■ The subject of my research : Frege and Hilbert – Two origins of modern philosophies of mathematics.
In this research I was concerned with two problems: (1) Is it correct for us to classify Richard Dedekind into a logicist like Gottlob Frege? (2) W ...
■ The subject of my research : Frege and Hilbert – Two origins of modern philosophies of mathematics.
In this research I was concerned with two problems: (1) Is it correct for us to classify Richard Dedekind into a logicist like Gottlob Frege? (2) What are the differences between Hilbertian axiomatic methods and Frege’s conceptions of axiomatics? I deal with (1) in the first year of my reasearch, and (2) in the second year of my reasearch.
■ The first stage: Frege and Dedekind – two kinds of logicism?
Dedekind asserted that real anaysis and arithmetic should be regarded as parts of logic in his Stetigkeit und Irrationale Zahlen(1972) and his Was sind und was sollen die Zahlen?(1888). But it is a little strange that many commentators are reluctant to classify him into a logicist. This situation is related with the formulation of logicism given by Rudolf Carnap(1930). He regarded logicism as the conjunction of two theses: (1) mathematical concepts can be reduced to logical concepts by definition. (2) Mathematical theorems can be reduced to logical theorems by logical proofs. Since Dedekind did not explicitly formulate logic as the system of logical definitions and proofs, his logicist tenet on arithmetic are sometimes regarded just as a rough assertion belongs to the pre-logicist era. But I think that this kind of interpretation of logicist movement in 19C mathematics based on Carnapian formulation is misguided. Because Carnapian formulation does not teach us anything about what is logic nature of mathematics. I think that although it might be admitted that Dedekind did not provide a explicit formulation of logical theory, he could be deserve to be regarded as a logicist for the following reasons: (1) Firstly, he agree with Frege in the following theses: that arithmetic is nor based upon spatial and temporal intuition nor upon empirical observations unlike geometry and natural sciences, and that our logical faculty consists of pure conceptual thoughts, and that our knowledge of arithmetic and real analysis could be constructed on the basis of pure conceptual thoughts just like logic. (2) Secondly, he think with Frege that conceptual thoughts are not based upon some banal conventions of language, but could provide us new knowledge of abstract domain because of its very creativity. So unlike Kant and many 20C logicists, they do not think that logicality imply triviality. (3) Finally, Dedekind regarded arithematic and real analysis as theories which could be provided on the basis of very general and abstract principles alone, and he emphasized that systems of arithmetical objects could be created by the reflection on the general conceptual principles alone. In this sense, his logicism is more abstract than Fregean logicism based on the theory of logical objects.
■ The second stage: Frege and Hilbert – axiomatic methods.
It is well known that Frege criticised Hilbert’s conceptions of axiomatic theories given in Grunlagen der Geometrie(1899). In Frege’s eye, Hilbert’s thoughts of basic concepts such as axioms, definitions, meanings, references, truths, and mathematical existences seemed ill-founded. Anyone who read their correspondences at first would have the impression that the one's philosophy of mathematics incompatibly conflicted with the other's one. So most commentators has tried to fix what is the crucial difference between them form which other differences are originated. In this research my concerns are somewhat different from this kind of previous investigations. I am concerned with Frege's reconstructions of Hilbertian metamathematics, because I think his reconstructions were performed on the basis of a very thorough and deep analysis of Hilbert's thoughts. In the course of my research, I have come to be finally certain with that such reconstructions should be regarded as a thorough research of the early Hilbertian proof theory and model theory upon Fregean type-theoretical background theory. In this point of view, I tried to answer the following three questions: (1) When we try to fix the meanings of words by Hilbertian method of implicit definitions, what should we regard as definienda and definientia? (2) When we strictly distinguish Hilbertian formal theories from contentual theories as their particular cases, what are the very final results of Frege's assessments of Hilbertian metamathematical proofs? (3) Why did Hilbert believe that his inference from the consistency of axioms of a formal theory to the existence of their model was valid? Why did Frege reject such an inference? My answer to (1)-(3) are the followings: Firstly, the objects of Hilbertian definitions are not primitive terms of contentual theories but logical structures of formal theories, and those structures are defined by their corresponding higher-order predicates. Secondly, Frege finally regarded Hilbertian metamathematical proofs rather as those to show the existence of properties of the higher-order logical relation as the reference of a Hilbertian formal theory than as those to show the existence of mutual relationships among the axioms of a contentual theory. Finally, it seems that, in Hilbert’s opinion, whenever some concepts or relations are logically compatible with one another we could construct some logico-mathematical model satisfy it. Although it is probable that Frege admit the constructiblity of such a logical model, he did not believe that it could be excluded that such a model be ill-functioning because of its emptiness.