must posit it not as a first structural moment of consciousness but as something prior to consciousness, as an “absolute causality”. The absolute I cannot be identical with the empirical I which is immediately aware of itself as the self-conscious I. The principle in which “thought and being are one” (VIPP, HKA 1/2, 86) is neither subject nor object—it is rather that “which cannot become a thing at all” (HKA 1/2, 90). Everything that can become an object is determined [bedingt], and hence cannot be the unconditional [das Un-Bedingte]. Since the absolute I, as Fichte discusses it, is grasped as an immediate certainty of both subject and object, as a fact that is identified with an act, it cannot be the unconditioned 41 In a letter exchange, Schelling says to Fichte: “I simply cannot imagine that for transcendental philosophy, reality is just something found, nor something found in conformity with immanent laws of intelligence; for in that case, though it may be found according to these immanent laws of intelligence of the philosopher, it would not be the laws of the object of philosophy, which is not that which finds reality, but is itself that which produces it; and truly for the philosopher himself, reality is not something simply found, but only for ordinary consciousness.” (HKA 2,1, 276)

Fichte’s original insight, Schelling’s observation, and Schelling’s point

195

principle, as it is “a determined I and thus, in Schelling’s terminology, an empirical I” (Nassar, 2012, p. 142). Thus, for Schelling, the Grundsatz is determined as the “unconditional positing of reality in itself through its own absolute power [Selbstmacht]” (ÜMFP, HKA I/1, 180). The absolute I, as absolute unconditionality, is still a self-relation, but a ‘self-relation’ in the sense of ‘self-determination’.42 Absolute power relates to itself insofar as it determines itself, as absolute causality and through freedom (VIPP, HKA 1/2, 101). This should make us hesitate. Does Schelling throw away the Kantian insights manifested in the Copernican revolution? We might ask: why should philosophy determine the possibility of a priori cognition from a non-epistemological, pre-subjective standpoint again? But this way of framing the issue would miss the actual import of Schelling’s observation and point. In a similar spirit to metaphysical readings of Kant, Schelling noticed that we cannot determine the possibility of a priori cognition from a purely epistemological standpoint, even if we concede that the objects conform to our forms of cognition and accept that we thereby mean the form and matter of cognition. Neither Schelling’s early philosophy nor his Naturphilosophie should be understood as naïve metaphysics or ontology. 43 It does not consist in a proposal that simply suggests that we should determine things and their essential properties without thinking about how they are determined in relation to our cognitive capacities.44 Rather than just positing self-consciousness as a brute fact that serves as a pre-reflexive condition 42 See (VIPP, HKA 1/2, 89). “The only ‘self’ there is, is misconstrued in both the transcendental- and the naturephilosophy when viewed as empirical, reflective, or transcendental consciousness; rather, the self as principle is the itself, the to auto, of the unconditioned (das Unbedingt [sic]) which cannot be a thing” (Grant, 2006, p. 16). 43 In that sense, I agree with Grant’s assertion that Schelling’s ‘overthrows the Copernican revolution’ (2006, p. 6). However, I do not think that this project “entails the systematic undoing of the critical revolution” (ibid., p. 5), but rather its consequent development. 44 An example of this would be to read Schelling as positing the existence of one class of things (e.g., powers) and then determine all things to share the essential properties that things of such nature exhibit. It has been argued that Schelling’s philosophy should be read as a power-based ontology (see Alderwick (2015)).

196

Schelling’s method of nature-construction

of consciousness in general, Schelling’s philosophy realises that the epistemological standpoint which Kant had proposed in order to solve the metaphilosophical challenge (and which had been developed by Maimon and Fichte) is ultimately grounded in metaphysical assumptions. It is not possible to show the possibility of a priori cognition—and thereby of a rational explanation of an infinitely intelligible world—without first assuming an unconditioned absolute activity. Philosophy, in order to be possible as a science, must be transformed into a special kind of metaphysics that determines the unconditioned of all science. Schelling’s early philosophy, as well as his Naturphilosophie (as we will see in the next section), is concerned with the possibility of a scientific philosophy and with the nature and role philosophy, in its propaedeutic function, must assume. In the course of his early works, Schelling develops a conception of propaedeutic philosophy as the science of the unconditioned per se, in which philosophy must assume an ontological standpoint in order to determine the possibility of science in general. In his writings between 1796 and 1798, Schelling grapples with the problem of the unconditioned and how, based on its adequate characterisation, philosophy can establish the possibility and actuality of ‘real knowledge’.45 This effort pushes him into a new direction, namely toward a philosophy that directly deals with the real we have knowledge of: the philosophy of nature. Still, in these early texts on Naturphilosophie, Schelling sticks to Kantian or Fichtean solutions.46 That is to say, in the Ideas for a Philosophy of Nature (1988/1797), the Treatise Explicatory of the Idealism in the Science of Knowledge (1797), and On the World Soul (1798), nature is genuinely explained as a structural moment of 45 This had already served as one way of addressing the problem in his early philosophy, where he states for example that “knowledge without reality is not knowledge” (HKA 1/2, 85). 46 His solutions qualify as Kantian because they are clearly influenced by the latter’s Metaphysical Foundations of Natural Science (2004b/1786) and his Critique of Pure Judgment (1790). Firstly, Schelling explains matter by virtue of attractive and repulsive forces, and secondly, he intends to overcome a mechanistic account of nature, bridging the gap between mind and world through the concept of the organism. His accounts are nevertheless of clear Fichtean descent, since natural products are explained as manifestations of the self-determining activity of the absolute I.

Naturphilosophie as science of the unconditioned

197

consciousness. In these writings, Schelling repeatedly expresses the view that “the assertion that spirit is not born from matter, but matter is born from spirit” (I/1, 273-4 [emphasis added]) holds true for any philosophy of nature. As he states in the Ideas (1797), the possibility of nature is equivalent to “the all-exclusive world of experience” (IPN, HKA I/5, 11).47 In 1799, these efforts come to an end as Schelling realises that his goal of developing Naturphilosophie must take into account his earlier reflections on the unconditioned and completely abandon the standpoint of consciousness.

3.3 Naturphilosophie as science of the unconditioned Rather than explaining nature with consciousness, Schelling’s new philosophical approach explains consciousness with nature; or, in his own words, consciousness and its objects are explained solely through the construction of “higher and necessarily unknown natural forces” (E, HKA 1/8, 31). By taking seriously his own assumptions about the proper subject matter of propaedeutic philosophy, i.e., as consisting in a pre-subjective, absolute causality, Schelling develops a new method that enables first philosophy to determine the possibility of all sciences, including itself: the method of nature-construction. This section is dedicated to explaining how Schelling translates his reflections on the unconditioned to the subject matter of nature, and consequently, how philosophy can treat of nature without conceiving nature as something inherently conditioned or given. In proximity to Maimon, Schelling proposes a philosophy of nature that is grounded in a first unconditioned principle (i.e., the Grundsatz) that must necessarily be posited into nature in order for it to be possible. While this principle is necessary for the possibility of philosophy and science, it is also impossible to immediately demonstrate its reality. Therefore, Schelling proposes to posit the ultimate Grundsatz of nature as a methodological principle, i.e., as 47 Also consider this quotation from his preface to the Ideas: “The pure theoretical philosophy concerns itself only with the investigation into the reality of our knowledge as such; it belongs, however, to the applied, under the name of a Philosophy of Nature, to derive from principles a determinate system of our knowledge (that is, the system of experience as a whole)”. (HKA 1/5, 61)

198

Schelling’s method of nature-construction

an “absolute hypothesis”, on the grounds of which only nature can be “constructed”. Thus, in order to explain the possibility of philosophy, propaedeutic philosophy must become Naturphilosophie, and in order to determine the unconditioned, Naturphilosophie must employ the method of the experiment of reason. As part of his lectures as newly appointed professor at the University of Jena,48 Schelling in March 1799 published his First Outline of a System of the Philosophy of Nature in March 1799, and shortly afterwards added to it an Introduction. Although he had published on philosophy of nature before 1799, and would continue to publish on this topic for most of his career,49 in this work Schelling expresses his conviction that he has arrived at a completely original position. In the Preface to the Outline, he notes that “[t]his treatise may surely be called a first outline, because no attempt of its kind has previously existed” (ESN, HKA I/3, 4). In agreement with this assessment, the following section will focus on Schelling’s (1799) system in its capacity as his first attempt to develop Naturphilosophie as an independent science, i.e., as grounded in a Grundsatz that ultimately reveals itself to be the Grundsatz of all science, making Naturphilosophie the first science he had been looking for in his early philosophy. The second step, i.e., turning Naturphilosophie into first philosophy, takes place mainly in the Introduction. Although 48

See Richards (2002, pp. 147-151) for historical background and the role of Goethe, who was a major factor in Schelling’s appointment and involvement in Jena. 49 Schelling’s first publication in this respect is his Timaeus interpretation of 1793. It has been argued that Schelling’s Naturphilosophie should be traced back to this text (see Buchheim (1987) and Baum (2000)). It has also been claimed that Naturphilosophie remains the central topic which guides and motivates all of Schelling’s later works (see Grant (2006, p. 14)). Schelling’s Naturphilosophie after 1799 goes hand in hand with developing his system of identity philosophy. This system heavily relies on a new conception of construction through intellectual conception, which he explicitly lays out in his essay On Construction (1802). As a consequence of his changed views on the possibility of intellectual intuition, Schelling drops the project of an experiment of reason. According to this later conception, philosophy can immediately demonstrate the reality of its concepts, and thus is not dependent on testing their validity through experiment; see Nassar (2013a, pp. 225-256) for some illuminating thoughts on this matter.

Naturphilosophie as science of the unconditioned

199

it is not known when exactly he drafted the Introduction, it can be placed roughly between April and June of the same year that he had published the Outline (1799). As he explicitly states at the outset, the Introduction serves as an illumination of the Outline, and as such does not follow a strict argumentative path (E, HKA 1/8, 40). Rather, it revisits the programme of philosophy by clarifying, step by step, the concept of Naturphilosophie as a science, and therefore qualifies as a metaphilosophical work in the narrower sense of this term. With regard to Schelling’s philosophical development, this work sticks out as the one which probably developed in closest collaboration with Johann Wolfgang von Goethe,50 which also explains why the Introduction has often been discussed in this context.51 For instance, the Introduction is discussed in relation to the development of some of Goethe’s core ideas, such as the concepts of intuitive intellect, metamorphosis, and archetype.52 Without disagreeing with this scholarship, I will explain Schelling’s methodological programme as deriving from his earlier philosophy and the metaphilosophical reflections of Kant and Maimon, as well as his particular philosophy of experiment. Although this programme fits well with Goethe’s thoughts on intuitive understanding, its point as a methodological solution to the metaphilosophical challenge is exactly that we do not need to postulate any special cognitive capacity. Let us come back to the beginning of this section. Schelling begins the Outline by raising the question of whether philosophy of nature can be described in such terms that it qualifies as a science of its own. In order to declare it a science, the philosopher of nature would have to present a Grundsatz for philosophy of nature; “[t]he question arises as to what

50 Due to Goethe’s efforts, Schelling received Fichte’s old post at Jena University and prepared his first lectures about Naturphilosophie. During the course of their collaboration, Goethe invited Schelling to join him in October 1799 to discuss the Outline and Introduction (see Nassar (2013a, pp. 193-5)). 51 See Nassar (2010, 2013a) and Richards (2002, 2006). Nassar, in particular, shows that Förster’s account would have done well to include, rather than dismiss, Schelling’s work in relation to Goethe’s conception of anschauender Verstand. 52 See for instance (HKA 1/8, 33, 71).

200

Schelling’s method of nature-construction

extent unconditionedness might be ascribed to nature” (HKA 1/7, 77).53 In continuity with his early science of the unconditioned, Schelling builds up his new account by first exposing it to the metaphilosophical challenge: how can Naturphilosophie be conceptualised as a science? And in continuity with his early philosophy, he suggests that nature should be understood in analogy to the absolute I: as unconditioned, or absolute activity. To further determine what it would entail to treat nature as unconditioned, Schelling goes back to his previous definition of the nature of the unconditioned of science. According to that definition, remember, “[t]he unconditioned cannot be sought in any individual “thing” not in anything of which one can say that it “is” ” (ESN, HKA 1/7, 77). One way to say that something is unconditioned is to say “that it is not conditioned by anything else”. Schelling connects this negative definition to the linguistic meaning of the German word for unconditional (unbedingt), which also contains the meaning ‘un-thinged’ [un-bedingt] (ibid.).54 Schelling uses this to illuminate why the unconditioned of a science can never be an object or thing, i.e., a particular being. This is so because “what ‘is’ only partakes in being, and is only an individual form or kind of being” (ibid.). What is unconditioned is never countable, or otherwise identifiable as one thing that is opposed to another. Rather than being an instance of being that can be differentiated from other instances by virtue of sharing some attributes but not others, the unconditioned can only consist in the principle of being, in virtue of which such differentiation and individuation is possible. In his early philosophy, Schelling argued that the only way for something to be determined and not conditioned is for it to be self-determined. Hence, an unconditioned philosophy cannot begin with a thing or object that 53 “The subject which is to be the object of philosophy in a given instance must be viewed, in a word, as unconditioned” (E, HKA 1/8, 77). 54 “[S]ince the unconditioned cannot be thought under the predicate of being, it obviously follows that as principle of all being, it can participate in no higher being. For if everything that ‘is’ is only, as it were, the color of the unconditioned, then the unconditioned itself must everywhere become manifest through itself— like light that requires no higher light in order to be visible” (ESN, HKA 1/7, 77).

Naturphilosophie as science of the unconditioned

201

is determined by other things or objects, but only with something that self-determines, i.e., a principle of all being. So, what does this mean for Naturphilosophie? Consider the following remark: The philosopher of nature treats nature as the transcendental philosopher treats the self. Thus Nature itself is unconditioned to him. This is not possible, however, if we proceed from objective being in Nature. In philosophy of nature objective being is as little something originary as in transcendental philosophy. (ESN, HKA 1/7, 78)

In contrast to Kant, Maimon, and also Fichte before him, Schelling proposes to conceive of nature fundamentally neither as something given, nor as something objective that depends on the forms of cognition, or the epistemic subject. Nature as unconditioned is not for consciousness; it should not be explained as an object that is generated according to the rules of consciousness. Schelling’s point is exactly that if we want to explain how it is possible that something is for consciousness and how a priori cognition of it is possible, then we must approach that which is for consciousness, namely nature, from a scientific perspective, too. If there should be a science of nature, then this science must begin by asking whether “unconditionedness can be ascribed to Nature” (ESN, HKA I/7, 77), which again means that the first principle of nature “cannot be sought in any individual natural object” (ibid.). In order to explain how something is for consciousness, i.e., how we as epistemic subjects can have knowledge of objects, we need a principle that is pre-subjective.55 55

Hence, Förster’s criticism (2018, pp. 234-252) completely misses Schelling’s line of argumentation. Schelling does not adopt Fichte’s conception of an absolute knowing subject as the ground of everything, and hence also does not think that philosophy can proceed by virtue of an intellectual intuition relying on the subject’s special epistemic relation of its own constitutive mental acts. Förster’s objections, namely that (i) the subject and object cannot be(come) identical in Naturphilosophie, and that (ii) Fichtean intellectual intuition cannot be employed because it is only possible by virtue of an abstraction from everything that is objective (such as epoché), therefore miss the mark of Schelling’s project. In fact, I would argue that the specific methodological conception offered at least in the 1799 Introduction does not need to rely on any conception of intellectual intuition because it resolves the problem of nature through a methodological conception.

202

Schelling’s method of nature-construction

Now, from the standpoint of consciousness, “everything that exists [i.e., being as nature] is a construction of the spirit” (ESN, 1/7, 78’). From the unconditioned standpoint of nature, however, being itself [i.e., Nature] is nothing other than the constructing itself, or since construction is thinkable at all only as activity, being itself is nothing other than the highest constructing activity, which, although never itself an object, is the principle of everything objective. (ibid. [emphasis added])

For a start, the unconditional principle of all being, nature’s Grundsatz as “absolute unconditionality” (HKA I/1, 96) should be understood as the “highest constructing activity”. Already in his early philosophy, Schelling had separated the absolute activity, which is the unconditioned, from the conscious activity of the epistemic subject. To be unconditional means not to be a thing, i.e., not to be determined by anything else. The only way to understand a principle as unconditioned, then, is to understand it as consisting in absolute activity. Therefore, Nature as “the concept of all being [Inbegriff alles Seyns]” is the highest constructing principle of the sum total of natural objects. It would “be impossible to view Nature as an unconditioned, if the concealed trace of freedom could not be discovered in the concept of being itself”, and this means that in our science of nature, i.e., in Naturphilosophie, “Nature has to be viewed as absolutely active” (E, HKA 1/8, 79). The unconditioned principle of nature must not be posited from the perspective of a constructing consciousness, but from the pre-subjective standpoint of nature itself, resulting in a principle of nature as “its own legislator”, i.e., as self-determining, and as something that “suffices for itself”; that is only determined by itself (HKA 1/7, 81). In this way, we arrive at a conception of nature that enables a scientific understanding of its subject matter: To philosophize about nature means to create Nature. […] Thus we do not know nature as product. We know Nature only as active—for it is impossible to philosophize about any subject which cannot be engaged in activity. To philosophize about nature means to heave it out of its dead mechanism to which it seems predisposed, to quicken it with freedom and to set it into its own free development—to philosophize about nature means, in other words, to tear yourself away from the common view which discerns in nature only what “happens”—and which, at most views the act as a factum, not the action itself in its acting. (ESN, HKA 1/7, 78-79)

Naturphilosophie as science of the unconditioned

203

Because to philosophize about nature means as much as to create it, we must first of all find the point from which nature can be posited into becoming, (ESN, HKA 1/7, 67 [emphasis added])

Naturphilosophie is possible only if nature is understood as a productive principle whose self-determining activity generates natural products. Unlike in Kant’s philosophy, Schelling’s nature does not consist of facts which, as extensive and intensive magnitudes, are ordered spatiotemporally. Rather, we can only explain natural products if we begin from absolute activity as the unconditional principle that generates particular objects. Schelling also maintains a commitment to infinite intelligibility and complete explanation; after all, he defines the first principle of science as one that provides science with an ultimate ground. Unlike Maimon, however, he presents a strategy that enables the philosopher to treat philosophy as an independent science that need not be reduced to the cognitive activity of an infinite mind. He shows that any explanation that begins from the standpoint of consciousness must still assume a self-determining activity that knows of itself through an immediately certain fact. Naturphilosophie begins with this self-determining activity not as fact-act, and thus first moment of consciousness, but as a constructive but pre-subjective process. On these grounds, it becomes clear how Schelling arrives at the metaphilosophical thesis that, instead of a philosophy that begins from the standpoint of consciousness (as Kant, Maimon, and Fichte propose), only Naturphilosophie can assume the role of a propaedeutic or first philosophy. In the Introduction, Schelling clarifies how this philosophy’s mode of explanation differs from his previous drafts, namely insofar as it proposes to explain “the ideal […] from the real” (E, HKA 1/8, 29). The form of explanation endorsed by Naturphilosophie is not idealistic, since it does not explain nature as determined by the structure of our consciousness.56 Instead, “[t]he first maxim of all true natural science, 56 “It follows naturally from this that there is no place in this science for idealistic methods of explanation, such as transcendental philosophy is fitted to supply, since for it Nature is nothing more than the organ of self-consciousness, and everything in Nature is necessary merely because it is only through the medium of such a Nature that self-consciousness can take place. This mode of explanation, however, is meaningless for physics (and for our science which

204

Schelling’s method of nature-construction

[is] to explain everything by the forces of Nature”; reason itself is to be explained as “a mere play of higher and necessarily unknown natural forces” (E, HKA 1/8, 31). Just like Maimon’s rational dogmatism, Naturphilosophie suggests a monistic conception—nature is to be explained from one kind of principle. By setting out to explain everything from natural forces, however, Schelling pushes a naturalistic agenda. As “speculative physics”, Naturphilosophie is not opposed to natural science but continuous with it.57 Speculative physics delivers a conception of nature that grounds all science in a fundamental set of concepts and principles, and thereby explains the legitimacy of the theoryconstruction and empirical data production of the individual sciences. Still, Naturphilosophie is non-objective and is not concerned with “the surface of nature”, i.e., natural objects and phenomena (E, HKA 1/8, 33). To integrate empirical natural sciences into Naturphilosophie, the latter must consist in a system of science that includes a theory on how natural phenomena (i.e., the objective part of science) derive from nonobjective first principles. Schelling’s version of ‘there is no given’ does not conceive of nature as constructed by the mind, but rather nature itself is understood as a productive principle. His conception of nature as self-contained absolute, as autonomous realm of being that is structured by its own immanent principles, achieves such an explanation without reference to cognitive faculties. What’s more, it is exactly because of this revised conception of a monistic system that Naturphilosophie can said to appropriately realise the import of Schelling’s earlier methodological reflection, i.e., that “every science that is science at all has its unconditioned” (ESN, HKA 1/7, 77). While the individual sciences all have an unconditioned that is merely presupposed and the burden of justification shifted, Naturphilosophie comes with an immanent justification, i.e., one which is not a derivative of another realm of being occupies the same standpoint) as were the old teleological modes of explanation, and the introduction of universal reference to final causes into the science of nature, which was adulterated as a result. For every idealistic mode of explanation, dragged out of its own proper sphere and applied to the explanation of Nature, degenerates into the most adventurous nonsense, examples of which are well-known.” (E, HKA 1/8, 30-31) 57 Schelling also describes it as a “spinozism of physics”. (E, HKA 1/8, 30)

Positing an absolute hypothesis

205

such as cognition or consciousness (e.g., the unity of apperception or the absolute I). Despite the obvious parallels with his early propaedeutic philosophy, Schelling’s Naturphilosophie brings significant changes to his metaphilosophical outlook. In particular, instead of assuming that this first unconditioned principle is constitutive of all science, Schelling aligns his Naturphilosophie with the by now familiar family of metaphilosophies that assume a methodological principle as an ultimate ground. Schelling thinks that Naturphilosophie needs to answer both the quid juris and the quid facti. On his view, nature is only possible on the grounds of a methodological principle, namely the one positing nature as the ‘original duality between productivity and product’ (HKA 1/8, 34). All objects of science must be understood as results of nature’s self-production, and it is the job of Naturphilosophie to outline the universal structure of this process. In the next section, I will discuss why Schelling thinks that nature can only be explained in this way if it is grounded in a methodological principle.

3.4 Positing an absolute hypothesis In contrast to his early philosophy, Schelling realises that, although he has found the “point from which nature can be posited into becoming”, this point cannot be posited as the first constitutive principle of nature, but only as the “first postulate of all philosophy of nature” (ENS, HKA 1/7, 67). In other words, to conceive nature as absolute activity, we must assume or presuppose an unconditioned principle without being able to immediately show its reality. Rather, this first principle remains methodological in nature, and can only be assumed as problematic. In the Introduction, Schelling says that we must “either give up all attempt ever to arrive at a knowledge of [the final causes of natural phenomena], or else we must altogether put [a first principle] into Nature, endow Nature with [it]” (E, HKA 1/8, 34). Furthermore, that which we put into Nature [in Natur hineingelegt] has no other value than that of a presupposition (hypothesis), and the science founded upon it must be equally as hypothetical as the principle itself. (ibid.)

206

Schelling’s method of nature-construction

If Naturphilosophie begins with a methodological principle, then (at first) both the principle and the science have only a hypothetical status. In this case, however, Schelling’s first science would neither be able to answer quid juris? nor quid facti?—all Naturphilosophie would amount to would be one possible system of nature, whose actual application to nature is problematic. Simultaneously, this first principle would not satisfy Schelling’s criteria for an unconditioned, since hypothetical principles depend on further evidence: something else that demonstrates their reality. Consequently, there must be something more about this hypothesis with which the philosopher endows nature, and indeed there is: the first principle of Naturphilosophie has the character of an absolute hypothesis. In the Introduction, Schelling argues that the only way to circumvent the merely hypothetical status of principles ‘put into nature’ is to find a principle that as a presupposition is “as necessary as Nature itself” (E, HKA 1/8, 34). His thought seems to be very similar to Maimon’s. Maimon argued that the possibility of philosophy as a science was only explicable on the grounds of this one particular fiction, i.e., that of an infinite intellect which creates its objects of cognition. In a similar fashion, Schelling argues that the unconditioned principle of nature is not hypothetical in the sense that it is one out of many potential explanations under which nature becomes creatable and hence possible. On the contrary, it is not possible to begin Naturphilosophie without positing exactly this principle, which delivers the only possible description under which nature becomes constructible. In a way, then, the necessity of the principle follows from its possibility. Admittedly, this argument initially sounds a lot like an argument from conceptual necessity: the concept of nature necessarily entails the concept of universal duplicity. Just as Kant derives the thing-itself from coherently thinking the notion of appearance, it might seem that Schelling employs a form of transcendental argument that argues from the possibility of having one concept to its necessary conditions of possibility. And yet, this is not Schelling’s strategy. The possibility at stake is not of a conceptual but metaphysical nature. If nature is to be metaphysically possible, then nature) must be such that the principle of universal duplicity holds. We find evidence for this reading later on in the text,

Positing an absolute hypothesis

207

once Schelling engages with the a priori/ a posteriori distinction.58 He argues for reversing Kant’s argument, saying that “WE KNOW Nature as a priori, [because] Nature IS a priori” (E, HKA 1/8, 36). The insight into the a priori character of some knowledge claim is really an insight into the a priori character of the principle whose truth it claims. Schelling’s example is the principle of nature as an organic whole, in which “all things mutually bear and support each other”, and which must have been prior to its parts (“the whole could not have arisen from the parts” (ibid.)). Now, this principle is not a priori because we know nature as a priori, but because nature as an organic whole is a priori. The apriority claim is thus one of metaphysical priority. The theoretical status of this hypothesis can be further elaborated with reference to Maimon’s method of fictions. In a letter dating from 19th November, 1800, Schelling defends the nature and method of Naturphilosophie against Fichte’s doubts: But if you were to [...] then say that the philosophy which I call purely theoretical is precisely the science that you speak of in your letter, namely, one which would make nature alone its object through free abstraction, and then permit it to construct itself through a (justifiable) fiction [erlaubte Fiktion], this is entirely and absolutely my view. (HKA 2/1, 281, as translated in Vater & Wood (2013, p.45))

Schelling explains his method for Naturphilosophie as making use of a scientific fiction.59 Similarly to Maimon, “erlaubte Fiktionen”, for Schelling, denotes theoretical concepts or models that cannot be exhibited or proven to completion, but which must nevertheless be assumed for the sake of being able to explain the ground of natural appearances and the way in which the rational explicability of nature can be conceived. In response to Fichte’s criticism concerning his treatment of 58 Schelling’s claim in this respect is that apriority/aposteriority are not properties of judgments, but pertain to the “kind of knowledge of these judgments” (HKA 1/8, 35). Although scientists investigate natural phenomena through experience, they come to learn of the necessary character of a judgment of experience. Just because I learn from observation that bodies fall does not mean that the law of falling bodies is a posteriori, but rather that I can get insight into its a priori necessary character. 59 As far as I am aware, Schelling’s connection to this discourse about the method of fictions has escaped scholarly attention.

208

Schelling’s method of nature-construction

nature, i.e., as something that can be explained independently from consciousness, Schelling embraces the latter’s talk of scientific fictions, making a proposal for how this method should be applied to his system of Naturphilosophie.60 The philosopher ought to produce a scientific fiction through free abstraction, that is, by means of reasoning that is not constrained by intuition, and which exactly operates outside of the bounds of ordinary experience. Since Naturphilosophie doesn’t proceed from the standpoint of the self-reflective consciousness,61 it has to posit the useful fiction of nature as abstracted from consciousness, that is, a model of nature in which it is treated as if it were conceivable from an unconscious and pre-subjective perspective.62 This fiction is the absolute hypothesis of Naturphilosophie: the concept of nature as a universal duplicity of productivity and product (E, HKA 1/8, 34).63 Schelling’s idea is not that we can actually think nature from the standpoint of the unconscious, but rather that we posit this standpoint 60

Fichte was directly inspired by Maimon’s conception of useful and necessary fictions in philosophy, and in various places suggests an interpretation of the Wissenschaftslehre as grounded in a fictional principle. On Fichte’s use of the method of fictions, see Breazeale (2002) and Crowe (2008). 61 In contrast, Fichte’s point is exactly that one can only abstract until arriving at the I, which again is the one thing that one cannot abstract from. In other words, we can bracket everything about a thought, except for our thinking that thought. 62 I am not interested here in Schelling’s method of depotentiation, which becomes his main method for the Naturphilosophie, and which is also part of his identity system (1801-1803). For this system, he also suggests a different account of construction (1802) (see for instance Buchheim (1990), Whistler (2013, pp. 123-137), and Ziche (2011)). However, I would suggest that we should read this process as based on an employment of scientific fictions. Schelling agrees with Fichte’s methodological conviction that the first principle (the absolute I) is initially only posited (through abstraction) as a fiction; its complete representation remains an ideal to approach. Whistler and Buchheim’s engagements are especially interesting here, as they both (and in independent ways) connect depotentiation to the phenomenological method of epoché (bracketing). 63 Beiser (2002, p. 487) argues that, in the Introduction, Schelling does not understand the autonomy of nature as “mere fiction” any more. This description can easily fall short of Schelling’s account because it isn’t precise about the exact nature of that fiction. Although Schelling no longer argues for ‘nature’ as inexplicable, given matter in the Introduction, he does suggest that we ought to understand the first principle of nature as a fiction in a pragmatic sense.

Positing an absolute hypothesis

209

by approaching it, in just the same way as Maimon had worked with the idea of an infinite intellect. That is, although we cannot really take up this standpoint, we can employ it as a fiction and treat nature as if it were constructible from a pre-subjective and pre-reflexive perspective. Schelling’s reasons for choosing this fiction are as discussed above: philosophy is only possible if it can be grounded in an absolutely unconditional principle—otherwise we cannot make intelligible how objects are explicable according to scientific standards (i.e., scientific explanation is complete explanation, committed to infinite intelligibility, and thus must be grounded in a pre-subjective principle). Nature as this unconditioned can only be conceived as self-productivity. Further, in order for the first principle to explain itself, as well as all of its conditions, it must both explain nature as absolute activity and nature as the sum total of natural products. In contrast to Kant and Maimon’s model, this explanation must not begin with natural facts or objects: “[o]riginally, no individual being at all (as an accomplished fact) is present for us in Nature, for otherwise our project is not philosophy, but empirical investigation” (ENS, HKA 1/7, 78 [emphasis added]). While any science that investigates particular products would qualify as an empirical science, Naturphilosophie must explain how natural products arise from this absolute activity, and how they eventually become objects of consciousness—this is what it means to ‘create nature’ in a scientific manner. Naturphilosophie must begin with a first fictional principle which contains an original duality between productivity and product. Although fictional, this principle is neither arbitrary nor contingent. Rather, if one wants to conceive the possibility of nature, it is necessary that one conceive it as grounded in an original duality between productivity and product. That nature must be explained as being grounded in a ‘universal duplicity of productivity and product’ is not derived from contingent facts about the cognitive make-up of finite thinkers (e.g., that nature must be conceived as inhibited productivity in product because finite thinkers can only know of finite products), but from ‘the nature of Nature as unconditioned’ itself. Schelling differentiates ‘Nature’ from ‘world’ in order to distinguish between two conceptions of nature, the first of which is connected to philosophy’s perspective, and the second

210

Schelling’s method of nature-construction

of which is connected to the natural sciences. While the latter denotes nature as “world”, i.e., objective nature as the unified whole of natural products, the spatiotemporal order of the objects known to finite cognition, Schelling says of this conception of nature that it is a “mere product (natura naturata) [that] we call Nature as object (with which alone empiricism deals)” (E, HKA 1/8, 41). Nature as absolute activity, i.e., as a continuous and infinite process of self-determination, from which subjectivity and objectivity arise, Schelling calls “Nature as productivity (natura naturans)”, and also (very suggestively) “Nature as subject (with this alone all theory deals)” (ibid).64 As Natura naturans, nature is defined as greater than its parts, and as irreducible to the composite of its parts. Natura naturans (productive nature) is the underlying and formative principle of nature that determines, and is expressed in, all of nature’s parts, and is that which can only be an unconditional first principle because it can never become objective.65 On Schelling’s elaboration, [a]ssuming, for example, what must be assumed, that the sum of phenomena is not a mere world, but of necessity a Nature (that is, that this whole is not merely a product, but at the same time productive), it follows that in this whole we can never arrive at absolute identity, because this would bring about an absolute transition of Nature as productive into Nature as product, that is, it would produce absolute rest. Such a wavering of Nature, therefore, between productivity and product, will necessarily appear as a universal duplicity of principles, whereby Nature is maintained in continual activity, and prevented from exhausting itself in its product; and universal duality as the principle of explanation of Nature will be as necessary as the idea of Nature itself. (E, HKA 1/8, 34)

Thus, the unconditioned first principle must contain an original duality because otherwise we could neither explain how there is at the same time absolute activity and a multiplicity of natural objects. If we accept that nature is only possible if explained from an unconditional principle, 64 Schelling’s adoption of Spinoza’s terminology in this context is ultimately only a superficial one; his understanding of nature as an infinite productive force presupposes a different notion of contingency than Spinoza’s, i.e., that of intentional causation—nature acts and thereby determines itself freely (Ostaric 2014, p. 64). 65 Nassar argues that this underlying principle just is the principle of metamorphosis; see (2013a, p. 200; E, HKA 1/8, 33, 71).

Positing an absolute hypothesis

211

and this again means that explanation is only possible if nature can be conceived as some sort of absolute productivity, we must also accept that this productivity also somehow inhibits itself, otherwise we could not explain how a multiplicity of natural products, i.e., finite objects, arise from this productivity. At the same time, Nature must always be “a wavering […] between productivity and product” (E, HKA 1/8, 34), otherwise it would either never come into existence at all, by virtue of not individuating itself, or it would cease being productive by virtue of exhausting itself in its product. Therefore, Schelling determines nature’s unconditioned self-productivity as infinite productivity and infinite limitation. Their mutual dependence explains the necessity of an unconditioned principle that contains an “original duplicity of principles”. Still, to verify that his system is not “an empty play with concepts” (ESN, HKA 1/7, 79), but actually constitutes a version of the experiment of reason, consider the second, epistemological, part of Schelling’s argument. Naturphilosophie’s new way of conceiving nature as absolute activity still poses another problem, namely that of quid facti.66 That is to say, even if it is shows that nature is only possible as science if conceived as absolute activity, we still have to ask whether we can have any proof of the reality of this principle (quid facti?). Since his new philosophy clearly excludes the standpoint of consciousness and the transcendental subject that has access to the world, as this world is its own construction, it is not immediately clear why this conception of nature should actually apply to the order of the natural world, nor to the laws of our consciousness for that matter. Besides the fact that he considers this unconditioned principle to have methodological status, Schelling has some additional thoughts on how the conception of nature’s unconditioned is connected to our epistemic capacities, whereby he carves out a revolutionary route to a procedure by virtue of which philosophy can construct nature as an infinitely-becoming product of (natural) science. 66 “We have answered the first question (how unconditionedness may be ascribed to Nature) through the assertion that Nature has to be viewed as absolutely active. This answer drives us to the new question: how can Nature be observed as absolutely active, or more clearly expressed: in what light must the totality of nature appear to us, if it is absolutely active?” (ESN, HKA 1/7, 79).

212

Schelling’s method of nature-construction

3.5 The Method of Nature-Construction While the previous section revealed that Naturphilosophie has to posit the methodological fiction of Nature as infinite self-productivity, this section will establish the special procedure according to which Naturphilosophie proves the reality of this methodological principle. As you might have noticed in quotations presented earlier, Schelling’s methodological paradigm of construction is linked to a procedure that enables a very particular kind of cognition: a kind of cognition in which the act of cognising something is identical with the act of creating that something. This epistemological ideal of cognition, which is equivalent to how divine cognition has sometimes been described, is connected to the methodological paradigm of construction, which you might remember from our investigations of Kant and Maimon’s metaphilosophies. The programme of philosophical construction runs through the work of many Post-Kantian philosophers and, although it has been treated in several philosophical works, remains a fairly opaque point in German idealism scholarship.67 Schelling’s method of construction emerges in the context of Kant’s discussion of mathematical construction and Fichte’s method of constructing consciousness.68 Drawing from these resources, and meshing them with his own thoughts about the necessary nature of theoretical philosophy and his philosophy of science, Schelling develops an original method for philosophy that offers a new perspective on the possibility and task of philosophical science. Although scholars have agreed that “[c]onstruction is a key theme 67 To my knowledge, there are two book-length discussions of this method, i.e., Ende (1973) and Taureck (1979), which both look at different articulations of this programme within the philosophies of German idealism. 68 After Kant’s (and also Maimon’s) clear rejection of a philosophy that constructs its concepts, Fichte’s revolutionary account of the absolute I, which, by virtue of determining the form and content of particular objects, can be claimed to construct its concepts. Ultimately, the possibility of construction is grounded in the absolute I’s capacity for intellectual intuition, i.e., an immediate awareness of myself in my acting, which is only realised in the realm of the practical. Thus, the possibility of philosophical construction in the realm of the theoretical must remain an idea that the I strives to realise in the realm of the practical. On Fichte’s method of construction, see Neuhouser (2014), Breazeale (2001; 2014), and Schmid (2020).

The Method of Nature-Construction

213

throughout Schelling’s writings” (Nassar, 2013, p. 252), surprisingly little scholarship has been devoted to this issue. The existing scholarship mainly focuses on Schelling’s method of construction after 1800, in particular its function within the latter’s ‘identity philosophy’. From those who have engaged with his method of construction in Naturphilosophie before 1800 (i.e., in his Outline and Introduction), 69 most have failed to understand the originality and independence of Schelling’s method of philosophical construction. By framing his philosophy as a form of transcendental philosophy, and thus conceiving the construction of nature as ultimately consisting in a construction of consciousness, the conventional approach explains Schelling’s method as a misguided imitation of Fichte’s method of philosophical construction.70 Against these readings, I argue that Schelling makes a genuine methodological proposal for constructing nature (and not consciousness) from a non-subjective standpoint that still generates a knowledge of nature as self-produced. In order to understand how this method of constructing nature is possible, we need to understand how he connects the possibility of Naturphilosophie to its presentability [Darstellbarkeit]. In continuity with Kant, Schelling contends that “[w]e know only the self-produced; knowing, therefore, in the strictest sense of the term, is a pure knowing a priori” (E, HKA 1/8, 34). To know something a priori is to know that it is necessarily so. To know something with necessity, we need to know it as “self-produced” (HKA 1/8, 34). In Kant’s model, this meant that we know something because the object of knowledge is determined a priori in virtue of our forms of cognition. Under Maimon’s model, this conception has been radicalised to entail the view that all objects of cognition with regard to their form and 69 These include, amongst others, Ende (1973), Krings (1982; 1985), Löw (1979), and Rudolfi (2001). Another large chunk of work on Schelling’s method of construction focuses on its articulation in his essay “On construction” (1802), as well as its relation to identity philosophy; see Taureck (1979), Ziche (2011), and Breazeale (2014). 70 A notable exception is Nassar (2013a, pp. 202-9), who offers a characterisation of Schelling’s method of ‘successive construction’. Although she also connects this account to Kant’s description of mathematical construction and Schelling’s theory of experiment, I do not agree with her reconstruction of either. I will indicate how my account differs from hers in the relevant places.

214

Schelling’s method of nature-construction

matter are generated according to the rules of an infinite intellect. To know these objects is to know their rules of generation. Now, within Schelling’s model, we need an account of how it is possible “to get a glimpse of the internal construction of nature” (E, HKA 1/8, 33) without viewing nature as the construction of a mind that ultimately has the same structure as ours. Rather, it must explain how “all phenomena are correlated in one absolute and necessary law, from which they can be deduced; in short, that in natural science all that we know, we know absolutely a priori” (ESN, 1/7, 35). So, how can we explain nature as an order of infinite intelligibility that is grounded in one principle, but still understand this procedure as construction, i.e., as the creation of objects? Before fleshing out his mature conception of nature-construction in the Introduction, Schelling makes some insightful remarks in the Outline concerning the epistemic accessibility of nature’s absolute activity. There, he connects the possibility of philosophy and science in general not only to the positing of a first unconditional principle, but also to the possibility of its presentation: According to the first principle, an original duality must simply be presupposed in Nature. For it permits no further derivation, because it is the only condition under which an infinite is finitely presentable at all, i.e., the condition under which Nature is at all possible. (ESN, HKA 1/7, 81 [emphasis added])

Naturphilosophie is possible only if it presupposes the original duality between productivity and product, and it is on this presupposition only that the infinite (productivity) is finitely presentable (through its products). In a Kantian spirit, Schelling here ties the possibility of nature to the possibility of its being cognised by finite thinkers, equating “nature is possible” with “nature is presentable [darstellbar]”, which again is equivalent to saying that nature “is cognisable by finite thinkers”.71 How nature, as grounded in an original duality, can be 71 Also consider this quotation: “The Philosophy of Nature, so that it does not degrade into an empty play with concepts, must demonstrate a corresponding intuition for all of its concepts. Therefore, the question arises how an absolute activity (if there is such a thing in Nature) will present itself empirically, i.e., in the finite. —Possibility of the exhibition of the infinite in the finite—is the highest problem of all systematic science” (HKA 1, 7, 79).

The Method of Nature-Construction

215

thought of as presentable becomes clear when we look at a second sense in which Schelling describes this duality: the original duality between an infinite productivity and its self-limitation.72 Natural science, as we know it, deals with empirical objects, particular products of nature, and the universal laws that determine the way these objects enter (causal) relations. Only in philosophy are we required to show that all products are ultimately just “apparent products” and nothing more than momentary inhibitions of the absolute productivity of nature.73 Naturphilosophie understands the particular objects of natural science as created; they are conceived of as the products of an infinite productive activity, which never exhausts itself in its products. This is the unconditional principle which Schelling posits into nature. He states that this “[p]roductivity is originally infinite; thus even when a product comes to be, this product is only an apparent product. Each product is a point of inhibition, but the infinite still “is” in each point of inhibition” (ESN, HKA 1/7, 80).74 Nature, as self-determining activity, must be thought of as an activity that limits itself continuously and thus “appears as inhibited ad infinitum” (ESN, HKA 1/7, 81). From this perspective, it becomes clearer how Schelling intends to explain the “[p]ossibility of the exhibition of the infinite in the finite” on which the possibility of Naturphilosophie depends (ESN, HKA 1/7, 79).75 From the perspective of finite cognition, the only way to explain nature as grounded in an ultimate axiom is to conceive of nature as empirically72 Jacobs notes that Schelling had not used the notion of productivity in the first version of the Outline, but only in later ones, as well as in his handwritten notes (see (HKA 1/8, 10)). Nassar suggests that this change is connected to the latter’s engagement with Goethe, since the term appears from the beginning of the Introduction (2013a, p. 311). 73 “In the usual view, the original productivity of nature disappears behind the product. For us the product must disappear behind the productivity” (HKA 1/7, 78). 74 Schelling explains this point in a similar manner to Maimon’s explanation of qualities as infinitesimals, as is explained below. 75 In fact, Schelling asserts that the “[p]ossibility of the exhibition of the infinite in the finite—is the highest problem of all systematic science” and that the “subordinate sciences solve this problem in particular cases” (ESN, HKA 1/7, 79).

216

Schelling’s method of nature-construction

infinite: “Absolute activity cannot be presented [dargestellt]76 by a finite product, but only by an infinite one” (ESN, HKA 1/7, 79). Schelling argues that nature as can only be presented through a finitude which is never complete, i.e., which is itself infinite. In other words, it can only be presented by an infinite becoming, where the intuition of the infinite lies in no individual moment, but is only to be produced in an endless progression—in a progression, however, which no power of imagination can sustain. Therefore, reason determines either to obliterate the series, or to assume an ideal limit to the series which is so far removed that in practical employment one can never be compelled to go beyond it (as the mathematician does when he assumes an infinitely large or small magnitude). (ESN, HKA 1/7, 80) The genuine concept of an empirical infinity is the concept of an activity that is infinitely inhibited. But how could it be inhibited to infinity if it did not flow into infinity and if it did not deposit its whole infinity in every individual point of the line that it describes? (ESN, HKA 1/7, 81) Evidently every (finite) product is only a seeming product, if again infinity lies in it […]. Nature is absolutely active if the drive to an infinite development lies in each of its products. (ESN, HKA 1/7, 83)77

Here, Schelling uses a methodological idea that must seem very familiar to readers of Maimon. Although human cognition is finite cognition and is thereby bound to spatiotemporally structured experience of nature (i.e., experience of particular natural products), it has ways to produce a description of nature as an order of infinite intelligibility. So, like Maimon, Schelling needs a model to conceptualise how to determine qualities in a purely conceptual manner and to account for the fact that these qualities only appear to us as quantities.78 In contrast to Mai76 We will see later that ‘presenting’ has a special meaning in Schelling’s methodology. It denotes a particular mode of representation in which the representation and that which it is represented as are in some way identical, and in particular such that the former functions as an exemplification of, and thus makes intelligible, the latter. For more on this, see section 5. 77 Or: “For the philosopher, the points of inhibition will be signified by products; every product of this kind will represent a determinate sphere which Nature always fills anew, and into which the stream of its force incessantly gushes” (ESN, HKA 1/7, 82). 78 In the Outline, Schelling uses another mathematical example to describe the kind of intuition that natural philosophy must be capable of: “The

The Method of Nature-Construction

217

mon’s model, Schelling does not understand these qualities to be the ultimate elements of cognition but, so to speak, the ultimate elements of nature. In order to describe these elements from the standpoint of Naturphilosophie, Schelling appropriates the concept of infinitesimals. Natural products must be conceptualised as quantitative representations of nature’s elements, or “differentials”, which, as purely qualitative relations, are the rules from which these products are generated. As Maimon states with regard to an object being through intuition, “[w]ith respect to intuition = 0, the differential of any such object in itself is dx = 0, dy = 0 etc.; however, their relations are not = 0, but can rather be given determinately in the intuitions arising from them” (II: 32). Thus, nature’s self-productivity can be approached by virtue of integration, through particular natural products which present a determinate relation of nature’s simple actants. Natural products should be treated as representations of the activity of nature’s self-determination, which originates from points of relation between nature’s productive and self-limiting force.79 That is to say, “[e]ach formation is only the phenomenon of a determinate proportion ally infinite is only the external intuition of an absolute (intellectual infinity) whose intuition is originally in us, but which could never come to consciousness without external, empirical exhibition. The proof of this is that this intuition comes to the fore precisely when the empirically infinite series lying before the imagination is obliterated (“I blot you out, and you lie fully before me”)” (HKA 1/7, 79-80). Nassar (2012; 2013b) in this context draws an illuminating comparison to Spinoza’s account of intuitive knowledge (see e.g., E2p40s2), through which the mathematician “sees the idea that underlies and determines the numbers and their relations. The idea, therefore, is not an abstraction, but is immanently realized in numerical relations; the relations are singular manifestations of the idea” (2013b, p. 245). Schelling abandons this understanding when introducing his method of nature-construction through experiment in the Introduction. 79 Through its self-relating activity, nature must be seen as producing an infinite number of original actions, or “pure entelechy”: “every original quality, as quality (not as substrate, in which quality merely inheres), must be thought as pure intensity, pure action, then qualities generally are just the absolutely empirical factors in our knowledge of Nature, of which no construction is possible, and in respect to which there remains nothing for the philosophy of nature except the proof that they are the absolute limits of its construction” (HKA 1/8, 49).

218

Schelling’s method of nature-construction

of original actants. If evolution were complete, then this would be = to the universal dissolution into simple actants. Every product, therefore, is = to a determinate synthesis of actants” (SWI, 102). In Schelling’s model of Naturphilosophie, therefore, nature’s self-productivity takes the place of pure relationality, which in Maimon’s model was the infinite intellect.80 Products arise as a relation of forces or actants, and empirical appearances manifest in a determinate or finite synthesis of forces, instead of in a complete synthesis. So, translated into familiar vocabulary, Schelling employs a version of the method of fictions in order to arrive at a description of nature as infinite self-productivity which allows him to understand natural objects as products of nature’s self-construction. Further, “knowing the internal construction” of nature is only possible on the grounds of a second condition: it must be possible to intervene into nature through experimentation (E, HKA 1/8, 33-34). It is by virtue of this second condition that nature’s self-construction becomes accessible to us as finite cognisers, without having to conceptualise nature’s products as structured in virtue of the forms of our cognition. To understand this methodological conception of “constructing nature through experiment”, we must pay closer attention to that part of Schelling’s philosophy of science which deals with the nature of scientific experimentation. By understanding the experiment as “presenting concepts”, he finds a way to explain how science as whole can be understood as a construction that knows its objects because it creates them: on this view, all empirical experimentation must be understood as the creation of natural products, presenting nature as an infinite self-inhibition in its products.

3.6 Presenting nature through experiment First and foremost, Schelling reasons that an intervention into nature is itself only comprehensible if nature is constructed as something that scientists can “invade through freedom”. A key aspect of Schelling’s Naturphilosophie is not just his presupposition of the possibility of an 80 Interesting in this context is Bonsiepen’s interpretation, which connects Schelling’s natural qualities to Leibniz’s entelechies (1997, pp. 190-94, 277-8).

Presenting nature through experiment

219

intervention into nature, but that he makes the explanation of this possibility an integral part of this science. Naturphilosophie articulates the conditions under which an intervention into nature is possible; it establishes the a priori structure according to which nature produces the multiplicity of phenomena. By virtue of that, Schelling’s philosophy also determines the conditions under which nature can be created in artificial or experimental settings. Which roles and functions are occupied by experiments, what kind of evidence they produce—through which kinds of interactions with nature—depends on the very conception of nature that Naturphilosophie postulates. Since Schelling’s model determines nature as unconditioned, and thus as active and not passive, as an infinitely continuous process, and not as a collection of permanent products, experimentation has to be conceived as the controlled production of natural phenomena, and experimental actions as a form of transaction with nature. Through a particular understanding of Baconian interventionism, Schelling can show how experiments make transparent the rules of generation of natural products, and thereby give a tentative answer to the quid facti. In light of Schelling’s innovation in articulating the method of construction, this method should not serve to study nature qua the structure of human consciousness, but rather study the structure of consciousness and all of its possible objects qua the structure of nature. To do so, and to stay true to his pre-subjective explanation of nature, Schelling shifts the medium of construction from sensible intuition to to the real materiality of nature itself. Unlike Kant’s account, in which mathematics is able to construct its concepts because it can present their corresponding intuition, Schelling’s position has to rely on the scientific experiment as a medium of reflection and presentation. In an insightful example, he compares two ways in which we can have knowledge of a machine. On the one hand, we can know it by simple observation. All we can come to know from this is that this machine exists. On the other hand, Schelling supposes, we can have the same knowledge an inventor has of her machine. This knowledge is one of its ‘inner construction’, i.e., its conditions of possibility. Knowing its inner construction provides us with the knowledge to bring this machine about, to realise it according to its concept (E, HKA 1/8, 32). Since to “get a glimpse of the internal

220

Schelling’s method of nature-construction

construction of nature” is only possible through experimentation, it seems that Schelling allocates to the experiment—understood as active invasion into nature, and not as mere observation—the role of pure intuition. That is, through materialising experimental systems (i.e., their objects of study, their experimental apparatus, and the interaction between them), experimenters create artificial settings in which they intervene in order to reflect and ultimately present the concepts and principles that they set out to investigate.81 Consider how Schelling characterises experiments: Now, it would certainly be impossible to get a glimpse of the internal construction of Nature if an invasion of Nature were not possible through freedom. It is true that Nature acts openly and freely; its acts however are never isolated, but performed under the concurrence of a host of causes which must first be excluded if we are to obtain a pure result. Nature must therefore be compelled to act under certain definite conditions, which either do not exist in it at all, or else exist only as modified by others. —Such an invasion of Nature we call an experiment. Every experiment is a question put to Nature, to which it is compelled to give a reply. But every question contains an implicit a priori judgment; every experiment that is an experiment, is a prophecy; experimenting itself is a production of phenomena. […] We know only the self-produced; knowing, therefore, in the strictest sense of the term, is a pure knowing a priori. (HKA 1/8, 33 [emphasis added])

First of all, scientific experimentation is only possible if we can invade into nature “through freedom”. That is to say, it must be possible to somehow manipulate nature according to a plan. Experimenters must be able to approach nature with their questions, which contain “a priori judgments”, or “prophecies”, about how things are going to behave under different circumstances. It makes predictions about experimental outcomes and makes plans for experimental interventions in order to test the adequacy of these “prophecies”. Furthermore, experimentation is not concerned with nature as it is ‘found’. One of the cornerstones of Bacon’s revolutionary programme 81 Thus, I disagree with Nassar’s assessment of mathematical construction as entailing an immediate relation between concept and pure intuition. It is not that “the relation between the idea of nature and natural phenomena is not immediate in the same way that it is in geometry” (2013a, p. 206). Rather, the natural-philosophical experiment exploits the same kind of reflexivity within a different medium.

Presenting nature through experiment

221

of science was to reconceive nature as it figures within experimental contexts, and not as it is observed in common experience. Schelling agrees with the interventionists’ view that “in nature there is just complexity” (Hacking, 1983, p. 226), and in order for it to become observable and interpretable, nature has to be prepared and manipulated. Just as everyday consciousness is abstracted from, and is separated and bracketed in, philosophical experimentation, natural phenomena are artificially isolated and dissected in natural-scientific experimentation. Experimentation is possible because scientists can intervene into nature’s complexity by using experimental apparatus; because they can get nature “to act” in artificial settings through isolating, manipulating, and stabilising natural variables, their experimental actions make nature’s actions ‘visible’ and ‘legible’. According to Schelling, experiments should therefore be understood as “productions of phenomena”. Through invasion into nature, they establish the rules of production of natural products. To systematise Schelling’s conception of experiment under the naturalphilosophical model, it is helpful to set up an analogy to Kant’s conception of mathematical presentation.82 According to this analogy, the natural-philosophical experiment is best understood as a reflective performance, which simultaneously acts as the genetic description of that process, and which produces the object of its inquiry, and can therefore be called a ‘presentation’ in the proper sense. The structural analogy between these two instances of presentation is based on three elements: (i) performativity, (ii) reflexivity, and (iii) genetic explanation. (i) Schelling’s Naturphilosophie models both the experimenter and that which she experiments with as nature, and thus ultimately as a manifestation of the same constructive activity. Experiments are thus understood as a transaction between experimenter and nature. In line with Baconian interventionism, experiments are understood as performances rather than passive events of observation. Modelling nature as an ongoing process of individuation, grounded in the principle of universal duplicity entails conceiving of experiments as practices, sequences of acts, through which 82

I have argued for this claim in detail elsewhere; see Schmid (2018).

222

Schelling’s method of nature-construction

nature’s actions are “bounded by human activity” (Gooding, 1990, p. 66). By virtue of this transactional practice, nature is presenting itself; successful experiments as intermediate links become finite self-demonstrations of nature self-limiting—they are infinite activity in one point. They, too, view nature not as factum but as “action itself in its acting” (ESN, HKA 1/7, 79). (ii) The acts of preparation performed by the experimenter provide a structural analogue to the mathematician ‘reflecting on the rule of the performance while performing it’. The purpose of putting nature under artificial circumstances is precisely to reflect on the different ‘rules’ that inform and guide the natural phenomenon under investigation. Experiments are self-reflexive performances in that they reflect on the different ways in which conditions, circumstances, changes, and instruments interact with the experimental course. At the same time, reflexivity is made possible through a material articulation of laws or conditions that determine nature’s acts. In geometrical construction, the mathematician reflects on the rules of constructing a triangle through exhibition, i.e., by constructing a triangle in pure intuition. In experimental construction, nature, too, is brought to materialise the conditions that determine the genesis of its product or phenomenon. The rules of how nature produces a phenomenon are not only expressed in concepts and laws, but also through actual materiality. Understood from an active perspective, experiments exemplify what they determine. Through material reflection, the experiment produces a particular that expresses universality, i.e. a ‘material universal’.83 (iii) Finally, Schelling characterises experiments as ‘productions of phenomena’. Through the course of an experiment, setting up its design, its apparatus, and theory-construction, the experimenter learns about how different factors influence, and finally determine, the generation of phenomena. In Schelling’s terminology, the experimenter thereby reflects on nature’s act of producing 83

See Rheinberger (2001, pp. 116-117).

Presenting nature through experiment

223

this phenomenon. The suggested relation between the knowledge of objects and the ability to bring them about is a familiar one: “we only know the self-produced” (HKA 1/8, 34). For Kant, the mathematician had an epistemic means to reach an adequate definition of a triangle because triangles are objects of the kind that finite thinkers can intentionally [willkürlich] produce, which correspond with the concepts defining them. This productive performance also entails making explicit the genesis of objects:84 “science […] views its object in becoming, and as something that has yet to be accomplished” (HKA 1/8, 40). As with geometrical construction, successful experimentation produces understood nature [begriffene Natur] because the experimental process is simultaneously a material realisation of the genetic structure of the phenomenon. In its presentational function, the experiment demonstrates how products of nature come about. Thus, production here also means manufacture [Herstellung].85 Experiments, as re-enactments of genetic descriptions of the phenomena they investigate, simultaneously make their objects come into existence. I have argued elsewhere that Schelling was probably mainly thinking of a specific type of experiment, namely what is sometimes called a demonstration experiment.86 In contrast with experiments that test hypotheses, as well as some others that do not (e.g., explorative experiments87 ), demon84

Remember the analogy between the philosopher and the machine-builder: “The idea of knowledge is here taken in its strictest sense, […] we can be said to know objects only when they are such that we see the principles of their possibility, for without this insight my whole knowledge of an object, e.g., of a machine with whose construction I am unacquainted, is a mere seeing, that is, a mere conviction of its existence, whereas the inventor of the machine has the most perfect knowledge of it, because he is, as it were, the soul of the work, and because it preexisted in his head before he exhibited it as a reality” (E, HKA 1/8, 33). 85 Both Rheinberger (2001; 2003) and Schlenstedt (2000) stress the productive character of presentation [Darstellung]. 86 See Schmid (2018, pp. 19-20) for a more extensive account. On the demonstration experiment as public spectacle, see Golinski (1999) and Hochadel (2003). 87 See Steinle (2005). Using (amongst other things) the example of Ørsted’s

224

Schelling’s method of nature-construction

strative experiments have a presentative role: these experiments consist in a performance that creates or generates the genetic structure of the phenomenon, and it is the controlled material process—the observable coming-into-being in front of our eyes—that is the main function of this type of experiment. Nature is presented through an active intervention, and realised through an experimental apparatus and its controlled interaction with its environment. Is this then the missing link needed for Naturphilosophie to claim to have an answer mto the quid facti? Can scientific experimentation validate an application of its a priori concepts and principles? And what do the hypotheses tested through scientific experimentation have to do with the ultimate conditions of Naturphilosophie? This connection becomes even more opaque when we consider Schelling’s remarks about the relationship between philosophy and the experimental sciences, echoing the Baconian distinction between “science” and “history” (E, HKA 1/8, 39). As part of the realm of “pure empiricism” and “natural history”, experimentation cannot be part of science —the idea of an experimental science “is a mongrel idea” (ibid.) since empiricism only includes natural histories of facts, that is, “accounts of what has happened, under natural or artificial circumstances” (E, HKA 1/8, 40). Science, on the other hand, should consist of pure theory; everything that admits of a priori construction. “What is pure empiricism, is not science, and what is science is not empiricism” (ibid.). Consequently, science as it stands must be revised. Schelling demands to “strip empiricism of all its theory, and restore it in its original nakedness”, and for all theoretical aspects of natural science to be integrated into the project of speculative physics, or Naturphilosophie. While empiricism must “regard […] its object in being, as something already prepared and accomplished; science [must] view […] its object in becoming, and as something that has yet to be accomplished” (ibid.). That is, while natural history is concerned with nature’s products, which are collected as electromagnetism experiments, Steinle makes a point of showing that scientific progress is not always theory-driven, and that experiments can bring about conceptual innovations without knowing what exactly they are looking for, i.e., through accidental findings.

Presenting nature through experiment

225

facts, science views nature in its productivity, as the constructing activity from which every natural product is generated. Given this division of science and natural history, why would Schelling claim that it is through experiments that science gets a glimpse into the internal construction of nature? How could it be that scientists can ‘construct nature through experiments’? From the standpoint of theory, or just Naturphilosophie, the facts or objects with which natural history is concerned, and which it brings about, must be viewed as products of an infinite productivity, since each natural product (in virtue of its potentially infinite development) is itself a spatiotemporal representation of nature’s self-productivity. Schelling contends that the construction of Naturphilosophie can connect to the sheer infinity of natural products through a specific kind of experiment which, so to speak, coordinates the work between theory and history. These experiments establish a recognisable link between the multiplicity of natural phenomena, i.e., the sum total of natural products, and the ultimate conditions of nature; the principles and concepts grounded in the ultimate axiom of nature as productivity and product, through which nature is constructed. Now, we may indeed be quite certain that every natural phenomenon, through whatever number of intermediate links, stands in connection with the last conditions of Nature […]. To find out these links is the work of experimental research. Speculative physics has nothing to do but to show the need of these intermediate links; but since every new discovery throws us back upon a new ignorance, and while one knot is being loosed a new one is being tied, it is conceivable that the complete discovery of all the intermediate links in the chain of Nature, and therefore also our science itself, is an infinite task. (E, HKA 1/8, 36 [emphasis added])

Through some of its interventions, natural history produces “intermediate links” between an infinity of natural products and their ultimate conditions (E, HKA 1/8, 46). Since he mentions Bacon’s doctrine in the very same context in which he explains the nature and role of experiments as intermediate links, it makes sense to assume that by these “intermediate links” Schelling meant crucial, or “central”, experiments.88 88 “Thus, for example, it becomes very clear through the whole course of our inquiry, that, in order to render the dynamic organization of the Universe evident in all its parts, we still lack that central phenomenon of which Bacon already

226

Schelling’s method of nature-construction

He understands these as experiments of the type discussed above, i.e., performances that bring about natural phenomena that exemplify its universal properties by making manifest its relation to, and thereby its determination through, the whole of nature.89 As a consequence, the performance constitutive of an experiment is both the object and the medium of reflection on, and understanding of, nature.90 Scientific experiments are the ‘intermediate links’ that enable science to “arrive […] at the construction of that infinite becoming, the empirical exhibition [Darstellung] of an ideal infinity” (E, HKA 1/8, 46). Experiments are thereby assigned the role of a mediator [Vermittler] between the speculative construction of nature and the manifold of natural phenomena that it explains. As mediators, experiments are able to translate philosophical constructions into nature and, thereby, to ‘mesh’ conceptual ideal construction with the real materiality of nature. speaks, which certainly lies in Nature but has not yet been extracted from it by experiment” (E, HKA 1/8, 36). Ziche shares this assumption, yet the presentative function which this kind of experiment carries for Naturphilosophie escapes his attention (2010, pp. 170-1, 176-7). He fails to describe its essential function, i.e., “as the experiment wherein all those functions of matter, magnetism, electricity, etc., so run together in one phenomenon that the individual function is distinguishable”. (E, HKA 1/8, 36) 89 In a handwritten note to his personal copy, Schelling writes: “In order to make this interdependence fully evident, we need the central phenomenon, or central experiment, of which Bacon speaks oracularly – I mean the experiment wherein all those functions of matter, magnetism, electricity, etc., so run together in one phenomenon that the individual function is distinguishable – proving that the one does not lose itself immediately in the other, but that each can be exhibited separately, an experiment which, when it is discovered, will stand in the same relation to the whole of nature, as galvanism to organic nature.” (E, HKA 1/8, 84-85) 90 In his critical work on the validity of Schelling’s treatment and use of empirical findings of his time, and following Wild’s (1968) remarks on the specificity of the natural-philosophical experiment, Mutschler (1990) analyses Schelling’s notion of experiment as put forward in the Introduction. In order to connect the work of speculative physics and empirical science (85ff.), he proposes the following conception: instead of serving for the purpose of quantification, the natural-philosophical experiment serves the purpose of exhibiting ideational content in external intuition, i.e. exhibiting form through actual matter. Other works on the natural-philosophical experiment can be found in Daiber (2001), or Mitchell (2013).

Presenting nature through experiment

227

Philosophical constructions are therefore not exactly ‘verified’ through experiments, but rather are presented or exhibited; the question quid facti is therefore answered through an infinite process of presentation. Naturphilosophie proves the reality of its concepts and principles by virtue of producing an infinite object:91 “[t]his product is a finite one, but as the infinite productivity of Nature concentrates itself in it, it must have a drive toward infinite development.— And thus gradually, and through all the foregoing intermediate links, we have arrived at the construction of that infinite becoming, the empirical exhibition of an ideal infinity” (E, HKA 1/8, 46). Under Schelling’s model of nature, experimental facts are re-conceived as acts of nature, thus, through the lens of Naturphilosophie, these experiments are endowed with a presentational character. Consequently, experiments assign finite and material processes (natural phenomena) to philosophical constructions,92 thereby satisfying the second criterion of Naturphilosophie: to find a medium through which the finite in the infinite as an infinite process of becoming can be presented. What makes us certain that Naturphilosophie’s constructions actually apply to nature’s products is that the way in which they frame nature and construct the axioms of all individual sciences appropriately, and through which they inform experimenters on how to intervene into nature, that is, on what kind of properties natural products have, and how these connect and interact in order to individuate and differentiate nature into a multiplicity of qualities. A good example of how Schelling envisions this interaction between Naturphilosophie and natural history can be found in the experimental 91 As mentioned earlier, Schelling’s Introduction emerged during times of intensive discussion with his friend J.W. Goethe, see Richards (2006, p. 38f.). The latter accredited the experiment a crucial role in his theory, e.g., in “The Experiment as Mediator of Subject and Object” (1989/1823). In contrast to Schelling, he takes the experiment to provide ideational content through the quantification of similar experiences. Thus, he tries to conceptualise his intuitions, rather ‘intuiting his concepts’. Therefore, the two approaches differ significantly in whether they work ‘top-down’ (Schelling) or ‘bottom-up’ (Goethe), i.e. whether the concept is mediated with the actual materiality of nature or whether the manifold of phenomena is investigated for patterns and the intuition conceptualised. 92 Poser (1979, p. 132ff.) makes a similar point.

228

Schelling’s method of nature-construction

research on the phenomenon of electricity, and in particular the experiments conducted by Luigi Galvani, as well as Alessandro Volta and Joachim Ritter’s reactions to them. Named after its founding figure, Galvani, so-called “galvanic” research at first intended to show the existence of a vital force, i.e., “animal electricity”.93 Upon accidentally observing that frogs’ legs would twitch under the influence of lightning discharges, Galvani began to devise an experiment that could demonstrate that muscle contain a form of electricity, which is stored in living organisms in an analogous fashion to its storage in Leyden jars. Volta objected to these results through experiments which showed that Galvani’s experimental recreation of the effect, i.e., through touching the spinal cord attached to a dissected frog leg linked with a metal, could also be generated through yet another experiment which made no use of living organisms.94 In accordance with his findings, he reasoned that the electrical current observed on closing the electric circle is not generated by the living organism, but by the two dissimilar metals involved, namely the iron or brass on which the frog preparation was mounted and the probe of metal used to touch the frog muscle. He thereby developed an experimental apparatus which could generate the same effect, but instead of frog muscles and spinal cords just used water and metal. More precisely, he built a pile of alternating zinc and silver discs, each separated by brine-soaked cloth, and connected both ends of the pile with a wire, which came to be known as the first chemical battery. “[F]rom Volta’s recent experiments […] it becomes clear that in order to produce 93

See Galvani (1791). Although, for a long time, historians of science had thought of Galvani’s experiments as mistakes corrected by Volta, this view has been corrected as well in recent times: “historians have noted that neither Volta nor Galvani won this debate, because both were partly right. Some of Galvani’s phenomena were due to electricity generated by the “voltaic” pile, but some observations, such as the muscular contractions caused by forming a loop in which the cut end of a nerve touched the muscle to which it was attached, were independent of metals. Nevertheless, an active experimental effort by a number of scientists to repeat and extend Galvani’s observations faded after two decades, probably because of the failure to attain decisive new results. Meanwhile, the “Galvanic” currents generated by voltaic cells acquired an important role as a tool for investigating the nervous system” (Holmes, 2003, p. 234). 94

Presenting nature through experiment

229

electricity at all the mere contact and separation of two heterogeneous conductors is necessary” (ESN, HKA 1/7, 185). Now, how do these experiments succeed at presenting an intermediate link between “every natural phenomenon” and “the last conditions of Nature”? Volta’s experiment not only demonstrates how we must manipulate nature in order to bring about an electric current, but, as a performance which consists in the generation of a “universal” instance of nature, this experiment also demonstrates how the different forms, according to which natural products can be individuated and differentiated, are related. On Schelling’s view, experiments such Volta’s, and also Ritter’s, present the universal construction of nature insofar as they show how the generation of a kind of phenomenon is linked to the generation of another chemical phenomenon. More precisely, it shows that in the production of one material universal, which is supposed to exhibit one kind of quality, e.g., particular electrical qualities, the experimenter can also effectively manipulate another kind of quality, e.g., its chemical qualities. What Schelling means by this is brought out more clearly when considering one of Ritter’s galvanic experiments, which was in fact directly inspired by his reading of Schelling’s early Naturphilosophie.95 Ritter’s experiment was based on building so-called “galvanic chains”, which consisted of two differing metals (e.g., copper and zinc) and a conducting fluid (e.g., water). He realised that, when closing the chain to generate the electric current, he could observe this event as constantly accompanied by a chemical reaction of a specific type: each time electricity began to flow, the water turned milky and the zinc oxidised. It was this presentation of a mutual change in the chemical reactants that for Schelling made “the bearing which the electric relationship of bodies has upon that of their oxidizability […] intelligible” (HKA 1/5, 159). Now we should be in a better position to understand Schelling’s (at times arcane) descriptions of crucial experiments as intermediary links that serve to connect the multiplicity of natural phenomena or products to the last conditions of Nature. Schelling characterises central experiments as experiments “wherein all those functions of matter, magnetism, 95 See Ritter (1798; 1800). See Breidbach (2004) on the relationship between Ritter and Schelling.

230

Schelling’s method of nature-construction

electricity, etc., so run together in one phenomenon that the individual function is distinguishable” (E, HKA 1/8, 84). Ritter’s experiments exhibit the universal through the particular in two ways, since they not only demonstrate the genesis of an electric effect, but also the way in which this process maps on to the generation of a chemical effect.96 By presenting an experiment that serves to predict a change both in chemical and electric properties, Ritter also shows that one and the same experimental act is responsible for a difference in both electrical and chemical properties. This is how crucial experiments can provide experimental verifications of the particular constructions of Naturphilosophie. They serve to exemplify nature’s self-production through experiments in which the manipulation of natural products not only demonstrates one of their properties, but which simultaneously reveals the genetic structure of natural products in general. Since one form, or “potency”, of natural matter does not “lose itself immediately in the other, but […] each can be exhibited separately” (ibid.), Volta’s and Ritter’s galvanic experiments succeed at mapping the structure of electrical phenomena on to the structure of chemical phenomena. Thus, the natural-philosophical experiment not only isolates one thing as different from others, but simultaneously exhibits the different modes in which natural products in general are individuated, e.g., in virtue of their chemical and electrical properties. For instance, Schelling argues that experimenting with galvanic chains can show which processes of chemical differentiation go hand in hand with which processes of electrical differentiation. It should therefore also not surprise us that Schelling did not go along with the consequences of Volta’s objection to Galvani’s postulation of animal electricity. For reasons of Naturphilosophie, he had better motivation to argue for the existence of similar galvanic chains in living organisms, and hence that galvanism also presented an intermediate link between the forms that determine prop96 In 1820, Hans Christian Ørsted devised an experiment which for the first time could show that electrical currents create magnetic fields, thereby linking the phenomenon of electricity to the phenomenon of magnetism (see Ørsted (1998)). This is often referred to as the one big success of Schelling’s Naturphilosophie in its aspiration to predict or “divine” the crucial experimental phenomena yet to be found; see for example Friedman (2007).

Metaphilosophy as constructive and experimental practice

231

erties in organic, as opposed to inorganic, matter. Through generating particular material phenomena, crucial experiments not only exemplify universal and necessary laws but also exhibit the different forms under which nature can be determined, i.e., as individuated and differentiated. In doing so, crucial experiments serve to verify the absolute hypothesis of Naturphilosophie under a particular construction—under a concrete conceptualisation of nature’s forms of difference that arise from its self-productive activity.

3.7 Metaphilosophy as constructive and experimental practice Reviewing this conception of Naturphilosophie, we arrive at a new way to answer the metaphilosophical challenge of providing an appropriate philosophy of philosophy. As first science, philosophy has both a constitutive and normative role with respect to the scientific status of all others. More precisely, philosophy acts as a first science by grounding natural science in a methodological principle that allows the natural sciences to be conceived as an infinite collection of facts, united under one unconditioned principle. This principle theorises nature as a universal duplicity between productivity and product. By treating nature in this way, philosophy can then begin to construct a theoretical framework of its structure. The adequacy of these those theoretical constructions is not decided only their logical consistency, or their satisfaction of a rationalist standard, but by whether that they can tested through scientific experiments. A surprising similarity to Maimon’s metaphilosophical conception is that the hypotheses constructed as part of the system of Naturphilosophie can be, and are being, tested through experiments in the sciences. According to Schelling’s revolutionary philosophy of experiment, however, scientific experimentation can actually present (darstellen) nature-philosophical theories. What we have not paid attention to so far is the fact that this method not only reserves a special place and conception for experimentation in the natural sciences, but renders philosophy an experiment of reason. Namely, although “[t]his absolute hypothesis must bear its necessity within itself, […] it must, besides this, be brought to an empirical test”

232

Schelling’s method of nature-construction

(E, HKA 1/8, 35). The principles of Naturphilosophie, as a system grounded in an absolute hypothesis, are brought to an empirical test, in that they provide the model according to which not only scientific explanation, but also scientific experimentation—in short, any treatment of the object of natural science—is to take place. Should experimental practice produce examples of natural phenomena which cannot be made comprehensible through the construction of Naturphilosophie, e.g., experiments demonstrating that the chemical properties of objects cannot be distinguished from its physical properties, it will have to revise its description which explains science as possible, “for, inasmuch as all the phenomena of nature cannot be deduced from this hypothesis so long as there is in the whole system of nature a single phenomenon, which is not necessary according to that principle, or which contradicts it, the hypothesis is thereby at once shown to be false” (E, HKA 1/8, 35). Note, however, that this does not erase the methodological necessity of an absolute hypothesis. Natural science proceeds by virtue of hypothesis testing. And in order to be able to test hypotheses, natural science must always be grounded in an absolute hypothesis that is not testable.97 Naturphilosophie is the science that makes possible all other science by explaining how knowing something a priori, and that is, to be necessary, is possible. It grounds natural science by positing an absolute hypothesis and a system of principles that derive from it. We have seen that the demand for a unified system of nature that satisfies the standards of scientificity includes a first postulate, i.e. an ultimate presupposition.98 It is only based on this presupposition that nature can be understood as active, and consequently, as constructible: “[b]y this deduction of all natural phenomena from an absolute hypothesis, our knowing is changed into a construction of Nature itself” (SW 1, 3, 278).

97 I thus find Nassar to be slightly misleading in her assessment that “the absolute hypothesis must be tested to conform to the phenomena of nature” (2013a, p. 207). What is tested is a particular construction of nature, whose first principle is the absolute hypothesis. 98 “[P]recisely for that reason, it [the first principle] is the point at which our analysis (experience) can never arrive. It must be simply posited into Nature, and it is the first postulate of all philosophy of nature” (ESN, HKA 1/7, 67).

Metaphilosophy as constructive and experimental practice

233

Under the guidance of Naturphilosophie, science as such is transformed into one all-encompassing experiment: It is, therefore, conceivable that speculative physics (the soul of true experimentation) has, throughout all time, been the mother of all great discoveries in Nature. (E, HKA 1/8, 37 [emphasis added])

Schelling’s theory of philosophy and experimentation opens a new horizon for understanding between philosophy and the natural sciences, namely one in which the barrier between philosophy, the mathematical sciences, and the experimenting sciences is lifted. Naturphilosophie determines the a priori structure of each particular natural product and thereby simultaneously determines the unifying first principles of all sciences. Like Kant and Maimon before him, Schelling’s metaphilosophy adopts a method that is not completely different to, but rather continuous with, those of the natural sciences. Unlike the version of the metaphilosophy-first research programme that we considered previously, his methodological proposal blends not only the boundary between the normativity of philosophical theories and the sciences, but also meshes the experimental activity of the natural sciences with the construction of philosophical theories. In later writings, Schelling would still even note that “we can therefore compare all earlier aspirations of philosophy (since Cartesius) to the experiment in the natural sciences” (HKA I, 10, 227).

4 Experiments of reason

Through its individual studies of Kant’s propaedeutic method, Maimon’s method of fictions, and Schelling’s method of nature- construction, this book has developed the premises needed to establish the plausibility of the metaphilosophy-first claim. All three philosophical programmes agree in their metaphilosophical goal: to develop a method for establishing the possibility and nature of philosophy as a scientific discipline. Moreover, Kant, Maimon, and Schelling show a similar methodological approach. Their methods are developed through an intensive and detailed engagement with the sciences of their day, and in particular with those sciences that were still in their formative stages. In these concluding remarks, I want to offer some justifications for the view that each procedure provides a version of what I categorise—following Kant—under the notion of an “experiment of reason”. By this I mean that each version of the metaphilosophy-first programme considered in this book advances a specific conception of philosophical a priori experimentalism, whether implicitly (as in Maimon) or explicitly (as in Kant and Schelling). I will consider this new conception of philosophical experimentation with respect to one central aspect: the relationship between philosophical and scientific (or ‘real’) experimentation. Kant frames the history of theoretical philosophy as a production site of metaphysical principles. Metaphysical principles purport to explain the basic structure of nature and determine the ways in which things are necessarily connected to each other. On Kant’s view, this family of philosophical programmes faces the following problem: unlike the sciences, metaphysics has no rigorous procedure. This methodological defect results in an inability to show how metaphysical principles are to be discovered, and thus also why they would be able to ground the necessary

Experiments of reason

235

and universal character of scientific statements. Kant contends that this impasse can only be resolved through conceiving of theoretical philosophy as a business that begins with metaphilosophical investigations: as metaphilosophy-first. Only if metaphysics is studied through the lens of metaphilosophy can it be transformed into a science. For Kant, this means that we must study the nature and possibilities of human cognition in order to reach a verdict on the nature and limits of metaphysical cognition. One salient feature of his metaphilosophical programme is the premise that ‘metaphilosophy-first philosophies’ can only succeed at their task if they adopt a rigorous, scientific procedure. I have argued that this feature is only one of several, which together allow for the identification and characterisation of a specific family of research programmes. The other features that I identified included the presumed continuity between philosophical and scientific methods, which was achieved through the prior analysis of scientific theory and practice, as well as the explicit use of analogies between both sets of methods, with the goal of explaining the workings of one through the workings of the other. Having studied all three proposals in isolation, I now propose that they can jointly be described under the umbrella term of “experiments of reason”. Kant, Maimon, and Schelling developed metaphilosophical methods that can be unified not only in virtue of their shared aims and goals, but also by the kind of argumentation they employ. Beginning with Kant’s experiment of pure reason, it is possible to trace a lineage of metaphilosophical programmes that all advance a specifically (Post-)Kantian brand of experimental philosophy. Unlike Early Modern or Newtonian experimentalist philosophies, this new methodological conception devises a type of philosophical experimentation which proceeds in a priori manner, while still retaining its experimental character. In doing so, it proposes a move away from the empiricist and inductive argumentations of earlier experimental philosophers, as well as from the rationalist and deductive argumentations of earlier speculative philosophers. Kant, Maimon, and Schelling conceive of a new method of argumentation that—in more familiar terms—shows how armchair methods must not be opposed to experimental methods. And it is in virtue of their properties as a priori experiments that these experiments of reason can

236

Experiments of reason

fulfil their metaphilosophical purpose without running into regresses or having been partial or circular all along. They test metaphilosophical theories through different varieties of a priori experimentation, thereby only exhibiting the kind of circle that is inherent to any experimental procedure that tests (and potentially confirms) hypotheses. In what follows, I reconsider each of the previously discussed methods, with regard to their character as experiments of reason, to see how they evade the issues typically associated with a priori, experimental metaphilosophical methods. To do so, I will focus on how each philosopher conceives the relationship between philosophical experimentation and (empirical) experimentation in the sciences. Framing the issue in this way, it will not only be revealed why all of these methods introduce a form of experimental philosophy, but also why with this new conception of experimental philosophy comes with a new understanding of how science, as a whole, must be conceived.

4.1 Kant’s experiments of pure reason We have seen that, in the B Preface, Kant described the relationship between reason and nature as similar in certain ways to that between “an appointed judge” and their “witnesses”, “who compels [them] to answer the question [they] put to them” (Bxiii). According to his revolutionary paradigm, the objects of nature must conform to the a priori principles that govern the cognition of objects of nature. And by experimenting with a priori principles and concepts, metaphilosophy tries to arrive at an articulation of the representational framework responsible for the possibility of different kinds of cognition, including empirical cognition and the kinds of cognition used in metaphilosophical inquiry itself. The articulation of this framework entails an articulation of the conditions which constitute the objects of scientific cognition in general. As a result, Kant’s propaedeutic method also serves to articulate the conditions on the grounds of which it becomes intelligible why natural objects can successfully figure in scientific explanations—namely because of a network of mental or a priori structures that determines the concept of cognising and knowing an

Kant’s experiments of pure reason

237

ject. The relation between philosophical experiments and empirical experiments must then be conceived in this way: experiments of reason establish the first-order laws which yield the structure of natural objects in general, while the real experiments of the natural sciences, step by step, establish those second-order laws which determine the special relationships between different classes of the same objects.1 Consequently, Kant’s experiments of reason, through their chemistrystyled procedure of decomposing and recombining pure elements of cognition, i.e., those elements making up the inner structure of scientific cognition in general, not only establish the possibility of philosophical but also of real experimentation. Philosophical experimentation is engaged in a form of hypothetical reasoning that can arrive at best explanations through the generation of a growing network of consistent concepts that flow from each other, while remaining unified under the same theoretical assumptions. If successful, Kant’s experiment of reason confirms a theory of cognition which describes the constitution of the space of reason as such. Metaphilosophy-first thereby experimentally establishes those theoretical elements which condition any attempt at constructing theories that are possible and justified.2 Real experiments, on the contrary, test the products of hypothetical reasoning through the experimental generation of actual instances of natural phenomena. Whether these instances can serve as evidence for the universal and necessary judgments of natural science depends on whether the experiment of reason has successfully presented the a priori structures in virtue of which scientific experience of objects is constituted. For a proposal which refers to Bacon’s Instauratio Magna, this conception will seem incoherent. Without doubt, perhaps the most important contention of Bacon’s philosophy of experiment is its insistence that experiments should not invest theoretical hypotheses but rather manipulate nature to produce large data collections of its history. On the Baconian view of experimental philosophy (both in the domain of 1 Obviously, in this description I am disregarding the proper place and function of the Metaphysical Foundations of Nature (2004b/1786), in which Kant establishes the a priori metaphysical principles of the particular objects of Newtonian physics by positing the empirical concept of matter. 2 See Vanzo (2012, pp. 77-81).

238

Experiments of reason

natural science and philosophy), it seems, scientific inquiry is experimental in virtue of empirical observation and data gathering, relegating to hypothesising and theory-construction a role in an indefinite future of scientific history. Kant’s conception of experimentation generally seems to go against the grain of this programme, since it embraces the necessity of prior theory-construction and, even more, since it explicitly argues for a form of experimentation that proceeds in an armchair or a priori manner. In some of his remarks about real experimentation, Kant clearly states that any observation or manipulation of nature presupposes hypothesising or theory-construction. And he recommends the adoption of the same paradigm of experimentation in propaedeutic philosophy in order to invest it with a proper method. Things become clearer when we consider that Kant’s conception of experimentation also takes important cues from Newtonian science. On Newton’s conception, what makes science experimental is not just its empirical methods of fact-collection and systematic observation, but its method of experimental reasoning, that is, the ways in which it tests the applicability of highly idealised mathematical models to physical phenomena under highly idealised circumstances. On his view, the success of the sciences depended both on the use of idealisations and modelsystems for theory-construction, as well as for the artificial production of natural phenomena, to decide between alternative explanations of these phenomena. Historically, scientific theories were also verified using what we would today label as thought experiments (e.g., Galileo’s falling bodies experiment), or through the collection of observations that support the mathematical models applied (e.g., Copernicus, Kepler). Indeed, Kant’s favourite example of a revolutionary experiment—Copernicus’ heliocentric theory of planetary motion—is not a hypothesis verified through physical experiments, but a premium example of mathematical modelling that is later adjusted and perfected through Kepler’s work (who bases his modelling on better observations). It’s also worth noting that the point of Kant’s short experimental history is not to show how these experiments developed a life of their own, sometimes managing to make us aware of a phenomenon that transcended concurrent theories and hypotheses. Rather, it contends that all of these experiments are

Kant’s experiments of pure reason

239

products of the hypothesising and theorising of these scientists, as is evident in the examples cited in the B Preface.3 Now, following Kant’s metaphilosophy-first, metaphilosophy must adopt an experimental method to transform metaphysics from a battlefield of opposing metaphysical theories into a production-site of adequate theories about the nature of philosophy itself, which for Kant means to establish a correct theory about philosophical cognition. It then seems that what characterises Kant’s conception of philosophical experiment is, on the one hand, that metaphilosophy is conceived as methodologically similar to the sciences, namely insofar as it constitutes an experimental discipline. On the other hand, however, metaphilosophy also remains dissimilar to the sciences, insofar as his propaedeutic philosophy can only experiment within the a priori realm of pure reason. As philosophers after Kant would notice, he thereby created a rift between what it means to experiment with thought and what it means to experiment with nature. My claim is not that Kant takes metaphilosophical procedure to consist in thought experiments. Rather, he holds that metaphilosophical experiments cannot test their hypotheses outside the realm of thought. Kant theorises experimentation in philosophy, as well as experimentation in the sciences, as being dependent on the activity of human reason, which in both cases equips the experimental apparatus with the necessary questions to interrogate “its witnesses”. In the case of metaphilosophy, this apparatus is necessarily a priori, since experimental reasoning here inquires into the very structures which condition its operations. Contra earlier experimental philosophers, Kant makes explicit that philosophical speculation, if it is directed not at the object but at our cognition thereof, can successfully proceed in a priori manner. Moreover, since his method of reasoning consists neither in deduction from first principles, nor in induction from particular instances, he can also show how 3 “When Galileo rolled balls of a weight chosen by himself down an inclined plane, or when Torricelli made the air bear a weight that he had previously thought to be equal to that of a known column of water, or when in a later time Stahl changed metals into calx and then changed the latter back into metal by first removing something and then putting it back again, a light dawned on all those who study nature” (KrV, Bxiii).

240

Experiments of reason

metaphilosophical argumentation must neither be circular, partial, nor run into regress. Experiments of reason are experimental insofar as they do not presuppose or dogmatically assert principles, but experimentally test them for their validity. However, Kant remains cagey about what it is that these principles are tested against. If experiments of reason constitute active interventions into “the course of our mind” (as opposed to the course of nature), and if they entail an artificial generation of phenomena that can serve as evidence for the correctness of the hypothesised principles, what do they really consist in? Most likely, experiments of reason, as a type of thought experiment, generate instances of mental phenomena. These are produced through reflection, and as such they could be checked for “consistency” with the principles hypothesised. Yet, although Kant provides a theory of the objects of metaphilosophical investigation, that they can be compared to chemical elements and be investigated in similar ways, their exact status as products of experiments remains unclear. It seems, then, that even if Kant does provide a theory for how a priori experiments on thoughts proceed, he is less successful in demonstrating that their conclusions consist in more than possible experiences.

4.2 Maimon’s metaphysical modelling Interestingly enough, this end point of Kant’s proposal is what came under investigation through the lens of Maimon’s criticism. We have seen that Maimon presented an elaborated string of argumentation to show that, and to what extent, Kant’s conception of metaphilosophical experimentation must run into problems. His main criticism consists in the point that Kant’s experiment of reason can only establish the conditions according to which an object of cognition in general must be represented. Even if the latter’s experiment of reason can establish the conditions of possibility under which things can become objects of cognition in general, it cannot show how we get from “something as an object of discursive spatiotemporal cognition in general” to “this particular object that I am observing at a particular place and time”. This has devastating consequences for the relation between

Maimon’s metaphysical modelling

241

ation in thought and real experimentation, because although Kant’s experiment can show how experiments can establish consistent theories, it remains unclear whether it can say anything substantial about the actual manifold of observations and experiences that the experimenting natural sciences are based on.4 Maimon’s quid facti argument showed that scientists have no way of knowing with absolute certainty whether the models of cognition they implicitly rely on (e.g., in virtue of their modes of explanation) do in fact determine the “principles […] according to which alone the agreement among appearances can count as laws” (KrV, Bxiii). Moreover, through the quid juris argument, Maimon also showed that Kant is not only lacking the evidence necessary to prove the objective validity of his hypotheses, but also that, if we accept the latter’s Heterogeneity thesis, we do not even arrive at a consistent hypothetical framework. That is, Maimon argued that, on the grounds of Kant’s hypotheses, it is not possible to arrive at any consistent meta-epistemological framework which could serve to ground and legitimise scientific explanation and experimentation. Hence, he confronts Kant’s methodological solution with an external and an internal problem. First, given the methodological suppositions made by Kant’s metaphilosophical conception, his experiments of reason cannot contribute more than hypothetical theories of cognition. Second, even if we accept his theory, we run into contradictions, and consequently cannot even confirm the genuine possibility of the theory suggested. To put this into the context of this study, Maimon’s arguments raise doubts about two things: whether it is possible to construct hypotheses which allow for a consistent framework of how experimentation in thought is possible, and whether these conclusions about such reasoning can have any objective import? For Maimon, metaphilosophical experiments cannot move from universal representations to singular representations; or, to put this into 4 This also creates more general problems for Kant’s theory of real experimentation, since it can neither show whether (or how) actual observations and experiments might differ from possible ones—how experiments have a life of their own—nor can it show how there can be experiments that test theoryconstructs that are incompatible with our current ones (see Vanzo (2012, pp. 92-93)).

242

Experiments of reason

the experimental context, they cannot move from hypotheses about the general nature of objects to claims about the particular objects which instantiate them. Since they can only treat of general concepts and principles, while the objects of real experimentation concern actual and particular objects, philosophy cannot point to any of these particular instances in order to demonstrate the necessity and universality of its principles. For these reasons, Maimon inferred that metaphilosophy must take up the role of a production site of systematic fictions. As such, it can generate models of cognition which can explain how, that is, under what conditions, we could determine whether universal philosophical principles apply to singular representations of natural instances. At the same time, by assigning to these principles only the status of fictional principles, i.e., theoretical entities whose function is not to represent but to enable science to treat its objects in a rational manner, it avoids the problem of misrepresenting the modal status of its conclusions. In correspondence to his criticism of Kant, and his concomitant analysis of scientific practice, then, Maimon suggests a methodological programme for metaphilosophy that is grounded in accepting a radical difference between philosophical and real experimentation. Although his philosophical experiment sets up a way to close the gap between different heterogeneous forms of cognition, his proposal consolidates another gap: the gap between the hypothetical objects of theoretical philosophy and the empirical objects of the natural sciences. Since the experiments of theoretical philosophy can only establish the form and content of possible objects of cognitions, their method must be conceived as a form of a priori experimentation that is opposed to real experimentation. Maimon’s methodological emphasis lies on granting absolute normativity to the spontaneous and rational activity of reason; it pursues the task of metaphilosophy as turning science into a rational practice. His methodological solution differs from Kant’s in that it admits partiality regarding a rational standard of explanation, but avoids circularity insofar as it accepts the hypothetical or fictional status of its theories. The original point of Maimon’s metaphilosophy is that it is assigns itself the role of generating different theoretical frameworks, which may or may not serve to ground and systematise natural phenomena as they

Maimon’s metaphysical modelling

243

are collected and described by the natural sciences. Analogously to the modelling practice of the mathematised sciences, metaphilosophy generates fictions that provide methods for treating the objects of science as if they were intelligible without limits. However, these experiments can only ever provide models of how the realm of reason (which is intelligible without limits) can be made consistent with the realm of natural science (which is only intelligible up to a certain limit). At the same time, it is exactly through its ability to produce scientific fictions and imitate mathematical modelling practices that philosophy can hope to mediate between the two realms of experimentation and the scientific paradigms that govern their standards of truth and inquiry. Whether a proposed philosophical fiction is in fact successful is decided through the continuing progress of the sciences employing a style of explanation consistent with this model. Although it is not in the power of metaphilosophy itself to prove the objective validity of its conclusion (but rather their potential to be proven), thereby upholding the Kantian distinction between the realm of thought and real experimentation, its cooperation with the natural sciences opens up this exact possibility. Scientific explanation, according to Maimon, must satisfy the standard of explanatory completeness. It is is the same commitment which instructs him in his theory choice regarding the explanation of human cognition. We saw that Maimon’s account of cognition explains the form and content of cognition as homogenous and thus opens a possibility to understand the application of a priori forms to cognitive matter as rule-governed application. On grounds of this strategy, he develops a model on which a priori cognition can be warranted in issuing necessary and universal judgments, since on this model, we can explain how our forms of cognition apply to their content. Given this rationalist standard of explanatory completeness, as finite cognizers, we are never in possession of the complete concepts or representations of objects, and thus also never in the possession of the explanans necessary to produce a complete explanation. This is how Maimon’s rationalist commitment leads to the position of an “empirical skeptic”: it is by committing to a rationalist standard of what can count as scientific explanation, namely only a complete and unified explanation, that we will find our cognitive and scientific resources to be limited to a degree that makes such

244

Experiments of reason

explanation impossible, thus forcing us to take up the position of the empirical skeptic. While Kant subscribes to the view that the natural sciences depend on philosophy because the latter’s principles constitute the possibility of scientific experience, Maimon thinks natural sciences depend on philosophy exactly in the absence of such a constitutive relationship. In delivering consistent conceptions of objectivity, philosophy delivers systems of principles that provide the conditions given which we would be justified in assuming that, e.g., necessary and universal connections obtain between sensible appearances. Thereby, philosophical modelsystems yield different possible descriptions of how it is that science can explain one thing in terms of another, and only Maimon’s system – at least that is what he wants us to believe – can deliver a description that satisfies a rationalist standard of explanation. Overall, then, metaphilosophy is assigned the task of producing useful fictions that articulate theoretical frameworks on the grounds of which philosophy and science are coherently unified under one system. For Maimon, to answer the question whether metaphysics is possible as science, is to establish whether science as such is possible, and to determine whether science as such is possible, is to determine whether philosophy can construct a model of cognition, which simultaneously satisfies the rational standard of explanation and shows natural appearances to be of the right structure for rational inquiry to be possible. To satisfy this standard, philosophical method must become hypothetical and use fictions. On Maimon’s view, differential calculus and metaphilosophy should be conceived as similar because they both exemplify a mode of reasoning and demonstration that defines its concepts and theories through explicating the rules of production of their corresponding objects. Only when philosophical principles can be established in this way will they succeed at proving the reality of a particular conception of objectivity. However, he argues that a metaphilosophy-first approach can only develop a suitable description of a priori cognition if it postulates theoretical concepts and principles which are “impossible as objects of cognition”. Consequently, metaphilosophy must succumb to a practice of hypothetical model-building. In Maimon’s hands, theoretical philosophy morphs from the production site of theories of cognition into the production

Schelling’s experimental constructions

245

site of systematic fictions, and it is in virtue of Maimon’s explanatory rationalism that his philosophy must adopt an experimental character, assigning to the natural sciences the task of actually confirming the objective validity of its hypotheses. Metaphilosophy, therefore, must engage in a priori experimentation from the armchair, while the sciences do the dirty work of experimenting in the laboratory and providing philosophical hypotheses with content, thereby successively—but never completely—proving metaphilosophical theories. In Maimon’s case, this would be the thesis of explanatory rationalism and the theory of a priori cognition that accompanies it.

4.3 Schelling’s experimental constructions In continuity with the metaphilosophical efforts of his predecessors, Schelling develops another methodological solution to secure the scientification of theoretical philosophy: metaphilosophy must “construct nature through experiment”. In light of the relation between theory and experiment, his method of nature-construction pushes the paradigm of philosophical experimentation one decisive step further: metaphilosophy constructs the theory of those universal structures which natural products solidify as they are created in each real experiment. Through reconceiving the function of “crucial experiments” as mediators between the universal constructions of Naturphilosophie and the material manifold of phenomena provided through natural histories, Schelling forges a path along which we can conceive a unified science under the methodological paradigm of experimental philosophy. In contrast to Kant and Maimon, Schelling realised that if metaphilosophy explains things from the (subjective) standpoint of human consciousness, its intelligibility remains limited. This is so because things can only be determined as things that are “for consciousness”, as things whose explanation is dependent on a prior distinction between subject and object, which itself cannot be made intelligible by the same means. To address this, Schelling argued for a new experimental philosophy, one which establishes those a priori conditions from which the subject-object-divide only ever arises. To do so, he provided a revised

246

Experiments of reason

version of a metaphilosophy-first programme, which reframes its goal as consisting in the explanation of “everything from the forces of nature”. Unlike Kant and Maimon’s realisations of this research programme, he proposed that philosophical experimentation must not be conceived from the perspective of a subjective consciousness which cognises objects. Instead, philosophical experimentation has to be reconceived from the perspective of the self-producing nature which creates, and thus determines, itself through the manifold of its products. In contrast to Maimon’s modelling conception, Schelling’s metaphilosophical solution sees the usefulness of fictions exactly in the perceived fact that they permit philosophers to take up the standpoint of a constructive activity that is outside the structure of objective consciousness. What nature-philosophers need to posit is that the a priori conditions of science as such are not “the limits of appearances”, but the self-productive principle that is nature. Naturphilosophie can only experiment successfully—i.e., explain its objects without limits—if it grounds the explanation of nature in one absolute unconditioned principle. This principle is expressed through the useful fiction of an original duplicity between nature as productivity and nature as product. To conceive this principle as something that is actualised in nature, that does not remain within the hypothetical realm of philosophical thoughtconstructs, Schelling assigns it the status of an absolute hypothesis. This is as far as experimentation in thought goes. However, for Schelling, this only makes for one part of what philosophical experimentation consists of. Naturphilosophie not only conceives its construction of nature’s a priori structure as a hypothesis, but also advances a conception of how this hypothesis can be confirmed through the nature-philosophical experiment. In order to determine the framework that governs scientific inquiry and explanation, Schelling extends the experimental activity of Naturphilosophie into the sphere of the experimental activity of natural science. He observes that not just real experimentation, but also philosophical experimentation, depends on the creation of phenomena for its theoretical assumptions. This reconfiguration of the relation between philosophical and real experimentation is based on the idea that science must be remodelled and instructed as a whole, namely in such a way that

Conclusion

247

the confirmation of metaphilosophical theory through the generation of particular products of nature is conceivable. His methodological conception forges a path through which metaphilosophy can rid itself of the charges of partiality and circularity. Through nature-construction, Naturphilosophie—unlike Maimon’s philosophy—can not only hypothesise, but also validate, its rationalist construction of nature. By extending the experimental activity of metaphilosophy into the realm of nature, it can also prove its legitimacy experimentally, without running into a vicious circle. According to Schelling’s conception, it is possible to conceive the activity of the experimenting natural sciences as a controlled presentation of nature’s self-productive activity. The decisive move here is to understand the constructive activity of philosophy and that of nature not as different but, ultimately, as expressions of the same original natural activity. To Schelling, metaphilosophy-first is not mere experimentation with models, but experimentation with nature. According to his conception, philosophical experimentation encompasses experimentation with a creative force that can be made to reveal its creative activity in the controlled settings of real experimentation. Real experiments realise the experiments of Naturphilosophie since they not only present [darstellen] the special laws of nature but also, and more importantly, the general laws of nature. In that sense, Schelling blurs the boundaries between a priori and empirical experimentation, between experimentation in thought and experimentation in nature. For Schelling, ultimately, Naturphilosophie has generated a correct theory of objectivity if it confirms its a priori hypotheses throughout the actual process of real experimentation which, qua these fundamental principles becomes part of the experimental activity of reason in its self-investigation.

4.4 Conclusion Kant, Maimon, and Schelling’s metaphilosophical programs all present variants of what I want to describe as “experiments of reason”. In pursuing philosophical projects that put metaphilosophy first, all of them partake in a new research paradigm of experimental philosophy:

248

Experiments of reason

of metaphilosophy which not only studies and justifies the experimental method but proceeds experimentally in so doing. All of them conceive metaphilosophy as a methodological discipline which is both experimental and speculative. This is made possible in two ways. On the one hand, they understand ‘nature’ as something that is not independent from ‘reason’, and contend that it is reason’s modelling activity which makes the possibility of philosophy and science conceivable. On the other hand, all three programs depart from an intense engagement with scientific practice, specifically one that is based on experimentation, idealisation, and modelling practices. Kant sees idealising practices in experimental chemistry and tries to transfer these procedures to the philosophical practice of experimenting with thought; Maimon encounters a similar practice in mathematical modelling and aims to transform philosophy into a model-based science; and Schelling understands natural science as the creation of phenomena, and assigns Naturphilosophie the role of generating the scientific paradigm thus verified. Its specific conception that ties the justification of metaphilosophical conceptions to their successful ‘performance’ within controllable settings prepares the ground for a conception of theoretical philosophy as pure activity [Tätigkeit] or practice. In conclusion, metaphilosophy-first finds its task in the explication of reason’s activity as it conditions or constructs the objective realm of nature. To become science, so Kant, Maimon, and Schelling agree, philosophy—via metaphilosophy—must engage in experiments with the very activity that generates and resolves all of its problems. Metaphorically speaking, reason must engage in a self-experiment. Only when it employs a method of this form can it hope to arrive at the secure course of a science. To become a science, philosophy must begin with experiments of reason.

Acknowledgements

Over the course of writing and revising this book, I have benefitted from the help and insights of many people, a few of whom I would like to thank here. At different points of this project, I profited greatly from conversations with Daniel Allemann, Dries Bostyn, Timon Böhm, Karim Bschir, Karin De Boer, Huub Brower, Bridger Ehli, Olivier Del Fabbro, Manuel Fasko, Fabienne Forster, Stephen Harrop, Lukas Hilgert, Brigitte Hilmer, Gunnar Hindrichs, Rebekka Hufendiek, Christian Jany, Ursula Klein, Victoria Lazlo, Muriel Leuenberger, Laura Marongiu, Conrad Mattli, Robbie Matyasi, Anne Meylan, Jan Müller, Nico Müller, Martin Münnich, Matthieu Queloz, Andrej Peter, Nicolas Porot, Will Ratoff, Thierry Schütz, Norman Sieroka, Marc Sommer, Alisha Stöcklin, Peter Thielke, Markus Wild, Kenneth Winkler, and Paul Ziche, who together have inspired many of the little twists and turns that together make up this dissertation. I am thankful to audiences at ETH Zurich, the International Kant Congress hosted by the Norwegian Kant Society, the Maimonides Centre for Advanced Studies in Hamburg, the Biennial Meeting of the North American Kant Society, the Virtual Roundtable for History of Philosophy, and the University of Basel. Throughout my time as a graduate student, I have been lucky to be an (official or unofficial) part of graduate programmes at ETH Zurich, Yale University, and the University of Basel, and I am indebted to these communities for creating in me the sense of belonging. I am also very thankful to Emma Bolton for her excellent proofreading and most helpful suggestions. And I thank Lukas Hilgert again, this time for typesetting. I am especially thankful to Michael Hampe, for introducing me to the special concerns of metaphilosophy, and for his openness and

250

Acknowledgements

willingness to engage with my evolving philosophical interests that were sometimes quite alien to his own. My special thanks also go to Paul Franks for our many conversations, his guidance during my wanderings in the region of Kant’s and Maimon’s philosophies, and for helping me to develop my own interpretations at my own pace. I am also very grateful to Robert Stern for his many constructive criticisms, which gave the Kant chapter a whole new direction. All of this would not have been possible if not for Arno Schubbach, who introduced me to the project and generously shared his ideas with me. I owe much of my interest in the nitty-gritty of the history of philosophy and science to his uninterrupted enthusiasm for this project. His support—not only as a teacher but also as a colleague—has enabled and inspired me from the first to the last draft. Finally, many thanks to Michael Della Rocca for encouraging me to not only think about my philosophical ideas, but also to talk about them, and trust that they are worth communicating. Mario Schärli, James Lewis, and Jake Rhode have been incredible friends and readers. Absolutely fearless, they cut through the thick woods of unnecessary repetitions, grandiose arguments, and thin conclusions, and saw light where there was only darkness. Without their friendship and inspiration, the ambient anxiety of grad life probably would not have had the equally amusing quality that made it such a valuable experience after all. The same goes to my non-academic friends and family who have always accompanied and supported me, unconditionally. I thank my cat Aristotle because every good acknowledgement section includes thanks to a non-human animal. Finally, I thank Cyrill, for his generosity, love, and acceptance. This book presents a revised version of my dissertation, which I defended at ETH Zurich in July 2020. I thank the Swiss National Foundation (SNF) for funding this project, along with research group “Begriffe und Praktiken der Darstellung in Philosophie, Chemie und Malerei um 1800” for granting me a PhD position. With kind permission of the publishers, materials from the following previously articles were integrated into this book: Portions of chapter two were previously published in Jelscha Schmid, 2021. “Maimon’s question for a scientific method in philosophy”, Philosophers’ Imprint, 21(36), 1-21; and sections of chapter three were previously published in Jelscha Schmid, 2018. “Schelling’s

Acknowledgements

251

method of Darstellung: presenting nature through experiment”, Studies in History and Philosophy of Science Part A, 69, 12-22.

References

Abbreviations Frequently cited works have been identified by the following abbreviations. Immanuel Kant (1724–1804) AA (1900– ). Kants gesammelte Schriften. (Preußischen Akademie der Wissenschaft, Ed.). Walter de Gruyter. KrV (A edition 1781/B edition 1787). Kritik der reinen Vernunft; Critique of Pure Reason, (P. Guyer & A. Wood, Trans. & Ed.) as Kant (1998). Critique of Pure Reason. Cambridge University Press. Prol (1783). Prolegomena einer jeden zukünftigen Metaphysik, die als Wissenschaft wird auftreten können; Prolegomena to any future metaphysics: That will be able to come forward as science, (M. Friedman, Trans.) in Kant, I. (2004a). Theoretical Philosophy after 1781. (H. Allison, P. Heath, P. Guyer & A. Wood, Eds.). Cambridge University Press, pp. 29–170. KpV (1788). Kritik der praktischen Vernunft; Critique of Practical Reason, (M. J. Gregor, Trans.), in Kant, I. (1996). Practical Philosophy. (M. J. Gregor, P. Guyer & A. Wood, Eds.). Cambridge University Press, pp. 133–272. MNS (1790). Metaphysische Anfangsgründe der Naturwissenschaften; Metaphysical Foundations of Natural Science, (M. Friedman, Trans. & Ed.), as Kant (2004b). Kant: Metaphysical foundations of nature. Cambridge University Press. KdU (1790). Kritik der Urteilskraft, in: Akademieausgabe (AA).

Abbreviations

253

Salomon Maimon (1753–1800) (1965–1976). Gesammelte Werke. (V. Verra, Ed.). Hildesheim: Olms. For an improved version edition of Maimon’s first major published work in philosophy, see Maimon, S. (2004). Versuch über die Transzendentalphilosophie. (F. Ehrensperger, Ed.). Meiner. Tr Versuch über die Transzendentalphilosophie (1790), (N. Midgley, H. Somers-Hall, A. Welchman, & M. Reglitz, Trans. & Eds.), as Maimon, S. (2010). Essay on Transcendental Philosophy. Continuum. PhWB (1790). Philosophisches Wörterbuch, oder Beleuchtung der wichtigsten Gegenstände der Philosophie, in alphabetischer Ordnung, von Salomon Maimon. Unger. KB (1791). Baco und Kant. Schreiben des H. S. Maimon an den Herausgeber dieses Journals. Berlinisches Journal für Aufklärung, Bd. Vll/2, 99–122. AARW (1792). Ankündigung und Aufforderung zu einer allgemeinen Revision der Wissenschaften. Einer Königlichen Akademie der Wissenschaften zu Berlin vorgelegt. Deutsche Monatsschrift. Bd. III, 42–52. LG (1793). Salomon Maimon’s Lebensgeschichte. Von ihm selbst gcschrieben und herausgegeben von K. P. Moritz. Zweiter und letzter Theil. Friedrich Vieweg dem altern. Strf (1793). Streifereien im Gebiete der Philosophie. Vieweg. BNO (1793). Bacons von Verulam Neues Organon. Aus dem Lateinischen iibersetzt von George Wilhelm Bartoldy. Mit Anmerkungen von Salomon Maimon. Gottfried Carl Nauck. Logik (1794). Versuch einer neuen Logik oder Theorie des Denkes. Nebst angehängten Briefen des Philalethes an Aenesidemus. Ernst Felisch. KUmG (1797). Kritische Untersuchungen iiber den menschlichen Geist oder das höhere Erkenntniß- und Willensvermogen von Salomon Maimon. Gerhard Fleischer dem Jüngern. PG (1797). Pragmatische Geschichte des Begriffs von Philosophie, und Beurtheilung der neuern Methode zu philosophiren. GW

254

References

phisches Journal einer Gesellschaft Teutscher Gelehrter. Bd. Vl/2, 150-181. Johann Gottlieb Fichte (1762–1814) GA (1964– ). J.G. Fichte: Gesamtausgabe der Bayerischen Akademie der Wissenschaften, (R. Lauth, H. Jacobs, & H. Gliwitzky, Eds.). Stuttgart-Bad Canstatt: Frommann-Holzboog. Friedrich Wilhelm Joseph Von Schelling (1775–1854) HKA (1976– ). Werke: Historisch-kritische Ausgabe. (H. M. Baumgartner, W. G. Jacobs, & H. Krings, Eds.). Frommann-Holzboog. AEIW (1796–1797). Abhandlungen zur Erläuterung des Idealismus der Wissenschaftslehre, Treatises Explicatory of the Idealism of the Doctrine of Science, in: Pfau, T. (1994). Idealism and the endgame of theory: Three essays by F.W.J. Schelling (T. Pfau, Ed. & Trans.). Albany: State University of New York Press, pp. 61–138. ÜMFP (1794). Über die Möglichkeit einer Form der Philosophie überhaupt; Concerning the Possibility of a Form of Philosophy in General, in: Schelling (1980). The unconditional in human knowledge: Four early essays 1794–6 (F. Marti, Ed. & Trans.). Bucknell University Press, pp. 35–58. VIPP (1795). Vom Ich als Prinzip der Philosophie oder über das Unbedingte im menschlichen Wissen; Of the I as Principle of Philosophy or of the Unconditioned in Human Knowledge, (F. Marti, Trans.), in: Schelling (1980), pp. 63–128. IPN (1797). Ideen zu einer Philosophie der Natur; (E. E. Harris, & P. Heath, Trans.) as Schelling, F W. J. (1988). Ideas for a Philosophy of Nature. Cambridge University Press. ESN (1799). Erster Entwurf eines Systems der Naturphilosophie; (K. R. Peterson 2004, Trans. & Ed.), as Schelling, F W. J. (2004). First Outline of a System of the Philosophy of Nature. State University of New York Press, pp. 1–192. E (1799). Einleitung zum Ersten Entwurf eines Systems der Natur philosophie. First Outline of a System of the Philosophy of Nature

Other texts

255

(2004). (K. R. Peterson 2004, Trans. & Ed.). State University of New York Press, pp. 193–232.

Other texts Alderwick, C. (2015). Freedom and powers in Schelling’s metaphysics [Doctoral Dissertation, University of Sheffield]. White Rose eTheses Online. https://etheses.whiterose.ac.uk/9336/1/Freedom%20and%20 Powers% 20in%20Schelling%27s%20Metaphysics.docx Allison, H. E. (2004). Kant’s transcendental idealism: An interpretation and defense. Yale University Press. Ameriks, K. (1992). The critique of metaphysics: Kant and traditional ontology. In P. Guyer (Ed.), The Cambridge companion to Kant (pp. 249–279). Cambridge University Press. Anstey, P., & Vanzo, A. (2012). The origins of Early Modern experimental philosophy. Intellectual History Review, 22(4), 499–518. Anstey, P. R. (2002). Locke, Bacon and Natural History. Early Science and Medicine, 7 (1), 65–92. Anstey, P., & Vanzo, A. (2016). Early Modern Experimental Philosophy. In J. Sytsma & W. Buckwalter (Eds.), A Companion to Experimental Philosophy (pp. 87–102). Blackwell. Atlas, S. (1964). From Critical to Speculative Idealism. M. Nijhoff. Bacon, F. (1905). The philosophical works of Francis Bacon (J. Spedding, R.L. Ellis, & J.M. Robertson, Eds.). G. Routledge & Sons. (Original work published 1765) Bacon, F. (2000). The New Organon (L. Jarine, & M. Silverthorne, Eds.). Cambridge University Press. (Original work published 1620) Bartha, P. (2019). Analogy and Analogical Reasoning, In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Spring 2019 ed.), Stanford University. Baum, M. (2000). Die Anfänge der Schellingschen Naturphilosophie. In C. Asmuth, A. Denker, & M. Vater (Eds.), Schelling: Zwischen Fichte und Hegel (pp. 95–112). Grüner.

256

References

Beiser, F. C. (1987). The Fate of Reason: German Philosophy From Kant to Fichte. Harvard University Press. Beiser, F. C. (2002). German idealism: The struggle against subjectivism, 1781–1801. Harvard University Press. Beiser, F. (2003). Maimon and Fichte. In G. Freudenthal (Ed.), Salomon Maimon: Rational dogmatist, empirical skeptic (pp. 233-248). Springer, Dordrecht. Bensaude-Vincent, B., & Blondel, C. (2016). Science and spectacle in the European Enlightenment. Routledge. Bergman, S. H. (1967). The philosophy of Solomon Maimon. Magnes Press, Hebrew University. Bertoloni M. D. (2016). The Rise of Modern Science Explained: A Comparative History, written by H. Floris Cohen. Early Science and Medicine 21(4), 363–368. Bethune, J. E. D. (1829). The Life of Galileo Galilei: With illustrations of the advancement of experimental philosophy. William Clowes. Bird, Graham (1962). Kant’s Theory of Knowledge: An Outline of One Central Argument in the Critique of Pure Reason. Humanities Press. Bird, G. (2017). Kant’s theory of knowledge: An outline of one central argument in the “Critique of Pure Reason” (1st ed.). Routledge. Boerhaave, H. (1753). Elementa chemiae (2nd ed.). Paris. Bokulich, A. (2009). Explanatory fictions. In M. Suárez (Ed.), Fictions in science: Philosophical essays on modeling and idealization (pp. 91–109). Routledge. Bonsiepen, W. (1997). Die Begründung einer Naturphilosophie bei Kant, Schelling, Fries und Hegel: Mathematische versus spekulative Naturphilosophie. Vittoria Klostermann. Boyle, R. (2008). Designe About Natural History. In M. Hunter, & P. R. Anstey (Eds.), Robert Boyle project occasional papers 3. Robert Boyle Project. (Original work published 1684) Brakel, J. (2006). Kant’s legacy for the philosophy of chemistry. In E. Scerri, D. Baird, & L. McIntyre (Eds.), Philosophy of chemistry. Boston Studies in the Philosophy of Science, vol. 242 (pp. 69–91). Springer. Brandom, Robert (2002). Tales of the Mighty Dead: Historical Essays in the Metaphysics of Intentionality. Harvard University Press.

Other texts

257

Bransen, J. (1991). The antinomy of thought: Maimonian skepticism and the relation between thoughts and objects. Kluwer. Breazeale, D. (1996). The Theory of Practice and the Practice of Theory: Fichte and the “Primacy of Practical Reason”. International Philosophical Quarterly 36(1), 47–64. Breazeale, D. (2001). Fichte’s Conception of Philosophy as a ‘Pragmatic History of the Human Mind’ and the Contributions of Kant, Platner, and Maimon. Journal of the History of Ideas, 62(4), 685–703. Breazeale, D. (2002). Fichte’s philosophical fictions. In D. Breazeale, & T. Rockmore (Eds.), New essays on Fichte’s later Jena Wissenschaftslehre (pp. 175–208). Northwestern University Press. Breazeale, D. (2003). Reinhold gegen Maimon über den Gebrauch der Fiktionen in der Philosophie. In M. Bondeli, & W. Shrader (Eds.), Die Philosophie Karl Leonhard Reinholds (pp. 123–151). Rodopi. Breazeale, D. (2014). Men at Work: Philosophical Construction in Fichte and Schelling. Europhilosophie Online. http://www.europhilosophie.eu/recherche/IMG/pdf/ Plenar-Breazeale-beitrag.pdf. Breazeale, D. (2018). Philosophy and “the method of fictions”: Maimon’s proposal and its critics. European Journal of Philosophy, 26(2), 702–716. Breidbach, O. (2004). Schelling und die Erfahrungswissenschaft. Sudhoffs Archiv für Geschichte der Medizin und der Naturwissenschaften, 88(2), 153–174. Brown, J. R. (1991). The laboratory of the mind: Thought experiments in the natural sciences. Routledge. Buchheim, T. (1987). Die Spekulation des Werdens in Platons ‘Timaios’. In R. Löw (Ed.), IOKEIOSIS: Festschrift für R. Spaemann zum 60. Geburtstag (pp.23–43). Weinheim. Buchheim, T. (1990). Das ‘objektive Denken’ in Schelling’s Naturphilosophie. Kant-Studien, 81(3), 321–338. Burtt, E. A. (1925). The metaphysical foundations of modern physical science; A historical and critical essay. Kegan Paul, Trench, Trubner. Butts, R. (1961). Hypothesis and explanation in Kant’s philosophy of science. Archiv für Geschichte der Philosophie, 43(2), 153–170.

258

References

Butts, R. (1962). Kant on hypothesis in the “Doctrine of Method” and the Logik. Archiv für Geschichte der Philosophie, 44(2), 185–203. Buzzoni, M. (2011). Kant und das Gedankenexperiment. Über eine kantische Theorie der Gedankenexperimente in den Naturwissenschaften und in der Philosophie. Deutsche Zeitschrift für Philosophie, 59(1), 93–107. Buzzoni, M. (2018). Kantian accounts of thought experiments. In M. T. Stuart, Y. Fehige, & J. R. Brown (Eds.), The Routledge companion to thought experiments (pp. 327–341). Routledge. Buzzoni, M. (2019). Thought experiments in philosophy: A NeoKantian and experimentalist point of view. Topoi, 38(4), 771–779. Camp, E. (2020). Imaginative frames for scientific inquiry. Metaphors, telling facts, and just-so stories. In A. Levy, & P. Godfrey-Smith (Eds.), The Scientific Imagination (pp. 304–336). Oxford University Press. Carnap, R. (1932). Überwindung der Metaphysik durch logische Analyse der Sprache. Erkenntnis, 2, 219–241. Carrier, M. (1986). Die begriffliche Entwicklung der Affinitätstheorie im 18. Jahrhundert. Newtons Traum—und was daraus wurde. Archive for History of Exact Sciences, 36(4), 327–389. Carrier, M. (1990). Kants Theorie der Materie und ihre Wirkung auf die zeitgenössische Chemie. Kant-Studien, 81(2), 170–210. Carrier, M. (2001). The theory of matter of Kant and his views on chemistry. In E. Watkins (Ed.), Kant and the Sciences (pp. 205–230). Oxford University Press. Carson, E. (1997). Kant on intuition in geometry. Canadian Journal of Philosophy, 27(4), 489–512. Carson, E. (1999). Kant on the method of mathematics. Journal of the History of Philosophy, 37(4), 629–652. Cartwright, N. (1983). How the laws of physics lie. Oxford University Press. Cassirer, E. (1974). Das Erkenntnisproblem in der Philosophie und Wissenschaft der neueren Zeit, Bd. 1. Wissenschaftliche Buchgesellschaft. Cassirer, E. (2000). Das Erkenntnisproblem in der Philosophie und Wissenschaft der neueren Zeit. Dritter Band. Die nachkantischen Systeme. Felix Meiner.

Other texts

259

Chakravartty, A. (2017). Scientific realism. In E. N. Zalta (Ed.), Stanford Encyclopedia of Philosophy (Summer 2017 ed.). Stanford University. https://plato.stanford.edu/entries/scientific-realism Chance, B. (2016). Kant and the discipline of reason. European Journal of Philosophy, 23(1), 87–110. Chalmers, A. (2012). Klein on the origin of the concept of chemical compound. Foundations of Chemistry, 14, 37–53. Cohen, H. (1877). Kants Begründung der Ethik. Dümmler. Cohen, H. (1885). Kants Theorie der Erfahrung. Dümmler. Conant, J. (2016). Why Kant Is Not a Kantian. Philosophical Topics 44(1), 75–125. Cooper, R. (2005). Thought experiments. Metaphilosophy, 36(3), 328–347. Copernicus, N., & Copernicus, N. (1973). De revolutionibus. Walter De Gruyter. (Original work published 1543). Crowe, B. (2008). Fichte’s fictions revisited. Inquiry, 51(3), 268–287. Currie, G. (1995). Imagination as simulation: Aesthetics meets cognitive science. In M. Davies, & T. Stone (Eds.), Mental Simulation (pp. 151–169). Blackwell. Daiber, J. (2001). Experimentalphysik des Geistes: Novalis und das romantische Experiment. Vandenhoeck & Ruprecht. De Boer, K. (2011). Transformations of transcendental philosophy: Wolff, Kant, and Hegel. Bulletin of the Hegel Society of Great Britain, 32(1-2), 50–79. De Boer, K. (2015). The vicissitudes of metaphysics in Kant and early Post-Kantian philosophy. Revista Portuguesa de Filosofia, 71(2–3), 267–286. De Boer, K., & Howard, S. (2018) A ground completely overgrown: Heidegger, Kant and the problem of metaphysics. British Journal for Philosophy, 27(2), 358–377. Della Rocca, Michael (2020). The Parmenidean Ascent. Cambridge University Press. Demeter, Tamàs (2012). Hume’s experimental method. British Journal for the History of Philosophy, 20(3), 577–599. Donovan, M. (1975). Lectures on the elements of chemistry, delivered in

260

References

the university of Edinburgh, by the late Joseph Black. The Edinburgh Review: Or Critical Journal, 3, 1–26. (Original work published 1806–1807) Douven, I. (2017). Abduction. In E. N. Zalta (Ed.), Stanford Encyclopedia of Philosophy (Summer 2021 ed.). Stanford University. https://plato.stanford.edu/entries/abduction/ Duffy, S. (2014). Maimon’s theory of differentials as the elements of intuition. International Journal of Philosophical Studies, 22(2), 228–247. Ehrensberger, F. (2004). Einleitung. In F. Ehrensberger (ed.), Maimon, Salomon: Versuch über die Transzendentalphilosophie (pp. VII– XLIX). Meiner. Ende, H. (1973). Der Konstruktionsbegriff im Umkreis des Deutschen Idealismus. Verlag Anton Hain. Engstler, A. (1990). Untersuchungen zum Idealismus Salomon Maimons. Frommann-Holzboog. Euclid, ., & Heath, T. L. (1956). The thirteen books of Euclid’s Elements. New York: Dover Publications. Falkenburg, B. (2018). Was beweist Kants Experiment der Vernunft? In V. Waibel, M. Ruffing, & D. Wagner (Eds.), Natur und Freiheit: Akten des XII. Internationalen Kant-Kongresses (pp. 641–660). De Gruyter. Farrington, B. (1964). The philosophy of Francis Bacon: An essay on its development from 1603–1609, with new translations of fundamental texts. Liverpool University Press. Fehige, Y. J.H. (2012). Experiments of pure reason: Kantianism and thought experiments in science. Epistemologia, 35(1), 141–160. Fehige, Y. J.H. (2013). The relativized a priori and the laboratory of the mind: Towards a Neo-Kantian account of thought experiments in science. Epistemologia, 36(1), 55–73. Fehige, Y. J.H., & Stuart, M. (2014). On the origins of the philosophy of thought experiments: The forerun. Perspectives on Science, 22(2), 179–220. Ferrarin, A. (2015). The powers of pure reason: Kant and the idea of cosmic philosophy. University of Chicago Press. Fichte, J. G. (1962-2012). J.G. Fichte: Gesamtausgabe der Bayerischen

Other texts

261

Akademie der Wissenschaften (E. Fuchs, R. Lauth, H. Jacobs, & H. Gliwitzky, Eds.). Frommann. Fichte, J. G. (2013). The philosophical rupture between Fichte and Schelling: Selected texts and correspondences (1800–1802) (M. G. Vater, & D. W. Wood, Eds. & Trans.). State University of New York Press. Fincham, R. M. (2016). Reconciling Leibnizian monadology and Kantian criticism. British Journal for the History of Philosophy, 23(6), 1033–1055. Fine, A. (1993). Fictionalism. Midwest Studies in Philosophy, 18(1), 1–18. Förster, E. (2018). Die 25 Jahre der Philosophie (3rd ed.). Vittorio Klostermann. Franks, P. (2003). What should Kantians learn from Maimon’s scepticism? In G. Freudenthal (Ed.), Salomon Maimon: rational dogmatist, empirical skeptic (pp. 200–232). Kluwer. Franks, P. (2005). All or nothing. Systematicity, transcendental arguments, and skepticism in German idealism. Harvard University Press. Franks, P. (2014). Skepticism after Kant. In J. Conant, & A. Kern (Eds.), Varieties of skepticism: Essays after Kant, Wittgenstein, and Cavell (pp. 17–58). De Gruyter. Franks, P. (2017). Finality and form in Kant’s critical philosophy. Hylomorphism Workshop, June 15, unpublished. Franks, P. (2020). From World-Soul to Universal Organism: Maimon’s Hypothesis and Schelling’s Physicalization of a Platonic- Kabbalistic Concept. In G. Anthony Bruno (Ed.), Schelling’s Philosophy: Freedom, Nature, and Systematicity (pp. 71–92), Oxford University Press. Freudenthal, G. (2006). Definition and construction. Salomon Maimon’s philosophy of geometry. Max Planck Institute. Freudenthal, G. (2010). Maimon’s philosophical program: Understanding versus intuition. International Yearbook of German Idealism, 8, 83–105. Freudenthal, G. (2019). Overturning the narrative: Maimon vs. Kant. Discipline Filosofiche, 29:1, 47–68. Friedman, M. (1985). Kant’s theory of geometry. The Philosophical Review, 94(4), 455–506.

262

References

Friedman, M. (1992). Kant and the exact sciences. Harvard University Press. Friedman, M. (2007). Kant—Naturphilosophie—Electromagnetism. In R. M. Brain, R. S. Cohen, & O. Knudsen (Eds.), Hans Christian Ørsted and the Romantic legacy in science (pp. 135–158). Springer. Friedman, M. (2012). Kant on geometry and spatial intuition. Synthese 186: 231–255. Friedman, M. (2013). Kant’s construction of nature. A reading of the metaphysical foundations of natural science. Cambridge University Press. Fulkerson-Smith, B. A. (2013). Bacon’s Illuminating Experiments and Kant’s Experiment of Pure Reason. In S. Bacin, A. Ferrarin, C. La Rocca, & M. Ruffing (Eds.), Kant und die Philosophie in weltbürgerlicher Absicht: Akten des XI. Kant-Kongresses 2010 (pp. 455–466). De Gruyter. Galileo, G. (1957). Excerpts from the Assayer. In S. Drake (Ed.), Discoveries and opinions of Galileo (pp. 231–80). Doubleday & Co. Galileo, G. (1967). The Dialogue Concerning the Two Chief World Systems. (S. Drake, Trans.). (2nd edition). University of California Press. (Original work published 1632). Galvani, L. (1791). De vibrus electricitatis in motu musculari commentarius. Ex Typographia Instituti Scientarium. Gaukroger, S. (2001). Francis Bacon and the transformation of EarlyModern philosophy. Cambridge University Press. Gaukroger, S. (2010). The collapse of mechanism and the rise of sensibility: Science and the shaping of Modernity, 1680–1760. Oxford University Press. Gaukroger, S. (2016). The natural and the human: Science and the shaping of Modernity, 1739–1841. Oxford University Press. Gava, G. (2015). Kant’s synthetic and analytic method in the Critique of Pure Reason and the distinction between philosophical and mathematical syntheses. European Journal of Philosophy, 23(3), 728–749. Gava, G. (2018). Kant, Wolff, and the method of philosophy. In D. Garber, & D. Rutherford (Eds.), Oxford studies in Early Modern philosophy (Vol. 8) (pp. 271–303). Oxford University Press. Gendler, T. S. (1998). Galileo and the indispensability of scientific

Other texts

263

thought experiment. The British Journal for Philosophy of Science, 49(3), 397–424. Gendler, T. (2004). Thought experiments rethought—and reperceived. Philosophy of Science, 71(5), 1152–1163. Gière, R. N. (1988). Explaining science: A cognitive approach. Chicago University Press. Glock, H-J. (2008). What is analytic philosophy? Cambridge University Press. Glock, H-J. (2013). Replies to my commentators. Journal for the History of Analytical Philosophy, 2(2), 35–42. Gloy, K. (1996). Kants Philosophie und das Experiment. In Y. Kato, & G. Schönrich (Eds.), Kant in der Diskussion der Moderne (pp. 64–91). Suhrkamp Verlag. Godfrey-Smith, P. (2006). The strategy of model-based science. Biology and Philosophy, 21(5), 725–740. Goethe, J. W. (1989). Der Versuch als Vermittler von Objekt und Subjekt. In E. Beutler (Ed.), Gedenkausgabe der Werke, Gespräche und Briefe (Vol. 16, pp. 844–855). Artemis Verlag. (Original work published 1823) Goh, K. (2014). Reinhold and the transformation of philosophy into science. In M. C. Altman (Ed.), The Palgrave handbook of German idealism (pp. 243–263). Palgrave Macmillan. Goldenbaum, U. (2016). The geometrical method as a new standard of truth, based on the mathematization of nature. In G. Gorham, B. Hill, E. Slowik, & C. K. Waters (Eds.), The language of nature: Reassessing the mathematization of nature. Minnesota University Press. Golinski, J. (1999). Science as public culture: Chemistry and Enlightenment in Britain, 1760–1820. Cambridge University Press. Gooding, D. C. (1990). Experiment and the making of meaning. Springer. Grant, I. H. (2006). Philosophies of nature after Schelling. Continuum. Guéroult, M. (1929). La philosophie transcendantale de Salomon Maimon. F. Alcan. Guicciardini, N. (2004). Dot-age: Newton’s mathematical legacy in the Eighteenth Century. Early Science and Medicine, 9(3), 218–256.

264

References

Guyer, P. (1987). Kant and the claims of knowledge. Cambridge University Press. Guyer, P. (2000). Absolute idealism and the rejection of Kantian dualism. In K. Ameriks (Ed.), The Cambridge companion to German idealism (pp.43–64). Cambridge University Press. Hacking, I. (1975). The emergence of probability. Cambridge University Press. Hacking, I. (1983). Representing and intervening. Cambridge University Press. Hall, A. R. (1963). From Galileo to Newton. Dover Inc. Hegel, G. (2010). Preface to the First Edition. In D. Di Giovanni (Ed.), Georg Wilhelm Friedrich Hegel: The Science of Logic (Cambridge Hegel Translations, pp. 7–11). Cambridge University Press. Heidegger, M. (1956). Was ist das—die Philosophie? Günter Neske. Heidegger, M. (1990). Kant and the problem of metaphysics (R. Taft, Trans.) (4th ed.). Indiana University Press. Henrich, D. (1967). Fichtes ursprüngliche Einsicht. Vittorio Klostermann. Henrich, D. (1969). The proof-structure of Kant’s transcendental deduction. The Review of Metaphysics, 22(4), 640–659. Hintikka, J., & Remes, U. (1974). The method of analysis. Springer. Hobbes, T. (1997). Elemente der Philosophie. Erste Abteilung: Der Körper (K. Schuhmann, Ed.). Felix Meiner. Hochadel, O. (2003). Öffentliche Wissenschaft. Elektrizität in der deutschen Aufklärung. Wallstein Verlag. Hoffmann, T. S. (2010). ‘Darstellung des Begriffs’. Zu einem Grundmotiv neueren Philosophierens im Ausgang von Kant. In H. Busche, & A. Schmitt (Eds.) Kant als Bezugspunkt philosophischen Denkens. Festschrift für Peter Baumanns zum 75. Geburtstag (pp. 101–116). Königshausen & Neumann. Holmes, F. L. (2003). The physical sciences in the life sciences. In M. J. Nye (Ed.), The Cambridge history of science: Vol. 5. The Modern physical and mathematical sciences (pp. 219–236). Cambridge University Press. Hooke, R. (1664). Micrographia. Jo Martyn & Co. Hooke, R. (1705). The Posthumous Works of Robert Hooke, ...

Other texts

265

taining His Cutlerian Lectures, and Other Discourses, Read at the Meetings of the Illustrious Royal Society. ... Illustrated with Sculptures. To These Discourses is Prefixt the Author’s Life, ... Publish’d by Richard Waller. Sam. Smith and Benj. Walford. Hoyningen-Huene, P. (2008). Thomas Kuhn and the chemical revolution. Foundations of Chemistry, 10, 101–115. Hume, D. (1975). David Hume: enquiries concerning human understanding and concerning the principles of morals (P. H. Nidditch, Ed.) (3rd ed.). Oxford University Press. Hume, D. (2011). David Hume: A treatise of human nature, Volume 1. Texts (D. Norton, Ed.). Oxford University Press. (Original work published 1739-40) Hunter, M. (2007). Robert Boyle and the Early Royal Society : A Reciprocal Exchange in the Making of Baconian Science. The British Journal for the History of Science, 40(1), 1–23. Jacobi, F. H. (2019). David Hume über den Glauben oder Idealismus und Realismus. Ein Gespräch (1787). Felix Meiner Verlag. (Original work published 1787). Jalobeanu, D. (2016). “The marriage of physics with mathematics”: Francis Bacon on measurement, mathematics, and the construction of a mathematical physics. In G. Gorham, B. Hill, E. Slowik, & C. K. Waters (Eds.), The language of nature: Reassessing the mathematization of natural philosophy in the Seventeeth Century (pp. 51–80). Minnesota University Press. https://doi.org/10.5749/j.ctt1d390rg.5 Jesseph, D. M. (2008). Truth in fiction: origins and consequences of Leibniz’s doctrine of infinitesimal magnitudes. In U. Goldenbaum, & D. Jesseph (Eds.), Infinitesimal differences: Controversies between Leibniz and his contemporaries (pp. 215–234). De Gruyter. Kalin, M. G. (1972). Kant’s transcendental arguments as Gedankenexperimente. Kant-Studien, 63(1–4), 315–328. Karsten, W. J. G. (1783). Anleitung zur gemeinnützlichen Kenntnis der Natur. Halle. Katz, M., & Sherry, D. (2013). Leibniz’s Infinitesimals: Their Fictionality, Their Modern Implementations, and Their Foes from Berkeley to Russell and Beyond. Erkenntnis 78(3), 571–625.

266

References

Kauferstein, C. (2006). Transzendentalphilosophie der Mathematik: Versuch einer systematischen Rekonstruktion der Leitlinien einer Philosophie der Mathematik in Kants ”Kritik der reinen Vernunft” und Maimons “Versuch über die Transzendentalphilosophie”. IbidemVerlag. Kim, S.-H. (2008). Bacon Und Kant: Ein Erkenntnistheoretischer Vergleich Zwischen Dem Novum Organum Und der Kritik der Reinen Vernunft. Walter de Gruyter. Kimhi, I. (2018). Thinking and being. Harvard University Press. Kitcher, P. (1990). Kant’s transcendental psychology. Oxford University Press. Kitcher, P. (1986). Projecting the order of nature. In R. E. Butts (Ed.), Kant’s Philosophy of Physical Science (pp. 201–235). Springer. Klein, U. (1994a). Origin of the concept of chemical compound. Science in Context, 7(2), 163–204. Klein, U. (1994b). Verbindung und Affinität: Die Grundlegung der Neuzeitlichen Chemie an der Wende vom 17. Zum 18. Jahrhundert. Birkhäuser Verlag. Klein, U. (1995). E. F. Geoffroy’s table of different ‘rapports’ observed between different chemical substances—A reinterpretation. Ambix, 42(2), 79–100. Klein, U. (1996). The chemical workshop tradition and the experimental practice: Discontinuities within continuities. Science in Context, 9(3), 251–287. Klein, U., & Lefèvre, W. (2007). Materials in Eighteenth-Century science: A historical ontology. MIT Press. Kreis, G. (2019). The idea of transcendental analysis: Kant, Marburg Neo-Kantianism, and Strawson. British Journal for the History of Philosophy, 27(2), 293–314. Krings, H. (1982). Kann man die Natur verstehen? In W. Kuhlmann, & D. Böhler (Eds.), Kommunikation und Reflexion: Zur Diskussion der Transzendental-Pragmatik: Antworten auf Karl-Otto Apel (pp. 391–378). Suhrkamp. Krings, H. (1985). Natur als Subjekt. Ein Grundzug der spekulativen Physik Schellings. In R. Heckmann, H. Krings, & R. W. Meyer (Eds.), Natur und Subjektivität: Zur Auseinandersetzung mit der

Other texts

267

Naturphilosophie des jungen Schelling (pp. 111–128). FrommannHolzboog. Kroner, Richard (1921). Von Kant Bis Hegel. J.C.B. Mohr. Kuhn, T.S. (1962). The structure of scientific revolutions. University of Chicago Press. Kühne, U. (2005). Die Methode des Gedankenexperiments. Suhrkamp. Kühne, U. (1977). The essential tension. Chicago University Press. Kuntze, F. (1912). Die Philosophie Salomon Maimons. Carl Winter. Lachterman, D. (1992). Mathematical construction, symbolic cognition and the infinite intellect: Reflections on Maimon and Maimonides. Journal of the History of Philosophy, 30(4), 497–522. Lakatos, I. (1970). Falsification and the methodology of scientific research programmes. In A. Musgrave (Ed.), Criticism and the Growth of Knowledge (pp. 91–196). Cambridge University Press. Thomas Land (2014). Spatial representation, magnitude and the two stems of cognition. Canadian Journal of Philosophy, 44(5-6), 524–550. Lazerowitz, M. (1964). Studies in metaphilosophy. Routledge and Kegan Paul. Leibniz, G. W. (1963). Leibniz, Mathematische Schriften. 7 vols [= GM]. (C. I. Gerhardt, Ed.). georg Olms. (Original edition published 1849–63). Leibniz, G. W. (1965). Die philosophischen Schriften. 7 vols. (C. I. Gerhardt, Ed.). Georg Olms. (Original edition published 1875–90). Leibniz, G. W. (1989). Philosophical essays [= PE] (R. Ariew, & D. Garber, Trans.). Hackett Publishing Company. Levey, S. (2008). Archimedes, infinitesimals and the Law of Continuity: On Leibniz’s fictionalism. In U. Goldenbaum, & D. Jesseph (Eds.), Infinitesimal differences: Controversies between Leibniz and his contemporaries (pp. 107–133). De Gruyter. Lipton, P. (2000). Inference to the Best Explanation. In W. H. NewtonSmith (Ed.), A companion to the philosophy of science. Blackwell. Locke, J. (1975). An Essay Concerning Human Understanding (P. H. Nidditch, Ed.). Clarendon Press. (Original work published 1690) Löw, R. (1979). Qualitätslehre und Materienkonstruktion. Zur systematischen Aktualität von Schellings Naturphilosophie. In L. Hasler

268

References

(Ed.), Schelling. Seine Bedeutung für eine Philosophie der Natur und der Geschichte (pp. 99–106). Froomann-Holzboog. Macquer, P. J. (1749). Eléments de chimie-théorique. Chez Jean-Thomas Herissant. Mahoney, E. P. (1998). The mathematical realm of nature. In D. Garber, & M. Ayers (Eds.), The Cambridge history of Seventeenth-Century philosophy (Vol. 1, pp. 702–55). Cambridge University Press. Maimon, S. (1965–1976). Gesammelte Werke (V. Verra, Ed.). Olm. Maimon, S. (2000). Giv’at ha-Moreh (S. H. Bergman, & N. Rotenstreich, Eds.). Israel Academy of Sciences and Humanities. Maimon, S. (2010). Essay on transcendental philosophy (N. Midgley, H. Somers-Hall, A. Welchman, & M. Reglitz, Trans.). Continuum. (Original work published 1790) Martens, R. (2000). Kepler’s philosophy and the New Astronomy. Princeton University Press. McDowell, J. (1994). Mind and World. Harvard University Press. McLear, C. (2015). Two kinds of unity in the Critique of Pure Reason. Journal of the History of Philosophy, 53(1), 79–110. McMullin, E. (1984). Two ideals of explanation in natural science. Midwest Studies of Philosophy, 9(1), 205–220. McMullin, E. (1985). Galilean idealization. Studies in History and Philosophy of Science Part A, 16(3), 247. McNulty, M. B. (2014a). Kant on chemistry and the application of mathematics in natural science. Kantian Review, 19(3), 393–418. McNulty, M. B. (2014b). Kant’s philosophy of chemistry [Doctoral dissertation, UC Irvine]. eScholarship. https://escholarship.org/uc/item/7qw2r7hp McNulty, M. B. (2015). Rehabilitating the regulative use of reason: Kant on empirical and chemical laws. Studies in History and Philosophy of Science Part A, 54, 1–10. McNulty, M. B. (2017). What is chemistry, for Kant? Kant Yearbook, 9(1), 85–112. Mendelssohn, Moses (1785). Morgenstunden oder Vorlesungen über das Dasein Gottes. Christian Friederich Voss. Mercer, C. (2019). The contextual revolution in Early Modern philosophy. Journal of the History of Philosophy, 57(3), 529–548.

Other texts

269

Mitchell, R. (2013). Experimental life—Vitalism in Romantic science & literature. Johns Hopkins University Press. Moore, A. (2010). The transcendental doctrine of method. In P. Guyer (Ed.), The Cambridge companion to Kant’s Critique of Pure Reason (pp. 310–326). Cambridge University Press. Møller, Sofie C. (2013). The Court of Reason in Kant’s Critique of Pure Reason. Kant-Studien 104(3), 301-320. Møller, Sofie (2020). Kant’s Tribunal of Reason: Legal Metaphor and Normativity in the Critique of Pure Reason. Cambridge University Press. Mutschler, H.-D. (1990). Spekulative und empirische Physik— Aktualität und Grenzen der Naturphilosophie Schellings. Kohlhammer. Nassar, D. (2010). From a philosophy of self to a philosophy of nature: Goethe and the development of Schelling’s Naturphilosophie. Archiv für Geschichte der Philosophie, 92(3), 304–321. Nassar, D. (2012). Spinoza in schelling’s early conception of intellectual intuition. In E. Förster, & Y. Melamed, Spinoza and German idealism (pp. 136–155). Cambridge University Press. Nassar, D. (2013a). The Romantic absolute. Being and knowing in Early German Romantic philosophy, 1795–1804. Chicago University Press. Nassar, D. (2013b). Intellectual intuition and the philosophy of nature. In J. Haag, & M. Wild (Eds.), Übergänge—diskursiv oder intuitive? Essays zu Eckart Försters “Die 25 Jahre der Philosophie” (pp. 235–258). Vittoria Klostermann. Nassar, D. (2016). Nature as unconditioned? The critical and systematic function of Schelling’s early works. In K. Gjesdal (Ed.), Debates in Nineteenth-Century European philosophy (pp. 121–132). Routledge. Natorp, P. (1912). Kant und die Marburger Schule. Kant-Studien, 17, 193–221. Neuhouser, F. (1990). Fichte’s theory of subjectivity. Cambridge University Press. Nersessian, N.J. (1992). In the theoretician’s laboratory: Thought experimenting as mental modeling. PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association, 1992, 291–301. Newton, I. (1984). The optical papers of Isaac Newton: The optical

270

References

lectures: Vol. 1. 1670–1672 (A. Shapiro, Ed.). Cambridge University Press. Newton, I. (1999). The Principia: Mathematical principles of natural philosophy: A new translation (I. B. Cohen, & A. Whitman, Trans.) University of California Press. Nisenbaum, K. (2018), For the love of metaphysics. Oxford University Press. Nowack, L. (1992). The idealizational approach to science: A survey. In J. Brenzinski and L. Nowack (Eds.), Idealization III: Idealization and truth (pp. 9–63). Rodopi. Olson, M. (2018). On the significance of the Copernican revolution: Transcendental philosophy and the object of metaphysics. Con-Textos Kantianos, 44(5), 528–542. Ørsted, C. H. (1998). Selected scientific works of Hans Christian Ørsted (K. Jelved, A. D. Jackson, & O. Knudsen, Ed. & Trans.). Princeton University Press. Ostaric, L. (2014). The concept of life in early Schelling. In L. Ostaric (Ed.), Interpreting Schelling (pp. 48–70). Cambridge University Press. Overgaard, S., Gilbert, P., & Burwood, S. (2013). An introduction to metaphilosophy. Cambridge University Press. O’Shea, J. R. (2019). On Sellar’s exam question trilemma: Are Kant’s premises analytic, or synthetic a priori, or a posteriori? British Journal for the History of Philosophy, 27(2), 402–421. Paul, L. (2012). Metaphysics as modeling: The Handmaiden’s tale. Philosophical Studies, 160(1), 1–29. Pettersson, H. (2019). Metaphilosophical themes: Naturalism, rationalism and the conceptions of philosophy [Doctoral dissertation, University of Turku]. UTUPub. https://www.utupub.fi/handle/10024/147021. Pippin, R. (1988). Fichte’s Contribution in Fichte and Contemporary Philosophy. Philosophical Forum 19(2–3), 74–96. Pollok, K. (2010). ‘The transcendental method’: On the reception of the Critique of Pure Reason in Neo-Kantianism. In P. Guyer (Ed.), The Cambridge companion to Kant’s Critique of Pure Reason (pp. 346–379). Cambridge University Press.

Other texts

271

Poser, H. (1979). Spekulative Physik und Erfahrung. Zum Verhältnis von Experiment und Theorie in Schellings Naturphilosophie. In L. Hasler (Ed.), Schelling – Seine Bedeutung für eine Philosophie der Natur und Geschichte (pp. 129–138). Frommann-Holzboog. Posy, C. J. (1992). Kant’s philosophy of mathematics. Kluwer Academic Publishers. Powers, J. C. (2012). Inventing chemistry. Herman Boerhaave and the reform of the chemical arts. Chicago University Press. Prauss, G. (1974) Zur Problematik der Dinge an Sich. In G. Funke (Ed.), Akten des 4. Internationalen Kant-Kongresses: Mainz, 6.–10. April 17974, Teil 2: Sektionen 1,2 (pp. 222–239). De Gruyter. Pringé, H. (2018). Maimon’s criticism of Kant’s doctrine of mathematical cognition and the possibility of metaphysics as a science. Studies in History and Philosophy of Science, 71, 35–44. Proops, I. (2003). Kant’s Legal Metaphor and the Nature of a Deduction. Journal of the History of Philosophy 41(2), 209–229. Proops, I. (2021). The fiery test of the Critique. Oxford University Press. Reese, William L. (1990). Morris Lazerowitz and Metaphilosophy. Metaphilosophy 21(1-2): 28–42. Reinhold, K. (1795). Versuch einer Neuen Theorie des menschlichen Vorstellungsvermögens. 2. Auflage. Widtmann und Mauke. Rescher, N. (2014). Philosophy in philosophical perspective. Lexington Books. Reinhold, K. J. (2011). Essay on a New Theory of the Human Capacity for Representation. De Gruyter. (Original work published 1789). Rheinberger, H.-J. (2001). Experimentalsysteme und epistemische Dinge— Eine Geschichte der Proteinsynthese im Reagenzglas. Wallstein Verlag. Rheinberger, H.-J. (2003). Präparate—‘Bilder’ ihrer selbst. Eine bildtheoretische Glosse. In H. Bredekamp & G. Werner (Eds.), Bildwelten des Wissens. Kunshistorisches Jahrbuch für Bildkritik (pp. 9–19). Akademie Verlag. Richards, R.J. (2002). The Romantic conception of life—Science and philosophy in the age of Goethe. University of Chicago Press. Richards, R.J. (2006). Nature is the poetry of mind, or how Schelling solved Goethe’s Kantian problems. In M. Friedman, & A. Nordmann

272

References

(Eds.), The Kantian legacy in Nineteenth Century science (pp. 27–50). MIT Press. Richardson, A. (2006). ‘The fact of science’ and the critique of knowledge: exact science as problem and resource in Margburg NeoKantianism. in M. Friedman, & A. Nordmann (Eds.), The Kantian legacy in Nineteenth Century science (pp. 211–226). MIT Press. Ritter, J. W. (1798). Beweis, dass ein beständiger Galvanismus den Lebensprozess im Thierreich begleite. Nebst neuen Versuchen und Bemerkungen über den Galvanismus. Weimar. Ritter, J.W. (1800). Beyträge zur nähern Kenntniss des Galvanismus und der Resultate zu seiner Untersuchung. Ersten Bandes erstes und zweytes Stück. Frommann. Rockmore, T. (1996). Fichte’s antifoundationalism, intellectual intuition and who one is. In D. Breazeale, & T. Rockmore (Eds.), Fichte: historical contexts/contemporary controversies (pp. 79–94). Humanities Press. Rockmore, T. (2019). German idealism, epistemic constructivism, and metaphilosophy. Epistemology & Philosophy of Science, 56(4), 139– 154. Rorty, R. (1967). Linguistic turn: Recent essays in philosophical method. Chicago University Press. Rosenberg, J. F. (1975). Transcendental arguments revisited. Journal of Philosophy, 72(18), 611–624. Rudolphi, M. (2001). Produktion und Konstruktion—Zur Genese der Naturphilosophie in Schellings Frühwerk. Frommann-Holzboog. Sandkaulen, B. (1990). Ausgang vom Unbedingten. Über den Anfang in der Philosophie Schellings. Vandenhoeck und Ruprecht. Schelling, F. W. J. (1976). Historisch-Kritische Ausgabe (H. Fuhrmans, W. G. Jacobs, & H. Krings, Eds.). Frommann-Holzboog. Schelling, F. W. J. (1980). The unconditional in human knowledge: Four early essays 1794–6 (F. Marti, Ed. & Trans.). Bucknell University Press. Schelling, F. W. J. (1994). Idealism and the endgame of theory: Three essays by F.W.J. Schelling (T. Pfau, Ed. & Trans.). State University of New York Press.

Other texts

273

Schelling, F. W. J. (2004). First outline of a system of the philosophy of nature (K. R. Peterson, Trans.). State University of New York Press. Schlenstedt, D. (2000). Darstellung. In K. Barck, M. Fontius, B. Steinwachs & F. Wolfzettel (Eds.), Ästhetische Grundbegriffe 1 (pp. 831–875). Stuttgart: Metzler Verlag. Schmid, J. (2018). Schelling’s method of Darstellung: Presenting nature through experiment. Studies in History and Philosophy of Science Part A, 69, 12–22. Schmid, J. (2020). Es ist so, weil ich es so mache: Fichtes Methode der Konstruktion. Fichte-Studien, 48(2), 389–412. Schmid, J. (2022). The experiment of pure reason: a chemical reading [unpublished manuscript]. University of Basel. Schönecker, D., Schulting, D., & Strobach, N. (2011). Kants kopernikanisch- newtonianische Analogie. Deutsche Zeitschrift für Philosophie, 59(4), 497–518. Schubbach, A. (2017). Kants Konzeption der geometrischen Darstellung. Kant-Studien, 108(1), 19–54. Schultze, G. E. (1996/1792). Aenesidemus oder über die Fundamente der Herrn Professor Reinhold in Jena gelieferten Elementar-Philosophie: Nebst einer Verteidigung des Skeptizismus gegen die Anmassungen der Vernunft. Meiner. Sellars, W. (1967). Some remarks on Kant’s theory of experience. The Journal of Philosophy, 64(20), 633–647. Sellars, W. (2002). Kant’s transcendental metaphysics. Sellars’ Cassirer Lectures notes and other essays (Jeffrey F. Sicha, Ed.). Ridgeview Publishing Company. Senderowicz, Y. (2003). Maimon’s ‘Quid Facti’ Argument. In G. Freudenthal (Ed.), Salomon Maimon: Rational dogmatist, empirical skeptic (pp. 176–199). Kluwer. Serck-Hanssen, C. (2017). Rediscovering the Critique of Pure Reason as a propaedeutic to metaphysics: What Heidegger saw and McDowell missed. In A. M. Rasmussen, & M. Gabriel (Eds.), German idealism today (pp. 215–230). De Gruyter. Shabel, L. (2003). Mathematics in Kant’s critical philosophy: Reflections on mathematical practice. Routledge.

274

References

Shapin, S., & Schaffer, S. (1983). Leviathan and the air-pump: Hobbes, Boyle, and the experimental life. Princeton University Press. Shapiro, A. E. (2004). Newton’s “Experimental Philosophy.” Early Science and Medicine, 9(3), 185–217. Shi-Hyong, K. (2008). Kant und Bacon: ein erkenntnistheoretischer Vergleich zwischen dem “Novum Organum” und der “Kritik der reinen Vernunft”. De Gruyter. Snow, D. E. (1996). Schelling and the end of idealism. State University of New York Press. Socher, A. P. (2006). The radical Enlightenment of Solomon Maimon: Judaism, heresy, and philosophy. Stanford University Press. Spinoza, B. D., & Curley, E. M. (1988). The collected works of Spinoza. Princeton University Press. Stahl, G. E. (1718). Zufällige Gedancken und nützliche Bedencken über den Streit von dem sogenannten Sulphure, und zwar sowol dem gemeinen, verbrennlichen, oder flüchtigen, als unverbrennlichen, oder fixen. In Verlegung des Weisenhauses. Stahl, G. E. (1720). Anweisung zur Metallurgie, Oder der metallischen Schmelz- und Probier-Kunst. Nebst dessen Einleitung Zur GrundMixtion derer unterirrdischen mineralischen metallischen Cörper. Alles mit gründlichen Rationibus, Demonsstrationibus und Experimentis nach denen Becherischen Principiis ausgeführet. Caspar Jacob Ensseln. Stahl, G. E. (1730). Philosophical principles of universal chemistry. J. Osborn and T. Longman. Stahl, G. E. (1734). Zymotechnia fundamentalis oder allgemeine Grund Erkänntniss der Gährungskunst … aus dem lateinischen ins Teutsche übersetzt. Regensburg Johann Leopold Montag. Stanford, P. K. (2016). Instrumentalism: Global, local, and scientific. In P. Humphrey (Ed.), The Oxford handbook of philosophy of science (pp. 318–336). Oxford University Press. Stang, N. (2022). Kant’s transcendental idealism. In E. N. Zalta (Ed.), Stanford Encyclopedia of Philosophy (2021 ed.). Stanford University. https://plato.stanford.edu/entries/kant-transcendental-idealism/ Stern, B. (1999). Transcendental arguments. Oxford University Press.

Other texts

275

Steinle, F. (2005). Explorative Experimente. Ampère, Faraday und die Ursprünge der Elektrodynamik. Franz Steiner Verlag. Sticker, M. (2017). Experiments in ethics? Kant on chemistry, means of education and practical philosophy. Idealistic Studies, 46(1), 41–64. Strawson, P. (1966). The bounds of sense. Philosophy, 42(162), 379–382. Strawson, P. (1964). A function for thought experiments. In T. S. Kuhn (Ed.), The essential tension: Selected studies in scientific tradition and change (pp. 240–265). University of Chicago Press. Strawson, P. F. (1973). Different conceptions of analytical philosophy. Tijdschrift Voor Filosofie, 35(4), 800–834. Stroud, B. (1968). Transcendental arguments. Journal of Philosophy, 65, 241–256. Suárez, M. (2008). Fictions in science: Philosophical essays on modeling and idealization. Routledge. Taureck, B. (1979). Das Schicksal der philosophischen Konstruktion. Oldenbourg. Thielke, P. (2001a). Getting Maimon’s goad: Discursivity, skepticism, and Fichte’s idealism. Journal of the History of Philosophy, 39(1), 101–134. Thielke, P. (2001b). Discursivity and causality: Maimon’s challenge to the second analogy. Kant-Studien, 92(4), 440–463. Thielke, P. (2003). Intuition and diversity: Kant and Maimon on space and time. In G. Freudenthal (Ed.), Rational dogmatist, empirical skeptic (pp. 89–124). Kluwer Academic Publishers. Thielke, P. (2008). Apostate rationalism and Maimon’s Hume. Journal of the History of Philosophy, 46(4), 591–618. Thielke, P. (2014). Rationalism, empiricism, and skepticism: The curious case of Maimon’s ‘coalition-system’. In M. C. Altman (Ed.), The Palgrave handbook of German idealism (pp. 222–242). Palgrave Macmillan. Thielke, P., & Melamed, Y. (2008). Salomon Maimon. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Summer 2021 Edition). Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/sum2021/entries/maimon/. Tiles, J. E. (1993). The experiment as intervention. British Journal for the Philosophy of Science, 44(3), 463–475.

276

References

Tonelli, G. (1976). Analysis and synthesis in XXVIII. Century science prior to Kant. Archiv für Begriffsgeschichte, 20, 178–213. Tonelli, G., & Chandler, D. H. (1994). Kant’s “Critique of Pure Reason” within the tradition of modern logic: A commentary on its history. Georg Olms Verlag. Vahid, H. (2011). Scepticism and varieties of transcendental arguments. Logos & Episteme, 2, 295–411. Vaihinger, H. (1935). The philosophy of ‘as if’. Kegan Paul, Trench, Trubner & Co., LTD. Van Fraassen, B. (1980). The scientific image. Oxford University Press. Vanzo, A. (2012). Kant on experiments. In J. Maclaurin (Ed.), Rationis Defensor. Springer. Vollrath, E. (1962). Die Gliederung der Metaphysik in eine Metaphysica generalis und eine Metaphysica specialis. Zeitschrift für philosophische Forschung, 16(2), 258–284. Weisberg, M. (2007a). Three kinds of idealization. Journal of Philosophy, 104(12), 639–659. Weisberg, M. (2007b). Who is a modeler? British Journal for the Philosophy of Science, 58(2), 207–233. Westphal, K. (2017). Thought experiments, epistemology & our cognitive capacities. In M. T. Stuart, Y. J. H. Fehige, & J. R. Brown (Eds.), The Routledge companion to thought experiments (pp. 128–149). Routledge. Whistler, D. (2013). Schelling’s theory of symbolic language. Oxford University Press. Wild, C. (1968). Reflexion und Erfahrung. Verlag Karl Alber. Williamson, T. (2007). The philosophy of philosophy. Blackwell. Williamson, T. (2017). Model-building in philosophy. In R. Blackford, & D. Broderick (Eds.), Philosophy’s future: The Problem of philosophical progress (pp. 159–172). Wiley. Wolff, C. (1996). Discursus praeliminaris de philosophia in genere. Einleitende Abhandlung über Philosophie im allgemeinen. FrommannHolzboog. Wood, D. (2012). Mathesis of the mind. Rodopi. Wood, P. B. (1990). The natural history of man in the Scottish Enlightenment. History of Science, 28(1), 89–123.

Other texts

277

Ziche, P. (2010). ‘Höherer’ Empirismus: Passive Wissenschaft, letzte Tatsachen und experimentelle Philosophie bei F. W. J. Schelling. In J. Stolzenberg, & F. Rush (Eds.), Philosophie und Wissenschaft / Philosophy and science (pp. 165–184). De Gruyter. Ziche, P. (2011). Das System als Medium. Mediales Aufweisen und deduktives Ableiten bei Schelling. In W. G. Jacobs, J. Stolzenberg, & V. Waibel (Eds.), System und Systemkritik um 1800 (pp. 147–168). Felix Meiner Verlag. Zubersky, A. (1925). Salomon Maimon und der kritische Idealismus. Meiner.