Intuitionistic Proof Versus Classical

Truth The Role of Brouwer s Creative

Subject in Intuitionistic Mathematics
1st Edition Enrico Martino (Auth.)
Visit to download the full and correct content document:
https://textbookfull.com/product/intuitionistic-proof-versus-classical-truth-the-role-of-br
ouwer-s-creative-subject-in-intuitionistic-mathematics-1st-edition-enrico-martino-auth/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Recent Advances in Intuitionistic Fuzzy Logic Systems

and Mathematics Said Melliani

https://textbookfull.com/product/recent-advances-in-
intuitionistic-fuzzy-logic-systems-and-mathematics-said-melliani/

Generalized Intuitionistic Multiplicative Fuzzy

Calculus Theory and Applications Shan Yu

https://textbookfull.com/product/generalized-intuitionistic-
multiplicative-fuzzy-calculus-theory-and-applications-shan-yu/

Visualizing Mathematics The Role of Spatial Reasoning

in Mathematical Thought Kelly S. Mix

https://textbookfull.com/product/visualizing-mathematics-the-
role-of-spatial-reasoning-in-mathematical-thought-kelly-s-mix/

Rethinking Luxury Fashion: The Role of Cultural

Intelligence in Creative Strategy Thomaï Serdari

https://textbookfull.com/product/rethinking-luxury-fashion-the-
role-of-cultural-intelligence-in-creative-strategy-thomai-
serdari/
Proof Technology in Mathematics Research and Teaching
Gila Hanna

https://textbookfull.com/product/proof-technology-in-mathematics-
research-and-teaching-gila-hanna/

The Classical Music Industry Routledge Research in the

Creative and Cultural Industries 1st Edition Dromey

https://textbookfull.com/product/the-classical-music-industry-
routledge-research-in-the-creative-and-cultural-industries-1st-
edition-dromey/

Proof Complexity Encyclopedia of Mathematics and its

Applications 1st Edition Jan Krají■ek

https://textbookfull.com/product/proof-complexity-encyclopedia-
of-mathematics-and-its-applications-1st-edition-jan-krajicek/

A Transition to Proof An Introduction to Advanced

Mathematics Textbooks in Mathematics 1st Edition Neil
R. Nicholson

https://textbookfull.com/product/a-transition-to-proof-an-
introduction-to-advanced-mathematics-textbooks-in-
mathematics-1st-edition-neil-r-nicholson/

Yearning for the impossible : the surprising truth of

mathematics Second Edition Stillwell

https://textbookfull.com/product/yearning-for-the-impossible-the-
surprising-truth-of-mathematics-second-edition-stillwell/
Logic, Epistemology, and the Unity of Science 42

Enrico Martino

Intuitionistic
Proof Versus
Classical Truth
The Role of Brouwer’s Creative Subject
in Intuitionistic Mathematics
Logic, Epistemology, and the Unity of Science

Volume 42

Series editors
Shahid Rahman, University of Lille III, France
John Symons, University of Texas at El Paso, USA

Editorial Board
Jean Paul van Bendegem, Free University of Brussels, Belgium
Johan van Benthem, University of Amsterdam, The Netherlands
Jacques Dubucs, CNRS/Paris I, France
Anne Fagot-Largeault, Collège de France, France
Bas van Fraassen, Princeton University, USA
Dov Gabbay, King’s College London, UK
Karel Lambert, University of California, Irvine, USA
Graham Priest, CUNY, USA
Gabriel Sandu, University of Helsinki, Finland
Göran Sundholm, Universiteit Leiden, The Netherlands
Heinrich Wansing, Ruhr-University Bochum, Germany
Timothy Williamson, Oxford University, UK
Logic, Epistemology, and the Unity of Science aims to reconsider the question
of the unity of science in light of recent developments in logic. At present, no single
logical, semantical or methodological framework dominates the philosophy of
science. However, the editors of this series believe that formal techniques like, for
example, independence friendly logic, dialogical logics, multimodal logics, game
theoretic semantics and linear logics, have the potential to cast new light on basic
issues in the discussion of the unity of science.
This series provides a venue where philosophers and logicians can apply specific
technical insights to fundamental philosophical problems. While the series is open
to a wide variety of perspectives, including the study and analysis of argumentation
and the critical discussion of the relationship between logic and the philosophy of
science, the aim is to provide an integrated picture of the scientific enterprise in all
its diversity.

More information about this series at http://www.springer.com/series/6936

Enrico Martino

Intuitionistic Proof Versus

Classical Truth
The Role of Brouwer’s Creative Subject
in Intuitionistic Mathematics

123
Enrico Martino
FISPPA Department
University of Padua
Padua
Italy

ISSN 2214-9775 ISSN 2214-9783 (electronic)

Logic, Epistemology, and the Unity of Science
ISBN 978-3-319-74356-1 ISBN 978-3-319-74357-8 (eBook)
https://doi.org/10.1007/978-3-319-74357-8
Library of Congress Control Number: 2017964580

© Springer International Publishing AG, part of Springer Nature 2018

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by the registered company Springer International Publishing AG part
of Springer Nature
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Così è (se vi pare)
So it is (if you think so)
Luigi Pirandello
To Livia
Preface

I was stimulated by Göran Sundholm to collect some of my old papers on

Intuitionism, which turn back to the eightieths and the ninetieths of the past century.
In those years, there was a lively philosophical debate between classical and
intuitionistic logicians and mathematicians. I was especially interested in the works
of Dummett, Troelstra, van Dalen, Sundholm, Veldman and others.
A main peculiarity of my research is a deep analysis of the main tenets of the
pioneers of Intuitionism: Brouwer and Heyting. In particular, it is analysed
Heyting’s explanation of the intuitionistic meaning of logical constants and
Brouwer’s idealisation of the creative subject as grounds of intuitionistic truth.
Besides, it is upheld the importance of the role of certain imaginary acts of
choice performed by an ideal agent for explaining the notion of reference not only
in intuitionistic but also in classical logic.
A crucial question, discussed in the work, is the following: to what extent
succeeds the intuitionistic perspective to avoid the classical realistic notion of truth?
My answer is that a form of realism is hidden in the idealisation of Brouwer’s
creative subject. In fact, in order to exploit the role of the creative subject, we need
to think him as if he were a real being: the mere imagine of him in our mind would
not be able to perform the actions required by his role.
Some papers of the present collection are written together with Daniele Giaretta
and Gabriele Usberti.
I’m grateful to Göran Sundholm for his interest in my work on Intuitionism.
I thank the Springer editor Shahid Rahman. Besides of thank Ali Mohammed,
Stephen O’Reilly and Nisha Keeran. A particular thank to my colleague Vittorio
Morato for preparing the manuscript in LATEX.

Padua, Italy Enrico Martino

June 2017

ix
Contents

1 Brouwer, Dummett and the Bar Theorem . . . . . . . . . . . . . . . . . . . 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Terminology and Symbolism . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Dummett’s Argument . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Critique of Dummett’s Argument . . . . . . . . . . . . . . . . . . . . . . . 4
1.5 Limits of the Eliminability of f-inferences . . . . . . . . . . . . . . . . 10
1.6 Final Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2 Creative Subject and Bar Theorem . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1 The Creative Subject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 The Creative Subject and Existential Statements . . . . . . . . . . . . 19
2.3 Equivalence of (PIE) and (BIM ) . . . . . . . . . . . . . . . . . . . . . . . . 21
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3 Natural Intuitionistic Semantics and Generalized
Beth Semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2 Generalized Beth-Models and Natural Models . . . . . . . . . . . . . 23
3.3 Generalized Natural Models . . . . . . . . . . . . . . . . . . . . . . . . . . 24
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4 Connection Between the Principle of Inductive Evidence
and the Bar Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.1 Inductive Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5 On the Brouwerian Concept of Negative Continuity . . . . . . . . . . . . 31
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
5.2 The Negative Continuity Theorem . . . . . . . . . . . . . . . . . . . . . . 32
5.3 A Proof of NCP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5.4 Weak and Strong Negation . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

xi
xii Contents

5.5 The Role of Time in Brouwer’s Argument . . . . . . . . . . . . . . . . 34

5.6 Brouwer’s Argument and Solipsism . . . . . . . . . . . . . . . . . . . . . 35
5.7 NCP and Lawless Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5.8 Revising NCP with the Help of the Creative Subject . . . . . . . . 40
5.9 Extensional Functions and Intensional Choice Sequences . . . . . 41
5.10 Troelstra’s Abstraction Process and NCP . . . . . . . . . . . . . . . . . 42
5.11 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
6 Classical and Intuitionistic Semantical Groundedness . . . . . . . . . . . 47
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
6.2 Construction of Model M . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
6.3 Axiomatisation of T . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
6.4 The Aczel–Feferman Intensional Operator . . . . . . . . . . . . . . . . 50
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
7 Brouwer’s Equivalence Between Virtual and Inextensible
Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
7.2 Reconstruction of Brouwer’s Paper of 1927 . . . . . . . . . . . . . . . 54
7.3 Comment on Brouwer’s Text . . . . . . . . . . . . . . . . . . . . . . . . . 55
7.4 How Brouwer Misinterpreted Himself . . . . . . . . . . . . . . . . . . . 57
7.5 A Minor Mistake in the Cambridge Lectures . . . . . . . . . . . . . . 60
7.6 On Posy’s Reconstruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
8 An Intuitionistic Notion of Hypothetical Truth for Which Strong
Completeness Intuitionistically Holds . . . . . . . . . . . . . . . . . . . . . . . 63
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
8.2 Symbolism and Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . 64
8.3 The Failure of Strong Completeness for Natural Semantics . . . . 65
8.4 Hypothetical Truth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
8.5 Remarks on Hypothetical Truth . . . . . . . . . . . . . . . . . . . . . . . . 67
8.6 Generalized Beth Semantics . . . . . . . . . . . . . . . . . . . . . . . . . . 69
8.7 Connection Between Hypothetical Semantics
and GB-Semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 70
8.8 A Strong Completeness Proof for GB-Semantics . . . . . . . . . . .. 71
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 73
9 Propositions and Judgements in Martin-Löf . . . . . . . . . . . . . . . . . . 75
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
9.2 Propositions and Judgements . . . . . . . . . . . . . . . . . . . . . . . . . . 75
9.3 Truth and Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
9.4 Metaphysical Realism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Contents xiii

10 Negationless Intuitionism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
10.1 Natural Semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
10.2 Failure of Strong Completeness . . . . . . . . . . . . . . . . . . . . . . . . 86
10.3 Second-Order Negationless Semantics . . . . . . . . . . . . . . . . . . . 89
10.4 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
11 Temporal and Atemporal Truth in Intuitionistic Mathematics .... 97
11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 97
11.2 Tenselessness and Classical Truth . . . . . . . . . . . . . . . . . . .... 98
11.3 Potential Intuitionism as a Subsystem of Epistemic
Mathematics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
11.4 Temporal Truth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
12 Arbitrary Reference in Mathematical Reasoning . . . . . . . . . . . . . . 113
12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
12.2 Some Objections to TAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
12.3 TAR as Embodied in the Logical Concept of an Object . . . . . . . 117
12.4 The Ideal Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
12.5 Arbitrary Reference and Impredicativity . . . . . . . . . . . . . . . . . . 122
12.6 Plural Reference Versus Sets . . . . . . . . . . . . . . . . . . . . . . . . . . 124
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
13 The Priority of Arithmetical Truth over Arithmetical
Provability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
13.2 Orthodox Versus Non-orthodox Intuitionism . . . . . . . . . . . . . . 135
13.3 The Constructive Notion of a Process . . . . . . . . . . . . . . . . . . . 138
13.4 Computational Realism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
14 The Impredicativity of the Intuitionistic Meaning
of Logical Constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
15 The Intuitionistic Meaning of Logical Constants
and Fallible Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
15.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Origin of the Essays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Subject Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
Chapter 1
Brouwer, Dummett and the Bar Theorem

with P. Giaretta

Abstract It is criticised Dummett’s refutation of Brouwer’s dogma. It is argued that

his criticism rests on an erroneous interpretation of Brouwer’s idea of “canonical
proof”.

1.1 Introduction

In Dummett (1977), Elements of Intuitionism, hereafter E I , Dummett gives a refuta-

tion of the “Brouwer’s dogma”, the famous assumption on which Brouwer based his
proof of the bar theorem, by means of an original argument involving the intuition-
istic notion of “proof containing inferences with infinitely many premisses”. In the
present article, we criticise Dummett’s argument, which rests on what seems to us
an incorrect interpretation of Brouwer’s idea of “canonical proof”, and we propose
an alternative interpretation of it. Notwithstanding our critique, we thank Dummett
for having stimulated our reflection on the foundation of Intuitionism.

1.2 Terminology and Symbolism

α, β, γ , . . . are variables for choice sequences. α(n) is the finite sequence of the first
n terms of α:

α(n) = α(o), . . . , α(n − 1) (α(0) = )

u, v, w, . . . are variables for finite sequences of natural numbers. If u =
u 0 , . . . , u n , the length of u (the number of its terms), which we will denote by
u ), is n + 1. l() = 0
l(
If u = u 0 , . . . , u n  and v = v0 , . . . , vm , we let u ∗v be u 0 , . . . , u n , v0 , . . . , vm .
If n is a natural number, we let uˆ  n be u ∗ n.
uˆ
 n, for any n, will be called a successor of u and u will be called the predecessor
of every uˆ  n.
We will use “ u is extended by v”, in symbols u ≥ v, to mean that ∃w(  v = u ∗ w). 
© Springer International Publishing AG, part of Springer Nature 2018 1
E. Martino, Intuitionistic Proof Versus Classical Truth, Logic, Epistemology,
and the Unity of Science 42, https://doi.org/10.1007/978-3-319-74357-8_1
2 1 Brouwer, Dummett and the Bar Theorem

We will use the term “spread” to refer to a function s defined on all finite sequences
and with values in {0, 1} such that:

s() = 0, u ) = 0 → ∃ks(uˆ
s(  k) = 0, s(uˆ
 k) = 0 → s(
u) = 0

u ) = 0, we say that u is admissible for s, and s can be thought of as the

If s(
tree whose nodes are the admissible finite sequences. In the following we will write
u ) = 0 or u ∈ s.
indifferently s(
If ∀ u ) = 0, s is the universal spread.
u s(
We will say that a sequence α belongs to a spread s, α ∈ s, if ∀ns(α(n)) = 0.
A sequence α belongs to a finite sequence u, α ∈ u, if ∃nα(n) = u.
Let u ∈ s and R be a species of nodes of s. We say that R bars u if every choice
sequence of s belonging to u meets R, formally if ∀αα∈s (α ∈ u → ∃nα(n) ∈ R).
We say that R is monotonic if u ∈ R → ∀k(uˆ  k ∈ s → uˆ
 k ∈ R).
We say that R is decidable if ∀ u u∈s (
u ∈ R ∨ u ∈
/ R).
Let R be a species of nodes of s. We call F -closure of R (in s) the species R F
inductively defined as follows:

(i) u ∈ R → u ∈ R F ;
(ii) ∀u u∈s (∀k u˚
 kuÙk∈s ∈ R F → u ∈ R F ).

Following Kleene and Vesley (1965), hereafter referred to as FIM, we say that a node
u, belonging to s, is inductively barred by R if u ∈ R F . It is evident, by induction
on the construction of R F , that if R bars inductively u, then R bars u.
At this point bar theorem, more explicitly the bar-induction theorem (shortly B I ),
monotonic bar theorem, decidable bar theorem, with regard to a spread s and to a
species of nodes R, can be expressed as follows:

• B I (s, R): If R bars , then R inductively bars .

• B I M (s, R): If R is monotonic and R bars , then R inductively bars .
• B I D (s, R): If R is decidable and R bars , then R inductively bars .

In the following, for simplicity, we will refer mostly to the universal spread and
simply say that R satisfies the bar theorem (and we will write B I (R)).
We will write simply B I instead of ∀R B I (R) (analogously for B I M and B I D ).
B I M and B I D are usually assumed as axioms in treatments of intuitionistic analy-
sis. It is known that B I M → B I D and that, under the hypothesis of ∀α∃x-continuity,
B ID → B IM .
B I is instead certainly false, as the following Kleene counterexample (K C) shows.
Let P(x) be an arbitrary decidable predicate on the natural numbers and R be
the species of nodes defined by: k ∈ R ↔ P(k),  ∈ R ↔ ¬∀x P(x). Obviously
R bars , but until we have a proof of ∀x P(x) ∨ ¬∀x P(x), we cannot assert that
 ∈ R F .
1.2 Terminology and Symbolism 3

Brouwer gave various formulations and proofs of the bar theorem or of particular
cases of it in Brouwer (1924b, 1927, 1954), hereafter referred to as 1924b, 1927,
1954. The hypothesis of decidability or monotonicity of R never occurs explicitly in
Brouwer’s proofs. Yet the hypotheses of 1924b and of 1927 imply the decidability
of R. On the other hand, in the formulation of 1954, which seems to be the most
general, there is no hypothesis on the barring species and so, for K C, this formulation
is certainly wrong.
All three proofs are based on the assumption, which we will call “Brouwer’s
dogma” (B D), that a proof of a node being barred can always be reduced to a
“canonical proof” (c.p.), where only inferences of the following three forms occur:
• η-inferences: u ∈ R
R bars u
• ζ -inferences: R bars u
R bars uˆ
k
R bars uˆ
 0, . . . , R bars uˆ
 k, . . .
• F -inferences:
R bars u
(Indeed Brouwer does not mention explicitly the η-inferences.) Then the proof
proceeds by acknowledging the ‘eliminability of the ζ -inferences’, i.e. the possibility
of further reducing the canonical proof to one in which only η- and F -inferences
occur. Finally in Brouwer’s argument, translated into our notation, the conclusion is
obtained by observing that a so-reduced proof of ‘R bars ’ can be transformed into
a proof of ‘ ∈ R F ’ by a simple substitution of ‘ u ∈ R F ’ for every occurrence of
‘R bars u’.

1.3 Dummett’s Argument

Dummett describes a c.p., to which B D refers, as a proof “expanded into its ‘fully
analysed’ form, that is, a form, in which every step has been broken down into a
sequence of steps each of which is as short as possible” (Dummett 1977, p. 94).
After he gives a more detailed description:
Brouwer’s notion of an analysis of a proof appears to be this: that whenever, in the course
of the proof, we appeal to some operation as yielding a result of a certain kind, then, in the
analysed form of the proof, that operation will actually be carried out. Thus the appearance
of a universally quantified statement in a proof, for instance the statement

∀k R bars uı
k
signifies our recognition that a certain operation will, when applied to any element of the
domain (in this case, to any natural number k), yield a proof of the corresponding instance
(here, of the statement ‘R bars uı k’). In the fully analysed proof, therefore, the universal
quantification does not appear: the operation is actually applied to each element of the
domain, yielding a proof of the corresponding instance, and that which formerly was inferred
from the universally quantified statement now appears as following from the individual
instances taken together. (Dummett 1977, p. 96)
4 1 Brouwer, Dummett and the Bar Theorem

The proofs, in which, according to Brouwer, only η-, ζ - and F -inferences occur,
should be of this type. After having considered a c.p. in the form of a well-founded tree
(i.e. of a finite paths tree), Dummett starts to expound Brouwer’s proof by describing
the process of elimination of the ζ -inferences. We quote here only the description of
the case of a ζ -inference which is preceded by a η-inference:

since this case, as we will see, turns out to be particularly critical.

[…] we consider that path in the proof-tree which leads from the conclusion ‘R bars uı  k’
of our ζ -inference to the conclusion ‘R bars ’ of the whole proof. Each of the statements
occurring on this path is of the form ‘R bars v’ for some finite sequence v which is either
an initial segment or an extension of u. Moreover, in the passage from one statement to the
next, the length of the finite sequence mentioned is either increased or diminished by 1, and
the length of the sequence mentioned at the end of the path is 0. Hence, somewhere along
the path there must occur the statement ‘R bars u’. We accordingly replace the entire proof
of that occurrence of ‘R bars u’ by a derivation of it by means of an η-inference. (Dummett
1977, pp. 97–98)

Finally, after having expounded the obvious conclusion of the proof (which we
mentioned at the end of the last paragraph), Dummett observes that since this proof
makes no reference either to the decidability or to the monotonicity of R, it is certainly
wrong because of K C. He concludes that what is wrong must be B D, the only
unjustified assumption of the proof, and he considers K C a refutation of it.
It comes as no surprise that consideration of this example shows that the flaw in Brouwer’s
proof lies in his unsupported assertion that any fully analysed proof that a species bars  (or
any other finite sequence u) can contain only η-, ζ - and F -inferences. (Dummett 1977, p.
98)

1.4 Critique of Dummett’s Argument

According to Dummett’s conclusion, there should not be only η-, ζ - and F -inferences
in a c.p of ‘R bars ’, where R is the barring species of K C. On the other hand the
proof-tree:
1.4 Critique of Dummett’s Argument 5

would seem to contain only inferences of the mentioned types. But Dummett rejects
this obvious objection on the basis of a strange principle of constructivity.
He says that, since an intuitionistic proof must be understood as a mere mental
act, “we cannot consider it as, so to speak, already having certain features, or dis-
playing certain patterns, which we have failed to notice” (Dummett 1977, p. 100).
He concludes that, with regard to the species R of K C, though it is obviously true
that for a fixed k we can decide whether P(k) or ¬P(k) and so deduce ‘R bars k’
according to the case, from ‘k ∈ R’ by means of an η-inference or from ‘R bars
’ by means of a ζ -inference, yet in the whole proof of ‘R bars ’, understood
as a unique global mental act, the various k have not been individually analysed
and therefore the various formulae ‘R bars k’ are not deduced in it by Brouwer’s
inferences.
Dummett gives the name “Θ-inference” to the “analysed version” of an inference
of the type:

R bars u ∨ R bars uˆ
k
ˆ
R bars u k
that is the inference which we obtain from it by replacing the premiss by a determinate
one of the two disjoints, and maintains that all that can be said about the c.p. of ‘R bars
’, considered as global mental act, is that the formulae ‘R bars k’ are deduced
by means of Θ-inferences. Therefore the c.p. at issue would have, according to
Dummett, not the aspect of the proof-tree 2 but the aspect:

So the Θ-inferences would constitute a new type of inference which can occur in
a c.p., and, according to Dummett, they would not be reducible to those of Brouwer.
6 1 Brouwer, Dummett and the Bar Theorem

We think, on the contrary, that the proof-tree 2 is intuitionistically no less correct

than proof-tree 3. Certainly, in the mental act constituting the proof 2, the various
k are not completely analysed; what is present in that mental act is only the mere
possibility of analysing them and then of deducing ‘R bars k’, according to the
case, by a suitable Brouwer inference. But this is the best we can demand from
an intuitionistic point of view! In accordance with the conception of the potential
infinite, the fact that the proof of ‘R bars ’ consists of the infinitely many proofs
of the various ‘R bars k’ can only be understood in the sense that the knowledge
of a generative process of the infinitely many proofs of the premisses of the last
F -inference is a constituent of the proof. This can be said with reference to the
proof-tree 2 as well as to proof-tree 3.
Dummett himself explains very correctly the intuitionistic meaning of an inference
with infinitely many premisses:
Thus the only way of understanding the idea of an inference from denumerably many pre-
misses A(0), A(1), . . . which is consistent with a constructivist outlook proves to coincide
exactly with the intuitionistic interpretation of an inference from ∀n A(n). (Dummett 1977,
pp. 96–97)

But then he goes on:

An intuitionistic proof involving inferences from universally quantified statements really is,
therefore, what Brouwer maintains, a representation of a more analysed proof containing
inferences from infinitely many premisses. (Dummett 1977, p. 97)

Now, it is not clear exactly what Dummett means. The expression “representation
of a more analysed proof containing inferences from infinitely many premisses” and
what he says about a “fully analysed proof” suggest the notion of a proof, where
the infinitely many premisses A(0), A(1), . . . are all visibly present. But this notion
does not have any sense from an intuitionistic point of view, according to which, as
Dummett himself recognises, to draw a conclusion from the infinitely many premisses
A(0), A(1), . . . can only mean to draw it from the premiss ∀n A(n). There is no way
in which a proof of ∀n A(n) can have a “fully analysed version” such that the various
proofs of the A(k) occur explicitly in it. It may be, of course, that the knowledge
of the generative process makes us a priori aware that the infinitely many proofs
are very similar, in which case the “generic” proof of A(k) would be visualisable,
or that, on the contrary, the form of a proof of A(k) essentially depends on k, in
which case it would be difficult to conceive the global configuration of the infinitely
many proofs. But this is a distinction which, though suggestive, rests on a concept of
“visualisability” which, in our view, cannot satisfactorily characterise the soundness
of an intuitionistic proof, both because it is too vague and, above all, because it seems
to concern the intuitability of the actual infinite rather of potential infinite.
However, even if someone wanted to adopt such a standard of evaluation, it would
not seem possible for him to consider the proof-tree 3 more acceptable than proof-
tree 2. The same argument used by Dummett against proof-tree 2 can be used against
proof-tree 3: in this proof, conceived as global mental act, it is not known what
the various uk are. Only after having analysed a single k, it is possible to establish
1.4 Critique of Dummett’s Argument 7

whether uk = k or uk =  and so to execute the corresponding Θ-inference. But,
without such an analysis, all that can be said is that inferences of the type:
R bars k ∨ R bars 
R bars k

and not their analysed forms, occur in that proof. It turns out that one can have a global
intuition of the proof in question only if one gives up demanding a representation of
its “fully analysed” form in the sense expounded by Dummett.
Dummett attributes to Brouwer the idea of “fully analysed proof” which he
expounds. We do not know to which of Brouwer’s articles he refers. However, it
seems to us that at least in the articles 1924b, 1927, 1954, in which he discusses
the bar theorem, Brouwer means by “analysis of a proof” something different. The
intuition on which the dogma rests seems to be the following: the knowledge of the
fact that a node u is barred must, after all, either be reduced to the knowledge that
u ∈ R or be obtained through the knowledge that its predecessor or all its successors
are barred. That is, it is always possible to transform any given proof P of‘R bars u’
into a new proof P of one of the following types:

P and P0 , . . . , Pk , . . . are simpler proofs than P . The analysis of P in order to

obtain P consists in explicating, by reflection on P, the implicit reference to the
nodes surrounding u and not, as Dummett seems to believe, in actually carrying out
those constructions which had been acknowledged feasible in P. In case of proof-
tree 6 P consists of a generative process of the proofs Pk , which are not necessarily
canonical, and of the final F -inference. In its turn, every Pk can be transformed into
a Pk which, if different from proof-tree 4, splits into simpler proofs, and so on.
These considerations suggest the following precise inductive definition of c.p.
8 1 Brouwer, Dummett and the Bar Theorem

Definition 1.1 (a) u ∈ R is a c.p. of ‘R bars u’;

R bars u
(b) If P is a c.p. of ‘R bars u’, then for every k, P is a c.p. of ‘R bars
R bars uˆ
k
uˆ
 k’;
P0 , . . . , Pk , . . .
(c) If, for every k, Pk is a c.p. of ‘R bars uˆ
 k’, then is a c.p. of
R bars u
’R bars u’.
By generalising the Definition 1.1, it is possible to obtain the general concept of
an intuitionistic proof-tree. To be precise, we state first of all the following definition
of “inductively defined tree” (i.d.t.):
Definition 1.2 (a) {} is an i.d.t.
(b) If T0 , . . . , Tn is a finite sequence of i.d.t., then the species formed by  and by
k ∗ u, where u ∈ Tk (0 ≤ k ≤ n), is an i.d.t.;
(c) If T0 , . . . , Tk , . . . is an infinite sequence of i.d.t., then the species formed by 
and by k ∗ u, where u ∈ Tk (0 ≤ k), is an i.d.t.
Then the definition of “proof-tree” (p.t.) concerning a species A of axioms and a
species I of inference rules is as follows:
Definition 1.3 A p.t., concerning A and I, is a couple T, c, where T is an i.d.t.
(called the support of the p.t.) and c is an application which associates a formula with
every node of T so that
(a) if u is a terminal node, then c( u ) ∈ A;
(b) if u is not terminal, then c( u ) is the conclusion and the various c(uˆ
 k), for all k
such that uˆ  k ∈ T , are the premisses of an instance of one of the inference rules
of I.
We think that this is the precise and correct explication of the intuitionistic concept
of p.t. A confirmation that this was, in substance, the concept of c.p. which Brouwer
had in mind in his exposition of the bar theorem seems to be the fact that the Defi-
nition 1.2 is quite analogous to Brouwer’s definition of well-ordered species and, in
a footnote in 1927, Brouwer mentions explicitly the analogy between mathematical
proofs and well-ordered species:
Just as, in general, well-ordered species are produced by means of the two generating oper-
ations from primitive species [cf. (Brouwer 1926, p. 451)], so, in particular, mathematical
proofs are produced by means of the two generating operations from null elements and ele-
mentary inferences that are immediately given in intuition (albeit subject to the restriction
that there always occurs a last elementary inference). (Brouwer 1927, p. 460)

An alternative way of characterising a p.t. is to take as support a barred spread

instead of an i.d.t. This is the way followed, in substance, by Kleene in FIM in the
formalisation of Brouwer’s proof of the bar theorem (FIM, pp. 65–67) and seems
to be implicitly adopted by Dummett in his informal description of p.t. In particular
Dummett does not care at all that the tree is inductively defined and demands only
that it is well-founded:
1.4 Critique of Dummett’s Argument 9

To be a proof, it must be well-founded: if the proof is conceived as arranged in tree form,

every branch must be finite. (If it were possible to form an infinite sequence of propositions,
beginning with the conclusion of the proof, each subsequent proposition being one of the
premisses upon which the preceding one depended, then we should have no reason to accept
as true any proposition in the sequence, including the conclusion of the ‘proof’. This resem-
bles Aquinas’s denial of the possibility of an infinite regress in causes). (Dummett 1977, p.
95)

We want to point out a weak point of this approach. We define a barred spread
(b.s.) as a couple s, R, where s is a spread and R a species of nodes of s barring
the vertex. We define a p.t. with support s, R in the following way:

Definition 1.4 A p.t. with support s, R, concerning A and I, is a term s, R, c,
where c is an application associating a formula with every node of s so that
(a) if u ∈ R, then c(u ) ∈ R;
(b) c(u ) is the conclusion and the various c(uˆ
 k), for all k such that uˆ
 k ∈ s, are the
premisses of an instance of an inference rule I.

We say that a p.t. (both according to the Definition 1.3 and the Definition 1.4)
is sound if, whenever the axioms are true (under a certain interpretation) and the
inference rules are truth-preserving, the conclusion (formula at the vertex) is true.
We say that an i.d.t. (a b.s.) is sound if every p.t. having it as support is sound.
Now, it follows immediately, by induction on its construction, that every i.d.t. is
sound.
Is a b.s. always sound?
Let us suppose that s, R is sound. Then, if we take as A the species of the
formulae ‘ u ∈ R F ’, where u ∈ R, as I the species of the F -inferences, as c the
application associating the formula ‘ u ∈ R F ’ with every u ∈ s, the p.t. s, R, c
turns out to be sound, from which follows  ∈ R F . So B I (s, R).
Let us suppose, vice versa, B I (s, R). Let s, R, c be any p.t., with support s, R,
concerning a set A of true axioms (under a certain interpretation) and a set I of truth-
preserving inference rules. Then, by induction on the construction of R F , it follows
that, for every u ∈ R F , c(u ) is true (under the interpretation in question). Since
 ∈ R F by hypothesis, we conclude that c() is true. So s, R is sound.
Therefore
Corollary 1.1 A b.s. s, R is sound if and only if B I (s, R).
Since, by KC, BI does not hold in general, it follows that a s, R is not generally
sound.
Of course, soundness can be obtained by imposing suitable conditions on s, R,
but, by Corollary 1.1, the problem of fixing such conditions is exactly equivalent
to the problem of determining the validity conditions of the bar theorem. Thus, in
order to prove the last one, to use the notion of proof based on a b.s. leads to a
petitio principii. In F I M it is explicitly demanded (translating into our terminology)
that s, R satisfies B I (s, R); equivalently, it is possible to consider a notion of p.t.
defined as in Definition 1.4 but with reference to a support s, R inductively barred
10 1 Brouwer, Dummett and the Bar Theorem

rather than simply barred. But then, because of the inductive definability of R F which
the notion of inductive barredness refers to, the notion of proof-tree according to the
Definition 1.4 is essentially reduced to that of the Definition 1.3.
Therefore, henceforth, a p.t. will always meant according to the Definition 1.3 and
a c.p. according to the Definition 1.1. By assuming such references, Brouwer’s dogma

B D(R): if R bars , there is a c.p. of ‘R bars ’

acquires a precise meaning.

For convenience of exposition, we will state BD(R) in a slightly different way.
Let us consider the species IndR inductively defined as follows:

Definition 1.5 (a) u ∈ R → u ∈ I nd R;

(b) u ∈ R → uˆ  k ∈ I nd R;
(c) ∀k uˆ
 k ∈ I nd R → u ∈ I nd R.

That is IndR is the closure of R by η-, ζ - and F -inferences. In other words, IndR
is the species of the nodes u for which there exists a c.p. of ‘R bars u’. Then B D(R)
can be expressed in the form:

B D(R): if R bars , then  ∈ I nd R.

Now, since, for the species R of K C, it is quite self-evident that  ∈ I nd R (i.e.

that the proof-tree 2 is a c.p. in our sense), BD(R) holds and so, contrary to what
Dummett maintains, K C does not refute Brouwer’s dogma.

1.5 Limits of the Eliminability of ζ -inferences

Since, as we have seen, for the R of K C B D(R) holds, but B I (R) does not, the
ζ -inferences certainly cannot be eliminated from the c.p. of ‘R bars ’. In fact, the
proof-tree 2 clearly shows that the process for eliminating the ζ -inferences described
by Dummett (which we quoted in the Sect. 1.3) is not constructive: in order to apply
it, we should replace the whole c.p. by the proof

if ζ -inferences occur in the original proof-tree 2, and we should obtain as c.p of ‘R

bars ’
1.5 Limits of the Eliminability of ζ -inferences 11

if ζ -inferences do not occur in proof-tree 2. Since proof-trees 7 and 8 differ in all

nodes of length 1, it follows that, until we have proved ∀x P(x) ∨ ¬∀x P(x) and so
found out whether ζ -inferences occur in proof-tree 2, no node of length 1 of the c.p.
can be built.
Therefore KC does not refute BD but refutes the unconditional eliminability of
the ζ -inferences from a c.p.
The hypotheses of monotonicity or of decidability on R serve precisely to warrant
such eliminability. In fact we have:

Lemma 1.1 (a) u ∈ I nd R ∧ R is monotonic → u ∈ R F .

(b)  ∈ I nd R ∧ R is decidable →  ∈ R F .

Proof (a) It suffices to prove that R F is closed by ζ -inferences. In fact, by induction

on R F , we can show that if u ∈ R F , then u ∈ R or ∀k uı
k ∈ R F . So, because R
ˆ
is monotonic, it follows in any case that ∀k u k ∈ R .F

(b) We say that u is prebarred by R if there exists a v ≥ u such that v ∈ R . Let us

prove, by induction on IndR, that

u ∈ I nd R → u ∈ R F ∨ u is prebarred by R.

Let us suppose that u ∈ I nd R. If u ∈ R, the conclusion follows directly.

If u = uˆ
 k and v ∈ I nd R, then, by the inductive hypothesis, v is prebarred, in
which case u is too, or v ∈ R F − R, in which case ∀k vık ∈ R F and so, in particular,
u ∈ R F .

If, finally, ∀k uˆ
 k ∈ I nd R, then we can decide, by virtue of the decidability of
R, whether u is prebarred by R. In the positive case the conclusion follows directly.
In the negative case we can show, by the inductive hypothesis, that, for every k,
uˆ
 k ∈ R F or uˆ k is prebarred; in this last subcase, since u is not prebarred, uˆ
k∈ R
and so, in every one of the two subcases, uˆ  k ∈ R F whence u ∈ R F .
It seems to us that this completely clarifies the relationship between K C and
Brouwer’s proof.
12 1 Brouwer, Dummett and the Bar Theorem

1.6 Final Considerations

Of course, the fact that KC does not refute BD does not say anything about the
reliability of the latter. Yet it can easily be proved that BD is exactly equivalent to
B IM :

Theorem 1.1 B D ↔ B I M

(where B D = ∀R B D(R), B I M = ∀R B I M (R))

Proof Let us assume B D and let R be a monotonic barring species. By B D  ∈

I nd R and so, by Lemma 1.1a,  ∈ R F . Vice versa, let us assume B I M . If R is a
species barring , its monotonic closure R ζ bars  too, and so  ∈ R ζ F = I nd R.

Therefore, since B D and B I M are equally reliable, Dummett’s claim that it is

convenient to assume B I M directly as an axiom, in order to avoid the problematic
character of B D, turns out to be quite groundless. Dummett makes a further obser-
vation in favour of his claim. Referring to the formalisation of the dogma in F I M,
he says:
In order to formalize it, it is necessary to assume, as an axiom, a formalization of the
statement ‘If R bars , then there is a proof of “R bars ” which uses only η-, ζ - and
F -inferences’. Formalization of this statement is quite straightforward; a proof-tree can
obviously be represented as a dressed spread. However, standing on its own, such an axiom
appears quite ad hoc; especially when, in order to be able to exploit the well-founded character
of the proof-tree, we should also have to assume axiomatically a principle of transfinite
induction within such proof-trees (the general principle of transfinite induction can be proved
equivalent to the principle of Bar Induction): it is therefore preferable to assume the principle
of Bar Induction at the outset as an axiom or axiom schema. (Dummett 1977, pp. 102–103)

However, the assumption that the p.t.’s satisfy a principle of transfinite induction,
which is essentially equivalent to the bar theorem, does not depend on the formali-
sation but on the notion of p.t. adopted by Kleene. This notion is explicitly intended
to satisfy the above-mentioned principle, but, as we have already noted, there is no
need to impose the same requirement on the notion of p.t. of Definition 1.3, since it
is a natural consequence of its construction that the latter notion satisfies it.
We have two further remarks to make about the idea, suggested by Dummett, that
B D would be an ad hoc axiom. First of all B D seems to express correctly Brouwers
intuition, already mentioned above, that the knowledge that a node u ∈ / R is barred
must be based on an examination of its surrounding nodes. On the contrary, this
intuition is lost in the bar theorem, in which no explicit reference is made to the
predecessor of u. In the second place, even if no more evidence is attributed to B D
than to B I M , or one prefers to ignore the problem of evidence, B D seems to us
interesting in itself. While, in fact, the bar definition involves the notion of infinite
sequence, this notion does not occur at all in the definition of IndR, which is stated
entirely in terms of finite sequences. Thus B D says, in effect, that the bar notion,
even for an R such that B I (R) does not hold, can be expressed in terms of finite
sequences. In this sense, B D can be regarded as a generalisation of B I M .
1.6 Final Considerations 13

We want in conclusion to make some comments on Brouwer’s original articles.

We shall consider first whether the mistaken idea that the ζ -inferences can, in
general, be eliminated goes back to Brouwer. Brouwer summarises the elimination
process in 1924a, and in 1954 without ever considering the case of a ζ -inference
preceded by an η-inference, that is, the following situation:

This case is not even considered in Heyting’s description, concerning a finitary

spread, in Heyting (1956, pp. 42–44). We can guess that in Brouwer (1924a, 1927)
and in Heyting (1956), its treatment has been omitted because the implicit decidability
of the barring species (besides, in Heyting, the finitary character of the spread)
makes it obvious. But this is not so in Brouwer’s (1954) in which no hypothesis
on R occurs and thus, as we have seen, the elimination of the ζ -inferences is illicit.
Brouwer’s argument, paraphrased and translated into our terminology, is in substance
the following:
if in the given proof of ‘R bars ’ it has not been demonstrated that  ∈ R, then ‘R bars ’
has been deduced by an F -inference; so, for every k, ‘R bars k’ must have been proved
before ‘R bars ’ and thus cannot have been deduced by a ζ -inference.

Here Brouwer seems indeed to overlook, probably because of the lack of precision
of his statement, the possibility, pointed out by K C, that ‘ ∈ R’ is not known but
can nevertheless be used to construct, for some k, a proof of ‘R bars k’.
Our last observation concerns Brouwer’s statement in 1954 that the barring species
R is “not necessarily predeterminate”. If, as we believe, this means that R can depend
on choice sequence parameters, it seems to us that this possibility makes the validity of
B D more problematical. Let us consider, in fact, the following example. We assume
that the universe of choice sequences is built by a “creative subject” who, at every
stage of knowledge n, conceives one and only one sequence αn (the supposition that
the creative subject conceives only one sequence per stage was put in question because
of Troelstra’s paradox, but this can be avoided by means of a suitable distinction of
reference levels. Cf. Troelstra (1969)). Since such a universe, though numerable, is
potentially as rich as the universe of all sequences, we conjecture that the bar theorem
holds for it (within the limits within which it holds for the universe of all sequences).
Now, if R is the species of the nodes α(n + 1), for n ≥ 0, then R obviously bars
 and is decidable (because a node u of length m belongs to R if and only if m = 0
and u = α m−1 (m), and yet  ∈ / R F , since R F = R).
14 1 Brouwer, Dummett and the Bar Theorem

In general, it seems to us that B D is fairly plausible, provided that R can be

defined without reference to the concept of infinite sequence. In fact, as we have
already noted, B D expresses the eliminability of this concept from the bar notion,
i.e. the possibility of describing the species of the nodes barred by R without using the
concept of infinite sequence. Now, this possibility seems to us rather unreliable (as
well as not very interesting) if the concept in question is already essentially involved
in the definition of R.

References

Brouwer, L. (1924a). Bemerkungen zum Beweis der gleichmässigen Stetigkeit voller Funktionen.
Verhandelingen der Koninklijke Nederlandsche Akademie van Wetenschappen te Amsterdam, 27,
644–646.
Brouwer, L. (1924b). Beweis, dass jede volle Funktion gleichmässig stetig ist. Verhandelingen der
Koninklijke Nederlandsche Akademie van Wetenschappen te Amsterdam, 24, 189–193.
Brouwer, L. (1926). Zur Begrüdndung der intuitionistischen Mathematik III. Mathematische
Annalen, 96, 451–488.
Brouwer, L. (1927). Über Definitionsbereiche von Funktionen. Mathematische Annalen, 97, 60–75.
English translation in From Frege to Gödel, Cambridge MA, 1967, pp. 446–463.
Brouwer, L. (1954). Points and spaces. Canadian Journal of Mathematics, 6, 1–17.
Dummett, M. (1977). Elements of intuitionism. Oxford: Clarendon Press.
Heyting, A. (1956). Intuitionism: an introduction. Amsterdam: North-Holland Publishing Company.
Kleene, S. C., & Vesley, R. E. (1965). The Foundations of intuitionistic mathematics. Amsterdam:
North-Holland Publishing Company.
Troelstra, A. (1969). Principles of intuitionism. Berlin: Springer.
Chapter 2
Creative Subject and Bar Theorem

Abstract In the present article, a reasonably precise description of Brouwer’s notion

of “creative subject” is proposed and an axiom is introduced which is conceptually
equivalent to the bar theorem.

2.1 The Creative Subject

The idea of the creative subject occurs in various writings of Brouwer in a somewhat
vague manner. Several logicians, among them Kreisel (1967), Troelstra (1969) and
Dummett (1977), have recently tried to analyse this concept by proposing some
axioms, which are however rather controversial, as Dummett’s discussion shows.
The creative subject, which we will call Σ, carries out his mathematical activity in
time, which we assume divided in ω distinct states of knowledge. The main feature of
the theory of the creative subject consists in using the fact that knowledge increases in
time by explicitly referring to the stage in which Σ gets to know a certain proposition.
Such a reference is expressed by the propositional operator n . If A is a formula,
n A stands for: “at stage n, Σ has evidence for A”.
The kind of axioms for n we put forward essentially depends on our idealisation
of Σ.
In the literature, two essentially different conceptions of n A were proposed:
the strict one, according to which Σ can at any stage know only a finite number of
propositions, and the wider one, according to which at any stage Σ is allowed to
know infinitely many propositions.
The wider conception is proposed, among others things, by Troelstra (1969) as a
possible solution of a paradox of diagonalization in the strict theory (see Sect. 2.3),
which he pointed out.
We begin with some critical observations on the wider conception. The following
axioms are usually assumed both in the strict theory and in the wider one:

(1.1) ∀n(n A ∨ ¬ n A)
(1.2) ∀n∀m(n A →n+m A)
(1.3) ∃n n A ↔ A

© Springer International Publishing AG, part of Springer Nature 2018 15

E. Martino, Intuitionistic Proof Versus Classical Truth, Logic, Epistemology,
and the Unity of Science 42, https://doi.org/10.1007/978-3-319-74357-8_2
16 2 Creative Subject and Bar Theorem

These axioms are not sufficient to yield a precise conception of the creative sub-
ject, but they seem to be minimal requirements; they are implicitly used by Brouwer
himself in the well-known constructions of counterexamples for some classical the-
orems.
(1.1) is usually justified in the following manner: the predicate n A is decidable
since at the stage n Σ knows whether he has evidence for A or not. We will shortly
discuss such a justification.
(1.2) says that the knowledge of Σ is cumulative: at every stage, he knows again
(or he maintains the knowledge of) what he knew at the preceding stages.
(1.3), from left to right, says that Σ reasons in an intuitionistically correct man-
ner, that is, he knows only intuitionistically true propositions. From right to left, it
seems to express the “solipsistic” conception of intuitionistic mathematics: the true
propositions are only those which are known to the creative subject. If one does not
want to commit oneself to this last statement, (1.3) can be weakened by replacing it
by:
(1.3 ) ∃n n A → A
and

(1.3 ) A → ¬¬∃n n A

(1.3 ) (called by Kreisel the Principle of Christian Charity) says that if A is true, it
cannot be excluded that, in some stage, Σ comes to know it.
In the wider conception it is allowed that, if n A and if B is a “immediate
consequence” of A, then n B. Apart from the difficulty of generally characterising
in a sufficiently precise manner the vague notion of immediate consequence, it seems
that the following axioms for it are hardly objectionable (see Dummett 1977). Let
P(x) be a predicate on natural numbers:

(1.4) n ∀x P(x) → ∀x n P(x)

(1.5) n P(m) →n ∃x P(x)
(1.6) n (A ∨ B) →n A∨ n B

In (1.6), it is understood that A and B do not contain choice parameters nor any
information which is not available at stage n (see Sect. 2.2).
Now, we see at once that from (1.1)–(1.6) we can deduce some intuitionistically
incorrect propositions.
In fact, let P(x) be a decidable predicate (without choice parameters), such that
∀x(P(x) ∨ ¬P(x)). From (1.3) it follows that, for some n,

n ∀x(P(x) ∨ ¬P(x))

whence, by (1.4) and (1.6),

∀x(n P(x)∨ n ¬P(x)).

2.1 The Creative Subject 17

It follows that, if ∃x P(x), then there exists a k such that n P(k) and so, by 1.5,
n ∃x P(x). Therefore, the following equivalence holds

∃x P(x) ↔ n ∃x P(x).

Thus, since, by (1.1), n ∃x P(x) ∨ ¬ n ∃x P(x), we have ∃x P(x) ∨ ¬∃x P(x).

But this cannot be intuitionistically acceptable for an arbitrary P(x).
In my opinion this incoherence arises from the fact that (1.1) is acceptable only
if, in order to have n A, it is required that Σ is aware of having evidence of A. Now,
the evidence of ∀x P(x) does not involve (unless we greatly stress the idealisation
of Σ) the conscious evidence of all single instances P(k). Thus, (1.4) holds only
if evidence is meant as implicit, not necessarily conscious, evidence. Hence, the
incompatibility of (1.1) and (1.4) results.
Since Σ is no real subject but an idealised subject, we could idealise him so that the
evidence of ∀x P(x) involves the conscious evidence of each single P(k). But such
an idealization would attribute to Σ superhuman powers which would allow him to
know some classically but not intuitionistically true propositions. Since intuitionistic
truths are to be humanly knowable, we must be very cautious in idealising Σ.
Therefore, since (1.4) seems to be a minimal requirement of the wider conception,
we hold that it forces the axiom (1.1) to be abandoned.
Then, at the same stage, Σ can implicitly know infinitely many propositions. If,
for instance, at stage n he knows the soundness of the axioms and of the inference
rules of a certain system F, then at the same stage n he has implicit evidence of all
theorems of F (but he is not necessarily able to decide whether a given formula is a
theorem, as (1.1) requires). Indeed, one could even assume that every stage is closed
with respect to “analytic” deductions and that the passage from one stage to another
is characterised by an increase of external information (see Posy 1977). But we think
that in that case n A should be more properly be read: “At the stage n, Σ can
have evidence of A”. The potential interpretation serves the purpose of avoiding that
superhuman powers (as that one of deducing actually all what is deducible from the
available information) are attributed to Σ. The closure with respect to the analytic
deductions seems to be in agreement with Grzegorczyk’s semantics, (cf. Grzegorczyk
1964).
Nevertheless, it seems to me that such a conception has not the effect of taking the
idea of the creative subject seriously, but rather of replacing it by other notions, such
as “deducibility (in an intuitive sense) from certain given information”, (Van Dalen,
1978). On the contrary, we believe that the originality and the strength of the theory
of the creative subject consist just in the possibility of exploiting the concept of
conscious evidence, which Brouwer has already partially exploited by his implicit
assumption of (1.1). We hold that new interesting results are obtainable not just
by suppressing (1.1) but rather by adding to it new axioms intended to exploit the
concept more deeply.
Therefore, we will abandon the wider conception in favour of the strict one,
according to which the request of the conscious evidence is certainly plausible. This
request can be made even more explicit by stating, for instance, that in order to
18 2 Creative Subject and Bar Theorem

have n A, Σ has to write a proof of A at the stage n. We here obviously pass

over the inadequacy of natural language to express an intuitive proof; our statement
serves only the purpose of clarifying metaphorically the meaning of the adjective
“conscious”. What is of importance is that, proving only finitely many propositions
at each stage, Σ has the possibility of explicitly directing his attention at each one
of them.
The axiom (1.2) results to be obvious. Without committing ourselves to the solip-
sistic conception, we will accept (1.3) with the following justification: since A is
intuitionistically true, if it has been proved by someone, we suppose that, as soon as
a proof of A is found (by someone), it is communicated to Σ who includes it among
the propositions of which he has evidence (at the stage he is in at the moment of
communication).
We now intend to propose a modification of the strict conception in order to
obviate, at least in part, the drawback of the invalidity of (1.4) and of other axioms
which seem to hold only in the large conception.
We replace (1.4) with the scheme
(1.4 ) ∀n(n ∀x P(x) →n P(k)).
If (1.4 ) is understood in the sense that all its instances should be simultaneously
true, then it is certainly not acceptable in the strict conception, since it presents the
same difficulties as (1.4). Therefore, we propose to understand the validity of an
axiom scheme in the following sense: for each instance, it is possible “to programme
Σ” at the stage 0 so that the instance in question is true. By programming Σ we
mean “to instruct Σ so that if he happens to be in certain favourable conditions,
he performs certain deductions which interest us”. In this sense, (1.4 ) is thoroughly
acceptable: whenever we have fixed P and k, we can instruct Σ so that, if he deduces
∀x P(x) at some stage, he takes care of deducing P(k) at the same stage.
Similarly for (1.5), whenever P is fixed, Σ can be instructed to deduce ∃x P(x)
as soon as he has deduced P(m) for some m. Likewise for (1.6) (with the necessary
restrictions on the choice parameters).
In general, we admit the following principle:

(1.7) Whenever, for a fixed proposition A, we recognise that if we were in a certain

state of knowledge s we could prove A, we can suppose that, whenever Σ is
in the state s, he actually deduces A.

Some other axioms which turn out to be sound are the following:

(1.8) ∀n(n (A → B) → (n A →n B)

(1.9) ∀n(n (A ∧ B) →n A∧ n B)
On the contrary, we observe that
(1.8 ) (A → B) → ∀n(n A →n B)
does not hold, since until we know a proof of A → B (or of ¬(A → B)) we cannot
instruct Σ to deduce B from A.
Another random document with
no related content on Scribd:
Jack and Thomason climbed to the middle of the bed and sat down
awkwardly, both looking in the same direction, like rowers in a boat.
“And remember you have paddles in your hands,” reminded Bonnie
May.
“I have a paddle,” responded Jack.
“I ain’t,” objected Thomason.
“Oh, yes, you have,” declared Jack, “one just like mine.” He took a
stroke with an imaginary paddle, held suitably.
“Well—I have a paddle,” conceded Thomason.
Bonnie May then was helped to the “bridge,” beside Clifton.
Clifton began. He was not quite sure about the lines, but he recalled
the situation clearly enough. “Best go below, my daughter,” were the
words which filled the room with a ringing effect. “I have not seen a
gull since the second watch ended, and they do not hide from
ordinary storms. I fear we may be caught in a tempest.”
“Look at them!” she screamed. “Look! Look!”
Bonnie May clasped her hands in a frenzy of earnestness. Her
words came with intense eloquence: “Let me stay with you, father. I
fear no storm while I am by your side.”
Her voice filled the room with tones which were intense, even,
resonant, golden.
Mrs. Harrod, regarding her incredulously, put out a hand and
touched Flora on the arm. No one else stirred.
There came Clifton’s response: “But, child, I tell you Davy Jones’s
locker fairly gapes in gales like this. I bid you go below.”
The response came with even greater intensity: “But tell me first,
father: Would a raft live in such a sea as this?”
So the rather silly lines were repeated, back and forth. But they
scarcely seemed silly. The two players were putting a tremendous
earnestness into them, and the “audience” felt no inclination at all to
smile.
The two players came to the point in the story where the ship struck
a rock, and their intensity was more than doubled. The raft began its
part in the scene, but nobody looked at it for a time.
Clifton was trying to compel Bonnie May to consent to board the raft.
He had seized her arm roughly and was threatening her. She
screamed her refusal. Then it came time for her to behold the
murderous looks on the faces of the two men on the raft.
“Look at them!” she screamed. “Look! Look!” She pointed at the raft,
her eyes wide with terror. The “audience” could not refrain from
looking at the raft.
Jack and Thomason were wielding their paddles with great vigor.
Jack had also begun to lurch from right to left, as a man might do in
a storm-tossed raft. Thomason, catching the drift of things, was
imitating him.
And then, unfortunately, Thomason’s bed gave way. With an ear-
splitting crash it collapsed, just as Bonnie May screamed: “Look!
Look!”
And of course it was at that precise instant that Mrs. Baron came
rushing into the room.
 CHAPTER XXVIII
AFTER THE CURTAIN WAS LOWERED

Mrs. Baron had returned from her calling expedition earlier than
she had expected to. She had had a feeling that something might go
wrong. Prescience is really a wonderful thing.
Now as the poor lady stood within Thomason’s room she was quite
terrified. For the moment there had been a dreadful din. And now,
looking at Thomason, she caught the rebellious expression in his
round, innocent eyes. She saw that he had brass rings in his ears.
Unfortunately she did not associate the brass rings with the window
blinds. And his face was horribly streaked. His right leg was sticking
up in air quite inelegantly, and he was clawing at some other
unspeakable person in an effort to regain his equilibrium.
And then there was Bonnie May, with an insane light in her eyes.
And behind Bonnie May was a smirking creature who grinned
maliciously at Mrs. Baron, as if he and she shared some guilty secret
in common. Certainly she did not know the man.
Moreover, there stood Flora, looking unspeakably demure, with the
man Addis by her side. Addis was looking as if her arrival had
provoked him. His look seemed to say: “If you don’t like it, why don’t
you run along?”
Mrs. Baron did not stop to take in any of the others. At first she was
speechless, as the saying is, though she was trying to shape certain
comments which she meant to direct at Bonnie May.
She opened her mouth once and again quite helplessly. Then she
found her voice.
“You little—limb of Satan!” The words came with difficulty. In that
instant her features looked quite unlovely. Bonnie May might have
told her that elderly people ought never, under any circumstances, to
become violently angry. But Bonnie May was in no condition to utter
elemental truths.
“You awful little—wretch!” added Mrs. Baron. “No sooner do I turn my
back than you disgrace me! You open my door to—the whole street!”
Bonnie May was blinking rapidly. She was very pale. If you dreamed
that you were finding large sums of money, and some one threw a
bucket of cold water on you, and you woke up to find yourself in the
poorhouse—that perhaps fairly describes her mental state.
She had not been quite sorry that the bed collapsed. Some of the
secondary cells in her brain had been warning her, as she stood on
the “bridge,” that the third act could scarcely be made to come to a
true climax. She couldn’t be projected into the sea really. She would
have to step tamely down from the table and begin to talk in a
commonplace fashion.
Under favorable conditions the collapse of the bed would have been
a relief.
But now she stood looking at Mrs. Baron trying to reach her soul
through her angry eyes. She shrank so from being humiliated before
her friends—the old and the new. If Mrs. Baron, who had been so
kind in many unimportant ways and times, could only spare her now!
“If you will permit me, madam—” began Clifton.
“Who are these—gentlemen?” demanded Mrs. Baron, still wrathfully
regarding Bonnie May—Bonnie May and no other.
“They are my friends,” said Bonnie May. “They have known me
always. And really, you know, we weren’t doing anything wrong!”
Clifton had assisted her to the floor; and now, after an appealing step
in Mrs. Baron’s direction, and the swift conclusion that nothing she
could do would save the situation, she broke into tears and
staggered from the room.
“Bonnie May!” called Clifton, with overflowing solace in his tone. He
ran after Bonnie May. The other actor, casting brass rings and red
bandanna to the floor, followed.
“Emily Boone!” The voice was Mrs. Harrod’s. “I think you might
blame us, if it’s all so terrible. We encouraged her. We enjoyed it.”
Mrs. Baron now turned toward the assembled group. She seemed
dazed. “I—I didn’t know you were here!” she said, her voice
trembling weakly. And then—“I don’t care! What would any woman
do, coming home and finding strangers and—and such a scene in
her house?”
“We invited them in, mother,” confessed Baron weakly.
“Yes,” echoed Flora, “they were old companions of Bonnie May’s,
and we thought it would be nice to invite them in!”
“And I suppose you invited—him in, too?” retaliated Mrs. Baron,
indicating Addis by a scornful, slight movement of her head.
The effect of this upon Flora was most distressing. Could her mother
so far forget herself as to reveal family differences in the presence of
Mrs. Harrod and the McKelvey girls? Her wounded eyes fairly
begged for mercy.
Addis promptly came to her relief.
“No, she didn’t, Mrs. Baron. I just dropped in.” His voice, by reason
of its bigness and calmness, had the effect of making every one in
the room feel how petty and needless had been the unpleasantness
which Mrs. Baron’s arrival had created. His hair seemed more
bristling than ever as he added: “If you will permit me, I’ll bid you
good day.” He made a rather stiff bow, which was meant to include
every one in the room, and turned to go.
But here Mrs. Harrod interfered again. “Peter!” she called.
The uttering of the unfamiliar given name created profound surprise
in certain minds.
“Peter!” she repeated. “I won’t have you go away like that. I want you
to know Mrs. Baron better than you seem to know her. She doesn’t
mean half she says. Emily, tell him I’m right!” She looked
commandingly at Mrs. Baron. It was evident that she had a nature
which was not to be subdued by trivial mishaps.
Mrs. Baron flinched. “Who is Peter?” she demanded feebly.
“If you don’t know, I advise you to cultivate your son’s friends. Do
you mean that you don’t know Peter Addis? Why, he’s been like a
son of mine. You ought to have known how fond I and the colonel
are of him. I’m surprised you’ve never met him at our house.”
“I never did,” said Mrs. Baron, swallowing with difficulty.
“Well, for goodness’ sake let’s go down-stairs—please excuse me for
suggesting, Emily, in your house—and behave ourselves. I suppose
we’ve all been at fault—all except that delightful child. I’m going to
find her and tell her so!”
“It was so funny!” declared the elder Miss McKelvey, appealing
tremulously to Mrs. Baron, and patting her on the arm. She thought
of laughing, which was, she believed, the easiest thing to do in all
sorts of circumstances.
Mrs. Harrod’s brain was working energetically. She had been reading
various faces, and she realized that even yet Mrs. Baron had not
spoken to Addis. She drew conclusions. On the way down-stairs she
kept Addis close to her.
“Do you know, Peter,” she said, in large, cheerful tones, “I think it’s
downright shabby for you to neglect us as you have been of late. I
miss those old evenings so!—when you and the colonel used to
come in from hunting, and sit down and eat like two famished boys,
and bring the atmosphere of outdoors with you. Do you remember
how the dogs used to slip into the house, in spite of the colonel’s
scolding, and put their heads on your knees while you ate supper?
Those were the occasions that made a home worth having.”
Addis, entirely satisfied with the turn affairs were taking, responded
eagerly: “I certainly do remember. I’ve often wondered if the colonel
had Queenie yet. There was a dog for you!”
“Oh, no! Queenie’s been dead over a year. It’s Prince and Hector,
now—Queenie’s puppies. The colonel says they’re every bit as
smart as their mother was. I wish you’d come out soon. On a
Sunday, if you’d rather find us alone. We’ll sit out under the grape-
arbor. You know the grapes are just getting ripe. Those little vines
have grown up beautifully. The colonel always has his bottle of what-
do-you-call-it out there, and his pipe, and I send the servants away
and prepare a little lunch——”
They were in the sitting-room now, too eagerly engaged in their
conversation to think of sitting down, and Mrs. Baron was waiting
humbly to regain control of the situation.
Mrs. Harrod was not unmindful of her old friend’s discomfort; but she
had an idea she was engaged in giving a patient a dose of medicine,
and that she ought to be careful that none of it was spilled.
“If you’ll excuse me,” said Mrs. Baron, now thoroughly dejected, “I’ll
look for Bonnie May. I think I ought to have a talk with her.”
She had heard every word that Mrs. Harrod had spoken to Mr. Addis.
And she had heard enough.
She went to Bonnie May’s room. She was too confused to realize
that Flora accompanied her. But as she stood staring miserably into
the empty room she heard Flora’s comforting voice.
“She’s probably down-stairs, mother, with—with her friends.”
Flora went to the stairway and called. There was no response. She
listened, anxiously turning her eyes toward her mother; but there
was no sound of voices on the floor below.
“They wouldn’t have remained in the house a minute,” declared Mrs.
Baron, who was now frankly remorseful.
“But Bonnie May—she may have gone back to talk to Mrs. Shepard,”
suggested Flora. They could hear Mrs. Harrod’s voice, pleasantly
masterful. She had introduced Addis to the McKelvey girls, now that
she happened to think of it, and they were slipping eager gusts of
laughter and disconnected phrases into the conversation.
Mrs. Baron and Flora went down-stairs and appealed to Mrs.
Shepard.
Bonnie May had gone out, Mrs. Shepard said. She had come down-
stairs and telephoned something in great haste, and then she had
induced her two gentleman friends to go away. An automobile had
come quite promptly, and she had gone away in it.
Mrs. Baron turned away from her daughter and rested her hand
against the wall at the foot of the staircase. Her attitude spelled
repentance and fear.
She went up into the child’s room, and Flora followed close enough
to hear a low, tremulous cry of despair.
“I wouldn’t, mother!” soothed Flora, whose eager voice brought Mrs.
Harrod and the others.
Mrs. Baron was standing beside a little worktable and a chair that
were Bonnie May’s. Her face was quivering. “I’m a disagreeable old
creature,” she declared. “I don’t deserve to have any happiness.”
One hand fumbled with a handkerchief, which she lifted to her eyes.
From the other, slowly relaxing, a handful of roses and ridiculous
little silk butterflies fluttered slowly to the floor.
“I want you all to leave me—please!” she begged. “I’m not fit to be
seen.” She put forth a hand to Mrs. Harrod. “Do come back again
soon,” she begged. “And you, too,” she added, extending her hand
to the McKelvey girls. And then, as she dabbed her discolored eyes,
she concluded with—“And you, too!” She glanced aside, but her
hand went out to Addis.
Then she disappeared into her own room, and softly closed the door.
Flora’s eyes were shining as she escorted the party down-stairs.
“She’s only gone to visit friends,” she declared. “She’ll be back.”
The McKelvey girls burst from the front door ahead of the others.
They were cheerful creatures who were not to be depressed long by
the scenes they had just witnessed.
Flora, standing in the hall to let the others pass, heard them
shrieking joyously: “Oh, what a lovely new car you’ve got, Mrs.
Harrod,” and then she heard Mrs. Harrod explaining, as she
emerged into the sunlight: “A birthday present from the colonel.”
They had all passed out now except Addis, and when Flora opened
the door a little wider for him he stood still an instant and looked out.
The others were out there inspecting Mrs. Harrod’s new car.
Then he took Flora’s hand in his and closed the door firmly and
securely.
It was fully a minute before the door was opened again, and Addis
descended the steps alone.
Mrs. Harrod and the McKelvey girls forgot the new machine
immediately. They were all looking at Peter Addis. And they were all
thinking precisely the same thing, namely, that they had never in all
their lives seen a man who looked more extraordinarily handsome
and happy.
 CHAPTER XXIX
THE MANSION IN SHADOW

When Bonnie May did not return to the mansion that night the fact
was not commented upon by any member of the family. It was not
quite remarkable that she should spend the night with the
Thornburgs. That was where she had gone, of course.
It is true that Mrs. Baron was decidedly uncomfortable. The rupture
that had occurred was more serious than any that had preceded it.
Possibly she had gone too far. There was the possibility that Bonnie
May might nurse a very proper grievance and decide that it was
pleasanter to live with the Thornburgs than to continue her residence
at the mansion.
In brief, she might refuse to come back. That was Mrs. Baron’s fear.
It was a fear which hurt the more because she was unwilling to
speak of it.
However, when the next day passed and night came, Baron took no
trouble to conceal his anxiety—for still Bonnie May had not returned.
He called up the Thornburgs by telephone. Was Bonnie May there?
He asked the question very affably. Yes, came back the reply—in an
equally affable tone—she was there. Would he like to speak to her?
No, she need not be troubled; he merely wished to be sure she was
there.
Baron believed, without expressing his belief to any one, that it
would be a mistake to manifest anxiety about the late guest—or
probably the temporarily absent guest. So it came about that one
day followed another, and Bonnie May did not come back, and the
several members of the family pretended that nothing was specially
wrong.
It was Mrs. Baron who first thrust aside a wholly transparent
pretense.
“That’s the trouble with that Thornburg arrangement,” she said at
dinner one day, apropos of nothing that had been said, but rather of
what everybody was thinking. “I don’t blame her for being offended;
but if the Thornburgs were not making efforts to keep her she’d have
been back before now. On the whole, we were really very good to
her.”
“Oh, I shouldn’t worry,” declared Baron briskly. “She’ll be back. If she
doesn’t come before long I’ll go over there and—and tole her back.”
A second week passed—and she had not returned. And now her
absence was making a distinct difference in the mansion. The dinner
and sitting-room conversations became listless; or during the course
of them a tendency toward irritability was developed.
One day Mrs. Baron sought her son alone in his attic. Said she: “Do
you suppose she’s not coming back at all?” She looked quite wan
and bereft as she asked the question.
Baron felt remorseful. “Of course she is,” he assured her. “I’m going
over to the Thornburgs’. I’m going to see about it.”
Bonnie May was acting foolishly, he thought. The Thornburgs were
not keeping faith. Yet it was a difficult matter for him to make a clear
case against either Bonnie May or the Thornburgs, and he was by
no means comforted by a little event which transpired one morning.
He encountered the two actors as he was leaving the mansion, and
his impulse was to speak to them cordially. But in returning his
greeting they manifested a well-simulated faint surprise, as if they
felt sure Baron had made a mistake. They nodded politely and
vaguely and passed on.
In his mind Baron charged them angrily with being miserable cads,
and he was the more angry because they had snubbed him in such
an irreproachable fashion.
Even Baron, Sr., became impatient over the long absence of Bonnie
May. Realizing that his usual practise of watching and listening was
not to be effective in the present instance, he leaned back in his
chair at dinner one evening and asked blandly: “What’s become of
the little girl?”
And Mrs. Baron made a flat failure of her effort to be indifferent. Her
hand trembled as she adjusted her knife and fork on her plate. “Why,
I don’t know,” said she. “You know, she has two homes.” But she was
afraid to attempt to look anywhere but at her plate.
Baron was astounded by the utter dejection which his mother tried to
conceal. Why, she loved the child—really. She was grieving for her.
And that evening he emerged from the house with much grimness of
manner and made for the Thornburgs’.
The dusk had fallen when he reached the quiet street on which the
manager lived. Street-lamps cast their light among the trees at
intervals. In the distance a group of children were playing on the
pavement. Before the Thornburg home silence reigned, and no one
was visible.
Yet as Baron neared the approach to the house he paused abruptly.
He had been mistaken in believing there was no one near. In the
heavy shadow of a maple-tree some one was standing—a woman.
She was gazing at the lower windows of the Thornburg residence.
And there was something in her bearing which seemed covert,
surreptitious.
He, too, looked toward those windows. There was nothing there
beyond a frankly cheerful interior. He could see no one.
What was the woman looking at? He glanced at her again, and a
bough, swaying in the breeze, moved from its place so that the rays
from a near-by lamp shone upon the figure which appeared to be
standing on guard.
She was overdressed, Baron thought. Under an immense velvet hat
weighted down with plumes masses of blond hair were visible. Her
high, prominent cheek-bones were not at all in keeping with the
girlish bloom which had been imparted to her cheeks by a too
obvious artifice. She had caught up her skirt lightly in one hand, as if
the attitude were habitual, and one aggressively elegant shoe was
visible.
He had paused only momentarily. Now he proceeded on his way,
passing the woman in the shadow with only half the width of the
sidewalk between her and him.
He had recognized her. She was the woman who had stood in the
theatre that night talking to Thornburg—who had visited Thornburg in
his office. Could she be Miss Barry? Baron wondered.
A maid let him into the house and drew open a sliding door,
revealing the lighted but empty drawing-room. She took his card and
disappeared.
He sat for a time, counting the heavy minutes and listening intently
for sounds which did not reach him. Then the manager and his wife
entered the room, both bending upon him strangely expectant
glances.
Baron arose. “I’ve taken the liberty—” he began, but Thornburg
instantly swept all formalities aside.
“That’s all right,” he said. “Keep your seat.” Then, obviously, they
waited for something which they expected he had come to say.
But he was listening for the sound of Bonnie May’s voice. He
seemed almost absent-minded to the man and woman who were
intently regarding him.
Then Thornburg, plainly afraid of offending his guest by a too
impulsive or impatient word, fell back upon commonplaces. He
concluded that he must wait to hear what Baron had come to say.
“You’ve heard about Baggot’s good luck?” he asked.
“I think not,” replied Baron, not at all cordially.
“His play. They’re getting ready to put it on in Chicago. His people
have a theatre there that’s not engaged just now. There’s to be an
elegant production—first-class people and everything. Baggot’s gone
on to look after the rehearsals. We ought to have it here by the first
of the year—or earlier, if a number two company is organized.”
“I hadn’t heard,” said Baron. “I haven’t seen Baggot lately.” With
intention he spoke listlessly. Thornburg wasn’t coming to the point,
and he didn’t intend to be played like a fish.
An uncomfortable silence fell again, and again Baron found himself
listening intently.
And then he could bear the suspense no longer. He leaned toward
Thornburg with animation. “Look here, Thornburg,” he said, “I don’t
believe you’re playing fair!”
“You might explain that,” responded the manager curtly.
“You know what the agreement was. I don’t believe she’d stay away
like this unless she’d been restrained.”
Thornburg’s only response was a perplexed frown. It was Mrs.
Thornburg who first took in the situation. She arose, painfully
agitated, and faced Baron. “Do you mean that she isn’t at your
house?” she demanded. Her voice trailed away to a whisper, for
already she read the answer in his eyes.
Baron sank back in his chair. “She hasn’t been for weeks,” he
replied.
Thornburg sprang to his feet so energetically that the caller followed
his example. “I thought it was you who wasn’t playing fair,” he said.
And then he stared, amazed at the change in Baron’s manner.
The younger man was rushing from the room. There had come to
him unbidden the picture of the two actors who had snubbed him in
front of his house—a recollection of their studied aloofness, their
cold, skilful avoidance of an encounter with him. They had taken her!
But at the door he paused. “But I telephoned to you,” he said,
remembering. “You told me she was here.”
“She was here the day you telephoned. She went away the next
day.”
Baron frowned. “She went away—where?”
“She went in the machine. Of course we supposed——”
Thornburg hurried to the telephone and was speaking to his
chauffeur, in a moment. “Oliver? Come to the house a moment,
Oliver—and hurry.”
He replaced the receiver and hurried back to meet the chauffeur.
The soldierly appearing young chauffeur was standing at attention
before them in a moment.
“We want to know if you can remember where you took Bonnie May
the last time she left the house.”
“Perfectly, sir. She asked me to stop at the Palace Theatre. She said
she was expecting to meet a friend there. And she told me I was not
to wait—that she wouldn’t need the car again that afternoon.”

Fifteen minutes later Baron was ringing the bell of the house next to
the mansion. He couldn’t recall the two actors’ names, but he
described them. He wished to see them on urgent business.
But they had paid their bill and gone away. The woman who met
Baron at the door was sure they had said something about finishing
their engagement at the Folly and about leaving the city.
As Baron turned away from the door it seemed to him that the street
had suddenly gone empty—that the whole world was a haunted
wilderness.
 CHAPTER XXX
“THE BREAK OF DAY”

“Mr. Victor Baron, please.”

An usher with an absurdly severe uniform and a frankly cherubic
countenance had pushed aside the hangings and stood looking into
the Baron box in the Barrymore Theatre.
It was the night of the first performance of Baggot’s play, “The Break
of Day,” in Thornburg’s theatre, and the Barons were all present—by
special and urgent invitation.
Baron had been studying the aisles full of people, eagerly seeking
their seats, and listening to the continuous murmur which arose all
over the house. But when he heard his name called he arose and
slipped out into the shadows.
“Mr. Thornburg sends his compliments and asks if you’ll be good
enough to visit him in his office for a few minutes.” Thus the cherubic
usher.
The Barrymore office was off from the lobby, but it commanded a
view not only of the street but also of the procession of men and
women who passed the ticket-office.
Thornburg had left the door open, and Baron, approaching, caught
sight first of a considerable expanse of dazzling white shirt-front and
then of the manager’s ruddy, smiling countenance. Evidences of
prosperity were all about. A procession of motor-cars continued to
stop before the theatre to deposit passengers. Throughout the lobby
there was the shimmer of costly fabrics worn by women, the flashing
of jewels, the rising and falling of gusts of laughter and a chaos of
happy speech. And everywhere there was the glitter of onyx panels
and pillars, and the warmth of hooded lights, and the indefinable
odor of fine raiment and many delicate perfumes.
Thornburg seized Baron’s hand and shoved the door to with his foot.
Happiness radiated from him. “I’ve a secret to tell you,” he began. “I
want you to be one of the first to know.”
“Let’s have it!” responded Baron, trying to reflect a little of the
manager’s gayety.
“You’ll remember my telling you that I had a little daughter by my first
wife?”
“I remember.”
“I’ve found her again!”
“Ah, that’s fine!”
“And that isn’t all. You’re going to see her to-night.”
Baron waited.
“She’s the girl they’ve been making all that fuss about in Chicago—
who’s been known only as ‘The Sprite.’ She’s got the leading part in
‘The Break of Day.’”
Baron felt his way cautiously. He couldn’t mar such superb
complacency, such complete happiness. “And Mrs. Thornburg—” he
began haltingly.
“God bless her, it’s all right with her. She knows, and she’s as happy
as I am.”
Baron shrunk back with a sense of utter loss. “Thornburg,” he said, “I
want you to tell me—is the little girl the daughter of—of Miss Barry?”
The manager clapped a heavy hand on Baron’s shoulder. “No,” he
responded. And after a moment’s almost pensive reflection he
regained his buoyant manner and resumed. “I’d like you to meet her.
Between acts, or after the play. You and your family. She’s young. I
think a little attention, especially motherly attention, will mean a lot to
her just now. Of course she mustn’t be worried to-night; but suppose
we make up a little party, after the performance, and make her feel
that she’s got friends here?”
Baron couldn’t think of refusing. “I’d have time to pay my respects, at
least,” he agreed. “And I’ll put the case before my mother and the
others, just as you have stated it. I think perhaps she’ll consent.”
“That’s a good fellow. I’ll be looking for you,” concluded Thornburg,
and then he joyously shoved Baron out of the office.
The footlights were being turned on and the asbestos curtain lifted
as Baron returned to his seat. Then the orchestra began to play, and
under cover of the music Thornburg’s secret and his invitation were
passed on to Mrs. Baron and to the others in the box.
Baron did not catch his mother’s response, and she did not repeat it.
She had turned to listen to the music. For the moment the orchestra
was commanding a good deal of attention. A cycle of popular
melodies was being played, and under the spell of the singing violins
the outside world was being made to recede into the distance, while
the mimic world became real.
Men and women forgot that out on the winter streets, only a few
yards from them, there was passing that disinterested throng which
always passes the door of every theatre; the eager, the listless, the
hopeful, the discouraged, and that sprinkling of derelicts who have
no present drama at all, but who are bearing inevitably on toward the
final tragedy.
The orchestra completed the popular melodies; and after a brief
interval the leader rapped his music-rack with his baton to enjoin
attention. Then he lifted his hand as if in benediction over a player to
his left, and a wood-wind instrument announced a new theme—
penetratingly, arrestingly. Then the strains of “The Ride of the
Valkyries,” with their strident and compelling quality, filled the theatre.
Baron was startled by the touch of a hand on his shoulder. Baggot
was leaning toward him. “That’s to create the right atmosphere,” he
whispered, nodding toward the orchestra. “It’s to put the idea of the
supernatural into everybody’s mind, you know.” He withdrew then.
Baron thought that was just like Baggot—to be explaining and
asserting himself, as if he were doing it all. He was glad to be rid of
him. He wanted to feel, not to think. Then he realized that the