by Ádám Tamás Tuboly
Eino Kaila’s recently translated Human Knowledge: A Classic Statement of Logical Empiricism from 1939 is an important document both in the history of analytic philosophy and the history of logical empiricism in particular. Kaila discusses all the relevant topics that featured in the discussions of the Vienna Circle in the early 1930s and provides a neat summary with his own historical narrative and critical remarks.
The book was first published in Finnish as Inhilmillinen tieto: Mitä se on ja mitä se ei ole (Human Knowledge: What it is and what it is not), then in Swedish and has only now been translated (by Anssi Korhonen) and edited (by Juha Manninen, Ilkka Niiniluoto and Georg A. Reisch). As Human Knowledge: A Classic Statement of Logical Empiricism it constitutes the sixth volume in the series Full Circle: Publications of the Archive of Scientific Philosophy. The editorial work of the book was carried out with great care and has resulted in a well-rounded edition with a very fluent translation. It contains a 15-page introductory article by Juha Manninen and Ilkka Niiniluoto which summarizes the content of the book and gives an overview of Kaila’s life and work, thereby helping the reader to contextualize Human Knowledge.
Human Knowledge was conceived as “an academic textbook in scientific philosophy as well as a systematic introduction to ‘logical empiricism’ for professional philosophers” (p. xxv.). As Rudolf Carnap noted in a letter to Kaila in 1940, “[…] the book is very suitable as an introduction to the conceptions of empiricism […]” (p. x.), but it should be mentioned that Kaila’s book fulfills also its other goal, namely to be a viable academic textbook.
Kaila’s treatise divides into three main parts, each of which builds on the previous one, but since Kaila always summarizes his achievements, the parts and their chapters also stand on their own feet if one is interested only in one of the three main topics. The goal of the first part (“Theory Formation”) is to show that “[t]he entire development of European science can be considered a […] compromise between the demands of reason and experience, whereby these two lines of thought, whose origins are to be found in quite opposite directions, increasingly converge and finally coincide […] within the movement known as logical empiricism” (p. 30.). Therefore in the first part of the book Kaila develops a systematic narrative about the history of philosophy and science relying on the notion of invariance.
Invariance is a sort of regularity and lawfulness in nature and it is the “aim of the human pursuit of knowledge” (p. 3.). Kaila shows that, in most cases, knowledge is concerned with the variously depicted invariances in nature – i.e., sameness between phenomena – like physics’ laws of nature, the biological laws of biology, equations in mathematics; but even in our everyday, pre-scientific life, “perception is constantly on the lookout for these invariances […]” (p. 4.).
After introducing the notion of invariances in the first chapter of Part 1, Kaila frames his investigations with some general remarks. He links the notion to Ernst Mach’s methodological principle about the ‘economy of thought’ according to which the aim of formulating scientific theories and laws is to provide continuously more parsimonious – i.e., more economical and simpler – descriptions of the world we are dealing with. As it turns out, invariances are the best candidates for economical and structured descriptions, and with the detection of regularities predictions also become much more simple and reliable. Kaila also demonstrates the links between the notion of invariance and the ideas of rationalization and isomorphism – the former is a pre-requisite for theory formation (a certain abstraction from empirical data) with the latter being the “[…] invariance of logical structure” (p. 13.) At this point, while Kaila seems to rely heavily on the structuralist point of view of Carnap’s Der logische Aufbau der Welt, he actually transcends it in some respects when he claims (after introducing the epistemological ideas of Kant, Schopenhauer and Bergson) that “it is wrong to say that we know nothing of things-in-themselves; after all, we do know their structure” (p. 14.).
The rest of Part 1 is a systematic investigation into history of ideas, philosophy and science. Kaila shows (in chapter 2) ‘how the search for invariances created Greek science’, and he does so by analyzing the underpinnings of Euclidean geometry and its axiomatic system, astronomical systems from the Greeks to Newton, the invariance postulates of British empiricism (which he calls “psychological empiricism” (p. 27.)), and the mechanistic conception of nature (from the ancient atomists to Newton).
In chapter 3 and 4 Kaila introduces various conceptions of knowledge. In the case of Plato, we find for the first time “[…] philosophy in the strict, scientific sense of the word” (p. 37.) since in all of his dialogues he was after invariances and took the subject of philosophy knowledge itself. The insights and aims of Plato, according to Kaila, could also be found in Galileo, Kepler, Descartes and Leibniz, all of whom were struggling with the Aristotelian conception of nature and knowledge. Plato and all of his intellectual successors were relying on mathematics and the importance of numbers for knowledge, both of which are connected to structures and invariances. By contrast, Aristotle denied the possibility of the mathematization of nature since he thought the mathematics is just another branch of science which deals with a definite feature of beings; hence the method of mathematics could not coincide with the method of nature.
This Aristotelian attitude set back the search for invariances until the new astronomical Platonist has appeared in the sixteenth century. As these scholars used the insights of Plato they often made the same assumptions: they thought that being is identical with ‘being unchanging’ and their efforts usually resulted in the duplication of the world: invariances are to be found in a different, more real, but hidden, transcendent realm. “These reminds us that we are still far away from Leibniz and logical empiricism and their insight that reason reveals not the Real Being ‘behind’ illusions, but rather invariances that exist in the relations between them” (p. 59.).
Instead of re-telling Kaila’s narrative in every detail, let us just mention his conclusion: logical empiricism seems to be the culmination of two different traditions, whether these be dubbed empiricism/rationalism or Platonism/Aristotelianism, all of which have put too much weight either on sensation or reason. Kaila also investigates Kant at various points, but as it turns out, since logical empiricism rejected the Synthetic A Priori, Kant was not much of a particular influence. (Actually, as the recent rehabilitation of Carnap and logical empiricism has shown, Kant was a much more important influence than has traditionally been thought.) Logical empiricism, after all, is just the tradition that showed how we can combine the demands of reason and sensation; that is, mathematical invariances and empirical invariances.
Part 2 of Human Knowledge subsequently is concerned with the formal truth of theories, while Part 3 is about the empirical truth of theories. Formal truth divides into two subclass, namely logical truth (chapter 6) and mathematical truth (chapter 7). Kaila is following the well-known logical empiricist conception that the statements of logic and mathematics are analytic and a priori. To support this idea, he starts with the definition of analyticity, which states that “a sentence is analytical if it follows from mere definitions” (p. 119.). At this point we find what is, in many respects, one of the most outstanding sections of the book.
First of all, Kaila disentangles the definition of analyticity with great care and in a comprehensive and systematic way: he devotes one section to each of the core concepts, namely “follows” (= entailment), “sentence” and “definition”. Secondly, when Kaila is about to define the notion of “follows”, he introduces in a textbook-like manner both propositional and predicate calculi, their semantical features and all the usual problems associated with them: decidability, completeness, soundness, etc. These sections of the book (pp. 92-116.) are of great value for anyone with an interest in modern formal logic who is looking for a short introduction and overview.
Last but not least, Kaila presents a picture of LE, and mostly of Carnap and his Logische Syntax der Sprache (1934), that is very close to the rehabilitated picture of logical empiricism which has recently been put forward by historians of philosophy. Analytic sentences are the ones which follow from definitions: but with respect to definitions we are free to create them any way we like. (There are, of course, certain constraints from the side of philosophy of science and methodology: we need the best tools for the search of invariances and prediction.) This means that (i) from different definitions we can derive different sentences, and (ii) from the same definitions we can derive different sentences if we define the notion of “follows” in a different manner. So what counts as analytic depends on our definitions, logic and language. As Kaila puts it:
Insofar as logical truths are analytical sentences, they become relative truths, in a sense. There is not just one logic; there are many logics – even in principle, indefinitely many logics […]. (p. 117.) Which one is correct, classical or intuitionistic propositional calculus? – This is an improper question. […] We cannot ask which interpretation is correct, because it is a matter of arbitrary decision which meaning we give to a sign. (pp. 119. Italics added.)
In these passages Kaila revives the words of Carnap from his Syntax book, where he claims: “In logic there are no morals. Everyone is at liberty to build up his own logic, i.e. his own form of language, as he wishes.” (§17)
Kaila appears to admit another undervalued aspect of Carnap’s theory of analyticity – i.e. in different languages different sentences counts as analytic. In his “Two Dogmas of Empiricism” (1963), Willard von Orman Quine characterized the logical empiricist view as treating analytic sentences as those “which we held true come what may”. Quine has shown, however, that analytic sentences, if needed, are also revisable. Interestingly, as has recently been noted, Carnap also pointed out in Syntax that analytic sentences are not immune to revision (§82), and the same follows from Kaila’s presentation: if need be, we can revise our definitions (= logic and language), and accordingly the range of analytic sentences will also vary.
In the final part of the book (Part 3) Kaila considers the empirical truth of theories. He deals with the notions of verification and falsification and points out how these ideas changed during the 1930s, mainly in the works of Carnap. Kaila’s sensitivity to these changes and the detailed presentations of the historical evolution of logical empiricists are important in themselves, but we will note here only one aspect of them.
Quine was notorious about his holism, namely that “[…] our statements about the external world face the tribunal of sense experience not individually but only as a corporate body” (“Two dogmas”, p. 41.). Accordingly, empirical meaningfulness comes into the picture not on the level of isolated sentences but of the whole system, the whole theory. Therefore experience could not refute any particular sentence on its own – “Any statement can be held true come what may, if we make drastic enough adjustments elsewhere in the system” (“Two dogmas”, p. 43.).
Anyone who is familiar with the history of twentieth century analytic philosophy has heard of the story of how Quine refuted logical empiricism and especially Carnap. In Human Knowledge, Kaila appears to accept the points of Quine’s criticism (p. 154. Italics added):
‘Experimentum crucis’—crucial experiment—was the term that people often used to refer to such singular instances that they took to be decisive. Strictly speaking, however, there are no such experiments. Suppose, for instance, that we have derived from some physical theory an empirical consequence and then come to see that experience does not correspond to the theory; in such a case we have always made a number of assumptions falling outside the sphere of the theory itself and concerning, say, the construction of the instruments we use in the experiment or the reliability of the observer himself. A negative test result can always be put down to such factors. We should note in particular that insofar as we can speak of experience refuting a theory, what gets refuted is the entire theory, and not just some particular statement belonging to it.
What we should bear in mind is that this passage is from 1939, while Quine formulated his ideas in the early 1950s. But we should not therefore think that Kaila was the one who first refuted Carnap and logical empiricism, since Carnap had formulated just the same ideas in his 1934 Syntax with which Kaila was familiar and which he used as a main reference point. Carnap claims (§82) that “[…] the test applies, at bottom, not to a single hypothesis but to the whole system of physics as a system of hypotheses (Duhem, Poincaré).” (For completeness, one should perhaps also mention at this point the naturalism of Neurath.)
In the remaining portions of Part 3, Kaila discusses the so-called protocol-sentence debate (though he does not say so explicitly) which is about whether the phenomenal or physical language should be taken as fundamental in epistemology and science. Contrary to the later conception of Carnap (from “Testability and Meaning”), Kaila contents himself with the idea of translation between these two types of languages and their relative merits. Chapter 8 contains some well-known remarks about the meaningfulness of metaphysics, while Chapter 9 takes on the particular problems of physical theories, such as the philosophical questions of the theory of light and quantum theory, determinism vs. indeterminism and how we can handle all of these from the point of view of logical empiricism.
Chapter 9, finally, touches upon the problem of “other minds”. Kaila shows that the logical behaviorism of logical empiricists is not the same as the usual “quite primitive, mechanistic and materialistic behaviorism” (p. 205.) which would like to claim that there is no consciousness or mental sphere:
That the sentences of psychology are interpreted as f-sentences [physical-sentences] in this way means that all intersubjective knowledge of reality speaks one and the same language everywhere; this language is f-language [physical-language], the language whose expressions refer to things, their properties, and relations in ‘physical reality’. (p. 205)
That is, if one is to talk about the mental states of others in a scientific manner and would like to make cognitively and factually meaningful statements, then one has to formulate one’s claims in our intersubjective world and discourse where anyone can test the relevant scientific statements so they can acquire the requisite public weight: a physicalist/behaviourist theory of other minds is capable of just that. It could not do more than this, but that is enough for Kaila and the logical empiricists.
In the final passages of the book, Kaila imagines some possible objections against this view and formulates his responses. The third objection in particular is worth mentioning (pp. 207-208):
Q: I do understand that the principle of logical behaviorism is a necessary consequence of the requirement of intersubjectivity, that is, the postulate that my sentences describing my own experiences must have some factual content for other people. […] But I do not see that this postulate itself of intersubjectivity has been given any justification. It seems to me that I am free to deny it. In that case, what would you do? What would you do, if I now tell you that I want to talk to myself, in a monologue and not to you, in a dialogue?
A: What you renounce in your words, that you accept in your deeds. I may learn about your theoretical rejection of the principle of intersubjectivity only if you execute it in practice. You can renounce this principle in one way only, by remaining silent. At most you can say, with Prince Hamlet: ‘The rest is silence’.
This passage shows that if one were to refute intersubjectivity as a prerequisite of science and philosophy one has two options: either one tries to argue against it – in which case, one is already engaged in intersubjective-scientific discourse – or (ii) one remains silent. So according to Kaila, logical empiricism is simply that tradition which embraced the idea of intersubjectivity (and also the correlative requirement of testability) in its most strict sense and explored all its consequences.
In summary, Human Knowledge, and the fact of its recent translation into English, is of great importance. For one, it is an interesting document of the logical empiricist movement from 1930s Finland and demonstrates Kaila’s remarkably systematic and comprehensive style. (As Carnap already put it in a letter: “I especially welcome the fact the you [Kaila] pay attention to the historical connections. For many readers this will be very welcome, since our existing publications the historical connections are almost ignored.”, p. x.) Yet, the recent translation is also important in a different sense.
In the 1930s several books about logical empiricism appeared in the English-speaking world. Two of them – Susan Stebbing’s Logical Positivism and Analysis (along with her reviews of Carnap) and Julius R. Weinberg’s An Examination of Logical Positivism – were quite close to the current and rehabilitated picture of loigcal empiricism and Carnap, but they were simply forgotten. The third one, Alfred J. Ayer’s Language, Truth and Logic made its way into the history of analytic philosophy and shaped prevalent views about logical empiricism and Carnap for several generations – it contributed much to the unified, simplistic, but erroneous received view which a lot of scholars are now trying to revise. However, if Kaila’s book has been translated in the 1940s, as Carnap was suggesting at the time (p. x), perhaps the history of twentieth century analytic philosophy, and of logical empiricism and Carnap in particular, would have been quite different.
Kaila’s book, now available in English, is a remarkable historical document, which is finally receiving wider dissemination, and which will no doubt have a detectable influence on the philosophy of science, language and epistemology, and its historiography.
Eino Kaila: Human Knowledge. A Classic Statement of Logical Empiricism
(Transl. by Anssi Korhonen, ed. by Juha Manninen, Ilkka Niiniluoto & George A. Reisch)
Price: US$ 49.95
Chicago: Open Court, xxvii+217 pages, paperback.
Ádám Tamás Tuboly is a PhD student in the Department of Philosophy, University of Pécs, Hungary; he mainly works on problems in the history of analytic philosophy, especially in the historical reception of Rudolf Carnap.
Acknowledgments: The author would like to thank the help of David Steele (from Open Court Publishing), and Axel Gelfert for his help with an earlier draft of this review.
(c) 2015 The Berlin Review of Books.