The Oxford Handbook of German Philosophy in the Nineteenth Century
Page 104
After the middle of the century, however, it became apparent that a merely introspective psychology is itself at odds with the standards of the sciences properly speaking. In consequence, a debate about a new form of skepticism emerges in philosophy starting in the 1870s: if the foundations of the sciences that philosophy is looking for are conceived psychologistically, skepticism follows. Note the difference from a standard skepticism of a Cartesian variety: according to the standard skepticism, we start with our mental representations and ask whether something corresponds to them. Thereby we both challenge and reassure our trust in our (scientific) knowledge within philosophical reasoning. In the new version, we start with the sciences which have a credibility of their own. However, philosophy itself can undermine this credibility by leading us to the wrong (psychologistic) type of basis of the scientific reasoning. In the case of standard skepticism we proceed within one theoretical framework of the mind. In the new debate, the question of which framework to choose is itself the central issue.
28.6 HELMHOLTZ
In my discussion of skepticism and epistemology in the post-Hegelian period, I want to focus particularly on the work of Hermann Helmholtz and the Marburg branch of neo-Kantianism. Helmholtz fully endorses both the experimental method and the content of the sciences. Part of his work can be read as an effort to provide counterparts to claims originating from Kant’s transcendental aesthetic within the natural sciences. The second tendency of nineteenth-century philosophy I have just mentioned is implicitly present in Helmholtz’s reasoning, too, since a feature of the mind can only be spelled out in a naturalistic theory if it has been transferred to the empirical domain in the first place. Within the empirical domain, Helmholtz is particularly interested in experiments that show that sensations cannot reasonably be considered as inner images that mirror the properties of the objects that affect us (a common sense claim that would apply to the standard cases where we are not dealing with the familiar phenomenon of misperception).18 The experiments reveal, for example, that the same visual sensation can be caused both by light waves and by an electrical current. Despite the similarity in the sensation the effect can be traced back to two different physical sources. In addition, the same cause can produce both the impression of light through the eye and the impression of heat through the skin. Even though the origin is the same, the sensory content differs. A third experiment shows that our vision of colors indicates similarities where they cannot be found under any other description.
These concerns about the accuracy of the mapping function underlying our sensory content tend to undermine our everyday life approach to reality. Shouldn’t we stop trusting our senses? This line of thought would invite a form of skepticism if these concerns were not dependent on the acceptance of the scientific experiments in the first place. However, Helmholtz does not want to rely only on reminding us of the nature of an experiment. He also tries to provide a transcendental argument with (limited) anti-skeptical force. Unlike the content of sensations, at least the principle of causality must necessarily be correct, since, whenever I deal with empirical information as it is present to me, I will have to deal with several instances of this information. If I do so in a way that conforms to scientific method, I have to rely on inductive reasoning. Inductive reasoning, in turn, already presupposes causality in its foundation (as the philosophy of science of the day claims). Thus, our belief in causality could never be challenged by considerations of a type which is similar to those Helmholtz uses to undermine our trust in sensory content. Causality would always be presupposed. As scientifically minded persons, we cannot help presupposing causality (H, 47ff.).
The power of the argument is limited, because Helmholtz believes that it is compatible with a version of skepticism that is even more thoroughgoing than the experimental challenge: the dream hypothesis. That is, the scenario in which I could live in a highly consistent dream.
In order to explain Helmholtz’s reasoning, I have to introduce more of Helmholtz’s own premises. Helmholtz conceives of the perceptual situation more generally as consisting of two elements. On the one hand, there is a volitional impulse. In a non-skeptical setting the impulse directs my body in space. On the other hand, there is the perceptual information I receive in response to my action. (Despite Helmholtz’s apparent Kantianism, this is obviously an empirical transformation of Fichte’s action-based approach to knowledge (H, 41).) In the skeptical scenario, of course, the bodily movement would not actually occur. I would be limited to the sphere of the dream. Nevertheless, the sensory content would still appear not to be directly brought about by the volitional impulse; both would still be correlated in a regular way. In a first effort, Helmholtz argues that even in this case we would have to assume a causal nexus underlying the two subjective elements in the perceptual situation. We would call it real (H, 45). It correlates volitional impulse and sensation and it regulates their appearances. However, this reasoning is compatible with the fact that the causes could be totally different from what the sensory content, which seems to map its causes, tells us. According to Helmholtz, this remaining challenge cannot be addressed on a priori grounds; but again we can deal with the challenge by referring to the standards of science.
Below the level of the transcendentally justified validity of causality, the scientific worldview offers different levels of criteria we can use for evaluation. First, there are constraints such as simplicity. (Does the dream hypothesis really generate the simplest explanation of the consistency of what I see?) Secondly, hypotheses should in principle be open to empirical testing. This is why Helmholtz so carefully chose his model of a correlation between an impulse and the appearing empirical sensations. The model is meant to be similar in structure to the scientific hypotheses we subject to experimental verification. (A radical skeptic would probably insist that the process of testing itself just happens in a dream even though he cannot provide a convincing explanation for the correlation.) Finally there are pragmatic grounds for accepting a hypothesis: in order to act, we have to accept various claims without being able to wait until science produces results. According to Helmholtz, a hypothesis which is not sensible to any of these criteria is just empty (H, 46).
28.7 NEO-KANTIANISM
Neo-Kantianism properly speaking appears in the late 1850s. That is, it emerges in the same period as Helmholtz’s writings. Some of the early writers, for example, Jürgen Bona Meyer, described the new role of Kantianism by referring to a cultural setting that bears similarities to Pyrrhonian skepticism: the materialist worldview and strict idealism are opposed to each other, thereby mutually defeating each other’s claims.19 Transcendental philosophy was meant to reveal that both views lack a proper foundation, whereas its own point of view is considered to be agnostic.
In the mature period, however, neo-Kantianism is less concerned with the rivalry of ideologies, but instead with the relation between philosophy and science. This is particularly true for Cohen’s second edition of Kant’s Theory of Experience20 and Paul Natorp’s writings in the 1880s. According to Cohen’s famous formula, philosophy’s reflection presupposes the “fact of science” (C, e.g. 577). Thereby, Cohen does not want to relate philosophy to the historical fact that there is a body of scientific theories. Nor does Cohen mean that accepting the validity of a specific set of scientific theories is a premise of his reasoning. Those claims would invite questions such as: what theories in particular is Cohen referring to? Can he possibly avoid referring to a certain stage within the historical development of the relevant theory? Rather, Cohen’s slogan recommends that we recognize that the sciences provide their own standards of objectivity. In this respect, the contribution of the sciences has priority. Critique, in contrast, “means the warning: not to treat philosophy and mathematics or natural sciences (…) on the same footing” (C, 578; translated by the author). The task of philosophy is merely “to understand and verify post factum how the objects and laws of mathematical experience are constituted” (C, 578).
/> Three aspects of Cohen’s approach are relevant for our topic. (1) When Cohen characterizes his project as Erkenntniskritik (critique of knowledge),21 he refers to a term—Erkenntnis—that played a crucial role in Kant’s own theory, too. However, the meaning of the term has changed dramatically. As we have already seen, in the terminology of the eighteenth century cognitions are a particular type of mental representations. We discover those representations when we turn our attention inwards. (This turn inwards, of course, is also the starting point for a potential skeptical reasoning.) For Cohen, Erkenntniskritik is philosophy of science in the modern sense. It focuses on the common foundations of the mathematical sciences.
(2) It cannot be the task of philosophy to put the entire body of our (scientific or mathematical) knowledge into brackets or to introduce standards of objectivity from the outside in a next step; thereby I mean standards that originate in philosophy only. (Hegel’s effort to relate his idea of speculative knowledge to the content of science is an example of introducing standards externally.) Of course, according to Cohen, too, philosophy is dealing with the a priori, but a critique of knowledge refers to the a priori only insofar as it is already present in the sciences themselves. Cohen supports this claim by his reading of the Kantian a priori forms: for Cohen, space, time, and the categories are properly understood as methods, in particular as the methods of mathematics and mechanics.
(3) Philosophy is not concerned with cognitive faculties, that is, our powers of representation, either. Cognitive faculties are a topic of psychology. If philosophy were based on a merely psychological analysis it could not avoid covering up the a priori status of some elements of the sciences. (Thus a strategy like Fries’ is incoherent.) In addition, philosophy would continuously struggle with the problem that its results may be grounded on a merely “human, individual consciousness” (C, 77).
In relating psychologistic tendencies in philosophy to a new threat of skepticism, Cohen is a forerunner of Husserl’s much better known criticism in the Prolegomena to his Logical Investigations. Husserl argues in the following way: if philosophy is dealing with the foundations of objectivity, then a psychologistic philosophy would end up grounding the possibility that something is true for me. If this is the only type of validity one can account for, relativism follows, which, in turn, invites skeptical consequences.22 Cohen conceives of a new aspect of critique as the project of drawing the boundaries of our knowledge: critique has to limit the claims of psychology.
This new, neo-Kantian way of setting up Erkenntniskritik is even more striking in the work of Cohen’s student and colleague Paul Natorp. His influential paper “On the Objective and Subjective Grounding of Knowledge”23 provides a clear-cut explanation of the philosophical strategy Natorp recommends. The distinction mentioned in the title of the paper is the following: when we are looking for a basis for the sphere of knowledge (that is, Erkenntnis), we can either focus on the cognitive processes, attitudes, and experiences (Erlebnisse) of the subject insofar as it is a knowing subject; or we can focus exclusively on the content of knowledge. Natorp is quite aware that Kant’s talk of sensibility and the synthetic activities of the understanding has frequently been considered as a source of inspiration for the first point of view. In section 28.3 of this chapter, we have also seen that Fichte’s meta-philosophical approach can certainly be considered to be an effort to spell out this point of view in a methodological and self-reflective way. The first option corresponds not merely to the psychologism Cohen is criticizing. According to Natorp, however, any promising effort to explain why knowledge can fulfill the demands of objectivity and thus establish a relation to the object has to abstract completely from its subjective origin. Starting from the activity of knowing, no path leads to an understanding of why our knowledge can be valid (or invalid). Among the reasons why the subjective strategy fails, the following is crucial: it violates the claim that “what grounds something not only must not, but cannot belong to another genus than what is grounded” (N, 250). The philosopher is left with the opposed “objective” grounding knowledge.
In his own version of this strategy, Natorp tries to turn the claim that we cannot rely on anything that has a different genus than what is to be grounded into a positive theory. Thereby he is not making a general (and rather empty) statement like the following:
Grounding knowledge means to prove that a specific content is objective. This proof, however, cannot help but proceed from another content of knowledge, thus something of a similar type. For otherwise no inferential relations could hold. The process or experience of knowledge, in turn, always already happens behind the back of this (and every) reasoning. Reflecting, and thereby transforming the process itself into the content of knowledge, does not help; a gap between grounding and what is to be grounded would reoccur within the reasoning.
Natorp has a much more specific application of the claim about the similar genus in mind. His account of objectivity of content is bound to the (different) areas of science. Content is objective insofar as it can be related to the general laws that hold within the relevant area, for example, physics. Here, we are dealing with those laws that figure in the scientific explanation of the corresponding phenomenon. In order to do so, the (scientific) laws must have the same genus as that which is explained. If the relation to the law can be established, content is objective; if the explanation is successful, the content in question is accepted as actually determining the object—we have acquired knowledge. According to Natorp, what grounds knowledge (Erkenntnis) is nothing other than the relationship between the particular and the universal that is to be found within knowledge. Natorp’s line of reasoning justifies a neo-Kantian claim I have already mentioned earlier. In Erkenntnistheorie (or Erkenntniskritik, as Cohen has called it) there is not much room for the subject’s reflective turn inwards. In performing this reflection, the subject would stand over and against all content. The subject would be in a position to put the content into brackets, thereby starting to question whether things really are as they appear to be. Following Natorp, endorsing this reflective turn is simply based on a misunderstanding of philosophical strategy, even though it is a frequent one.
Given this background, it is highly interesting to consider how Natorp interprets Descartes in his well-known book Descartes’ Theory of Knowledge (Erkenntnistheorie),24 projecting Descartes’ writings on the neo-Kantian program. Surprisingly, his reading is a charitable one. Natorp accomplishes this by silently stripping the genuine challenge from skepticism. Let us have a look at how this happens in detail. First, Natorp claims that Descartes’ treatise on method, the Regulae, are meant to precede the Meditationes. The methodological principles of the Regulae, however, “fully acknowledge the entitlement of scientific experience” (D, 98). In addition, they are based on the mathematical apparatus and they are free from metaphysics in their core. In other words: the Regulae (with some concessions in detail) fit the “agnostic” Erkenntnistheorie which presupposes the fact of mathematical science. Natorp reads the Meditationes in the same vein. He identifies, for example, Descartes’ cogito with Kant’s apperception (D, 42); only the final step is missing, namely to claim that the unity of apperception is exactly the same as the unity of law (which, in turn, is the basic principle of Natorp’s own theory).
However, Natorp’s way of reading Descartes has to face certain limits. One certainly cannot accommodate Descartes’ proof of the existence of God in the third Meditation in a roughly Kantian program. Therefore, the proof cannot serve as a device to calm our skeptical worries either. How, then, does Natorp deal with this crucial part of Descartes’ reasoning? He starts by projecting the Cartesian God onto Kant’s “transcendental ideal” (B599/A571ff.). The ideal is based on an idea produced by reason. It provides us with the unified ground of a complete determination of all things. The ideal, however, has just a regulative status. This means that it offers orientation, but it is by no means constitutive of the objectivity of knowledge. This leads to the decis
ive question about Natorp’s reading. Suppose Natorp can reasonably believe that God can be replaced by the transcendental ideal. What, then, are the skeptical doubts in the earlier Meditations, according to Natorp, that is, those doubts that the way back to God was meant to dissolve? Bear in mind that Natorp’s reading is meant to be both charitable and consistent.
Given this intention, these doubts could not have been doubts about the objectivity of knowledge, since a God reduced to the transcendental ideal would not allow a response to such doubts anyway. In turn, Natorp first mentions the Cartesian version of the doubts; then, he translates these doubts into the doubt whether we can grasp the objects other than in their dependence on our forms of cognition (D, 44). In other words: the skeptical question now is whether we have access to things-in-themselves. Unlike the Cartesian doubts, this doubt need not be dissolved in the first place; it is fully acceptable. Thus the book on Descartes reveals again that the neo-Kantian strategy does not assign a genuine role to skepticism about the body of our empirical knowledge. Unlike in Descartes, skepticism and secure knowledge are no counterparts. The task of validating knowledge is only a task for the sciences, not for the philosopher.
Overall, the intellectual situation can be summarized as follows: in the second half of the nineteenth century, there are primarily two different ways of responding to the new status of the sciences. The first option is to argue within the sciences and to treat skeptical hypotheses in a similar way to other scientific claims. This is Helmholtz’s approach. Neo-Kantianism, in contrast, distinguishes between philosophy and the sciences. Nevertheless, philosophy starts from the “fact of science.” Its task is merely to reflect post factum on the basic elements of the sciences such as the use of category terms or the infinitesimal calculus. In doing so, neo-Kantianism does not leave much room for a skeptical discussion that is external to the sciences, either. This is particularly true for skepticism of the traditional variety. If there is a danger of skepticism at all, it derives from psychologism. This shows how much the philosophical scene has changed between the late eighteenth and early nineteenth century. For both Kant and Hegel, the task of responding to (Humean, Cartesian, and Pyrrhonian) skepticism did play an important role in how they conceived of their respective philosophical projects. In Kant’s case, the relevant part of his philosophy is what has been called a “cognitive semantics.”25 In Hegel’s case, it is a philosophical theory of knowledge.