EMBEDDING LOGICAL EMPIRICISM INTO THE HISTORY OF EPISTEMOLOGY: EINO KAILA ON HUMAN KNOWLEDGE

ESSAY REVIEW EMBEDDING LOGICAL EMPIRICISM INTO THE HISTORY OF EPISTEMOLOGY: EINO KAILA ON HUMAN KNOWLEDGE Eino Kaila. Human Knowledge: A Classic Statement of Logical Empiricism. Trans. Anssi Korhonen. Ed. Juha Manninen, Ilkka Niiniluoto, and George A. Reisch. Chicago: Open Court, 2014. Pp. 288. $49.95 (paper). The name of philosopher and psychologist Eino Kaila (1890 1958) is well known among people who are interested in logical empiricism and its history. His writings, however, have received only limited attention. This can only partially be due to linguistic barriers; after all, most of his writings on the philosophy of science have appeared in German, and some of them are available in English (see Eino Kaila, Reality and Experience: Four Philosophical Essays, ed. R. S. Cohen with intro. by G. H. von Wright, Vienna Circle Collection 12 [Dordrecht: Reidel, 1979]). The current book, which is now available in English for the first time, was first published in Finnish in 1939 and translated into Swedish in the same year by G. H. von Wright, a student of Kaila. Even in 1979, von Wright espoused the view that Kaila s book was the best introduction to logical empiricism that had been written. Carnap wanted to arrange for an English translation of the book to be published in the Library of Unified Science but was hindered by the war. If he had been successful and an English translation of the book had been published, the so-called received view of logical empiricism may never have become as widespread as it did. The literal translation of the Finnish title would be Human knowledge: What it is and what it is not. The editors of the current volume decided in favor of HOPOS: The Journal of the International Society for the History of Philosophy of Science, vol. 6 (Spring 2016). 2152-5188/2016/0601-0006$10.00. 2016 by the International Society for the History of Philosophy of Science. All rights reserved. 148

Essay Review SPRING 2016 Human Knowledge: A Classic Statement of Logical Empiricism. Presumably, the subtitle is intended to indicate that the philosophical theses presented in the book (75 years after the original publication) are to be regarded as constituting a step in the history of philosophy of science in the twentieth century namely, in the classic period of logical empiricism. In this sense, the subtitle is well suited. But, it could also create the misleading impression that Kaila s systematic introduction to logical empiricism (xxvi) does nothing more than systematically rehash canonical theorems of logical empiricism in the form of a textbook. Such a book could of course be historically interesting and worthy of translation: a systematic presentation of logical empiricism that was already available in Finnish in 1939 would greatly enrich the stock of resources available to HOPOS scholars doing research on the philosophy of science of the twentieth century. However, it would be thoroughly mistaken to expect a textbook-style presentation of familiar themes from logical empiricism. Indeed, the book holds a number of surprises in store for the reader. In characterizing his philosophical position, Eino Kaila used the term ein logischer Empirismus (in German) as early as 1926. On Moritz Schlick s invitation, he visited Vienna in 1929, 1932, and 1934 and entered into a lively exchange with the members of the Vienna Circle. In his writings from this period, he engaged intensively with Frege, Russell, Carnap, Schlick, Popper, Gödel, Reichenbach, and others. It is clear from the titles of several of his writings in these years that the development of his philosophical views was shaped by reflection on the movement in Vienna and Berlin (e.g., Logical Positivism: A Critical Study, Annales Universitatis Aboensis B 13 [1930], On the System of Concepts of Reality: A Contribution to Logical Empiricism, Acta Philosophica Fennica 2 [1936], On the Physical Concept of Reality: Second Contribution to Logical Empiricism, Acta Philosophica Fennica 4 [1941]). The first surprise that one encounters in this book is that it exudes the same freshness as the early logical empiricists exuded in their discussions of fundamental questions about modern science. It has the charm of a time capsule that has been washed from that world that is passed onto the shore of the twentyfirst century. In this sense, the book really does deserve to be counted among the classical texts of early logical empiricism. The second and probably greater surprise that one encounters in the book is an idea that differs markedly from the views one finds in other classical writings from the 1920s and 1930s. Specifically, for Kaila, the term logical empiricism is not primarily used to designate a movement in twentieth-century philosophy but, rather, a characteristic of modern science since Galileo. According to Kaila, the most central achievements of logical empiricism, understood in this sense, can only be properly appreciated against the background of 149

HOPOS The Journal of the International Society for the History of Philosophy of Science the epistemology that Galileo overcame. This way of thinking leads toward a hybrid systematic-historical orientation, which we do not find elsewhere in early logical empiricism except perhaps to some extent in Edgar Zilsel and Philipp Frank. But, there are obvious similarities to Ernst Cassirer s The Problem of Knowledge (1906 20), as well as to his Substance and Function (1910) not least in the role played by Aristotle s account of substances in Kaila s reconstruction. According to Kaila, the Aristotelian understanding of substances in logic and metaphysics was precisely the impediment that had to be overcome, in successive historical stages, in order for the modern conception of knowledge to become established. Kaila s book contains 10 chapters, which are organized into three parts: part 1 (Theory Formation), part 2 (The Formal Truth of Theory), and part 3 (The Empirical Truth of Theory). In the first chapter, Kaila defines and explicates a conception of human knowledge that then serves as a red thread through the entire investigation. Here, too, a deep relationship between Kaila and Cassirer s philosophy of science becomes visible. According to Kaila, the search for invariances is a core characteristic of the human mind that is also found in aesthetic experience (5) and in prescientific thinking: the so-called laws of nature and even the so-called physical objects emerge from that search for invariances (3). In his book, Kaila attempts to show that the human search for knowledge can indeed be regarded as a search for invariances; and that knowledge itself, being the result of this search, can be understood as the discovery and representation of such invariances. If, furthermore, we succeed in showing that other conceptions of human knowledge in fact presuppose this self-same search for invariances, but err because of some prejudices and misconceptions, we will have accomplished something (17). A further surprise is then in store in the second and third chapters, where it becomes clear just how highly Kaila esteems the ancient Greeks: The dialogues of Plato, the best of which remain fresh as morning dew, reveal to us the entire depth of the struggling human spirit. Everything to which the ancient Greeks put their minds grew to become exemplary. Thus they created European art and European science (19). It is not only Euclidean geometry, the conception of an axiomatic system, and the notion of a theory as a hypotheticodeductive system that we owe to the Greeks but also the decisive step taken by transforming astronomy into a science. This step becomes apparent in the following famous question, attributed to Plato: What are those completely regular circular motions that, if assumed, would save planetary phenomena? The remarkable phrase to save the phenomena ; in Greek sozein tai phainomena occurs repeatedly thereafter. In its Latinized form apparentia salvare it is even used by Copernicus in his great work On the Revolution of Celestial Spheres. 150

Essay Review SPRING 2016 The phrase means: Which invariances would define the motions of stars in such a way as to render their wanderings only apparent? This was the beginning of scientific astronomy (25). Similarly, with respect to scientific philosophy: And it is really only with Plato that philosophy in the strict, the scientific sense of the word is born: that is, with Plato the eye of the intellect is now turned inward so that the concept of knowledge itself becomes an object of scientific consideration. In this way, the theory of theories, the theory of knowledge, is born (37). But, Kaila also observes that the Greeks set up significant barriers to science and to scientific philosophy for example, through their prejudice that true knowledge can only be knowledge of the unchangeable (38) and that the unchangeable must be conceived as being thing-like (51). In Kaila s view, these prejudices are philosophically interesting insofar as they are linked to a particular concept of human knowledge that pervades all of ancient Greek thought and was definitely established by Aristotle and... dominated all European scientific thought until the fifteenth and sixteenth centuries (27). In the atomists and in Plato, there was still a powerful impetus to the development of mathematical natural science (40) namely, the ideas of mathematical invariances and structural laws. But these impulses could not be developed further, because Plato considered transient objects to be mere reflections of true reality : Plato regards all mathematical natural science as no more than a probable story, which is not even supposed to be valid as full, exact account. The requirement of verification by experience is weak. It is at this point the Plato s heir, Aristotle enters the stage, taking the main role in the development of scientific ideas (41). The great achievements for which Aristotle deserves such high praise, such as turning toward empirical study and taking the first steps in formalizing logic, came at a high price (46). Rather than freeing the search for invariances from its peculiarly Greek limitations, Aristotle rejected, in the name of experience, the requirement imposed on it by Plato. This fact underlies one of the remarkable features of the Aristotelian conception of knowledge: it rejects the very idea of rationalizing concept and theory formation. This means, above all, that attempts to create a mathematical natural science are given up (42). These attempts were taken up again in the sixteenth and seventeenth centuries, and it was in Galileo that a conception of knowledge finally arose that combined both the search for invariances and the requirement of verification. In the fourth chapter, The Galilean Conception of Knowledge, Kaila characterizes the decisive step toward the modern concept of knowledge as follows: the most important kind of unchangeable thing is no longer to be found in things themselves, but rather in the relations between changes; it is to be found, that is, in relational invariances or functions, not in substances (51). 151

HOPOS The Journal of the International Society for the History of Philosophy of Science In this formulation, and elsewhere, Kaila s proximity to Cassirer becomes apparent. Like Cassirer, he also emphasizes that this new concept of knowledge obviated the characteristically Aristotelian need for a foundation in everyday commonsense thinking: Now, though, human thought took a decisive step as consciously developed relational concepts, often quite detached from everyday experience, came to occupy an ever greater role in the comprehension of reality (52). Kaila s understanding of the Galilean concept of knowledge is not limited to Galileo himself. Rather, Kaila presents this concept as a dense network of relations and thoughts extending from Leonardo da Vinci to Galileo, Newton, Descartes, and Leibniz. At the same time, this network is focused around two points: the Platonic and Aristotelian concept of knowledge, on the one hand, and the concept of knowledge found in logical empiricism, on the other. In Kaila s view, the Galilean concept of knowledge not only incorporated and reconfigured crucial aspects of Greek science but also took up new aspects that link it with the concept of knowledge in logical empiricism. Let s mention some of these features: Galileo s remark that the way of demonstration is distinct from the way of discovery (53); Galileo s way of demonstration, which consists in finding out what the consequences given assumptions have (55); the Leibnizian distinction between logical truths as analytic and empirical truths as synthetic (57); and the way Leibniz s philosophy overcame the imaginary opposition between thinking and extension, or between matter and spirit (66). In the fifth chapter, Kaila situates the problem of induction in the conceptual and historical context established in the previous chapters. Taking up the Galilean distinction between the way of demonstration and the way of discovery, he asks: Can the way of discovery, that is the search for invariances, be subjected to a number of exact rules? (69). In discussing Mill, Reichenbach, Leibniz, Hume, and Bayes, Kaila addresses the induction problem in connection with the labels enumerative induction, causal induction, inductive probability, relative simplicity, relative simplicity and the explanatory value of a theory, and justification of induction (71 84). His discussion ranges over countless examples from the history of science, including Kepler, Newton, Mendeleyev, Fresnel, and Poisson. According to Kaila, scientific imagination plays a decisive role in major scientific discoveries: it seems that this kind of scientific discovery [i.e., Kepler s laws of planetary motion] can be subjected to mechanical rules no more than the genius of an artist (71). A theory is created through free invention regulated by the search for invariances, and this way of discovery cannot marked out in advance. Only the way of demonstration can be regulated (76). 152

Essay Review SPRING 2016 After this chapter on induction, which concludes part 1, Kaila moves on to part 2, where he presents the fundamental concepts of formal logic in chapters 6, Logical Truth, and 7, Mathematical Truth. It is against this backdrop that he discusses Kant s doctrine of synthetic a priori statements. Drawing on Reichenbach and Carnap, he impugns Kant s view that there are synthetic a priori truths, and in so doing he establishes a bridge to the third part of the volume. The third part deals with the concept of truth in logical empiricism. According to Kaila, logical empiricism can be characterized by four main theses. The first two of these are a matter of consensus among philosophers who feel committed to this philosophy of science. The first was that all a priori sentences are analytic; in negative terms and this formulation is directed specifically against Kant there are no sentences that are both synthetic and a priori. The second main thesis was that every factual sentence must have some determinate consequences with respect to experience; or, what comes to the same, every statement concerning reality must have a determinate factual content (144). The second thesis is, in Kaila s terms, the principle of testability. Kaila observes that this principle can already be found both in Aristotle and in Leibniz and that the empiricists endorsed different versions of it. The logical empiricists achievement was to bring it into a clear form. In contrast to the first two theses, the third main thesis was a matter of controversy among the logical empiricists. This thesis says that every theory concerning reality must be translatable into the language of experience, that is, must be capable of being transformed into an equivalent form in which there occur, besides logical constraints, only empirical predicates (145). According to Kaila, the thesis is particularly controversial with respect to the question as to whether the translation in the language of experience can proceed wordby-word or sentence-by-sentence or perhaps even in a freer manner (146). Wittgenstein and Schlick like Hume would have endorsed the radical view that it must be possible to verify or falsify every single factual claim. This was the radical view of so-called logistic neopositivism, the first phase in the development of logical empiricism (148). In this connection, Kaila makes a point of referring explicitly to his own criticism of this view, which he published in 1930. In contrast to the early phase as Kaila demonstrates in his discussion of Carnap, Popper, Reichenbach, and Hempel logical empiricism arrived in later years at a holistic view of the testability principle. Pierre Duhem, Kaila observes, was one of the earliest advocates of Logical Empiricism (154). One essential reason why the testability principle must be interpreted in a holistic manner, according to Kaila, is that every well-developed scientific 153

HOPOS The Journal of the International Society for the History of Philosophy of Science theory works with rationalizations. As mentioned above, Kaila criticizes the Aristotelian concept of knowledge for precisely this reason namely, for leaving out a crucial component of scientific thinking by rejecting the idea of a rationalizing concept. In Kaila s view, general theoretical claims cannot be derived through experience but, rather, through the operation that he dubs rationalization. Through rationalization a regularity that experience has shown to be valid only within certain limits and sometimes with glaring exceptions becomes valid without restrictions and exceptions. Thus, rationalization adds to the regularity of experience and not only generalizes it, as induction does. We must therefore address the remarkable problems that are involved in the testability of rationalized systems (166). Drawing on Reichenbach, von Wright, Richard von Mises, and Hempel, Kaila demonstrates that the mathematical concept of probability is a typical result of rationalization (165). Since rationalization plays a decisive role, both in scientific thought and in prescientific thought (168), the principle of testability must be interpreted in a way that makes it applicable to rationalized theories. It is worth noting in this context that Edgar Zilsel s Das Anwendungsproblem (1916) may also have played a role, given that it, too, gave pride of place to the concept of rationalization. Kaila uses the expression das Anwendungsproblem (the problem of application) in numerous places in the context of his critique of Kant s account of synthetic a priori statements. In the passage in which Kaila introduces this term, he refers to section 13 of Reichenbach s Philosophie der Raum-Zeit-Lehre, where Reichenbach discusses Kant s conception of pure intuition but where Reichenbach, as far as I can see, does not actually use the term Anwendungsproblem. Although Kaila does not expressly mention Zilsel, he did know Zilsel s book well. The editors note that Kaila s book Der Satz vom Ausgleich des Zufalls und das Kausalprinzip (1925) was heavily influenced by Zilsel s book (xiv). Moreover, Kaila s terminology in other passages indicates the extent to which the philosophy of science of the twentieth century pervades his thinking: the fact that he uses the term idealization in the same sense as rationalization, for example, once again attests to the proximity of his thinking to that of Ernst Cassirer. Testability does not apply separately to each sentence of a rationalized theory. It is only as a whole that an idealized theory represents some region of experience; its sentences do not have the relatively transparent relation to experience that we find in inductive theories.... We must, then, give the principle of testability a broad interpretation, so that a theory in its entirety can be regarded as one sentence (170). In chapter 9, The Logic of Physical Theories, Kaila emphasizes the tremendous significance of rationalization for physical concepts and physical 154

Essay Review SPRING 2016 theories. In so doing, he focuses on the interpretation of microphysical theories (184) and physical determinism and indeterminism (189). The fourth thesis with which Kaila characterizes logical empiricism is the topic of chapter 10, The Principle of Logical Behaviorism (197). In this context, Kaila is able to draw on his comprehensive knowledge of psychology: before turning to philosophy, he had completed a doctorate in experimental psychology (in 1916), founded a lab for experimental psychology (in 1922), and (in the 1920s and 1930s) written a series of important monographs in psychology, such as Sielunelämäm rakenne (The structure of mental life, 1923) (xiii). During his stay in Vienna in spring 1934, Kaila worked on experimental psychology with Karl and Charlotte Bühler. That same year he published in Finnish his masterpiece in the psychology of personality, Persoonallisuus, which was soon translated into Swedish and English (xviii). Early on in the current volume, while explaining the significance of the concept of experiential sentences (148), Kaila had already introduced the distinction between phenomenal language or φ-language and physical language or f-language. At this point, he returns to that distinction and draws on Carnap s work from the early 1930s in order to define what he calls the principle of logical behaviorism : φ-sentences describing the so-called immediate experience of S1 are intersubjectively equipollent with f-sentences describing certain states of S1 s body.that a given φ-sentence and a given f-sentence are intersubjectively equipollent means that the former sentence, insofar as it is meant to be understood by other people, agrees in its factual content with the latter (204). Kaila emphasizes that this principle has little to do with Watson s quite primitive, mechanistic and materialistic behaviorism and also that logical behaviorism by no means wishes to claim that there is no consciousness (205). Of course, it becomes clear in this section that Kaila still sees many unsolved problems. He argues that the φ- language should be seen as fundamental with respect to f-language (198) and explains how he thinks that this provides a basis on which to investigate the problem of other minds (198), as well as the question of the intersubjective content of sentences (202). This brief chapter then closes with a discussion of three objections to the thesis of logical behaviorism (205). These open questions that come up at the end of the final chapter make it clear that Kaila intended his book as a contribution to a lively and ongoing philosophical movement, in which there was still much philosophical work to be done. They remind the reader that the book appeared in 1939, right around the middle of Kaila s career. The outstanding introduction to the book, written jointly by Juha Manninen and Ilkka Niiniluoto, is extremely helpful, providing a comprehensive look into Kaila s rich work straddling experimental psychology and scientific philosophy. And it underscores the magnitude of Kaila s 155

HOPOS The Journal of the International Society for the History of Philosophy of Science influence and in particular the influence of this book on philosophy in the Nordic countries. Kaila conceived the book both as a textbook of scientific philosophy for laypersons and university students and as a systematic introduction to logical empiricism for professional philosophers. In the two countries where it could most easily be read, it performed those functions for several decades and paved the way for analytic philosophy to become a dominant trend in the Nordic countries (ix). The book, which is now available in English some 75 years after its original publication, has retained much of its freshness. Perhaps it could also be used today as an excellent introduction to logical empiricism but with the proviso that this classification may be misleading, insofar as Kaila s book offers much more than what one would expect from an introduction to logical empiricism written in the second half of the twentieth century. In 1940, Carnap immediately recognized that the special quality of the book lay in the historical connections that Kaila had identified (x). Today, after several decades of HOPOS research, the historical side of Kaila s reconstruction of the process leading up to the concept of knowledge in logical empiricism is interesting in two respects. First, as a historical phenomenon, it is an example of a type of early logical empiricist thinking that aims to situate logical empiricism in the history of epistemology and of science. (As noted above, if Kaila s book had been published in English in the 1940s, as Carnap intended, our image of logical empiricism in the second half of the twentieth century would have been very different.) Second, the book is interesting as a provocation. Kaila s historical presentation has, in some respects, the character of a grand narrative. For years, historians of science and HOPOS scholars have spoken out against the tendency to indulge in grand narratives. And in contrast to the Whiggish tendencies to be found in grand narratives they have established a much more narrowly focused style of historical research, which avoids broad and sweeping historical narrative. There are excellent reasons for this. And there are also good reasons to avoid the sort of anachronistic interpretation that casts Galileo as an early logical empiricist. Be that as it may, I could not help but notice over and again while reading this book just how fruitful it is to see how Kaila links the debates that were ongoing in philosophy of science in the first half of the twentieth century to philosophical questions that were articulated by the ancient Greeks and the early Moderns. The conceptual riches that he thereby gains make it possible for Kaila to discuss the core themes of twentieth-century philosophy of science with a greater freedom than would have been possible if he had restricted himself to the philosophy of science of the nineteenth and twentieth centuries. Kaila s book is fundamentally shaped by this freedom and that is 156

Essay Review SPRING 2016 probably the main reason why the book appears surprisingly fresh and new even today. Thanks to John Michael for translating my text into English. Elisabeth Nemeth, University of Vienna 157