On the Brouwerian concept of negative continuity (Q1094422)

This article forms part of a still continuing debate on the meaning and validity of some arguments which \textit{L. E. J. Brouwer} used in his paper ``Über Definitionsbereiche von Funktionen'' [Math. Ann. 97, 60-75 (1926); English translation in: ``From Frege to Gödel'' (ed. J. van Heijenoort), 446-463 (1967; Zbl 0183.006)]. The mere fact that this debate goes on and on seems to confirm Brouwer's view that there is no secure way to put a mathematical argument into words. The main result of Brouwer's paper is the famous Continuity Theorem which says that, intuitionistically, every real function is continuous. Brouwer establishes this fact by first proving that every function from the closed real segment [0,1] to \({\mathbb{R}}\) is uniformly continuous. This follows from, firstly, a continuity principle for functions from the intuitionistic Baire space \({\mathbb{N}}^{{\mathbb{N}}}\) to the set \({\mathbb{N}}\) of natural numbers which Brouwer considers evident, and, secondly, the fan theorem. The fan theorem, in turn, is derived as a special case from the much more general bar theorem. Thus, Brouwer did go far in order to obtain the Continuity Theorem, and one naturally comes to ask if he could have gone a shorter way. It is extra puzzling that Brouwer, in the first paragraph of his 1926 paper, gives a short proof, involving no difficult principles like the bar theorem, of a somewhat weaker, so-called Negative Continuity Theorem. C. Parsons, when he wrote an introduction to Brouwer's paper for ``From Frege to Gödel'', posed the question why Brouwer did not directly derive the Continuity Theorem from his continuity principle. He guessed that Brouwer might have found this impossible on account of the necessity of a proper treatment of the reals as equivalence classes of, say, Cauchy sequences of rationals. In an unpublished report [\textit{W. Veldman}, ``On the continuity of functions in intuitionistic real analysis'', Mathematisch Instituut, Katholieke Universiteit Nijmegen, Report 8210 (1982)], the reviewer showed that, whatever Brouwer may have thought, the Continuity Theorem does follow from the continuity principle alone. This undermines the structure of Brouwer's paper as the Continuity Theorem itself is no more difficult to prove than the Negative Continuity Theorem preceding it. I agree with A. Heyting and A. S. Troelstra that Brouwer himself probably intended the Negative Continuity Theorem as a didactical example preparing the way for the main result of his paper. [Cf. \textit{A. S. Troelstra}, On the origin and development of Brouwer's concept of choice sequence, The L.E.J. Brouwer Centen. Symp., Proc. Conf., Noordwijkerhout/Holl. 1981, Stud. Logic Found. Math. 110, 465-486 (1982; Zbl 0522.03049).] The author thinks that Brouwer's Negative Continuity Theorem should be taken more seriously. As was emphasized by C. J. Posy, and understanding of this theorem is hampered by the ambiguity of negation in Brouwer's writings: at one time ``not A'' means ``A is absurd'' (strong negation), at another time ``not A'' means ``I have (as yet) no reason to assert A'' (weak negation). [Cf. \textit{C. J. Posy}, J. Philos. Logic 5, 91-132 (1976; Zbl 0349.02027).] Observe that weak negation is negation on a meta-level; this explains how the author may get rid of weak negations by introducing the creative subject. In the author's view, the Negative Continuity Theorem is a more fundamental result of intuitionistic mathematics than the Continuity Theorem, as the latter cannot be proved from the basic principles which suffice to establish the former. The author does not consider the continuity principle for functions from Baire space to \({\mathbb{N}}\) as such a basic principle. His basic principles are: a restricted form of the continuity principle, (a principle of open data for lawless sequences) and a result which usually is derived from other axioms, viz. the statement: \(\forall m\neg \forall \rho \exists n[\rho (n)=m]\). In his reconstruction of the Negative Continuity Theorem from these basic principles, and in his critical treatment of the continuity principle, the author exploits the theory of the creative subject and modern analyses of the notion of choice sequence. I doubt whether the author's reconstruction comes near to Brouwer's intentions. Without going into this question, I want to make two remarks. Firstly, from a constructive point of view, negative continuity, however one understands it precisely, is not a very useful property. For this reason, a Negative Continuity Theorem can never be an important theorem of intuitionistic mathematics. Secondly, everyone who reads the second basic principle of the author will feel inclined to seek some argument for it. This shows that this principle is not very basic, after all.