librationist

cum classical
theories of sets

draft: comments appropriate

Frode Alfson Bjørdal
............................
Copyright © 2024 Frode Alfson Bjørdal All rights reserved.
ii
iii
iv
 To my godson, Jon LW, 2/11/2012,
& my dear friend, nephew Jon VL, 2/17/1979 – 7/22/2011.

v
vi
Contents
1 Preface 1

2 Acknowledgements 3

3 Introduction 5

4 Librationism and its formal language 7

4.1 Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
4.2 Numeralism - the chiffer standpoint . . . . . . . . . . . . . . . . . . . . . 7
4.3 The inclusion of abstracts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.4 “=” and “∈” are not primitive in £ . . . . . . . . . . . . . . . . . . . . . . . 8
4.5 Metalinguistic conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

5 Semantics 13
5.1 On expression names, and their extension . . . . . . . . . . . . . . . . . . 13
5.2 Fairs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.3 Closure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.4 The range of librationist satisfaction . . . . . . . . . . . . . . . . . . . . . . 16
5.5 Relations between maxims, optima, plains and minors . . . . . . . . . . . 17
5.6 Orthodoxy, definiteness and paradoxicality . . . . . . . . . . . . . . . . . . 18
5.7 The non-triviality assumptions . . . . . . . . . . . . . . . . . . . . . . . . . 19

6 Maxims 21
6.1 Axioms & warrants, theorems & proofs. . . . . . . . . . . . . . . . . . . 21
6.2 The relations between the varieties of theses . . . . . . . . . . . . . . . . . 21
6.3 Arbitration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
6.4 Logic maxims . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
6.5 Maxims on truth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
6.6 Warrants of truth maxims . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

7 Optima 27

8 Plains 29

9 Inference modes 31
9.1 The simple inference modes . . . . . . . . . . . . . . . . . . . . . . . . . . 31
9.2 Involved inference modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

10 Alethic comprehension 35

11 The theory of identity 37

11.1 Co-essentiality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
11.2 The adequacy of identity as co-essentiality . . . . . . . . . . . . . . . . . . 38
CONTENTS CONTENTS

12 Alphabetologicality 41

13 Arithmetic 43

14 Shortcomings and redresses 46

14.1 Shortcoming related to existential instantiation . . . . . . . . . . . . . . . 46
14.2 An orthodox redress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
14.3 The Barcan failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
14.4 £ is omega-consistent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
14.5 More orthodox redresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

15 Classicalities and deviations 52

15.1 Facts on desiderata met by £ . . . . . . . . . . . . . . . . . . . . . . . . . . 52
15.2 The truth–conditionals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
15.3 £ is classic and paraclassical, but it is not paraconsistent . . . . . . . . . 53
15.4 Incompatability and complementarity . . . . . . . . . . . . . . . . . . . . 55

16 The Liar is Russell’s condition on his set 57

17 Librationist incompleteness phenomena 59

18 The reflective theory of comprehension 61

18.1 Responsible naiveté without revenge . . . . . . . . . . . . . . . . . . . . . 61
18.2 Argumenta ad paradoxo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

19 Manifestations 63
19.1 The manifestation sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
19.2 Quine atoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
19.3 The autocombatant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

20 Powersets are paradoxical lest as P ({𝑣|𝑣 = 𝑣}) 67

21 Non-extensionality and Ursets 69

22 Names and sets of Urelemente to transfinite orders 71

23 Heritors and regulars 73

24 Choice, power, potency and countability 75

24.1 The denumerable wellordering . . . . . . . . . . . . . . . . . . . . . . . . . 75
24.2 Function application notation . . . . . . . . . . . . . . . . . . . . . . . . . 75
24.3 The choice function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
24.4 The enumerator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
24.5 The enumeration postulates . . . . . . . . . . . . . . . . . . . . . . . . . . 77
24.6 Absolute countability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
24.6.1 e restricted to P ({𝑥|𝑥 ∈ 𝜔}) . . . . . . . . . . . . . . . . . . . . . . 78
24.6.2 e restricted to P(𝜔) . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

viii
CONTENTS CONTENTS

25 £ and the theories of vonsets 83

25.1 The potency vonset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
25.2 The Grothendieck vonset of w relative to v . . . . . . . . . . . . . . . . . . 84
25.3 Capture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
25.4 V and D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
25.5 Primitive theorems for classes . . . . . . . . . . . . . . . . . . . . . . . . . 88
25.6 The Tuple-lemmas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
25.7 The class existence theorem . . . . . . . . . . . . . . . . . . . . . . . . . . 89
25.8 The expansion lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
25.9 Proof that V is orthodox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
25.10 Proof that all members of V are orthodox . . . . . . . . . . . . . . . . . . 89
25.11 Global well ordering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

26 Space for librationist category theory? 91

27 The theory of vansets NF in £H(W) 93

28 𝐴 is true just if 𝐴 states the truth 97

Bibliography 97

ix
CONTENTS CONTENTS

x
 1 PREFACE

1 Preface

One cannot see it all from one point of view.

The investigations, which led to the librationist set theories, began in the spring of 1993.
It was a struggle to build upon imprecise thoughts to express beliefs precise enough
to be useful. But the author pressed on, as the investigations were very addictive, and
began giving talks already in 1996, and publishing unfinished ideas already in 1997, with
(Bjørdal 1998). Motivation to persist with the work was always found, so the hard work
continued over all those years, in between other rather taxing tasks.
The problems with the paradoxes in set theory are of great importance not only when
one attempts to find useful and philosophically reasonable foundational theories for the
formal sciences, but also when one seeks to account for a variety of important problems
in metaphysics, epistemology and other areas of philosophy.
Paradoxes, and related Cantorian foundational points of view, are especially important
in metaphysics. As (Grim 1991) stressed, and elaborated upon, standard thinking has
the awkward consequence that the world is not complete. As we shall see, librationism
challenges that dogma, as it has the consequence that there are only denumerably many
objects in the world; importantly, the validity of Cantor’s arguments for uncountability is
not challenged.
We can see the more comprehensive relevance of the paradoxes rather directly from the
fact that modal logics are very important philosophical tools for reasoning about ethics,
knowledge, and other central philosophical concepts. But threats of paradox undermine
the use of reasonable modal logics, with more than just a minimum of linguistic resources
and plausible closure principles, in order to reason about ethics and central philosophical
concepts. This is on account of such limitative results as were discovered by (Montague
1963), and investigated further by others, like (Friedman and Sheard 1987), and (Cantini
1996). A takeaway is for example that if a modal logic is expressive enough to license
the inference from Smith ought to sell his house to there is something Smith ought to do,
then Russell like paradoxes arise.
So paradoxes are pervasive in philosophy. The standard way to evade Montague like
limits is to put restrictions upon the linguistic resources. Librationist resolutions recom-
mend themselves, as such restrictions are not called for.
The focus in this essay will be upon the paradoxes in the context of mathematics, and
thus in theories of sets, and related theories, as category theory. A central result is
that the librationist set theoric extension £HR(D) of £ accounts for Neumann-Bernays-
Gödel set theory with the Axiom of Choice and Tarski’s Axiom. Moreover, £ succeeds
with defining an impredicative manifestation set W, die Welt, so that £H(W) accounts
for Quine’s New Foundations. Nevertheless, the points of view developed support the
view that the truth-paradoxes and the set-paradoxes have common origins, so that the
librationist resolutions of the set theoretic paradoxes are at the same time resolutions

1
 1 PREFACE

of the truth theoretic paradoxes. Both the librationist resolutions of the set theoretic
paradoxes and the truth theoretic paradoxes have non-trivial philosophical implications:
librationist set theories have the consequence that there are no absolutely uncountable
sets, and librationist truth theories allow the use of syntactical modalities in ways which
circumvent (Montague 1963).

2
 2 ACKNOWLEDGEMENTS

2 Acknowledgements

It was the best of times, it was the worst of times.

Charles Dickens

Several lectures on librationism were given, and essays published, over the years, and
the author often worried that too much material had seen the light of day.
The first lecture was Truth in Perspectives, for the Scandinavian Logic Symposium in Up-
psala, Sweden, in 1996. The next was Towards a Foundation for Type-Free Reasoning, in
Villa Lanna, Prague, published as an essay in (Bjørdal 1998), fifteen years before (Bjørdal
2012) which introduced the neologism librationism. (Bjørdal 2005), (Bjørdal 2006) and
(Bjørdal 2011) were published in the last interval.
During some semesters as from 2013 a number of seminaries were held, on a variety
of topics, for master and doctorate students of Philosophy, as Professor Colaborador
Voluntário, at Programa de Pós-Graduação em Filosofia na Universidade Federal do Rio Grande
do Norte, in Natal, Brasil: In the second semester of 2013 the seminary The librationist
Foundation for Reasoning was offered, and in the second semester of 2014 the seminary
A teoria libracionista das coisas.
Talks were held on the topic for annual LOGICA congresses, under the auspices of the
Czech Academy of Sciences, in the Czech republic, in 1997, 2004, 2005 and 2010. Asso-
ciated papers were published in the LOGICA Yearbook series as (Bjørdal 1998), (Bjørdal
2005), (Bjørdal 2006) and (Bjørdal 2011).
Other lectures were held for the Logic Colloquia, under the auspices of the Association
of Symbolic Logic: in Barcelona, 2011 – Helsinki, 2014 – Stockholm, 2017 and Prague,
2019; for World Congresses on Paraconsistent Logic: Melbourne, 2008 and Kolkata,
2014; for International Conferences on Non-classical Mathematics: Guangzhou, 2011
and Vienna, 2014; for Sociedade Brasileira de Lógica: Petrópolis, 2014 – Pirenópolis,
2017 and Salvador, 2022; for World Congresses on Logic, Methodology, and Philosophy
of Science: Helsinki 2014 and Prague, 2019; and for the World Congresses on Universal
Logic in Lisbon, 2010, Rio de Janeiro, 2013 and Istanbul, 2015. The Kolkata lecture was
reworked and published as (Bjørdal 2015).
Lectures were as well delivered for the conference on Logic, Reasoning and Rationality
at Centre for Logic & Philosophy of Science, Ghent University, Belgium, in 2010; for the
Moscow division of the Russian Academy of Sciences’s Steklov Mathematical Institute,
in 2014, and the Euler International Mathematical Institute at the Saint Petersburgh
division of Steklov Mathematical Institute, in 2015.
Most importantly, I delivered many lectures for the Seminary in Logic under the auspices
of the Mathematics Department at the University of Oslo, where I have had the honor
to talk about these matters since I started as Associate Professor in Philosophy there
in 1996. Those seminaries were, and are very useful, and I have learned a lot from
participating there with colleagues and advanced students

3
 2 ACKNOWLEDGEMENTS

All encounters were important for the mathematical and philosophical maturation of the
author, and resistance was usually useful. No one shares responsibility for undetected
errors, so the author will not now risk the names of others by thanking them especially.

4
 3 INTRODUCTION

3 Introduction

Nur wenn man nicht auf den Nutzen nach aussen sieht, sondern in der Mathematik selbst
auf das Verhältnis der unbenutzten Teile, bemerkt man das andere und eigentliche
Gesicht dieser Wissenschaft. Es ist nicht zweckbedacht, sondern unökonomisch und
leidenschaftlich. [. . . ] Die Mathematik ist Tapferkeitsluxus der reinen Ratio, einer der
wenigen die es heute gibt.

Robert Musil, in Der mathematische Mensch,

Mitteilungen der Deutschen Mathematiker-Vereinigung, Noº20, page 50, 1912.

It is presupposed that 𝐴 ∧ ¬𝐴 is a contradiction, and that a theory is inconsistent just

if it has contradictory theses. As per §15, £ is consistent and not contradictory. So the
librationist points of view are not dialetheic, for dialetheism is canonically characterized,
in (Priest, Berto, and Weber 2022), as a view which takes some contradictions to be true.
Moreover, £ is not a paraconsistent point of view, as the latter are not conservative in the
sense of Definition 15.3.13. Librationism, per Definition 15.3.15, may instead be taken
to offer an extraclassical and extracoherent point of view. To distinguish further, take
librationism to offer a bialethic point of view, and not a dialetheic one.
It will be showns in 28 that Librationism meets a challenge which it is difficult to see can
be met if one presupposes that contradictions, as 𝑝 ∧¬𝑝, are true, viz. to offer an account
of what a true sentence 𝑝 says, in a paradoxical situation, which its true negation ¬𝑝 does
not say in that situation.
A remark on designator is appropriate. One might hold that a theory is not a set theory
if it presupposes more linguistic resources than the language of set theory, understood
as first order logic plus the symbol ∈. This tenet is not followed here, and it is instead
presupposed that set theoretic reality should be investigated with such rescources which
best reveal it. As will beceome clear, we make use of set abstracts, and these are not elim-
inable, due to the fact that £ is a highly non-extensional theory. The symbol ∈, however,
is eliminable, by means of apposition.
As £ with additional assumptions interprets classical set theory, with global choice, it
seems misleading to hold that £ is not itself a set theory.

5
 3 INTRODUCTION

6
 4 LIBRATIONISM AND ITS FORMAL LANGUAGE

4 Librationism and its formal language

Die ganzen Zahlen hat der liebe Gott

gemacht, alles andere ist Menschenwerk.
Leopold Kronecker

4.1 Nomenclature
(Bjørdal 2012, p. 323) states "Librationism takes its name from the word “libration”, which
the reader is asked to look up if unfamiliar." Lunar librations were an inspiration.
After the publication of (Bjørdal 2012), £ was used for the librationist foundational sys-
tem. It will be indicated, as in §§25–27, when additional assumptions are made.
The pound sign £ is most prevalently used for the currency of Great Britain. It derives
from Latin libra pondo. Pondo is an adverb which means by weight. Libra was used for
the Roman pound - which was about 327 g, but also for scales and balances. Such scales
were an attribute to the Greek Goddess for Divine Justice, Themis, and for her daughter
Dike, who was the Goddess for Human justice. The roles of the attributes were thought
to be the weighing of the consequences of acts to find balance, and, therefore, justice.
The Goddess corresponding to Themis and Dike in the Roman religion was the blinded
Justitia, who also had a scale as attribute.
In the context of librationism, £ may be taken to symbolize the weighing and gauging of
balances between sentences, and perhaps most interestingly, from the librationst points
of view, in the case of sentences which are incompatible or complementary, in the sense
of Definitions 15.4.1 and 15.4.3.

4.2 Numeralism - the chiffer standpoint

The chiffers are the numbers-of-the-meta-language. The ordinal chiffers are defined à la
von Neumann by means of the meta mathematical variety theory, which one may take to
be the set-theory-of-the-meta-language. One must carefully distinguish ordinal chiffers
from corresponding ordinal numbers of the set theories expressed, and accounted for,
by the object language. The ciphers are numerals-of-the-meta-language, denoting finite
order chiffers. In the case of finite chiffers we underline the denoting cipher to contrast
with numerals of the set theories accounted for in the object language. So 0 e.g. denotes
the ordinal chiffer Zero. The natural chiffers are the finite ordinal chiffers, and the variety
of counting chiffers is the variety of the natural chiffers minus 0. The integer chiffers are
the natural chiffers extended with their negative counterparts.
The chiffer standpoint presupposed here is stronger than the point of view presupposed
by (Gödel 1931), which was that formulas, and expressions akin, may be correlated via
a coding with numerals denoting natural numbers. For the symbols and expressions of £
are taken to be counting chiffers, and their syntactical manipulations are accounted for
by the variety theory presupposed.

7
4.3 The inclusion of abstracts 4 LIBRATIONISM AND ITS FORMAL LANGUAGE

4.3 The inclusion of abstracts

The inclusion of abstracts is a trait shared with (Gandy, 1959), and with contributions to
the literature on non-classical set theories, including some which were at the time called
property theories1, as e.g. (Gilmore, 1974), and theories discussed by (Cantini, 1996),
and others, where abstracts were used because the principle of extensionality fails.

4.4 “=” and “∈” are not primitive in £

The formal language of £ is Polish, and without symbols for identity or membership.
A Polish ↓-connective is used, as per Definition 4.5.4.3. The membership relation can be
defined by means of apposition of terms, because there as a consequence of the Polish
policy are no parentheses in the formal language of £.
§11 shows that the identity of 𝑎 and 𝑏 can be delineated adequately by the statement
that 𝑏 is an element of all sets that have 𝑎 as an element, as in Definition 11.1.2.

4.5 Metalinguistic conventions

4.5.1. Definition Symbols of the meta language:
(1) Σ is the existential quantifier.
(2) Π is the universal quantifier.
(3) ∼ is negation.
(4) & is conjunction.
(5) r is disjunction.
(6) ⇒ is for implication.
(7) ⇔ is for bi-implication.
(8) [𝑥 : . . . ] is the set notation for use in the metalanguage.
(9) 𝜀 is the metalanguage symbol for membership.

1It seems that the term "property theory", despite seemingly having an origin with Kurt Gödel, became
unfortunate. The opening sentence of Roger Myhill’s article Paradoxes, in Synthese 60 (1984), 129-143,
is: “Gödel said to me more than once "There never were any set-theoretic paradoxes, but the property-
theoretic paradoxes are still unresolved"; and he may well have said the same thing in print.”
Remarks as this may have had such influence that some authors later used the term "property-theory",
for non-extensional set theories, which attempt to give more type-free accounts that approximate naive
abstraction in dealing with the paradoxes.
Nevertheless, there are now so many non-extensional set theories in the literature, beyond attempts to
deal with the paradoxes, that it seems unreasonable to consider them property theoretic, as opposed to
set theoretic.
Was Gödel aware of the contribution in (Scott 1961), or did he study (Friedman 1973). (Shapiro 1985)
is another witness to modern research into set theories without extensionality.

8
4.5 Metalinguistic conventions 4 LIBRATIONISM AND ITS FORMAL LANGUAGE

4.5.2. Definition (Other metamathematical symbols)

(1) === is the metamathematical identification and definition symbol.
(2) 𝛼, 𝛽, 𝛾, 𝛿, . . . are for ordinal numbers of the metalanguage.
(3) ≺, ⪯, ⪰, and ≻ are the orderings on the ordinal numbers of the meta language.
(4) 𝜇 is for the least operator of the metalanguage.

4.5.3. Definition (The finite order chiffers, and their integers)

(1) Ω === 0, 1, 2 . . . is the term for the finite order chiffers, i.e. the natural chiffers.
(2) Ω+ === 1, 2, 3... is the term for the positive integers, i.e. the counting chiffers.
(3) Ω− === −1, −2, −3 . . . is the term for the negative integers.
(4) Ω± === 0, 1, −1, 2, −2 . . . is the term for the integer chiffers.
(5) Ω− === 0, −1, −2 . . . is the term for tha nonpositive integers.

4.5.4. Definition (The symbols, their ciphers and chiffers)

(1) •
(2) v̈
(3) ↓
(4) ∀
(5) 𝝇
(6) c̈
(7) #
are the symbols, which stand for the chiffers denoted by the bijective base-2 ciphers
1, 21, 221, 2221, 22221, 222221 and 2222221, respectively.
4.5.5. Definition (Bijective base-2 cipher strings)
(1) Let 𝑛0 , 𝑛1 , 𝑛2 , 𝑛3 , . . . be base-2 cipher strings.
(2) ℓ(𝑛0 ) === ⌊𝑙𝑜𝑔2 (𝑛0 + 1)⌋ invokes the floor function ⌊ ⌋, and defines the length of the
bijective base-2 cipher needed to express chiffer 𝑛0 .
(3) Concatenation ⌢ is the function given by 𝑛 ⌢ 𝑛 === 𝑛 · 2ℓ(𝑛1 ) + 𝑛 .
0 1 0 1

(4) We know that ⌢, so defined, is associative.

(5) 𝑛0 ⌢ 𝑛1 is taken to be the denotatum of the apposition 𝑛0 𝑛1 .
(6) Just the ciphers 1 and 2 are the cipher strings of length 1.

9
4.5 Metalinguistic conventions 4 LIBRATIONISM AND ITS FORMAL LANGUAGE

(7) If 𝜎0 is a cipher string of length 𝑛 and cipher string 𝜎1 has length 1, then cipher
string 𝜎0 𝜎1 = 𝜎0 ⌢ 𝜎1 has length 𝑛 + 1.
4.5.6. Exercise: 𝜎 is a bijective base 2 cipher string just if it is a bijective base-2 cipher.
4.5.7. Definition (Expressionforms)
(1) An expression E is in symbolic form just if it is a string of symbols from Definition
4.5.4.1–7, formed according to the formation rules in §§4.5.9-4.5.11.
(2) Expression E is in cipher form just if it is a bijective base-2 cipher, which corresponds
with the symbolic form of E via coding of symbols into ciphers as in Definition 4.5.4.
(3) Expression E is in a presentable form just if Definitions as 4.5.12, 4.5.18 and 4.5.20
are abided by to such an extent that the expression is more understandable.
So the notion of presentable form is not precise.
Notice that presentable forms may lose information, as it may be translated back
again to various cipherforms, or symbolic forms; but we will assume that such
translation problems do not arise.
4.5.8. Definition (The underlines)To remind that expressions in the last analysis are
chiffers, denoted by ciphers, we in the remainder of this section underline, and
write variable, term, formula, sentence, constant, and so on. To ease the reading,
the underlines will not be used as from the next section.
4.5.9. Definition (Variables)
(1) v̈ is a variable.
(2) A variable succeeded by • is a variable.
(3) 𝑣 is variable v̈, and 𝑣
0 is variable v̈ ⌢ •.
𝑛+1 𝑛

(4) Nothing else is a variable.

(5) Variables are terms.

4.5.10. Definition (Primitive constants)

(1) 𝑐¥ is a primitive constant.
(2) A primitive constant succeeded by • is a primitive constant.
(3) 𝑐 is constant c̈, and 𝑐
0 is constant c̈ ⌢ •.
𝑛+1 𝑛

(4) Nothing else is a primitive constant.

(5) Primitive constants are terms without free variables, and so, per 4.5.15, constants.

4.5.11. Definition 𝑎𝑖 for arbitrary terms and A𝑖 for arbitrary formulas:

(1) If 𝑎 0 and 𝑎 1 are terms, 𝑎 1 𝑎 0 is a formula.

10
4.5 Metalinguistic conventions 4 LIBRATIONISM AND ITS FORMAL LANGUAGE

(2) If A0 and A1 are formulas, ↓A0 A1 is a formula.

(3) If A0 is a formula and 𝑣 0 is a variable, ∀𝑣 0 A0 is a formula.
(4) If A0 is a formula and 𝑣 0 is a variable, 𝝇𝑣 0 A0 is a term.
(5) Nothing else is a term or a formula.
(6) Just terms and formulas are expressions.

4.5.12. Definition Suppressing subscripts: When possible, 𝑎, 𝑏, 𝑐, . . . are written for 𝑎 0 , 𝑎 1 , 𝑎 2 , . . .,

while 𝑣, 𝑤, 𝑥, . . . are for 𝑣 0 , 𝑣 1 , 𝑣 2 , . . . , and 𝑚, 𝑛, 𝑜, . . . for 𝑛0 , 𝑛1 , 𝑛2 , . . . , along with A, B, C . . .
instead of A with subscripts. Other letters, or letter-like symbols, may be used for special
terms, or formulas.

4.5.13. Definition Binders, binds, ties and scopes:

(1) In ∀𝑣A, ∀ is the binder. 𝑣 is the bind of A and the tie of ∀. A is the scope of ∀.
(2) In 𝝇𝑣A, 𝝇 is the binder. 𝑣 is the bind of A and the tie of 𝝇. A is the scope of 𝝇.

4.5.14. Definition Free and bound variables:

(1) A variable occurrence in a formula, or term, is bound, just if it is a bind, or it is in
the scope of a binder with another occurrence as tie.
(2) Variable occurrences in a formula, or term, are free if not bound.
(3) A variable is free in a formula, or term, just if an occurrence is.
(4) A variable is bound in a formula, or term, just if an occurrence is.

4.5.15. Definition Sentences and constants:

(1) A term without free variables is a constant.
(2) A formula without free variables is a sentence.

4.5.16. Definition Substitution: If E is an expression, 𝑎 is a term and 𝑣 is a variable, E𝑣𝑎

is the expression obtained by substituting all free occurrences of 𝑣 in E with term 𝑎.

4.5.17. Definition Substitutability: Term 𝑎 is substitutable for variable 𝑣 in A just if, A is

atomic, or A is ↑ 𝐵𝐶 and 𝑎 is substituable for 𝑣 in both 𝐵 and 𝐶, or 𝐴 is ∀𝑤𝐵 and 𝑣 is not
free in 𝐵, or, 𝑤 does not occur in 𝑎 and 𝑎 is substitutable for 𝑣 in 𝐵.
4.5.18. Definition Postfixed variable vector notation:

E(𝑣, 𝑤, 𝑥) signifies that variables 𝑣, 𝑤 and 𝑥 are free in E.

4.5.19. Presentation resolve: A(𝑣, 𝑤, 𝑎) may be written for A(𝑣, 𝑤, 𝑥)𝑥𝑎 .

4.5.20. Definition Prefixed variable vector notation: Occasionally ∀𝔳®A is used for a sentence
which either is A, or for some 𝑛 > 0 and variables v0 . . . v𝑛−1 , ∀®𝑣 A is ∀v0 . . . ∀v𝑛−1 A.

11
4.5 Metalinguistic conventions 4 LIBRATIONISM AND ITS FORMAL LANGUAGE

4.5.21. Definition Parentheses, and defined operators for the object language:
(1) Delimiters for punctuation: (, ), [, ], . . .
(2) ¬A ===↓ AA
(3) (A ∧ B) ===↓ ¬A¬B
(4) (A ∨ B) === ¬ ↓ AB
(5) (A → B) === (¬A ∨ B)
(6) (A ↔ B) === (A → B) ∧ (B → A)
(7) ∃𝑣A === ¬∀𝑣¬A
(8) 𝑎 ∈ 𝑏 === 𝑏𝑎
(9) {𝑣|A} === 𝝇𝑣A

4.5.22. Definition Notation for binders restricted to set 𝑏:

(1) 𝐴 𝑏 and 𝑎 𝑏 signifiy that all variables bound in 𝐴 and 𝑎 are restricted to 𝑏.
(2) 𝑣 𝑏 is 𝑣.
(3) (𝑐 ∈ 𝑑) 𝑏 is 𝑐 𝑏 ∈ 𝑑 𝑏 .
(4) ¬𝐴 𝑏 is ¬( 𝐴 𝑏 ), ( 𝐴 ∧ 𝐵) 𝑏 is ( 𝐴 𝑏 ∧ 𝐵 𝑏 ), and so on for other connectives.
(5) {𝑣| 𝐴} 𝑏 === {𝑣|𝑣 ∈ 𝑏 ∧ 𝐴 𝑏 }.
(6) (∀𝑣) 𝐴 𝑏 === (∀𝑣)(𝑣 ∈ 𝑏 → 𝐴 𝑏 ).
(7) (∀®𝑣 ) 𝐴 𝑏 === is the sentence given by the least 𝑛 ≥ 0 such that

 
𝑛 > 0 & (∀𝑣 0 . . . ∀𝑣 𝑛−1 )(𝑣 0 ∈ 𝑏 ∧ . . . ∧ 𝑣 𝑛−1 ∈ 𝑏 → 𝐴 𝑏 )

 
𝑛 = 0 & 𝐴𝑏 .

12
 5 SEMANTICS

5 Semantics

Development of mathematics resembles a fast revolution of a

wheel: sprinkles of water are flying in all directions. Fashion –
it is the stream that leaves the main trajectory in the tangential
direction. The streams of epigone works draw most attention,
and they constitute the main mass, but inevitably disappear
after a while because they parted with the wheel. To remain
on the wheel, it is necessary to apply the effort in the direction
perpendicular to the main stream.

Vladimir Igorevich Arnold

The underlying theory of the meta language is variety theory Σ3 KPΩ, which is Kripke-
Platek variety theory, with Σ3 -collection, and the variety Ω of natural chiffers. Care should
be taken to not confuse the varieties of the meta language used to introduce £ with the
sets £ postulates the existence of.

5.1 On expression names, and their extension

5.1.1. Definition Expression names:
(1) If E is an expression, then E is its name.
(2) The semantic values of expression names is accounted for in §22.

5.2 Fairs
5.2.1. Definition: ⊩ is a function from initial sets Ξ, Ξ′, Ξ′′, . . . of formulas, and ordinals,
to sets of formulas. For any ordinal 𝛼, and formula A, we write (Ξ, 𝛼) ⊩ A for A𝜀 ⊩ (Ξ, 𝛼).
5.2.2. Definition Fairs: (Ξ, 𝛼) is fair, or a fair, just if for all formulas A and B:
(1) (Ξ, 𝛼) ⊩↓ AB just if neither (Ξ, 𝛼) ⊩ A nor (Ξ, 𝛼) ⊩ B.
(2) (Ξ, 𝛼) ⊩ ∀𝑣A(𝑣) just if (Ξ, 𝛼) ⊩ A𝑣𝑏 for all 𝑏 substitutable for 𝑣 in A .
(3) 𝛼 = 0 ⇒ ⊩ (Ξ, 𝛼) = Ξ, so (Ξ, 0) ⊩ A ⇔ A𝜀Ξ, and (Ξ, 0) ⊩ ¬A ⇔ A
𝜀Ξ.
(4) 𝛼 ≻ 0 ⇒ ((Ξ, 𝛼) ⊩ T A ⇔ Σ𝛾(𝛾 ≺ 𝛼 & Π𝛿(𝛾 ⪯ 𝛿 ≺ 𝛼 ⇒ (Ξ, 𝛿) ⊩ A))).
(5) If (Ξ, 𝛼) ⊩ T 𝑥 then ΣA(Formula(A) & 𝑥 = A ).
(6) ΠΞ[(Ξ, 𝛼) ⊩ T A ∨ T ¬A ] r


ΠΞ[(Ξ, 𝛼) ⊩ ¬T ¬A ] ⇔ ΠΞ[(Ξ, 𝛼) ⊩ ¬T A ] .

5.2.3. Remark: Diagonalization in £ does not come about as diagonalization in Peano

arithmetic, nor as in the modal provability logic GL with precisely □(□𝑝 → 𝑝) → □𝑝) as
characteristic axiom. For such reasons a symbolization E distinct from ⌜E⌝ is used for

13
5.3 Closure 5 SEMANTICS

the name of expression E. The diagonalization construction in £ allows quantifying into

named contexts, so that ∀𝑥∃𝑦T 𝑥 = 𝑦 is a well formed sentence.
5.2.4. Presentation resolve:

When the context allows, we write T A as an abbreviation for T A .

5.2.5. Remark: If formula A has one variable free, there may be different fairs Ξ and
Ξ ′ such that (Ξ, 0) ⊩ A and (Ξ ′, 0) ⊩ ¬A. So quantifying over fairs, as in §5.4, is like
quantifying over interpretations to define tautologicality of formulas in logics.
5.2.6. Theorem (Omega standard)

(Ξ, 𝛼) ⊩ ∃𝑣A ⇔ (Ξ, 𝛼) ⊩ A𝑣𝑏 for some 𝑏 substitutable for 𝑣 in A.

Proof:
(Ξ, 𝛼) ⊩ ∃𝑣A
⇕
(Ξ, 𝛼) ⊩ ¬∀𝑣¬A
5.2.2.1 ⇕
(Ξ, 𝛼) ⊮ ∀𝑣¬A
5.2.2.2 ⇕
for a 𝑏 substitutable for 𝑣 in A, (Ξ, 𝛼) ⊮ ¬A𝑣𝑏
5.2.2.1 ⇕
for a 𝑏 substitutable for 𝑣 in A, (Ξ, 𝛼) ⊩ A𝑣𝑏
□

5.3 Closure

5.3.1. Definition Cover, stabilization and closure:

(1) IN(𝛼, A, Ξ) just if Π𝛽(𝛼 ⪯ 𝛽 ⇒ (Ξ, 𝛽) ⊩ T A).
(2) OUT(𝛼, A, Ξ) just if Π𝛽(𝛼 ⪯ 𝛽 ⇒ (Ξ, 𝛽)
⊩ T A).
(3) IN(A, Ξ) just if Σ𝛼IN(𝛼, A, Ξ).
(4) OUT(A, Ξ) just if Σ𝛼OUT(𝛼, A, Ξ).
(5) STAB(A, Ξ) just if IN(A, Ξ) r OUT(A, Ξ).
(6) UNSTAB(A, Ξ) just if ∼ STAB(A, Ξ).
(7) 𝛼 covers Ξ just if : IN(A, Ξ) ⇒ IN(𝛼, A, Ξ).

14
5.3 Closure 5 SEMANTICS

(8) 𝛼 stabilizes Ξ just if 𝛼 covers Ξ, and (Ξ, 𝛼) ⊩ T A ⇒ IN(A, Ξ).

(9) The closure ordinal Ϙ is the least stabilizing ordinal.

5.3.2. Theorem (Herzberger 1980)

There is a closure ordinal.

Proof: Assume first that IN(A, Ξ), to presuppose

5.3.3. Definition: h(A) = 𝜇𝛼(IN(𝛼, A, Ξ)).

1. We first show that there is a covering ordinal:

We have

ΠA(IN(A, Ξ) ⇒ Σ𝛽(𝛽 = h(A))). (5.3.4)

So

ΠAΣ𝛽(IN(A, Ξ) ⇒ 𝛽 = h(A)). (5.3.5)

Π2 –collection and quantifier rules give us

ΠBΣYΠA(A𝜀B ⇒ Σ𝛽(𝛽𝜀Y&(𝛽 = h(A)))). (5.3.6)

Instantiate with B = [A : IN(A, Ξ)] to obtain

ΣYΠA(IN(A, Ξ) ⇒ Σ𝛽(𝛽𝜀Y&(𝛽 = h(A)). (5.3.7)

Let Z be a witness for (5.3.7), and define the least covering ordinal by means
of Π2 –separation,

𝜘 = [𝜈 : 𝜈𝜀Z & Ordinal(𝜈) & ΣA(IN(A, Ξ) & 𝜈 = h(A))]. (5.3.8)

Π2 -collection was invoked in the step from (5.3.5) to (5.3.6), and as Π𝑛 -collection
implies Σ𝑛+1 collection for Kripke–Platek theories, this justifies the choice of an un-
derlying variety theory at least as strong as Σ3 KPΩ for the meta language.2

2(Welch 2011) shows that KP + Σ3 -Determinacy is sufficient for the semantics for a commensurate
system AQI (Arithmetical Quasi Induction) introduced in (Burgess 1986), and (Hachtman 2019) shows
this equivalent to KP + Π21 -Monotone Induction. So a Σ3 -admissible ordinal is not necessary, but it may be
needed for the proof we use, which connects the coding of the formal language with the natural chiffers
of the meta theory. Welch has pointed out in private communication that a Σ2 -admissible ordinal, without
further assumptions, can be proven to be insufficient.

15
5.4 The range of librationist satisfaction 5 SEMANTICS

2. We next prove that there is a stabilizing ordinal:

Let [ 𝑓 (𝑛) : 𝑛𝜀Ω], by an adaptation of Cantor’s pairing function, be an enumeration

of all elements of UNSTAB(Ξ), where each element recurs infinitely often, so that if
B= 𝑓 (𝑚) and 𝑚 ≺ 𝑛𝜀Ω, then there is a natural number 𝑜, 𝑛 ≺ 𝑜𝜀Ω, such that 𝑓 (𝑜) = B.
Let 𝑔(0) = 𝜘 and 𝑔(𝑛 + 1) = the least 𝜈 > 𝑔(𝑛) such that

(Ξ, 𝜈) ⊩ 𝑓 (𝑛) ⇔ (Ξ, 𝑔(𝑛))

⊩ 𝑓 (𝑛)

Let ß = [𝛾 : Σ𝑚Σ𝜈(𝑚𝜀Ω & 𝜈 = 𝑔(𝑚) & 𝛾𝜀𝜈)]. It is obvious that ß is a limit ordinal
which covers Ξ. It is also clear that if 𝑚 ≺ 𝑛𝜀Ω then 𝑔(𝑚) ≺ 𝑔(𝑛). Since ß covers Ξ,
it suffices to show that (Ξ, ß) ⊩ T B entails that B is in STAB(Ξ), to establish that ß
stabilizes Ξ.
Suppose (Ξ, ß) ⊩ T B. It follows that

a) Σ𝜈Π𝜉 (𝜈 ⪯ 𝜉 ≺ ß ⇒ (Ξ, 𝜉) ⊩ B)

Since 𝑔 is increasing with ß as its range, we will for some natural number 𝑚𝜀Ω have
that 𝜈 ⪯ 𝑔(𝑚) ≺ ß, so that

b) Π𝜉 (𝑔(𝑚) ⪯ 𝜉 ≺ ß ⇒ (Ξ, 𝜉) ⊩ B)

Suppose B ∉ STAB(Ξ). By our enumeration of unstable elements where each term

recurs infinitely often, we have that B = 𝑓 (𝑛) for some natural number 𝑛, 𝑚 ≺ 𝑛 ∈ Ω.
It follows that 𝑔(𝑚) ≺ 𝑔(𝑛) ≺ ß. From a) and b) we can infer that (Ξ, 𝑔(𝑛)) ⊩ B, since
we supposed that (Ξ, ß) ⊩ T B. From the construction of function 𝑔, (Ξ, 𝑔(𝑛+1)) ⊩¬B,
contradicting b). It follows that (Ξ, ß) ⊩ T B only if B ∈ STAB(Ξ), so ß stabilizes Ξ.
3. The proof finishes with an appeal to Definition 5.3.1.9. □

5.4 The range of librationist satisfaction

5.4.1. Definition Satisfaction and satiation:
(1) Fair Ξ maximally satisfies A just if (Ξ, Ϙ) ⊩ T A.
(2) Fair Ξ optimally satisfies A just if (Ξ, Ϙ) ⊩ A.
(3) Fair Ξ plainly satisfies A just if (Ξ, Ϙ) ⊩ ¬T ¬A.
(4) Fair Ξ minorly satisfies A just if (Ξ, Ϙ) ⊩ ¬T ¬A ∧ ¬T A.

5.4.2. Definition:
(1) A is maximally satisfied just if for all fairs Ξ, (Ξ, Ϙ) ⊩ T A.
(2) A is optimally satisfied just if for all fairs Ξ, (Ξ, Ϙ) ⊩ A.

16
5.5 Relations between maxims, optima, plains and minors 5 SEMANTICS

(3) A is plainly satisfied just if for all fairs Ξ, (Ξ, Ϙ) ⊩ ¬T ¬A.

(4) A is minorly satisfied just if for all fairs Ξ, (Ξ, Ϙ) ⊩ ¬T ¬A ∧ ¬T A.

5.4.3. Definition Notation:

M
(1) |= A A is a maxim, as it is maximally satisfied.
O
(2) |= A A is an optimum, as it is optimally satisfied.
(3) |= A A is a plain, as it is plainly satisfied.
(4) |= A A is a minor, as it is minorly satisfied.
𝑚

5.5 Relations between maxims, optima, plains and minors

5.5.1. Definition of tautologies, antologies, determinates, standards:

(1) A is a tautology just if |= A.
M
(2) A is an antology just if |= ¬A.
(3) A is a determinate just if (|= A & 
|= ¬A) r (
|= A & |= ¬A).

(4) A is an indeterminate just if  |= ¬A.

|= A & 

M
(5) A is a standard, or a standard, just if |= (T A ∨ T ¬A ).
5.5.2. Exercise Show that (Ξ, 𝛼) ⊩ {𝑥|𝑥 ∈ 𝑥} ∈ {𝑥|𝑥 ∈ 𝑥} ↔ T {𝑥|𝑥 ∈ 𝑥} ∈ {𝑥|𝑥 ∈ 𝑥}.
5.5.3. Exercise Notice that {𝑥|𝑥 ∈ 𝑥} ∈ {𝑥|𝑥 ∈ 𝑥} is maximally satisfied by Ξ, or {𝑥|𝑥 ∈ 𝑥} ∉ {𝑥|𝑥 ∈ 𝑥}
is maximally satisfied by Ξ, and that {𝑥|𝑥 ∈ 𝑥} ∈ {𝑥|𝑥 ∈ 𝑥} consequently is indefinite, in
the sense of Definition 5.6.1.3.
5.5.4. Postulate
O
(1) |= A ⇒ |= A
M O
(2) |= A ⇒ |= A
(3) |= A ⇔ |= A & |= ¬A
𝑚

M
(4) |= A ⇔ |= A & 
|= ¬A

Proof:
(1): Use 7.1.6.
M O M O
(2): |= A ⇒ |= T A , given Definitions 5.4.2 and 5.4.3. So if |= A, |= T A holds. Also,
O M O
|= T A → A, which is 7.1.6. Thus, |= A only if |= A.

(3): By Definitions 5.4.2.3 and 5.4.2.4.

17
5.6 Orthodoxy, definiteness and paradoxicality 5 SEMANTICS

O
(4): Leftwards - |= A &  |=A ⇒ |= (T A ∨ T ¬A) on account of Postulate 6.3.1. So for all

fairs Ξ, (Ξ, Ϙ) ⊩ (T A ∨ T ¬A). As |= A, so that for all fairs Ξ, (Ξ, Ϙ) ⊩ ¬T ¬A, for all fairs
M
Ξ, (Ξ, Ϙ) ⊩ T A. So |= A.
M
Rightwards - Given |= A, for all fairs Ξ, (Ξ, Ϙ) ⊩ T A, so that by Definition 5.2.2.4, for all
Ξ, Σ𝛾(𝛾 ≺ 𝛼 & Π𝛿(𝛾 ⪯ 𝛿 ≺ 𝛼 ⇒ (Ξ, 𝛿) ⊩ A)). Consequently,

 
ΠΞ Π𝛾(𝛾 ≺ 𝛼 & Σ𝛿(𝛾 ⪯ 𝛿 ≺ 𝛼 ⇒ (Ξ, 𝛿) ⊩ A)) ,

M
so for all Ξ, (Ξ, Ϙ) ⊩ ¬T ¬A, and so |= A. Moreover, |= A & |= ¬A clearly entails absurdity,
M M
so |= A ⇒  |=¬A. In sum, |= A ⇒ |= A & |=¬A. Finish by joining the directions. □

5.5.5. Theorem: There are just minor and maximal tautologies. Optimal tautologies are
either minor or maximal, and minor tautologies are not maximal tautologies.

Proof: This follows from Theorems 5.5.4.3 and 5.5.4.4. □

5.6 Orthodoxy, definiteness and paradoxicality

5.6.1. Definition:
O
(1) A is orthodox just if |= ∀®𝑣 (T A ∨ T ¬A).
(2) Set 𝑎 is orthodox just if 𝑥 ∈ 𝑎 is orthodox.
(3) A is definite just if |= A or |= ¬A.
(4) A is apocryphal just if orthodox and indefinite.
(5) Set 𝑎 is apocryphal just if 𝑏 ∈ 𝑎 is apocryphal for some set 𝑏.
5.6.2. Remark: Some definite sentences are determinate, and some are indeterminate.
5.6.3. Remark: Set 𝑠 === {𝑣|𝑣 ∈ 𝑣} is apocryphal. For sentence 𝑠 ∈ 𝑠 is apocryphal, by
cause of its orthodoxy and the fact that it is indefinite because |̸ = 𝑠 ∈ 𝑠 and |̸ = 𝑠 ∉ 𝑠.
5.6.4. Definition: Formula A is paradoxical just if not orthodox. Given Definition 5.6.1.1,
O
this is the case just if |̸ = ∀®𝑣 (T A(®𝑣 ) ∨ T ¬A(®𝑣 )); so there is, given Definition 5.4.2.2, a fair Ξ
such that (Ξ, Ϙ) ⊩ ∃®𝑣 (¬T A(®𝑣 ) ∧ ¬T ¬A(®𝑣 )). By adapting Theorem 14.1.1, we find a vector
𝑎® for instantiation so that (Ξ, Ϙ) ⊩ (¬T A( 𝑎) ® ∧ ¬T ¬A( 𝑎)) ®
O
5.6.5. Definition: Sentence A is paradoxical just if not orthodox, just if |̸ = (T A ∨ T ¬A);
so there is, given Definition 5.4.2.2, a fair Ξ such that (Ξ, Ϙ) ⊩ (¬T A ∧ ¬T ¬A).
O
5.6.6. Definition: Set 𝑎 is paradoxical just if not orthodox just if |̸ = ∀𝑥(T 𝑥 ∈ 𝑎 ∨ T 𝑥 ∉ 𝑎);
so, given Definition 5.4.2.2, there is a fair Ξ such that (Ξ, Ϙ) ⊩ ∃𝑥(¬T 𝑥 ∈ 𝑎 ∧ ¬T 𝑥 ∉ 𝑎).
Consequently, given Theorem 14.1.1, for some term 𝑏, (Ξ, Ϙ) ⊩ (¬T 𝑏 ∈ 𝑎 ∧ ¬T 𝑏 ∉ 𝑎).

18
5.7 The non-triviality assumptions 5 SEMANTICS

O M
5.6.7. Fact: The proof of Theorem 6.3.2 shows that |= (T A ∨ T ¬A) ⇒ |= (T A ∨ T ¬A), so
M
Definition 5.6.1.1 entails that A is orthodox just if |= (T A ∨ T ¬A). But the latter should
not be used for defining orthodoxy, as the induced revision of Definition 5.6.4 would not
give the intended extension for the term ‘paradoxical’.

5.7 The non-triviality assumptions

5.7.1. Definition: A logical theory is trivial if all of its sentences are derivable.
The assumption that there are fair functions for variants of £ amounts to assuming that
the system under consideration is not trivial, and, consequently, consistent. It was shown
in §is a fair function for £ simpliciter just if Σ3 KP𝜔 is consistent.
There are fair functions for XXX if NBG+ Global Choice is consistent, and for £ under
much weaker assumptions.

19
5.7 The non-triviality assumptions 5 SEMANTICS

20
 6 MAXIMS

6 Maxims
{(𝑥, 𝑦)|𝑦 ≥ 𝑓 (𝑥)}

6.1 Axioms & warrants, theorems & proofs.

6.1.1. Definition:
(1) A warrant of an axiom is a semantic demonstration of it from Definition 5.2.2.
(2) A proof of a theorem is a demonstration of it from axioms and other theorems.

6.2 The relations between the varieties of theses

§5.4 gave the semantic distinctions between maximally, optimally, plainly and minorly
true. The corresponding syntactic notions are maxim, optimum, plain and menor. Here
“plain” is used as a noun, and occasionally as an adjective. The word “menor” is a variant
of “minor”, according to the Oxford English Dictionary, and the term “minor” is here used
as an adjective, while “menor” is used as a noun.
6.2.1. Postulate of the soundness: £ is sound for all theses of the treatise, as that is
checked individually. So

M O
⊢M A ⇒ |= A; ⊢O A ⇒ |= A; ⊢ A ⇒ |= A and ⊢𝑚 A ⇒ |= A.
𝑚

6.2.2. Axiom Relations between maxims, optima, plains and menors:

(1) ⊢M A ⇒ ⊢O A
(2) ⊢O A ⇒ ⊢ A
(3) ⊢𝑚 A ⇔ ⊢ A & ⊢ ¬A
(4) ⊢M A ⇔ ⊢ A & ⊢
¬A

6.3 Arbitration
6.3.1. Postulate (Optimal arbitration)
O
= (T A ∨ T ¬A ) r (|= A ⇔ |= ¬A).
|

W.6.3.1: Definition 5.2.2.6 states that

ΠΞ[(Ξ, 𝛼) ⊩ T A ∨ T ¬A ] r ΠΞ[(Ξ, 𝛼) ⊩ ¬T ¬A ] ⇔ ΠΞ[(Ξ, 𝛼) ⊩ ¬T A ] .
The right disjunct of Definition 5.2.2.6 amounts to |= A ⇔ |= ¬A, given Definitions 5.4.2.3
O
and 5.4.3.3, and |= (T A ∨ T ¬A ) is entailed by the left disjunct of Definition 5.2.2.6
via Definitions 5.4.2.2 and 5.4.3.2.

21
6.4 Logic maxims 6 MAXIMS

6.3.2. Theorem (Maximal arbitration)

M
|= (T A ∨ T ¬A ) r (|= A ⇔ |= ¬A).
O M
Proof: Confer Fact 5.6.7. It suffices to prove |= (T A ∨ T ¬A) ⇒ |= (T A ∨ T ¬A). Make
O
use of a disjunctive syllogism to obtain |= (T T A ∨ T T ¬A ) with Postulate 7.1.1 from
O
|= (T A ∨ T ¬A ), and use theorem ⊢O T T B → T (T B ∨ C) and disjunctive
O O
syllogism with |= (T T A ∨ T T ¬A ) to obtain |= T T A ∨ T ¬A . Use Definitions
M
5.4.2.1, 5.4.2.2, 5.4.3.1 and 5.4.3.2 and the results noted to conclude |= T A ∨ T ¬A
O
from |= T T A ∨ T ¬A . □

6.4 Logic maxims

6.4.1. Postulate Classical logic maxims:
(1) ⊢M 𝐴 → (𝐵 → 𝐴)
(2) ⊢M ( 𝐴 → (𝐵 → 𝐶)) → (( 𝐴 → 𝐵) → ( 𝐴 → 𝐶))
(3) ⊢M (¬𝐵 → ¬𝐴) → ( 𝐴 → 𝐵)
(4) ⊢M ∀𝑥( 𝐴 → 𝐵) → (∀𝑥 𝐴 → ∀𝑥𝐵)
(5) ⊢M 𝐴 → ∀𝑣 𝐴, provided 𝑣 is not free in 𝐴
(6) ⊢M ∀𝑣 𝐴 → 𝐴𝑣𝑏 , provided 𝑏 is substitutable for 𝑣 in 𝐴
(7) If ⊢M Γ belongs to (6.4.1.1−6.4.1.6), then so does ⊢M ∀𝑣Γ.

6.4.2. Remark: The role of a maximal inference mode, which allows the deduction from
⊢M (A → B) and ⊢M A to ⊢M B, is played by mode 9.2.5.
6.4.3. Remark: An induction, upon 6.4.1.7 and 9.2.5, proves generalization is a derived
inference mode. Compare the proof of Theorem 45.4 of (Hunter 1971, pp. 174–175).

6.5 Maxims on truth

6.5.1. Definition Russell’s paradoxical set:

r === {𝑥|𝑥 ∉ 𝑥}

6.5.2. Postulate Truth maxims:

(1) ⊢M T ( 𝐴 → 𝐵) → (T 𝐴 → T 𝐵)
(2) ⊢M T 𝐴 → ¬T ¬𝐴
(3) ⊢M (T r ∈ r ∨ T r ∉ r) → (T 𝐴 ∨ T ¬𝐴)
(4) ⊢M T 𝐴 ∨ T ¬𝐴 ∨ (T ¬T ¬𝐵 → T 𝐵)

22
6.6 Warrants of truth maxims 6 MAXIMS

(5) ⊢M T 𝐴 ∨ T ¬𝐴 ∨ (T 𝐵 → T T 𝐵)
(6) ⊢M T (T 𝐴 → 𝐴) → (T 𝐴 ∨ T ¬𝐴)
(7) ⊢M T (T 𝐴 → T T 𝐴) → (T 𝐴 ∨ T ¬𝐴)
(8) ⊢M ∃𝑣T A → T ∃𝑣A.
(9) ⊢M T ∀𝑣A → ∀𝑣T A
(10) ⊢M ∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) → ( 𝐴𝑣𝑎 → 𝐴𝑣𝑏 ), for 𝑎 and 𝑏 both substitutable for 𝑣 in 𝐴.
(11) ⊢M 𝔒({𝑥|A}) → (∀𝑥T A → T ∀𝑥A)

6.6 Warrants of truth maxims

Warrant 6.5.2.1: Suppose (Ξ, 𝛾) ⊩ T (A → B) and (Ξ, 𝛾) ⊩ T A. It follows that for some
ordinal 𝛿 and any ordinal 𝜖 such that 𝛿 ⪯ 𝜖 ≺ 𝛾, (Ξ, 𝜖) ⊩ (A → B) and (Ξ, 𝜖) ⊩ A. So,
on account of Definition 5.2.2.1, (Ξ, 𝜖) ⊩ B, and, consequently, (Ξ, 𝛾) ⊩ T B. So for any
ordinal 𝛾, (Ξ, 𝛾) ⊩ T (A → B) → (T A → T B). (Ξ, Ϙ) ⊩ T (T (A → B) → (T A → T B)) is a
M
consequence of this, so |= T (A → B) → (T A → T B).

Warrant 6.5.2.2: Assume (Ξ, 𝛾) ⊩(T A → ¬T ¬A). It follows that (Ξ, 𝛾) ⊩ (T A ∧ T ¬A). As
a consequence, (Ξ, 𝛾) ⊩ T A and (Ξ, 𝛾) ⊩ T ¬A. It follows that for some ordinal 𝛿 and any
ordinal 𝜖 such that 𝛿 ⪯ 𝜖 ≺ 𝛾, (Ξ, 𝜖) ⊩ A and (Ξ, 𝜖) ⊩ ¬A. But that is impossible.

Warrant 6.5.2.3: The postulate’s maxim somewhat extends (Bjørdal 2012). Let an ordi-
nal 𝛿 be monogamous just if a successor ordinal, so (Ξ, 𝛿) ⊩ T B just if (Ξ, 𝛿) ⊩ ¬T ¬B, for
M
any sentence B. |= (T r ∈ r ∨ T r ∉ r) → (T 𝐴 ∨ T ¬𝐴) holds simply because monogamous
ordinals are monogamous ordinals.

Warrant 6.5.2.4: Let an ordinal 𝛾 be reflected, just if (Ξ, 𝛾) ⊩ T 𝐵, provided (Ξ, 𝛾) ⊩

T ¬T ¬B. Any limit ordinal 𝜆 is reflected, for if B holds at all ordinals as from some ordinal
𝜇 below 𝜆 according to Ξ, then also ¬T ¬B holds at all ordinals as from 𝜇 below 𝜆 accord-
ing to Ξ. So limit ordinals are reflected, and successor ordinals are monogamous, in the
sense of Postulate 6.5.2.3. The content of 6.5.2.4 is thus that all ordinals are reflected or
monogamous, as for a monogamous successor ordinal 𝛿, ((Ξ, 𝛿) ⊩ (T A∨T ¬A), and if 𝛿 is
a reflected limit ordinal, (Ξ, 𝛿) ⊩ (T ¬T ¬B → T B). In either case, 6.5.2.4 is warranted.

Warrant 6.5.2.5: Let an ordinal 𝛾 be transitive just if for any A,

∃𝜃 (𝜃 ≺ 𝛾 & Π𝜉 (𝜃 ⪯ 𝜉 ⇒ (Ξ, 𝜉) ⊩ A)) ⇒ ∃𝜃 (𝜃 ≺ 𝛾 & Π𝜉 (𝜃 ⪯ 𝜉 ⇒ (Ξ, 𝜉) ⊩ T A)).

Precisely limit ordinals are transitive ordinals.

The content of Postulate 6.5.2.5 is that ordinals are transitive, or monogamous, in the
sense of Warrant 6.5.2.3. But that is true, as all ordinals larger than 0 are successor
ordinals or limit ordinals. So 6.5.2.5 has been warranted.

23
6.6 Warrants of truth maxims 6 MAXIMS

Warrant 6.5.2.6: At successor ordinals this holds, because there the consequent is true.
Let 𝜆 be a limit ordinal, and 𝜌 such that

Π𝜉 (𝜌 ⪯ 𝜉 ≺ 𝜆) ⇒ (Ξ, 𝜉) ⊩ T A → A,

so that (Ξ, 𝜆) ⊩ T (T A → A). Suppose there is some ordinal 𝜎 ≺ 𝜆 and 𝜌 ⪯ 𝜎 such that
(Ξ, 𝜎) ⊩ A. If so, (Ξ, 𝜆) ⊩ T A. If there is no ordinal 𝜎 ≺ 𝜆 and 𝜌 ≺ 𝜎 such that (Ξ, 𝜎) ⊩ A,
then (Ξ, 𝜆) ⊩ T ¬A. So 6.5.2.6 has been warranted.

Warrant 6.5.2.7: Postulate 6.5.2.7 holds at all successor ordinals, as the consequent al-
ways holds there. (Ξ, 𝜆) ⊩ T (T A → T T 𝐴) ⇒ Σ𝛿Π𝜖 (𝛿 ⪯ 𝜖 ≺ 𝜆 ⇒ (Ξ, 𝜖) ⊩ T A → T T 𝐴) if
𝜆 is a limit ordinal. But all ordinals 𝜖 in the interval from and including 𝛿 and less than 𝜆
will have a successor 𝜖 +1 which is also in the interval, so also (Ξ, 𝜖 +1) ⊩ T A → T T 𝐴. But
the latter statement has the consequence that (Ξ, 𝜖) ⊩ A → T 𝐴. So we have established
that for any limit 𝜆, (Ξ, 𝜆) ⊩ T (T A → T T 𝐴) → T (A → T 𝐴). Given postulate 6.5.2.2 and
contraposition, we obtain that (Ξ, 𝜆) ⊩ T (T A → T T 𝐴) → T (T ¬A → ¬A). At this point
is only takes postulate 6.5.2.6 to finish the warrant.

Warrant 6.5.2.8: Suppose (Ξ, 𝛾) ⊩ ∃𝑣T A. On account of Definition 5.2.2.2, (Ξ, 𝛾) ⊩ T A𝑣𝑏
for a 𝑏 substitutable for 𝑣 in A. So, on account of Definition 5.2.2.4 it follows that for an
ordinal 𝛿 and any ordinal 𝜖 such that 𝛿 ⪯ 𝜖 ≺ 𝛾, (Ξ, 𝜖) ⊩ A𝑣𝑏 for a 𝑏 substitutable for 𝑣 in
A. So on account of Definition 5.2.2.2, again, for an ordinal 𝛿 and any ordinal 𝜖 such that
𝛿 ⪯ 𝜖 ≺ 𝛾, (Ξ, 𝜖) ⊩ ∃𝑣A. So on account of Definition 5.2.2.4, (Ξ, 𝛾) ⊩ T ∃𝑣A.

Warrant 6.5.2.9: Let ordinal 𝛾 be such that (Ξ, 𝛾) ⊩ T ∀𝑣A. There is, consequently, an
ordinal 𝛿 such that for any ordinal 𝜖 fulfilling 𝛿 ⪯ 𝜖 ≺ 𝛾, (Ξ, 𝜖) ⊩ ∀𝑣A. So either 𝛾 = 𝛿 + 1 =
𝜖 + 1 or 𝛾 is a limit ordinal such that (Ξ, 𝜖) ⊩ ∀𝑣A for all ordinals 𝜖 such that 𝛿 ⪯ 𝜖 ≺ 𝛾.
In either case, (Ξ, 𝜖) ⊩ ∀𝑣A holds at any 𝜖 smaller than 𝛾 and at least as large as 𝛿. It
follows from Definition 5.2.2.2, that (Ξ, 𝜖) ⊩ A𝑣𝑏 , at any 𝜖 smaller than 𝛾 and at least as
large as 𝛿, for all 𝑏 substitutable for 𝑣 in A. So (Ξ, 𝛾) ⊩ T A𝑣𝑏 , for all 𝑏 substitutable for 𝑣
in A. So from Definition 5.2.2.2, again, (Ξ, 𝛾) ⊩ ∀𝑣T A. So (Ξ, 𝛽) ⊩ T ∀𝑣A → ∀𝑣T A holds
M
at any ordinal 𝛽. So (Ξ, Ϙ) ⊩ T (T ∀𝑣A → ∀𝑣T A), and consequently |= T ∀𝑣A → ∀𝑣T A.

Warrant 6.5.2.10: The warrant is in the proof of Theorem 11.2.1.5.

Warrant 6.5.2.11: – Notice that 6.5.2.11 is the Barcan postulate for orthodox formulas.
Assume

M
̸ = 𝔒({𝑥|A}) → (∀𝑥T A → T ∀𝑥A).
|

It follows, by Definitions 5.2.2 and 5.4.3, that for some fair function Ξ′:

(Ξ′, Ϙ) ⊩ ¬T (𝔒({𝑥|A}) → (∀𝑥T A → T ∀𝑥A)). (6.6.1)

24
6.6 Warrants of truth maxims 6 MAXIMS

Definition 5.2.2.4 has the consequence:

Π𝛾(𝛾 ≺ Ϙ ⇒ Σ𝛿 𝛾 ⪯ 𝛿 ≺ Ϙ & (6.6.2)

(Ξ′, 𝛿) ⊩ 𝔒({𝑥|A}) ∧ ∀𝑥T A ∧ ¬T ∀𝑥A )

Case 1/2 - 𝛿 is a limit: Suppose

(Ξ′, 𝛿) ⊩ 𝔒({𝑥|A}) ∧ ∀𝑥T A ∧ ¬T ∀𝑥A. (6.6.3)

Then, for all constants 𝑐, and all ordinals 𝜓 larger than a 𝜉 smaller than 𝛿,

(Ξ′, 𝜓) ⊩ 𝔒({𝑥|A}) ∧ A𝑥𝑐 ,

so as well

(Ξ′, 𝜓) ⊩ 𝔒({𝑥|A}) ∧ ∀𝑥A.

Also, however,
(Ξ′, 𝛿) ⊩ ¬T ∀𝑥A,

so that for some 𝜓 ⪯ 𝜙 ⪯ 𝛿,

(Ξ′, 𝜙) ⊩ ¬A𝑥𝑐 .

So
(Ξ′, 𝛿) ⊩ 𝔒({𝑥|A}) ∧ ∀𝑥T 𝐴 ∧ ¬T ∀𝑥A
cannot hold at a limit ordinal 𝛿.

Case 2/2 - 𝛿 = 𝛾 + 1 is a successor. Suppose

(Ξ′, 𝛿) ⊩ 𝔒({𝑥|A}) ∧ ∀𝑥T 𝐴 ∧ ¬T ∀𝑥A.

Then
(Ξ′, 𝛾) ⊩ ∃𝑥¬A,

so that there, by Theorem 5.2.6, is a constant 𝑐 for which

(Ξ′, 𝛾) ⊩ ¬A𝑥𝑐 .

However, as
(Ξ′, 𝛿) ⊩ ∀𝑥T 𝐴,

also
(Ξ′, 𝛾) ⊩ A𝑥𝑐 .

25
6.6 Warrants of truth maxims 6 MAXIMS

So
(Ξ′, 𝛿) ⊩ 𝔒({𝑥|A}) ∧ ∀𝑥T 𝐴 ∧ ¬T ∀𝑥A
cannot hold at a successor ordinal 𝛿.
Cases 1/2 and 2/2 entail that for any ordinal 𝛽, (Ξ, 𝛽) ⊩ 𝔒({𝑥|A}) → (∀𝑥T A → T ∀𝑥A).
M
So (Ξ, Ϙ) ⊩ T (𝔒({𝑥|A}) → (∀𝑥T A → T ∀𝑥A)), and so |= 𝔒({𝑥|A}) → (∀𝑥T A → T ∀𝑥A).
That warrants Postulate 6.5.2.11’s posit of ⊢M 𝔒({𝑥|A}) → (∀𝑥T A → T ∀𝑥A).

6.6.4. Remark: The semantic justification for some of the maxims of Postulates 6.5.2.1
– 6.5.2.11 can be lifted from (Bjørdal 2012)(340–341).
6.6.5. Remark: Postulates 6.5.2.6 and 6.5.2.7 originate with (Turner 1990).
6.6.6. Remark: The maxims of Postulates 6.5.2.7 and 6.5.2.8 were not included in (Bjørdal
2012), as the author thought they were both derivable. The warrant of Postulate 6.5.2.8
shows that this was correct for its maxim schema, but the warrant of Postulate 6.5.2.7
suggests that Postulate 6.5.2.3 is needed for its semantical justification.
6.6.7. Remark: Although the converses of Postulates 6.5.2.5 and 6.5.2.6 hold at limit
ordinals, they are not maxims, for we may at a sucessor 𝜎 have that

(Ξ, 𝜎) ⊩ (T ¬𝐴 ∨ T 𝐴) ∧ ¬T (T 𝐴 → 𝐴),

and it happen for {𝑥|𝑥 ∉ 𝑥} ∈ {𝑥|𝑥 ∉ 𝑥} at 𝜎 or 𝜎 +1. This contrasts with Remark 69.3.1.(ii)
in (Cantini 1996)(396).
6.6.8. Exercise: Let A be deferent just if for all fairs Ξ, (Ξ, Ϙ) ⊩ T A r (Ξ, Ϙ) ⊩ T ¬A.
Show that just deferent formulas are orthodox.
O M
6.6.9. Exercise: Prove that |= ∀®𝑣 (T A ∨ T ¬A) ⇒ |= ∀®𝑣 (T A ∨ T ¬A).
Remark on Exercise 6.6.9: Defining a formula A as orthodox just if |=M ∀®𝑣 (T A ∨ T ¬A),
instead of using Definition 5.6.1.1, is not advisable. For defining a formula as paradoxical
just if not orthodox, as in Definition 5.6.4, would then induce an unacceptable extension
for the term “paradoxical”.

26
 7 OPTIMA

7 Optima

Pax optima rerum, quas homini novisse datum est: pax

una triumphis innumeris potior: pax, custodire salutem et
cives aequare potens.

Silius Italicus

We have, as in Theorem 2 of (Bjørdal 2012, p. 342):

7.1. Postulate:
O
(1) |= T 𝐴 ↔ T T 𝐴
O
(2) |= T ¬T ¬𝐴 ↔ T 𝐴
O
(3) |= T (T 𝐴 → T 𝐵) → T ( 𝐴 → 𝐵)
O
(4) |= T ( 𝐴 → T 𝐴) ↔ T (T 𝐴 → 𝐴)
O
(5) |= ∀𝑥T 𝐴(𝑥) → T ∀𝑥 𝐴(𝑥)
O
(6) |= T 𝐴 → 𝐴

It is left as an exercise to warrant the optimal tautologies.

27
 7 OPTIMA

28
 8 PLAINS

8 Plains

Pure mathematics is, in its way, the poetry of logical ideas.

Albert Einstein

8.1. Postulate Plains:

(1) ⊢ T 𝐴 → 𝐴.
(2) ⊢ 𝐴 → T 𝐴.
(3) ⊢ T ∃𝑣 𝐴 → ∃𝑣T 𝐴.
(4) ⊢ ∀𝑣T 𝐴 → T ∀𝑣 𝐴.
(5) r ∈ r.
(6) r ∉ r.

8.2. Remark: Instances of the plains in Postulates 8.1.1 and 8.1.2 may be maxims or
minors. There are minor instances of Postulates 8.1.1 and 8.1.2 on account of Russell’s
paradoxical set, here denoted as in 8.1.5 and 8.1.6 by the r of Definition 6.5.1 on page
22. In the case of the attestor schema of Postulate 8.1.3, the failure of some maximal
versions follow from Corollary 14.1.5. The failure of the maximality for all instances of
Postulate 8.1.4 is shown in §14.3.
8.3. Exercise: As regards Postulates 8.1.5 and 8.1.6, prove that £ has ⊢ r ∈ r and ⊢ r ∉ r.

29
 8 PLAINS

30
 9 INFERENCE MODES

9 Inference modes

Recall that the valid inference from A to T A is the number theoretic

and meta theoretic principle that for all fairs Ξ, if A is in the variety
of formula numbers ⊢ (Ξ, Ϙ), then also T A is in ⊢ (Ξ, Ϙ).

9.1 The simple inference modes

Only ¬, T and one occurence of a formula variable are allowed in the formulas in the
antecedent and the consequent of the simple inference modes. Moreover, T may only
occur once in the antecedent, and in the consequent.
9.1.1. Postulate simple thetical inference modes:
1 ⊢ 𝐴⇒⊢T𝐴
2 ⊢ 𝐴 ⇒ ⊢ ¬T ¬𝐴
3 ⊢ ¬𝐴 ⇒ ⊢ T ¬𝐴
4 ⊢ ¬𝐴 ⇒ ⊢ ¬T 𝐴
5 ⊢T𝐴⇒⊢ 𝐴
6 ⊢ T 𝐴 ⇒ ⊢ ¬T ¬𝐴
7 ⊢ T ¬𝐴 ⇒ ⊢ ¬𝐴
8 ⊢ T ¬𝐴 ⇒ ⊢ ¬T 𝐴
9 ⊢ ¬T 𝐴 ⇒ ⊢ ¬𝐴
10 ⊢ ¬T 𝐴 ⇒ ⊢ T ¬𝐴
11 ⊢ ¬T ¬𝐴 ⇒ ⊢ 𝐴
12 ⊢ ¬T ¬𝐴 ⇒ ⊢ T 𝐴

The corresponding valid, simple maximal inference modes of Postulate 9.1.2 can be jus-
tified by the valid simple thetical inference modes in Postulate 9.1.1 on account of the
syntactical correlate of Theorem 5.5.4.4, which says that ⊢M 𝐴 just if ⊢ 𝐴 & ⊢¬𝐴.

The
inference mode of Postulate 9.1.2.1 is for example a consequence of the conjunction
of the modes provided by Postulates 9.1.1.1 and 9.1.1.9. The other dependencies are
straightforward to establish.

9.1.2. Postulate simple maximal inference modes:

1 ⊢M 𝐴 ⇒ ⊢M T 𝐴
2 ⊢M 𝐴 ⇒ ⊢M ¬T ¬𝐴

31
9.2 Involved inference modes 9 INFERENCE MODES

3 ⊢M ¬𝐴 ⇒ ⊢M T ¬𝐴
4 ⊢M ¬𝐴 ⇒ ⊢M ¬T 𝐴
5 ⊢M T 𝐴 ⇒ ⊢M 𝐴
6 ⊢M T 𝐴 ⇒ ⊢M ¬T ¬𝐴
7 ⊢M T ¬𝐴 ⇒ ⊢M ¬𝐴
8 ⊢M T ¬𝐴 ⇒ ⊢M ¬T 𝐴
9 ⊢M ¬T 𝐴 ⇒ ⊢M ¬𝐴
10 ⊢M ¬T 𝐴 ⇒ ⊢M T ¬𝐴
11 ⊢M ¬T ¬𝐴 ⇒ ⊢M 𝐴
12 ⊢M ¬T ¬𝐴 ⇒ ⊢M T 𝐴

9.2 Involved inference modes

9.2.1. Postulate Quantificational thetical modes:
1 ⊢ ∀𝑣T 𝐴 ⇒ ⊢ T ∀𝑣 𝐴
2 ⊢ ¬T ∀𝑣 𝐴 ⇒ ⊢ ¬∀𝑣T 𝐴
3 ⊢ T ∃𝑣 𝐴 ⇒ ⊢ ∃𝑣T 𝐴
9.2.2. Postulate The Barcan mode: Postulates 9.2.1.1 and 9.2.1.2 justify
⊢M ∀𝑣T 𝐴 ⇒ ⊢M T ∀𝑣 𝐴.

9.2.3. Remark: Quantificational thetical mode 9.2.1.3 does not enter such a combination
as do 9.2.1.1 and 9.2.1.2, for
⊢ ¬∃𝑣T A ⇒ ⊢ ¬T ∃𝑣A
is not a valid mode schema; so neither is ⊢M T ∃𝑣A ⇒ ⊢M ∃𝑣T A. This is clarified in the
limitative results of Theorem 14.1.3, its Corollary 14.1.5, and Theorem 14.1.5 in §14.1.

9.2.4. Postulate Thetical distributive modes:

1 ⊢M ( 𝐴 → 𝐵) ⇒ (⊢ 𝐴 ⇒ ⊢ 𝐵).
2 ⊢M ( 𝐴 → 𝐵) ⇒ (⊢ ¬𝐵 ⇒ ⊢ ¬𝐴).
3 ⊢ ( 𝐴 → 𝐵) ⇒ (⊢M 𝐴 ⇒ ⊢ 𝐵).

9.2.5. Postulate The maxim mode:

⊢M ( 𝐴 → 𝐵) ⇒ (⊢M 𝐴 ⇒ ⊢M 𝐵).

32
9.2 Involved inference modes 9 INFERENCE MODES

9.2.6. Remark:

Postulate 9.2.5 is entailed by Postulates 9.2.4.1 and 9.2.4.2.

9.2.7. Postulate Complex modes:

1 ⊢M T 𝐴 ⇒ ⊢M T (T 𝐴 ↔ 𝐴) ∧ T (T ¬𝐴 ↔ ¬𝐴) (The Tarski mode)
2 ⊢M T 𝐴 → 𝐴 ⇒ ⊢M T 𝐴 ∨ T ¬𝐴
3 ⊢M T ¬T ¬𝐴 ⇒ ⊢M T 𝐴
4 ⊢M T (T 𝐴 → T 𝐵) ⇒ ⊢M T ( 𝐴 → 𝐵)
5 ⊢ 𝐴 & ⊢ 𝐵 ⇒ ⊢ ¬T ¬𝐴 ∧ ¬T ¬𝐵
6 ⊢M 𝔒( 𝐴(𝑥)) ⇒ (⊢M ∃𝑥 𝐴 ⇒ ⊢M 𝐴𝑥𝑎 for some 𝑎 substitutable for 𝑥 in 𝐴).
7 ⊢M 𝐴𝑣𝑎 for any constant 𝑎 ⇒ ⊢M ∀𝑣 𝐴

Warrant 9.2.7.1: Clearly ⊢M T 𝐴 ⇒ ⊢M T ( 𝐴 ∧ T 𝐴) ∧ T ( 𝐴 ∧ ¬T ¬𝐴). It is librationistically

derivable that ⊢M T (( 𝐴 ∧ T 𝐴) → T ( 𝐴 ↔ T 𝐴)) and ⊢M T (( 𝐴 ∧ ¬𝑇¬𝐴) → T (¬𝐴 ↔ T ¬𝐴)),
so Postulate 9.2.5 suffices to finish.

Proof: (9.2.7.4) Suppose (Ξ, Ϙ) ⊩ T (T 𝐴 → T 𝐵). (i) Let 𝜌 be be a ordinal as from which
T 𝐴 → T 𝐵 holds, so that

Π𝜉 (𝜌 ⪯ 𝜉 ≺ Ϙ ⇒ (Ξ, 𝜉) ⊩ (T 𝐴 → T 𝐵).

Thus (Ξ, 𝜌 + 1) ⊩ (T 𝐴 → T 𝐵), and therefore (Ξ, 𝜌) ⊩ ( 𝐴 → 𝐵). Consequently, succeeding

successors will have T 𝐴 → T 𝐵 and 𝐴 → 𝐵. (ii) Let limit ordinal 𝜆 ≺ Ϙ, above 𝜌, have
T 𝐴 → T 𝐵, and 𝐴 → 𝐵 below, as from 𝜌. As 𝜆 ≺ Ϙ, from the assumption on 𝜌, (Ξ, 𝜆) ⊩
(T 𝐴 → T 𝐵). As (Ξ, 𝜆 + 1) ⊩ (T 𝐴 → T 𝐵), also (Ξ, 𝜆) ⊩ ( 𝐴 → 𝐵). (iii) By a repetition of (i)
and (ii) it follows that 𝐴 → 𝐵 holds as from 𝜌 below Ϙ, so that (Ξ, Ϙ) ⊩ T ( 𝐴 → 𝐵). □

Proof: (9.2.7.6) This is established on page 47, in the proof of Theorem 14.2.1. □

33
9.2 Involved inference modes 9 INFERENCE MODES

34
 10 ALETHIC COMPREHENSION

10 Alethic comprehension

If one, per impossibile, could have used naive comprehension for

truth, and for abstraction, the alethic comprehension principle
would have been true. Fortunately, one cannot justify the opposite
entailment from alethic comprehension to naive comprehension.

10.1. Postulate Alethic comprehension without parameters:

⊢M ∀𝑥(𝑥 ∈ {𝑦|A} ↔ T A 𝑥𝑦 ), where 𝑥 is substitutable for 𝑦 in A.

10.2. Theorem Alethic comprehension with parameters from 𝑣®:

⊢M ∀®𝑣∀𝑥(𝑥 ∈ {𝑦|A} ↔ T A 𝑥𝑦 ), where 𝑥 is substitutable for 𝑦 in A.

Proof: Appeal to 9.2.7.7 and Postulate 10.1. □

35
 10 ALETHIC COMPREHENSION

36
 11 THE THEORY OF IDENTITY

11 The theory of identity

To be is that there is
something in all your
essences.

A streamlining of sections 4 and 5 of (Bjørdal 2012, pp. 342–345) is obtained from the
inference modes 9.2.7.1 – 9.2.7.4, and as a result £ does not, as e.g. the comparable
systems studied by (Cantini 1996), need additional axiomatic principles for having well
behaved notions of identity in this section, or natural number in §13.

11.1 Co-essentiality
We define the identity relation by means of a notion of co-essentiality, which is similar
to the relation named membership congruency by Abraham A. Fraenkel and Yehoshua
Bar-Hillel, and discussed in (A. A. Fraenkel and Bar-Hillel 1973, p. 27), though not used
in the previous edition (A. A. Fraenkel and Bar-Hillel 1958).
11.1.1. Definition: Sets 𝑎 and 𝑏 are co-essential just if ∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢).
Regarding the term "co-essentiality", see (Forster 2019), which relates that (Hailperin
1944) "gave the first of a number of finite axiomatisations of NF now known. Many
of them exploit the function 𝑥 ↦→ {𝑦|𝑦 ∈ 𝑥} which is injective and total and is an ∈
-isomorphism. This function was known to Whitehead, who suggested to Quine that
{𝑦|𝑥 ∈ 𝑦} should be called the “essence” of x (a terminology clearly suggested by a view
of sets as properties-in-extension)."
11.1.2. Definition Identity via co-essentiality:

𝑎 = 𝑏 === ∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢)

The definiens in Definition 11.1.2 is a conditional, and not a biconditional. Theorem

11.2.1.4 justifies the other direction by instantiating with {𝑥|∀𝑢(𝑥 ∈ 𝑢 → 𝑎 ∈ 𝑢}.
The justification for the analogous definition ∗13 · 01 in Principia Mathematica, will not
justify Definition 11.1.2. For the symmetry of Definition 11.1.2, is in £shown by the proof
of 11.2.1.4 below, and without an appeal to predicativity, as in the proof of ∗13 · 01 by
Whitehead and Russell.
11.1.3. Lemma: ⊢M T (∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) → T ∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢)).

Proof: Suppose ⊢M ∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢). By instantiation we have:

⊢M ∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) →
(𝑎 ∈ {𝑣|∀𝑢(𝑎 ∈ 𝑢 → 𝑣 ∈ 𝑢)} → 𝑏 ∈ {𝑣|∀𝑢(𝑎 ∈ 𝑢 → 𝑣 ∈ 𝑢)}).

37
11.2 The adequacy of identity as co-essentiality 11 THE THEORY OF IDENTITY

But ⊢M 𝑎 ∈ {𝑣|∀𝑢(𝑎 ∈ 𝑢 → 𝑣 ∈ 𝑢)}, so that

⊢M ∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) → 𝑏 ∈ {𝑣|∀𝑢(𝑎 ∈ 𝑢 → 𝑣 ∈ 𝑢)}.

Finish with Alethic Comprehension and Postulate 9.1.2.1. □

11.1.4. Lemma: ⊢M T (𝑎 = 𝑏 → T 𝑎 = 𝑏)

Proof: Use Definition 11.1.2 and Lemma 11.1.3. □

11.1.5. Lemma: T (T 𝑎 ≠ 𝑏 → 𝑎 ≠ 𝑏)

Proof: Use Lemma 11.1.4, Postulates 6.5.2.2, 9.1.2.1 and logic. □

11.1.6. Lemma:

⊢M T (T ¬∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) → ¬∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢))

Proof: Use Lemma 11.1.5 and Definition 11.1.2. □

11.1.7. Lemma:

⊢M T ((T ∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢)) → ∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢))

Proof: Combine Lemma 11.1.6 with 7.1.4 to obtain

⊢M T (¬∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) → T ¬∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢)).

An instance of Theorem 6.5.2.2 is

⊢M T (T ¬∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) → ¬T ∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢)).

A hypothetical syllogism and contraposition now suffices to finish the proof. □

11.2 The adequacy of identity as co-essentiality

11.2.1. Theorem (Orthodoxy, equivalence and fungibility)
(1) ⊢M T 𝑎 = 𝑏 ∨ T 𝑎 ≠ 𝑏 Orthodoxy
(2) ⊢M 𝑎 = 𝑎 Reflexivity
(3) ⊢M 𝑎 = 𝑏 ∧ 𝑏 = 𝑐 → 𝑎 = 𝑐 Transitivity
M
(4) ⊢ 𝑎 = 𝑏 → 𝑏 = 𝑎 Symmetry
(5) ⊢M 𝑎 = 𝑏 → ( 𝐴𝑣𝑎 → 𝐴𝑣𝑏 ), with 𝑎 and 𝑏 substitutable for 𝑣 in 𝐴. Fungibility

Proof:

38
11.2 The adequacy of identity as co-essentiality 11 THE THEORY OF IDENTITY

1. Use Lemma 11.1.5 and Postulate 6.5.2.6.

2. Trivial
3. Trivial, given Definition 11.1.2
4. Clearly,

⊢M ∀𝑣(𝑎 ∈ 𝑣 → 𝑏 ∈ 𝑣) → (𝑎 ∈ {𝑤|∀𝑣(𝑤 ∈ 𝑣 → 𝑎 ∈ 𝑣)} → (𝑏 ∈ {𝑤|∀𝑣(𝑤 ∈ 𝑣 → 𝑎 ∈ 𝑣)}.

But
⊢M 𝑎 ∈ {𝑤|∀𝑣(𝑤 ∈ 𝑣 → 𝑎 ∈ 𝑣)},

so that by alethic comprehension,

⊢M ∀𝑣(𝑎 ∈ 𝑣 → 𝑏 ∈ 𝑣) → T ∀𝑣(𝑏 ∈ 𝑣 → 𝑎 ∈ 𝑣). (11.2.2)

An instance of Lemma 11.1.4 states:

⊢M ∀𝑣(𝑏 ∈ 𝑣 → 𝑎 ∈ 𝑣) → T ∀𝑣(𝑏 ∈ 𝑣 → 𝑎 ∈ 𝑣). (11.2.3)

By invoking 7.1.4 on equation 11.2.3 we obtain

⊢M T ∀𝑣(𝑏 ∈ 𝑣 → 𝑎 ∈ 𝑣) → ∀𝑣(𝑏 ∈ 𝑣 → 𝑎 ∈ 𝑣). (11.2.4)

Finish with a hypothetical syllogism with equations 11.2.2 and 11.2.4, and lastly
an appeal to co-essentiality Definition 11.1.2.
5. The promissory note issued in sentence Warrant 6.3.W10 of Postulate 6.5.2.10 on
page 24 of §6 is satisfied, and the mentioned Postulate is warranted.
Suppose, for 𝑎 and 𝑏 substitutable for 𝑣 in 𝐴, and fair function Ξ,

⊩ T (∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) → ( 𝐴𝑣𝑎 → 𝐴𝑣𝑏 )).

(Ξ, Ϙ)

On account of the validity of the mode of 9.2.7.3 we get

⊩ T ¬T (∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) ∧ 𝐴𝑣𝑎 ∧ ¬𝐴𝑣𝑏 ).

(Ξ, Ϙ)

It follows from Definition 5.2.2.1 that

(Ξ, Ϙ) ⊩ ¬T ¬T (∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) ∧ 𝐴𝑣𝑎 ∧ ¬𝐴𝑣𝑏 ).

On account of Postulate 6.5.2.1,

(Ξ, Ϙ) ⊩ ¬T ¬(T ∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) ∧ T 𝐴𝑣𝑎 ∧ T ¬𝐴𝑣𝑏 ).

39
11.2 The adequacy of identity as co-essentiality 11 THE THEORY OF IDENTITY

On account of the tautologicality of Lemma 11.1.7, we get

(Ξ, Ϙ) ⊩ ¬T ¬(∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) ∧ T 𝐴𝑣𝑎 ∧ ¬T 𝐴𝑣𝑏 ).

From alethic comprehension and existential generalization we obtain

(Ξ, Ϙ) ⊩ ¬T ¬(∀𝑢(𝑎 ∈ 𝑢 → 𝑏 ∈ 𝑢) ∧ ∃𝑢(𝑎 ∈ 𝑢 ∧ 𝑏 ∉ 𝑢)),

which is absurd. So Postulate 6.5.2.10 is tautological, and we are done. □

40
 12 ALPHABETOLOGICALITY

12 Alphabetologicality

That the universe was formed by a fortuitous concourse

of atoms, I will no more believe than that the accidental
jumbling of the alphabet would fall into a most ingenious
treatise of philosophy.

Jonathan Swift

Postulates 12.1 and 12.2 express, given Definition 12.3, that identity is an equivalence
relation which is neutral with respect to alphabetological variants.
12.1. Postulate The Lindenbaum-Tarski closure for identity: If classical logic proves that
∀𝑥( 𝐴(𝑥) ↔ 𝐵(𝑥)), then
⊢M {𝑥| 𝐴(𝑥)} = {𝑥|𝐵(𝑥)}.

12.2. Postulate Alphabetical indifference:

𝑦
{𝑥| 𝐴(𝑥)} = {𝑥|𝐵(𝑥)} → {𝑥| 𝐴(𝑥)} = {𝑦|𝐵(𝑥)𝑥 },

where 𝑦 is substitutable for 𝑥 in 𝐵.

12.3. Definition Alphabetologicality: Two sets are alphabetological variants of each
other just if they are identical on account of Postulates 12.1 and 12.2.
Postulates 12.1 and 12.2 compensate somewhat for the loss of extensionality in £, as
per §21, and secure such theorems as:

⊢M {𝑥| 𝐴(𝑥)} = {𝑦| 𝐴(𝑦) ∧ ∃𝑧(𝐵(𝑧) ∨ ¬𝐵(𝑧))}.

41
 12 ALPHABETOLOGICALITY

42
 13 ARITHMETIC

13 Arithmetic

The numbers may be said to rule the whole world of quantity,

and the four rules of arithmetic may be regarded as the complete
equipment of the mathematician.

James C. Maxwell

13.1. Definition
(1) ∅ = {𝑥|𝑥 ≠ 𝑥}
(2) 𝑎′ = {𝑥|𝑥 = 𝑎 ∨ 𝑥 ∈ 𝑎}
(3) 𝜔 = {𝑥|∀𝑦(∅ ∈ 𝑦 ∧ ∀𝑧(𝑧 ∈ 𝑦 → 𝑧′ ∈ 𝑦) → 𝑥 ∈ 𝑦)}
13.2. Theorem
(1) ⊢M ∅ ∈ 𝜔
(2) ⊢M ∀𝑥(𝑥 ∈ 𝜔 → 𝑥 ′ ∈ 𝜔)
(3) 𝜔 is orthodox
(4) ⊢M ∀𝑦(∅ ∈ 𝑦 ∧ ∀𝑧(𝑧 ∈ 𝑦 → 𝑧′ ∈ 𝑦) → ∀𝑥(𝑥 ∈ 𝜔 → 𝑥 ∈ 𝑦))
(5) ⊢M A(∅) ∧ ∀𝑥(A(𝑥) → A(𝑥 ′)) → ∀𝑦(𝑦 ∈ 𝜔 → A(𝑦))

Proof:
1. Combine alethic comprehension and the fact that

⊢M T ∀𝑦(∅ ∈ 𝑦 ∧ ∀𝑧(𝑧 ∈ 𝑦 → 𝑧′ ∈ 𝑦) → ∅ ∈ 𝑦)

2. This follows from alethic comprehension and

⊢M ∀𝑥(T (∀𝑦(∅ ∈ 𝑦 ∧ ∀𝑧(𝑧 ∈ 𝑦 → 𝑧′ ∈ 𝑦) → 𝑥 ∈ 𝑦)) →

T (∀𝑦(∅ ∈ 𝑦 ∧ ∀𝑧(𝑧 ∈ 𝑦 → 𝑧′ ∈ 𝑦) → 𝑥 ′ ∈ 𝑦))).

3. From logic:

⊢M ∅ ∈ 𝜔 ∧ ∀𝑥(𝑥 ∈ 𝜔 → 𝑥 ′ ∈ 𝜔) →
(∀𝑦(∅ ∈ 𝑦 ∧ ∀𝑥(𝑥 ∈ 𝑦 → 𝑥 ′ ∈ 𝑦) → 𝑎 ∈ 𝑦) → 𝑎 ∈ 𝜔).

By combining 1 and 2 we have

⊢M ∀𝑦(∅ ∈ 𝑦 ∧ ∀𝑥(𝑥 ∈ 𝑦 → 𝑥 ′ ∈ 𝑦) → 𝑎 ∈ 𝑦) → 𝑎 ∈ 𝜔).

Postulates 6.5.2.1 and 9.1.2.1, and alethic comprehension, give us

⊢M 𝑎 ∈ 𝜔 → T 𝑎 ∈ 𝜔.

43
 13 ARITHMETIC

9.2.5 along with Postulates 6.5.2.1, 6.5.2.2 and 6.5.2.6 give us

⊢M T 𝑎 ∈ 𝜔 ∨ T 𝑎 ∉ 𝜔

As 𝑎 was arbitrary, ⊢M ∀𝑥(T 𝑥 ∈ 𝜔 ∨ T 𝑥 ∉ 𝜔), and the proof is finished.

4. Immediate, given 3, as it is equivalent with
⊢M ∀𝑥(𝑥 ∈ 𝜔 → ∀𝑦(∅ ∈ 𝑦 ∧ ∀𝑧(𝑧 ∈ 𝑦 → 𝑧′ ∈ 𝑦) → 𝑥 ∈ 𝑦)).

5. For the following, compare (Cantini 1996, p. 356).

13.3. Definition: Let, for arbritrary sentence A(𝑥),
A′ (𝑥) === A(∅) ∧ ∀𝑥(A(𝑥) → A(𝑥 ′)) → A(𝑥)).

By logic,
⊢M A′ (∅) & ⊢M ∀𝑥(A′ (𝑥) → A′ (𝑥 ′)).

The inference mode of Postulate 9.1.2.1 and Postulate 6.5.2.10 entail

⊢M T A′ (∅) & ⊢M ∀𝑥T (A′ (𝑥) → A′ (𝑥 ′)).

By quantifier distribution and Postulate 6.5.2.1 we get

⊢M T A′ (∅) & ⊢M ∀𝑥(T A′ (𝑥) → T A′ (𝑥 ′)).

Alethic comprehension gives us

⊢M ∅ ∈ {𝑦|A′ (𝑦)} & ⊢M ∀𝑥(𝑥 ∈ {𝑦|A′ (𝑦)} → 𝑥 ′ ∈ {𝑦|A′ (𝑦)}).

Adjunction gives us
⊢M ∅ ∈ {𝑦|A′ (𝑦)} ∧ ∀𝑥(𝑥 ∈ {𝑦|A′ (𝑦)} → 𝑥 ′ ∈ {𝑦|A′ (𝑦)}).

4 and the inference of mode 9.2.5 give us

⊢M ∀𝑥(𝑥 ∈ 𝜔 → 𝑥 ∈ {𝑦|A′ (𝑦)}).

From 3 and 9.2.7.1 we have

⊢M ∀𝑥(T 𝑥 ∈ 𝜔 → 𝑥 ∈ 𝜔),

so that
⊢M ∀𝑥(T 𝑥 ∈ 𝜔 → 𝑥 ∈ {𝑦|A′ (𝑦)}).

Alethic comprehension gives us

⊢M ∀𝑥(T 𝑥 ∈ 𝜔 → T A′ (𝑥)),

which, combined with 9.2.7.4 establishes

⊢M ∀𝑥(𝑥 ∈ 𝜔 → A′ (𝑥))

44
 13 ARITHMETIC

Finish with an appeal to Definition 13.3, and rearrangement. □

45
 14 SHORTCOMINGS AND REDRESSES

14 Shortcomings and redresses

If all problems seem resolved, look in another direction!

§6 is supplemented with negative results, which to a large degree depend upon §11.

14.1 Shortcoming related to existential instantiation

Despite the important Theorem 5.2.6, which justifies
14.1.1. Theorem
for some term 𝑎
(Ξ, 𝛼) ⊩ ∃𝑥(¬T A ∧ ¬T ¬A) ==============⇒ (Ξ, 𝛼) ⊩ (¬T A𝑥𝑎 ∧ ¬T ¬A𝑥𝑎 ),

and consequently
14.1.2. Theorem Optimal existential instantiation:

If ⊢O ∃𝑥 𝐴, then ⊢O 𝐴𝑥𝑎 for some term 𝑎

There is, nevertheless, as pointed to in Remark 9.2.3, the following limitative result:
14.1.3. Theorem (Maximal lack of existential instantiation)

It may happen that ⊢M ∃𝑥 𝐴, and for no term 𝑎, ⊢M 𝐴𝑥𝑎 .

Proof: As the proof of Theorem 14.1.5. □

14.1.4. Corollary: Maximal existential instantiation, in the form

⊢M ∃𝑥 𝐴 ⇒ Σ𝑎 ⊢M 𝐴𝑥𝑎 , is not valid.

14.1.5. Theorem: The inference mode ⊢M T ∃𝑥 𝐴 ⇒ ⊢M ∃𝑥T 𝐴 is not valid.

M M
Proof: Let A be (𝑥 = ∅ ↔ r ∈ r). Obviously, |= T ∃𝑥A holds. Suppose that |= ∃𝑥T A. If
so (Ξ, Ϙ) ⊩ T ∃𝑥T A, and there is an ordinal 𝛾 such that (Ξ, 𝛽) ⊩ ∃𝑥T 𝐴 holds whenever
𝛾 ≺ 𝛽 ≺ Ϙ. Let limit ordinal 𝜆 satisfy 𝛾 ≺ 𝜆 ≺ Ϙ, so that (Ξ, 𝜆) ⊩ ∃𝑥T 𝐴. On account of
5.2.2.1 and 5.2.2.2, there is a term 𝑎 and an ordinal 𝛿 such that 𝑎 = ∅ ↔ r ∈ r holds at all
ordinals 𝜃 which satisfy 𝛿 ≺ 𝜃 ≺ 𝜆. But this is impossible, as 𝑟 ∈ 𝑟 holds at some of those
ordinals, and r ∉ r holds at others, whereas identity is orthodox. □

As stated in Remark 8.2, Theorem 14.1.5 entails that the attestor schema of Postulate
8.1.3 does not hold as a maxim, for, as its proof just showed, some instances of the
schema T ∃𝑣 𝐴 → ∃𝑣T 𝐴 are minor, i.e. paradoxical, truths.

46
14.2 An orthodox redress 14 SHORTCOMINGS AND REDRESSES

14.2 An orthodox redress

14.2.1. TheoremThe validity of 9.2.7.6 is shown, as announced on page 33:

⊢M 𝔒( 𝐴(𝑥)) ⇒ (⊢M ∃𝑥 𝐴 ⇒ ⊢M 𝐴𝑥𝑎 , for some 𝑎 substitutable for 𝑥 in A).

Proof:
Assume that 𝐴(𝑥) is orthodox, i.e. ⊢M T 𝐴(𝑥) ∨ T ¬𝐴(𝑥). (14.2.2)

By soundness,
M
⊢M ∃𝑥 𝐴 ⇒ |= ∃𝑥 𝐴, so for all fair functions Ξ, (Ξ, Ϙ) ⊩ T ∃𝑥 𝐴. (14.2.3)

As Ϙ is a stabilising ordinal, (Ξ, Ϙ) ⊩ ∃𝑥 𝐴. (14.2.4)

Given Definition 5.2.2 and Theorem 5.2.6, for a 𝑎, (Ξ, Ϙ) ⊩ 𝐴𝑥𝑎 . (14.2.5)
As 𝐴(𝑥) is orthodox, (Ξ, Ϙ) ⊩ T 𝐴𝑥𝑎 . (14.2.6)
M
So |= 𝐴𝑥𝑎 . (14.2.7)
So 9.2.7.6 is valid. (14.2.8)
□

14.3 The Barcan failure

As mentioned in Remark 8.2, it will be shown that the Barcan schema, in Postulate 8.1.4,
does not hold as a maxim, but only as a thesis.
The precursor to this negative result, in a truth theoretic context, is McGee’s paradox, in
(McGee 1985), which we adapt to our context. Compare (Cantini 1996, pp. 380–382)
and (Bjørdal 2012, p. 537).
First we decide upon some notions:
14.3.1. Definition: For 𝑟 in 14.3.1.5, recall Definition 6.5.1:

(1) 𝑎′ === {𝑥|𝑥 ∈ 𝑎 ∨ 𝑥 = 𝑎}.

(2) {𝑎, 𝑏} === {𝑥|𝑥 ∈ 𝑎 ∨ 𝑥 ∈ 𝑏}.
(3) {𝑎} === {𝑎, 𝑎}.
(4) 𝑎 𝜔 === {𝑢|∀𝑥(⟨∅, 𝑎⟩ ∈ 𝑥 ∧ ∀𝑦, 𝑧(⟨𝑦, 𝑧⟩ ∈ 𝑥 → ⟨𝑦′, {𝑣|𝑣 ∈ 𝑧}⟩) → 𝑢 ∈ 𝑥)}.
(5) t === {𝑥|𝑥 = 𝑟 ∧ 𝑥 ∉ 𝑥 ∧ ¬T 𝑥 ∈ 𝑥}.
(6) Use 0, 1, 2, . . . for the members of 𝜔.
(7) Let t0 === t and t𝑛+1 === {𝑣|𝑣 ∈ t𝑛 }.

47
14.3 The Barcan failure 14 SHORTCOMINGS AND REDRESSES

(8) B(t𝑖 ) === ∃𝑤(⟨𝑤, t𝑖 ⟩ ∈ t𝜔 ) → 𝑟 ∉ t𝑖

(9) B(𝑥) === ∃𝑤(⟨𝑤, 𝑥⟩ ∈ t𝜔 ) → 𝑟 ∉ 𝑥

14.3.2. Lemma: For any 𝑎, 𝑎 𝜔 is orthodox.

Proof: Adapt the proof of Theorem 13.2.3. □

14.3.3. Lemma:
(Ξ, 𝜆) ⊩ 𝑟 = 𝑟 ∧ r ∉ r ∧ ¬T r ∈ r just if 𝜆 is a limit.

Proof: For any successor ordinal 𝜒 + 1, (Ξ, 𝜒 + 1) ⊩ ¬T r ∈ r ↔ r ∈ r. Precisely at any limit

ordinal 𝜆, (Ξ, 𝜆) ⊩ r ∉ r ∧ ¬T r ∈ r. □

14.3.4. Theorem
Let 𝛼 ≺ Ϙ be a limit ordinal, and 𝛽 be 𝛼 + 𝜔 :
1. (Ξ, 𝛽) ⊩ ∀𝑥T B(𝑥)
2. (Ξ, 𝛽) ⊩ ¬T ∀𝑥B(𝑥).

Proof: 1. If (Ξ, 𝛽) ⊩ ¬∃𝑤(⟨𝑤, 𝑥⟩ ∈ t𝜔 ), it follows that (Ξ, 𝛽) ⊩ T B(𝑥) on account of

Lemma 14.3.2. If, on the other hand, (Ξ, 𝛽) ⊩ ∃𝑤(⟨𝑤, t𝑖 ⟩ ∈ t𝜔 ) we have that (Ξ, 𝛽) ⊩
T B(t𝑖 ), as there is a 𝛾 ⪰ 𝛼 + 𝑖 such that

∀𝛿(𝛼 ≺ 𝛾 ⪯ 𝛿 ≺ 𝛽 ⇒ (Ξ, 𝛿) ⊩ B(t𝑖 )).

So for any term 𝑦, (Ξ, 𝛽) ⊩ T B(𝑦), and so Ξ(𝛽) ⊢ ∀𝑥T B(𝑥).

2. Otherwise, (Ξ, 𝛽) ⊩ T ∀𝑥B(𝑥), and we would have (Ξ, 𝛿) ⊩ ∀𝑥B(𝑥) as from some
ordinal 𝛿 below 𝛽 and above 𝛼. Let 𝛿 === 𝛼 + (𝑛 + 1), for finite ordinal 𝑛 ⪰ 0, be such
an ordinal. A (Ξ, 𝛿) ⊩ B(tn ), by instantiation, this entails that (Ξ, 𝛼 + (𝑛 + 1)) ⊩ B(𝑡 𝑛 ).
As |= ∃𝑤(⟨𝑤, t𝑛 ⟩ ∈ t𝜔 ), it follows that (Ξ, 𝛼 + (𝑛 + 1)) ⊩ 𝑟 ∉ t𝑛 . As a consequence,
(Ξ, 𝛼 + 1) ⊩ 𝑟 ∉ t0 . But the latter entails (Ξ, 𝛼) ⊩ (𝑟 ≠ 𝑟 ∨ r ∈ r ∨ T r ∈ r) which
contradicts Lemma 14.3.3, as 𝛼 is presupposed to be a limit ordinal. □

14.3.5. Theorem
M
̸ = ∀𝑥T B(𝑥) → T ∀𝑥B(𝑥).
|

Proof: Theorem 14.3.4 with Definition 5.2.2 entail that for some 𝛽,

(Ξ, 𝛽) ⊮ ∀𝑥T B(𝑥) → T ∀𝑥B(𝑥).

It follows that
(Ξ, Ϙ) ⊮ T (∀𝑥T B(𝑥) → T ∀𝑥B(𝑥)),

and an appeal to Definition 5.4.3.1 finishes the proof. □

48
14.4 £ is omega-consistent 14 SHORTCOMINGS AND REDRESSES

14.3.6. Theorem:
⊬ M ∀𝑥T B(𝑥) → T ∀𝑥B(𝑥)

Proof: Appeal to soundness, i.e. in the case

M
⊢M ∀𝑥T B(𝑥) → T ∀𝑥B(𝑥) ⇒ |= ∀𝑥T B(𝑥) → T ∀𝑥B(𝑥),

and Theorem 14.3.5. □

14.3.7. Theorem: For some formula A,

⊢ ∀𝑥T A → T ∀𝑥A & ⊢ ∀𝑥T A ∧ ¬T ∀𝑥A

Proof: Let B(𝑥) in Theorem 14.3.6 be A, and combine with Postulate 8.1.4. □

14.4 £ is omega-consistent
Recall that Theorem 5.2.6 states that £ is omega-consistent.
(McGee 1985) famously isolated a rudimentary theory of truth which is consistent but
𝜔-inconsistent. (Friedman and Sheard 1987) proposed a more substantial a theory of
truth, which inherits the 𝜔-inconsistency property. (Halbach 1994) studied the Friedman
and Sheard logic, and found that its proof-theoretic strength is the same as the theory
of ramified analysis for all finite levels.
Given Theorem 14.3.7, an essential ingredient in the proof of McGee’s negative result
fails in £, viz. the statement that

∀𝑥(𝑥 ∈ 𝜔 → T A(𝑥) ) → T ∀𝑥(𝑥 ∈ 𝜔 → A(𝑥) ) (14.4.1)

in (McGee 1985, p. 399). Notice that ⊢M ∀𝑥(𝑥 ∈ 𝜔 → T A(𝑥) ) ↔ ∀𝑥T 𝑥 ∈ 𝜔 → A(𝑥) , so

14.4.1 follows from the Barcan-formula whose thesishood is denied by Theorem 14.3.7.
Moreover, exceptions to 14.4.1 in £ follow from Theorem 5.2.6 and Mcgee’s argument.

14.5 More orthodox redresses

Theorem 14.2.1 (Orthodox existential instantiation)

⊢M 𝔒( 𝐴(𝑥)) ⇒ (⊢M ∃𝑥 𝐴 ⇒ ⊢M 𝐴𝑥𝑎 for some 𝑎 substitutable for 𝑥 in A).

Proof: As on page 47. □

14.5.1. Theorem Orthodox attestor: If 𝐴(𝑥) is orthodox, then

⊢M T ∃𝑥 𝐴(𝑥) ⇒ ⊢M ∃𝑥T 𝐴(𝑥).

Proof: Appeal to Theorem 14.2.1, and existential generalization. □

49
14.5 More orthodox redresses 14 SHORTCOMINGS AND REDRESSES

14.5.2. Theorem The Barcan formula holds for orthodox formulas:

⊢M 𝔒(𝐵(𝑥)) ⇒ ⊢M (∀𝑥T B(𝑥) → T ∀𝑥B(𝑥)).

Proof: As on page 33. □

50
14.5 More orthodox redresses 14 SHORTCOMINGS AND REDRESSES

51
 15 CLASSICALITIES AND DEVIATIONS

15 Classicalities and deviations

On a signalé beaucoup d’antinomies, et le désaccord a subsisté, personne n’a été convaincu; d’une
contradiction, on peut toujours se tirer par un coup de pouce! Je veux dire par un distinguo.

Henri Poincaré

The Grundlagenkrise which struck the mathematical and philosophical communities as a

consequence of the paradoxes, showed one could not presuppose all pretheorethically
plausible comprehension principles in set theory or semantics.
In the following some facts which relate to desiderata fulfilled by £ will be expressed.
The reader may compare with the desiderata of (Leitgeb 2007) and (Sheard 2003), or
others, concerning theories on paradoxe. Some of the facts on desiderata follow from
§14.4, §15.2 and §15.3.

15.1 Facts on desiderata met by £

15.1.1. Fact: There are no type restricions imposed, and there is no language hierarchy.
15.1.2. Fact: Truth is compositional over ⊢M , and over similar set theoretic contexts. But
it is not compositional over ⊢, as there are cases such that ⊢ A and ⊢ B but not ⊢ A ∧ B.
15.1.3. Fact: Truth is a set, and so truth is as well a predicate. So it is a consequence
from the alethic comprehension principle of §10 that truth-paradoxes and set-paradoxes
are treated in the same way in £.
15.1.4. Fact: On account of results in §15.3, £ is classic in the sense that ⊢M A only if
classical logic does not prove ¬A, and if classical logic proves A then ⊢M A. Moreover, ⊢ B
if B is a thesis of classical logic, and if ⊢ B then classical logic does not prove ¬B.
15.1.5. Fact: £ is unswerving so that if A is a paradoxical sentence, then ⊢ A or ⊢ ¬A.
15.1.6. Fact: The outer veridical logic of £ is the set of theses which encapsulate truth
statements. ⊢M T A ∨ ¬T A is for example an instance of The Law of Excluded Middle
of classical logic in the outer veridical logic of £. The inner veridical logic of £ is the set
of theses which are encapsulated by a truth predicate. ⊢M T A ∨ ¬A is for example an
instance of The Law of Excluded Middle of classical logic in the inner verdidical logic of
£. It is a consequence of Fact 15.1.4 that the inner and outer veridical logics of £ are
classical, as they should be.
15.1.7. Fact: The truth of the truth conditionals in §15.2 has the consequence that the
outer veridical and inner veridical logics of T , see Fact 15.1.6, coincide in £, in the sense
of Definitions 15.2.1.7 and 15.2.1.10.
15.1.8. Fact: As related in §9, £ has novel inferential modes. The conjunction of these
may seem to be an amputation of the classical inferential principle modus ponens. But
they are in reality an extension of the classical inference rule modus ponens, as the maxim

52
15.2 The truth–conditionals 15 CLASSICALITIES AND DEVIATIONS

mode 9.2.5 serves all the purposes as modus ponens serves in classical logics, and all
classical logical theses are maxims of £The inference modes £ has beyond the maxim
mode helps engender novel theses.
15.1.9. Fact: A naive desideratum is that £ should obtain all truth-biconditionals, as in
Definitions 15.2.1.1 and 15.2.1.7, with their weak counterparts, by means of the inference
modes which £ endorses, as per §9. £ compensates for the fact that the statements of
Definitions 15.2.1.1, 15.2.1.2, 15.2.1.3 and 15.2.1.4 are not true with the truth of the
statements of Definitions 15.2.1.5 and 15.2.1.6, and with the fact that the inferential
modes exhibited in Definitions 15.2.1.7 and 15.2.1.10 can be used. A consequence of
this is that revenge paradoxicalities are not a threat. For more on this, see §18.
15.1.10. Fact: By §14.4, £ is omega-consistent, so it allows for standard interpretations.

15.2 The truth–conditionals

15.2.1. Definition:
1 Hale material truth adequacy: ⊢M T 𝐴 ↔ 𝐴
2 Hale material truthwards adequacy: ⊢M T 𝐴 ← 𝐴
3 Hale material truthly adequacy: ⊢M T 𝐴 → 𝐴
4 Weak material truth adequacy: ⊢ T 𝐴 ↔𝐴
5 Weak material truthwards adequacy: ⊢ T 𝐴 ← 𝐴
6 Weak material truthly adequacy: ⊢ T 𝐴 →𝐴
7 Hale formal truth adequacy: ⊢M T 𝐴 ⇔ ⊢M𝐴
8 Hale formal truthwards adequacy: ⊢M T 𝐴 ⇐ ⊢M𝐴
9 Hale formal truthly adequacy: ⊢M T 𝐴 ⇒ ⊢M𝐴
10 Weak formal truth adequacy: ⊢ T 𝐴 ⇔⊢ 𝐴
11 Weak formal truthwards adequacy: ⊢ T 𝐴 ⇐⊢ 𝐴
12 Weak formal truthly adequacy: ⊢ T 𝐴 ⇒⊢ 𝐴
15.2.2. Exercise £ obeys the formal and as well the weak material truthwards and truthly
adequacies of Definition 15.2.1. The other adequacies fail on acount of paradoxicalities.

15.3 £ is classic and paraclassical, but it is not paraconsistent

Let T be a theory.
15.3.1. Definition: T is adjunctive just if ⊢ 𝐴 & ⊢ 𝐵 ⇒ ⊢ 𝐴 ∧ 𝐵.
15.3.2. Definition: T is dejunctive just if ⊢ 𝐴 ∧ 𝐵 ⇒ ⊢ 𝐴 & ⊢ 𝐵.

53
15.3 £ is classic and paraclassical, but it is not 15
paraconsistent
CLASSICALITIES AND DEVIATIONS

15.3.3. Definition: T is cosistent just if for no 𝑝, 𝑇 ⊢ 𝑝 and 𝑇 ⊢ ¬𝑝.

15.3.4. Definition: T is consistent just if for no 𝑝, 𝑇 ⊬ 𝑝 ∧ ¬𝑝
15.3.5. Definition: T is contrasistent just if it is not cosistent.
15.3.6. Definition: T is contradictory just if it is inconsistent.
15.3.7. Definition: Let 𝜏 be classical logic.
15.3.8. Definition A is an antithesis of T just if ¬A is a thesis of T.
15.3.9. Definition S is a sedation of T iff no thesis of S is an antithesis of T.
15.3.10. Definition X is an extension of T just if all theses of T are theses of X.
15.3.11. Definition: Let 𝜏 be classical logic.
15.3.12. Fact. 3
That X is a proper extension of T holds just if
X is an extension of T and T is not an extension of X .

15.3.13. Definition Progressive, moderate and classic theories:

1 T is progressive just if it is a proper extension of 𝜏.
2 T is moderate just if it is a sedation of 𝜏.
3 T is classic just if it is progressive and moderate.
15.3.14. Definition T is coherent just if it is classic.
15.3.15. Definition T is extraclassical just if it is classic and contrasistent.
15.3.16. Definition T is extracoherent just if it is coherent and contrasistent.
15.3.17. Lemma £ is an extension of 𝜏.

Proof: Appeal to §6.4. □

15.3.18. Lemma 𝜏 is not an extension of £.

Proof: Given the solution to Exercise 8.3, £ has the paradoxical theses r ∈ r and r ∉ r. But
r ∈ r and r ∉ r are not theses of classical logic. □

15.3.19. Lemma £ is progressive.

Proof: £ is a proper extension of 𝜏 given Fact 15.3.12, Lemma 15.3.12 and Lemma 15.3.18.
An appeal to Definition 15.3.13.1 suffices to finish the proof. □

15.3.20. Lemma £ is moderate.

3For the following definition, and the notions involved here, compare with (Bjørdal 2015, p. 511).

54
15.4 Incompatability and complementarity 15 CLASSICALITIES AND DEVIATIONS

Proof: If 𝜏 proves ¬A, ⊢M ¬A as £ is progressive. Axiom 6.2.2.4, justified by Postulate

6.2.2.4, entails that if 𝜏 proves ¬A, ⊢ ¬A & ⊬ A. So a fortiori, if 𝜏 proves ¬A, ⊬ A. By
contraposition, if ⊢ A then 𝜏 does not prove ¬A. A is arbitrary, so no thesis of £ is an
antithesis of classical logic 𝜏. Consequently, £ is a sedation of 𝜏. An appeal to Definition
15.3.13.2 finishes the proof. □

15.3.21. Theorem £ is classic.

Proof: From Definition 15.3.13.3, as £ is progressive and moderate given Lemma 15.3.19
and Lemma 15.3.18. □

15.3.22. Exercise T is contrasistent just if for some 𝑝, T ⊢ 𝑝 and T ⊢ ¬𝑝.

15.3.23. Exercise If T is contrasistent and adjunctive then T is contradictory.
15.3.24. Exercise If T is contradictory and dejunctive then T is contrasistent.
15.3.25. Exercise If T is dejunctive and adjunctive, T is contradictory iff contrasistent.
15.3.26. Exercise If T is dejunctive and adjunctive, T is cosistent just if consistent.
15.3.27. Exercise Adjunction is not a valid inference mode in £.
15.3.28. Exercise £ is extraclassic and extracoherent, and so contrasistent.
15.3.29. Exercise Paraconsistent theories are not classic.
15.3.30. Theorem £ is not paraconsistent.
Proof: Given Theorem 15.3.21 and Exercise 15.3.29, paraconsistent theories are not
classic. But £ is classic by Theorem 15.3.21. □
15.3.31. Remark The pairs consistency & cosistency and contrasistency & contradiction
conflate in classical contexts, for classical systems are adjunctive and dejunctive.
15.3.32. Remark With proper comprehension, most paraconsistent theories are not even
moderate, as then some contradiction is a thesis.
15.3.33. Remark. The well-known non-adjunctive paraconsistent logic of (Jaskowski
1999) and (Jaskowski 1948), is moderate even with liberal comprehension principles.
But is it not conservative, and so not classic.

15.4 Incompatability and complementarity

15.4.1. Definition (Incompatability) The theses A and B of a consistent theory T are
incompatible just if T proves A, B, and ¬(A ∧ B).
15.4.2. Theorem (£ has incompatible theses) By the result of Exercise 8.3, £ proves
R ∈ R and R ∉ R. But £ is conservative, given §15.3. So the theses R ∈ R and R ∉ R of £
are incompatible, for given its conservativeness, £ proves ¬(R ∈ R ∧ R ∉ R).

55
15.4 Incompatability and complementarity 15 CLASSICALITIES AND DEVIATIONS

15.4.3. Definition (Complementarity) A and ¬A in a theory T are complementary just if

they are incompatible theses of T.
15.4.4. Corollary £ has complementary theses

56
 16 THE LIAR IS RUSSELL’S CONDITION ON HIS SET

16 The Liar is Russell’s condition on his set

Thus mathematics may be defined as the subject in

which we never know what we are talking about, nor
whether what we are saying is true.

Bertrand Russell

Frank Ramsey argued, in (Ramsey 1925, p. 20), that there is an essential difference
between syntactical paradoxes which “involve only logical or mathematical terms such
as class and number”, and semantic paradoxes, which “. . . are not purely logical, and
cannot be stated in logical terms alone; for they all contain some reference to thought,
language, or symbolism”.
Ramsey considered Russell’s paradox a canonical representative of syntactic paradoxes,
and the Liar he considered an archetypical semantic paradox.
In (A. A. Fraenkel and Bar-Hillel 1958, p. 5), the authors adjudged:
“Since (Ramsey 1925) it has become customary to distinguish between log-
ical and semantic (sometimes also called syntactic or epistemological) anti-
nomies.”
It is here nevertheless argued that one should take paradoxes, as the Liar-paradox, to be
so inextricably intertwined with set theoretical paradoxes, that one should not consider
them to be different in kinds.
Others reached the same conclusion, but on the basis of considerations different from
the ones adduced further below:
(Scott 1974)(1967) argued that the Zermelo axioms were justified by type theoretic rea-
soning:
“The truth is that there is only one way of avoiding the paradoxes: namely,
the use of some form of the theory of types. That was at the basis of both
Russell’s and Zermelo’s intuitions. Indeed the best way to regard Zermelo’s
theory is as a simplification and extension of Russell’s. (We mean Russell’s
simple theory of types, of course.) The simplification was to make the types
cumulative.” (Scott 1974)(208)
Alonzo Church, in (Church 1976), virtually equated Russell’s theory of types and Alfred
Tarski’s resolution of the Liar paradox, as he stated:
“In the light of this it seems justified to say that Russell’s resolution of the
semantical antinomies is not a different one than Tarski’s but is a special case
of it.”(Church 1976)(756)
The interest of Scott’s and Tarski’s points of view, for our purposes here, is that they,
jointly, take Tarski’s resolution of the alleged semantic paradoxes to be the same as Rus-

57
 16 THE LIAR IS RUSSELL’S CONDITION ON HIS SET

sell’s, and Zermelo’s, resolution of the, allegedly syntactical, set theoretic paradoxes.
In £ there are bridge principles, as for example per Theorem 16.1 and Definition 16.4, be-
tween given, supposedly syntactical paradoxes, and supposedly semantical paradoxes.
16.1. Theorem: There is a liar sentence L given by ⊢M L ↔ ¬T L .

Proof: By alethic comprehension,

⊢M r ∈ r ↔ T r ∉ r) . (16.2)

By negating both sides of the biconditional in 16.2, we get

⊢M ¬T r ∉ r) ↔ r ∉ r. (16.3)

16.4. Definition:
L === r ∉ r,

Substituting with L of Definition 16.4 in equation 16.3 gives the more canonical form for
the Liar sentence:

⊢M L ↔ ¬T L . (16.5)
□
16.5 is resolved as Russell’s paradox.
16.6. Proposition: Liar sentences, and variants, with provenances from classical Greek
philosophy, should be taken as given by maxims of Theorems as 16.1.
16.7. Theorem: ⊢ L, ⊢ ¬L, ⊢ T L , ⊢ T ¬L , ⊢ ¬T ¬L and ⊢ ¬T L .

Proof: We know that ⊢ r ∈ r and ⊢ r ∉ r, so from Definition 16.4, ⊢ L and ⊢ ¬L. Finish with
9.2.4.1 and 9.2.4.2. □

16.8. Observation: Each element in variety [L, T L , ¬T ¬L ] is incompatible with any

member of [¬L, T ¬L , ¬T L ] in £, and vice versa. Moreover, each element in variety
[L, T L , ¬T ¬L ] is complementary to precisely one member of [¬L, T ¬L , ¬T L ]
in £, and vice versa.

58
 17 LIBRATIONIST INCOMPLETENESS PHENOMENA

17 Librationist incompleteness phenomena

Kurt Gödel, when colleague John Bachall presented himself as a physicist at a

Princeton faculty dinner: "I don’t believe in natural science."

(Regis 1988, p. 58)

It is of interest to note that the usual proofs of Gödel’s incompleteness theorem typically
presuppose the theory is cosistent, so may consistent contrasistent theories in some
sense possibly finesse the limitation? Let us explore this cursorily in the context of £.
Observe first that T maximally obeys the Hilbert-Bernays-Löb derivability conditions in
the sense that for all A and B,

1. ⊢M A ⇒ ⊢M T A .
2. ⊢M T A → T T A .
3. ⊢M T A → B → (T A → T B ).

Take ⊢M T A to express not only that A is a true maxim, but as well that A is provable
as a maxim. Take therefore a thesis as ⊢M ∃𝑥¬T 𝑥 to expresses that £ is not trivial, and
⊢M ¬∃𝑥T T 𝑥 ∧ ¬T 𝑥 to express that £ does not prove a contradiction, or inconsistency.
Let us at this point restate 16.5: ⊢M L ↔ ¬T L .
If one supposes ⊢M L it follows that ⊢M T L from 1, and ⊢M ¬T L from 16.5. If one
supposes ⊢M ¬L it follows that ⊢M T ¬L from 1, and ⊢M T L from 16.5, so that one with
Postulate 6.5.2.2 has ⊢M T ¬L and ⊢M ¬T ¬L . As £ is maximally adjunctive so that
[⊢M A & ⊢M B] ⇒ ⊢M A ∧ B, in either case ⊢M T ¬L ∧ ¬T ¬L . So neither ⊢M ¬L nor ⊢M L,
but rather ⊢ ¬L and ⊢ L. So the sentence ¬L which is maximally incomplete, in the sense
that neither ¬L nor L is a maxim, is nevertheless a minor thesis.
It was pointed out, by means of Exercises 5.5.2 and 5.5.3, and Definition 5.6.1.3, that
neither ⊢ 𝑠 ∈ 𝑠 nor ⊢ 𝑠 ∉ 𝑠, if 𝑠 = {𝑥|𝑥 ∈ 𝑥}. But this is not a genuine incompleteness, as
neither |= 𝑠 ∈ 𝑠 nor |= 𝑠 ∉ 𝑠.
The author does not know that there is a sentence A such that ̸⊢M A and such that we
should want that ⊢M A, nor that there is a sentence B such that ⊬ B and such that we
should want that ⊢ B. Certainly, if 𝐶 is the statement that there is a certain inaccessible
cardinal larger or equal to the first hypothetized 1-inaccesible cardinal, it will be the case,
with the assumptions made, that even ̸⊢M CV , where V is as in §25, and the notation CV
as in Definition 4.5.22. It is not obvious to the author, however, at this point, that we
should want ⊢M CV . But for the record, if 𝐷 is the statement that there is a 0-inaccesible
cardinal, and that there for any 0-inaccesible is a larger 0-inaccesible, then ⊢M DV .

59
17 LIBRATIONIST INCOMPLETENESS PHENOMENA

60
 18 THE REFLECTIVE THEORY OF COMPREHENSION

18 The reflective theory of comprehension

Two paradoxes are better than one;

they may even suggest a solution.

Edward Teller

One may as a first approximation take the reflective theory of comprehension, i.e. of truth
and abstraction, which is supported by the librationist set theory, with its alethic compre-
hension principle, to be expressed above all by the inference modes of §9, and especially
the simple inference modes, for truth, and for set abstracts via alethic comprehension.

18.1 Responsible naiveté without revenge

The revenge problem is avoided as £ is unswerving, and has complementary theses"
Consider the Liar sentence L of Equation 16.5. If ⊢ L, it follows that ⊢ ¬T L via the
equation. However, it as well follows that ⊢ T L from ⊢ L and inference mode 9.1.1.1.
So ⊢ T L and ⊢ L state things as they are. Moreover, given ⊢ ¬T L , it follows that ⊢ ¬L
via inference mode 9.1.1.9. So ⊢ ¬L and ⊢ ¬T L state things as they are.
It is not a desirable option to prefer ⊬ L and ⊬ ¬L, for |= L and |= ¬L, and one should
attempt to have ⊢ B whenever |= B. Moreover, it has, as discussed in Fact 15.1.5, been
presupposed as desideratum that £ be unswerving, and decide paradoxical sentences.

18.2 Argumenta ad paradoxo

That an assumption in £ has the consequence that ⊢ A and ⊢ ¬A does not suffice as a
proof by contradiction against the assumption. Instead, if the considerations leading to
⊢ A and ⊢ ¬A cannot be extended to arrive at ⊢ A∧¬A, they just constitute an argumentum
ad paradoxo to show that ⊢ A and ⊢ ¬A are complementary theses of £.

61
18.2 Argumenta ad paradoxo 18 THE REFLECTIVE THEORY OF COMPREHENSION

62
 19 MANIFESTATIONS

19 Manifestations

There are very few theorems in advanced analysis which have been demonstrated in a logically
tenable manner. Everywhere one finds this miserable way of concluding from the special to the
general and it is extremely peculiar that such a procedure led to so few of the so-called paradoxes.

Niels Henrik Abel

We explain the manifestation set construction in §19.1, and will as from §25 see that
it facilitates £’s ability to be extended with strong set theoretic principles. In §19.2 we
show how we may obtain Quine atoms via orthodox manifestation sets. The foci in the
succeeding sections will be upon negative results: In §19.3 we account for the auto-
combative paradox. Next, in §20, we elucidate the virtually universal paradoxicality of
power sets. Finally the failure of extensionality in £ is discussed in §21, where it is shown
that all orthodox sets are distinct from, as well as co-extensional with infinitely many co-
extensional and pairwise distinct orthodox sets.

19.1 The manifestation sets

For the following construction, cfr. (Bjørdal 2012)(345–46), (Cantini 1996)(76), (Visser
1989)(695–96) and earlier literature referred to there. One may, plausibly, find that
Roger’s theorem and Kleene’s second recursion theorem are related, but the proof that
there are manifestation sets does not rely upon any presuppositions of computability.
19.1.1. Definition Kuratowskian ordered pairs:

⟨𝑎, 𝑏⟩ === {{𝑎}, {𝑎, 𝑏}}

19.1.2. Definition The manifestation set A of formula A(𝑣 0 , 𝑣 1 ):

(1) 𝑣𝜂𝑏 === ∃𝑣 2 (𝑣 2 = ⟨𝑣, 𝑏⟩ ∧ 𝑣 2 ∈ 𝑏)
{𝑣|𝑣𝜂𝑣 1 }
(2) 𝔞 === {𝑣 2 |∃𝑣 0 , 𝑣 1 (⟨𝑣 0 , 𝑣 1 ⟩ = 𝑣 2 ∧ A(𝑣 0 , 𝑣 1 )𝑣 1 }
(3) A === {𝑣|𝑣𝜂𝔞}

19.1.3. Theorem Manifest comprehension, for the manifestation set in Definition 19.1.2.3:

⊢M ∀𝑣(𝑣 ∈ A ↔ T T A(𝑣, A))

⊢M ∀𝑣(𝑣 ∈ A ↔ T T A(𝑣, A))

Proof: From Definition 19.1.2.3 and alethic comprehension,

⊢M 𝑐 ∈ A ↔ T 𝑐𝜂𝔞.

63
19.2 Quine atoms 19 MANIFESTATIONS

As a consequence of Definition 19.1.2.1 we have

⊢M T 𝑐𝜂𝔞 ↔ T ∃𝑣 2 (𝑣 2 = ⟨𝑐, 𝔞⟩ ∧ 𝑣 2 ∈ 𝔞)

From the two previous steps, Definition 19.1.2.2, alethic comprehension and 9.2.5 we
have

⊢M 𝑐 ∈ A ↔ T ∃𝑣 2 (𝑣 2 = ⟨𝑐, 𝔞⟩ ∧ 𝔗∃𝑣 0 , 𝑣 1 (⟨𝑣 0 , 𝑣 1 ⟩ = 𝑣 2 ∧ A(𝑣 0 , 𝑣 1 )𝑣 1

{𝑣|𝑣𝜂𝑣 1 }
))

It follows, by means of the theory of identity, that

⊢M 𝑐 ∈ A ↔ T T A(𝑐, 𝑣 1 )𝑣A1 ,

so that, on account of Definition 19.1.2.3 and Definition 4.5.18,

⊢M 𝑐 ∈ A ↔ T T A(𝑐, A).

Finish with universal generalization. □

19.1.4. Corollary Orthodox manifestation:

If A(𝑣 0 , 𝑣 1 ) is orthodox, ⊢M ∀𝑣 𝑣 ∈ A ↔ A(𝑣, A) .

19.1.5. Theorem(Comprehension for manifestation set with parameters)

Some manifestation sets have parameters, so if orthodox A has the free variables in 𝑣®:

⊢M ∀𝑣∀𝔳® (𝑣 ∈ A ↔ A( 𝔳®, 𝑣, A)).

Proof: Adjust Definition 19.1.2. For the notation, recall Definition 4.5.20. □

19.2 Quine atoms

The most elementary Quine atom is the manifestation set = of formula 𝑣 0 = 𝑣 1 . By means
of manifest comprehension,

⊢M ∀𝑣(𝑣 ∈ = ↔ T T 𝑣 = = ). (19.2.1)

As identity is an orthodox relation,

⊢M ∀𝑣(𝑣 ∈ = ↔ 𝑣 = = ). (19.2.2)

As identity is an equivalence relation,

64
19.3 The autocombatant 19 MANIFESTATIONS

⊢M = = =. (19.2.3)

So from equations 19.2.2 and 19.2.3,

⊢M = ∈ = (19.2.4)

19.2.5. Exercise: Prove that there are infinitely many distinct Quine atoms.

19.3 The autocombatant

In contrast to orthodox manifestation sets, many are paradoxical. This is for example the
case with the following quite heretical manifestations set ∉, which generates an infinity
of incompatible and complementary theses.
19.3.1. Theorem (The autocombative truths) ∉ is the manifestation set of formula 𝑣 0 ∉ 𝑣 1 ,
so that:
|= ∀𝑣(𝑣 ∈∉ ) & |= ∀𝑣(𝑣 ∉∉ ).

Proof: On account of Theorem 19.1.3:

⊢M ∀𝑣(𝑣 ∈∉ ↔ T T 𝑣 ∉∉),

so that by soundness
M
= ∀𝑣(𝑣 ∈∉ ↔ T T 𝑣 ∉∉).
|

If 𝜆 is any limit below the closure ordinal Ϙ, we will, for any term 𝑎, and any fair function
Ξ, have that (Ξ, 𝜆) ⊩ 𝑎 ∉∉; otherwise a contradiction would follow as 𝑎 ∉∉ would hold at
succeeding successor ordinals 𝜎, 𝜎 + 1 and 𝜎 + 2 below 𝜆. Consequently, we for such a
limit 𝜆 as well have that (Ξ, 𝜆 + 2) ⊩ 𝑎 ∈∉. From 5.2.2.2 we have that (Ξ, 𝜆) ⊩ ∀𝑣(𝑣 ∉∉) and
(Ξ, 𝜆 + 2) ⊩ ∀𝑣(𝑣 ∈∉)). As a result, (Ξ, Ϙ) ⊩ ¬T ¬∀𝑣(𝑣 ∈∉)) and (Ξ, Ϙ) ⊩ ¬T ¬∀𝑣(𝑣 ∉∉)). The
proof finishes by invoking Definitions 5.4.2.3 and 5.4.3.3. □

65
19.3 The autocombatant 19 MANIFESTATIONS

66
 20 POWERSETS ARE PARADOXICAL LEST AS P ({𝑉 |𝑉 = 𝑉 })

20 Powersets are paradoxical lest as P ({𝑣|𝑣 = 𝑣})

Das Wesen der Mathematik liegt in ihrer Freiheit.

Georg Cantor

Use standard notation, so that ⊢M 𝑎 ⊂ 𝑏 ↔ ∀𝑥(𝑥 ∈ 𝑎 → 𝑥 ∈ 𝑏), and posit

20.1. Definition The power set of 𝑎:

P (𝑎) === {𝑣|𝑣 ⊂ 𝑎}.

It turns out that a power set is paradoxical unless it is the power set of a maximally filled
set 𝑏 for which ⊢M ∀𝑥(𝑥 ∈ 𝑏).
20.2. Definition The universal set:

U === {𝑣|𝑣 = 𝑣}

20.3. Theorem 𝑎 is paradoxical if ̸⊢M ∀𝑥(𝑥 ∈ 𝑎 ↔ 𝑥 ∈ U) :

Proof: We use a case distincion to provide a distinct proof for the case where ⊢𝑚 ∃𝑣(𝑣 ∉ 𝑎).
(1) If ⊢M ∃𝑣(𝑣 ∉ 𝑎), use the autocombatant ∉, of Theorem 19.3.1, for which

⊢ ∀𝑣(𝑣 ∈∉) & ⊢ ∀𝑣(𝑣 ∉∉).

In this case ⊢∉∉ P (𝑎) and ⊢∉∈ P (𝑎), so P (𝑎) is paradoxical.

(2) If ⊢𝑚 ∃𝑣(𝑣 ∉ 𝑎), ⊢ U ∈ P (𝑎) and ⊢ U ∉ P (𝑎), so P (𝑎) is paradoxical. □

67
20 POWERSETS ARE PARADOXICAL LEST AS P ({𝑉 |𝑉 = 𝑉 })

68
 21 NON-EXTENSIONALITY AND URSETS

21 Non-extensionality and Ursets

It is impossible to be a mathematician without being a poet in soul.

Sofia Kovalevskaya

The principle of extensionality’s failure in type free theories is well known, and many have
contributed to the deposit of knowledge.
Let us first posit
21.1. Definition The principle of extensionality:
e
⊢M 𝑎 = 𝑏 → 𝑎 = 𝑏.

A particularly easy proof of the failure of the extensionality principle in £ is obtained by

making use of the fact that for any limit ordinal 𝜆,
e
Ξ(𝜆 + 1) ⊩ {𝑣|𝑣 = 𝑣} = {𝑣|𝑣 ∉ 𝑣} ∧ {𝑣|𝑣 = 𝑣} ≠ {𝑣|𝑣 ∉ 𝑣}.

e
As a consequence, there are sets 𝑎 and 𝑏 such that |̸ = 𝑎 = 𝑏 → 𝑎 = 𝑏, and so it follows, a
M e e M e
fortiori, that |̸ = 𝑎 = 𝑏 → 𝑎 = 𝑏, But ⊢M 𝑎 = 𝑏 → 𝑎 = 𝑏 ⇒ |= 𝑎 = 𝑏 → 𝑎 = 𝑏 is a soundness
e
requirement, so that ̸⊢M 𝑎 = 𝑏 → 𝑎 = 𝑏.
(Gilmore 1974) showed that a partial set theory it proves that there is an orthodox set
e
𝑎 such that 𝑎 = ∅ and 𝑎 ≠ ∅. (Bjørdal 2012, p. 345) relates Lev Gordeev’s more concise
proof of the same result as Gilmore’s, in the context of Explicit Mathematics, and some
on why it was published in (Beeson 1985), with acknowledgement.
Define Gordeev’s set with the manifestation theorem 19.1.3, so that one may posit
21.2. Definition (Via manifestation)∀𝑥(𝑥 ∈ 𝑔¤ ↔ T T (𝑥 = ∅ ∧ 𝑥 = 𝑔).
¤
21.3. Theorem: [Gordeev] 𝑔¤ is (i) orthodox, so ⊢M 𝑥 ∈ 𝑔¤ ↔ (𝑥 = ∅ ∧ 𝑥 = 𝑔),
¤ (ii) empty
and (iii) distinct from ∅.

Proof: As the proof of Theorem 4 in (Bjørdal 2012, p. 345): (i) 𝑔¤ is orthodox, on account
of the theory of identity. (ii) As ⊢M 𝑥 ∈ 𝑔¤ → (𝑥 = ∅∧𝑥 = 𝑔),
¤ ⊢M 𝑥 ∈ 𝑔¤ → 𝑔¤ = ∅, so 𝑔¤ is empty.
(iii) 𝑔¤ ≠ ∅, for else 𝑔¤ = { 𝑔}
¤ on account of Theorem 21.3 (i), which contradicts (ii). □

(Cantini 1996)(74), relates a proof, by Pierluigi Minari that we for any orthodox set 𝑎
may find a distinct orthodox set 𝑏 such that 𝑎 and 𝑏 are nevertheless co-extensional.
Theorem 5 (ii) in (Bjørdal 2012)(346), whose proof was left as an exercise, states the
result that Minari’s construction can be generalized, as in Theorem 21.4. This result
appears to be the most general non-extensionality result available, and we do not relate

69
 21 NON-EXTENSIONALITY AND URSETS

proofs of other results which are consequences.4 As Theorem 21.4 is proved below, the
mentioned exercise is solved.
21.4. Theorem: For orthodox set 𝑏, there are infinitely many pairwise distinct orthodox
and co-extensional sets, which are all co-extensional with 𝑏 and distinct from 𝑏.

Proof: Let orthodox 𝜐1 be given, and let 𝜐𝑛+1 be the manifestation set of

𝑖=𝑛
Û  𝑖=𝑛
Û 𝑖=𝑛
Ü 
𝑣 1 ∈ 𝜐𝑖 ∧ 𝜐𝑖 ≠ 𝑣 1 ∨ 𝜐𝑖 ∉ 𝜐𝑖 ∧ 𝜐𝑖 = 𝑣 1
𝑖=1 𝑖=1 𝑖=1

so that, by manifest comprehension and the logic of identity,

𝑖=𝑛
Û 𝑖=𝑛
 Û 𝑖=𝑛
Ü 
∀𝑣(𝑣 ∈ 𝜐𝑛+1 ↔ ( 𝑣 ∈ 𝜐𝑖 ∧ 𝜐𝑖 ≠ 𝜐𝑛+1 ∨ 𝜐𝑖 ∉ 𝜐𝑖 ∧ 𝜐𝑖 = 𝜐𝑛+1 ).
𝑖=1 𝑖=1 𝑖=1

If 𝑖=𝑛 (𝜐𝑖 = 𝜐𝑛+1 ), it follows that ⊢M 𝜐𝑛+1 ∈ 𝜐𝑛+1 ↔ 𝜐𝑛+1 ∉ 𝜐𝑛+1 , which is impossible. So
Ô
Ó𝑖=𝑛𝑖=1 e
𝑖=1 (𝜐𝑖 ≠ 𝜐 𝑛+1 ). Clearly, 𝑖=1 (𝜐𝑖 = 𝜐 𝑛+1 ). The process can be iterated, so we are done. □
Ó𝑖=𝑛

4Notice, however, that Theorem 21.4 may straightforwardly be extended to the result that there are 𝜔 · 2
pairwise distinct and co-extensional orthodox empty sets, and so on for larger countable ordinals.

70
 22 NAMES AND SETS OF URELEMENTE TO TRANSFINITE ORDERS

22 Names and sets of Urelemente to transfinite orders

Os números são as regras dos seres, e a

matemática é o regulamento do mundo.
Francisco Gomes Teixeira

22.1. Definition:
(1) Let 𝜐1 = 𝑔,
¤ as in Definition 21.2.
(2) For any 𝑛 ∈ Ω+ , 𝑛 is the 𝜐𝑛 of Theorem 21.4.
(3) 𝜐 𝜔 === {𝑥|∀𝑦(𝜐0 ∈ 𝑦 ∧ ∀𝑧(𝜐 𝑧 ∈ 𝑦 → 𝜐 𝑧+1 ∈ 𝑦) → 𝑥 ∈ 𝑦)}.
22.2. Theorem: [For manifestation set 𝜐𝜔 ]

⊢M ∀𝑥(𝑥 ∈ 𝜐𝜔 ↔ (𝑥 = 𝜐𝜔 ∧ ∃𝑦(𝑦 ∈ 𝜐 𝜔 ∧ 𝑥 = 𝑦))).

22.3. Remark: 𝜐𝜔 in Theorem 22.2 is orthodox as 𝜐 𝜔 in Definition 22.1.3 is orthodox.

22.4. Theorem: (i) 𝜐𝜔 is empty. (ii) ⊢M ∀𝑦(𝑦 ∈ 𝜐 𝜔 → 𝜐𝜔 ≠ 𝑦)}.

Proof: (i) If 𝑎 were an element of 𝜐𝜔 , ⊢M 𝑎 = 𝜐𝜔 ∧ ∃𝑦(𝑦 ∈ 𝜐 𝜔 ∧ 𝑎 = 𝑦). Given Theorem

21.4, all members of 𝜐 𝜔 are empty sets. Consequently, if 𝑎 were an element of 𝜐𝜔 then
𝜐𝜔 would be an empty set. So 𝜐𝜔 is an empty set. (ii) A rendition of Theorem 22.2 is
⊢M ∀𝑥(𝑥 ∉ 𝜐𝜔 ↔ (𝑥 = 𝜐𝜔 → ∀𝑦(𝑦 ∈ 𝜐 𝜔 → 𝑥 ≠ 𝑦))), so, as a consequence,

⊢M 𝜐𝜔 ∉ 𝜐𝜔 → (𝜐𝜔 = 𝜐𝜔 → ∀𝑦(𝑦 ∈ 𝜐 𝜔 → 𝜐𝜔 ≠ 𝑦)).

The proof finishes by invoking the maxim mode 9.2.5, as ⊢M 𝜐𝜔 ∉ 𝜐𝜔 on account of (i). □

22.5. Definition:
(1) 𝑐 is an urset just if 𝑐 = 𝑛 , for some 𝑛 ∈ Ω+ .
(2) In accordance with Definition 4.5.4, 2222221 , 222221 , 22221 , 2221 , 221 , 21
and 1 are the symbolic ursets: # , c¥ , 𝝇 , ∀ , ↓ , v¥ , • .
(3) The symbolic ursets are the atomic names, which denote the primitive symbols.
(4) Recall Definition 4.5.5.2 of ℓ(𝑛0 ) === ⌊𝑙𝑜𝑔2 (𝑛0 + 1)⌋, which uses 𝑙𝑜𝑔2 and the floor
function ⌊ ⌋, to define the length ℓ(𝑛0 ) of the bijective base-2 cipher needed to
express a given chiffer 𝑛0 .

71
 22 NAMES AND SETS OF URELEMENTE TO TRANSFINITE ORDERS

(5) So

ℓ(•) = ℓ(1) = 1
ℓ( 𝑣¥ ) = ℓ(21) = 2
ℓ(↓) = ℓ(221) = 3
ℓ(∀) = ℓ(2221) = 4
ℓ(𝜍) = ℓ(22221) = 5
¥ = ℓ(222221) = 6
ℓ( 𝑐)
ℓ(#) = ℓ(2222221) = 7.

(6) Given Definition 4.5.5, the joining of names is defined by positing

𝑛0 Z 𝑛1 === 𝑛0 𝑛1 === 𝑛0 ⌢ 𝑛1 === 𝑛0 · 2ℓ(𝑛1 ) + 𝑛1 .

(7) Given Definition 22.5.5, we may use Definition 22.5.6 to construe composite names
gramatically correct by joining names whilst obeying the formation rules of §4.5.
(8) For good ∈ {symbol, symbol string, variable, formula, constant, term, sentence},
N is a good name just if N is a good.
22.6. Caveat: In formula ∀𝑣T 𝐴 , is a term operating formula forming operator, so
the evaluation of T 𝐴 𝑣 is comparable with □ 𝐴𝑣𝑏 , where □ is any formula operating for-
𝑏

mula forming operator. So, for example, T 𝑣 = 𝑣 𝑣𝑏 is T 𝑏 = 𝑏 . A subtle substitution

function, e.g. as with (Smorynski 1977, 837 et passim) in the proof of Gödel’s incom-
pleteness theorem, is not needed, for there is no use of quantification into an opaque, or
otherwise “intensional”, context.
22.7. Fact (Ursets to any order)
Given Definition 22.1.2, 𝜐 𝜔 in Definition 22.1.3 serves as the set of the expression names
defined in §21. Given Theorems 22.2 and 22.4 (i), 𝜐𝜔 is another empty set distinct from
all members of 𝜐 𝜔 , and we may define a new omega ordered set of Ursets

𝜐 𝜔·2 === {𝑥|∀𝑦(𝜐𝜔 ∈ 𝑦 ∧ ∀𝑧(𝜐 𝑧 ∈ 𝑦 → 𝜐 𝑧+1 ∈ 𝑦) → 𝑥 ∈ 𝑦)}.

𝜐 𝜔·2 , and indeed 𝜐 𝛽 for any ordinal 𝛽, may serve as sets of Ursets, or Urelemente, for
whatever purpose one may have in mind, including that of naming extramathematical
individuals to equip £ with domains useful for applied mathematics and logic.

72
 23 HERITORS AND REGULARS

23 Heritors and regulars

A man is like a fraction whose numerator is what he is and whose denominator is what
he thinks of himself. The larger the denominator, the smaller the fraction.
Leo Tolstoy

Heritors and regulars are defined, and their behavior regulated so as to support the de-
velopment of the interpretation of NBG set theory of §25.
23.1. Definition: The Heritor is H === {𝑥|𝑥 = {𝑦|𝑦 ∈ 𝑥}}.
23.2. Definition: 𝑎 is an heritor just if ⊢M 𝑎 ∈ H.
23.3. Definition: H(𝑎) === 𝑎 = {𝑥|𝑥 ∈ 𝑎}
23.4. Theorem: The Heritor and heritors are orthodox.

Proof: The Heritor is orthodox by identity theory, and heritors by Postulate 6.5.2.6. □

23.5. Definition: 𝑎 is an hyposet of set 𝑏 just if 𝑎 = {𝑥|𝑥 ∈ 𝑎 ∧ 𝑥 ∈ 𝑏}.

23.6. Axiom:
⊢M H(𝑎) ∧ H(𝑏) ∧ 𝑎 ⊂ 𝑏 → 𝑎 = {𝑥|𝑥 ∈ 𝑎 ∧ 𝑥 ∈ 𝑏}.

23.7. Theorem:
⊢M H(𝑎) ∧ H(𝑏) ∧ 𝑎 = 𝑏 → 𝑎 = 𝑏.
E

Proof: An instance of Axiom 23.6 is ⊢M H(𝑏) ∧ H(𝑎) ∧ 𝑏 ⊂ 𝑎 → 𝑏 = {𝑥|𝑥 ∈ 𝑏 ∧ 𝑥 ∈ 𝑎}.

{𝑥|𝑥 ∈ 𝑎 ∧ 𝑥 ∈ 𝑏} = {𝑥|𝑥 ∈ 𝑏 ∧ 𝑥 ∈ 𝑎}, given §12, so just wed with the statement instance
H(𝑎) ∧ H(𝑏) ∧ 𝑎 ⊂ 𝑏 → 𝑎 = {𝑥|𝑥 ∈ 𝑎 ∧ 𝑥 ∈ 𝑏} of Axiom 23.6. □

23.8. Axiom:
⊢M H(𝑎) ∧ H(𝑏) ∧ 𝑎 ⊂ 𝑏 ← 𝑎 = {𝑥|𝑥 ∈ 𝑎 ∧ 𝑥 ∈ 𝑏},

so if 𝑎 is a hyposet of 𝑏, then 𝑎 and 𝑏 are heritors, and 𝑎 is a subheritor of 𝑏.

23.9. Theorem:

⊢M H(𝑎) ∧ H(𝑏) → (𝑎 ⊂ 𝑏 ↔ 𝑎 = {𝑥|𝑥 ∈ 𝑎 ∧ 𝑥 ∈ 𝑏}).

Proof: Invoke Axioms 23.6 and 23.8. □

23.10. Axiom (Heritors are hereditarily heritors)

⊢M H(𝑦) → ∀𝑥(𝑥 ∈ 𝑦 → H(𝑥)).

23.11. Observation: This section’s axioms do not commit to the existence of heritors.

73
 23 HERITORS AND REGULARS

23.12. Definition (Regular sets)

R(𝑥) === ∃𝑦(𝑦 ∈ 𝑥) → ∃𝑦(𝑦 ∈ 𝑥 ∧ ∀𝑧(𝑧 ∉ 𝑥 ∨ 𝑧 ∉ 𝑦))

23.13. Exercise Regular hereditarily orthodox sets are hereditarily regular.

Our attention below will be upon regular heritors.

74
 24 CHOICE, POWER, POTENCY AND COUNTABILITY

24 Choice, power, potency and countability

The axiom of choice is obviously true, the well-ordering principle

obviously false, and who can tell about Zorn’s lemma?
Jerry Bona

We show that the universe is countable. Theorem 20.3 is one of the important reasons
why that is so. Theorem 24.4.5 establishes that there is an orthodox bijection from the
set of natural numbers 𝜔 to the full universe U. §24.6 spells out in more detail how it is
that Cantor’s arguments, linked to power sets, are circumvented in £.

24.1 The denumerable wellordering

24.1.1. Definition:
h
Π𝑎Π𝑏 Constant(𝑎) ∧ Constant(𝑏) ∧ (Ξ, 𝛼) ⊩ 𝑎 ⊴ 𝑏

⇕

i
𝜇𝑥(𝑥𝜂Ω & 𝑥 ⪯ 𝑎 & (Ξ, 𝛼) ⊩ 𝑥 = 𝑎) ⪯ 𝜇𝑦(𝑦𝜂Ω & 𝑦 ⪯ 𝑏 & (Ξ, 𝛼) ⊩ 𝑦 = 𝑏)

24.1.2. Corollary:

(Ξ, 𝛼) ⊩ 𝑎 = 𝑏 ⇔ (Ξ, 𝛼) ⊩ 𝑎 ⊴ 𝑏 & (Ξ, 𝛼) ⊩ 𝑎 ⊵ 𝑏

24.1.3. Definition:

(Ξ, 𝛼) ⊩ 𝑎 ◁ 𝑏 ⇔ (Ξ, 𝛼) ⊩ 𝑎 ⊴ 𝑏 & (Ξ, 𝛼) ⊩ 𝑎 ≠ 𝑏

24.1.4. Axiom The wellordering:

(Ξ, 𝛼) ⊩ ∀𝑥, 𝑦(𝑥 ◁ 𝑦 ∨ 𝑥 = 𝑦 ∨ 𝑥 ▷ 𝑦)

24.1.5. Axiom The orthodoxy of the wellordering:

◁, and its cognate relations, are orthodox.

24.2 Function application notation

24.2.1. Definition:


𝑓 ‘𝑎 ≎ 𝑏 B ∀𝑥∀𝑦∀𝑧( (𝑥, 𝑦) ∈ 𝑓 ∧ (𝑥, 𝑧) ∈ 𝑓 → 𝑦 = 𝑧) ∧ (𝑎, 𝑏) ∈ 𝑓 .

24.2.2. Definition:
𝑏 ≎ 𝑓 ‘𝑎 B 𝑓 ‘𝑎 ≎ 𝑏

75
24.3 The choice function 24 CHOICE, POWER, POTENCY AND COUNTABILITY

24.2.3. Definition:
𝑥 ∈ 𝑓 ‘𝑎 B ∃𝑦( 𝑓 ‘𝑎 ≎ 𝑦 ∧ 𝑥 ∈ 𝑦)

24.2.4. Definition:
𝑓 ‘𝑎 ∈ 𝑥 B ∃𝑦( 𝑓 ‘𝑎 ≎ 𝑦 ∧ 𝑦 ∈ 𝑥)

24.2.5. Remark: The notation ≎ is used instead of =, for there are paradoxical functions
as e.g.
𝑔 = {(𝑥, 𝑦)|𝑥 = {∅} ∧ ((𝑟 ∈ 𝑟 → 𝑦 = ∅) ∧ (𝑟 ∉ 𝑟 → 𝑦 = {∅})},

for 𝑟 = {𝑥|𝑥 ∉ 𝑥}. For 𝑔 we do have that ⊢M ∀𝑥∀𝑦∀𝑧( (𝑥, 𝑦) ∈ 𝑔 ∧ (𝑥, 𝑧) ∈ 𝑔 → 𝑦 = 𝑧).


But, notice that (Ξ, 𝛼) ⊩ ({∅}, {∅}) ∈ 𝑔 just if (Ξ, 𝛼 + 1) ⊩ ({∅}, ∅) ∈ 𝑔. So we cannnot
write 𝑔‘{∅} = {∅} in the former case, and 𝑔‘{∅} = ∅ in the latter case. For identity is
an orthodox equivalence relations. So we use ≎ to avoid problems with the theory of
identity in exotic cases.
24.2.6. Remark: There certainly are sets of more orthodox functions so that a function
M
ℎ is an element in one of them only if |= ∀𝑥∀𝑦(ℎ‘𝑥 ≎ 𝑦 → 𝑥 = 𝑦).
24.2.7. Remark: The author discussed and introduced the notation ≎ in the article,
whose English translation is «“2+2=4” is misleading», in (Bjørdal 2008, pp. 55–66),
for such reasons which are adduced here.

24.3 The choice function

On account of ancient Greek διάλεξε, for was selected, we define ‘𝑤, the atled of 𝑤:
𝛿

24.3.1. Definition The choice function:

‘𝑤 ≎ {𝑥|(𝑥 ∈ 𝑤 ∧ ∀𝑦(𝑦 ∈ 𝑤 → 𝑥 ⊴ 𝑦)}.

𝛿

24.3.2. Definition Iterated choices from 𝑏:

𝑖=𝑛
Ø
𝑚 𝑖
𝑎≎ 𝑏 ⇔ ((𝑚 = 0 ∧ 𝑎 ≎ ‘𝑏) ∨ ∃𝑛(𝑛 ∈ 𝜔 ∧ 𝑚 = 𝑛 + 1 ∧ 𝑎 ≎ ‘(𝑏 \
𝛿 𝛿 𝛿 𝛿
𝑏))).
𝑖=0

24.4 The enumerator

Given Axiom 24.1.5, the orthodoxy of 𝜔 and U, and the fact that 𝑛
𝑤 is orthodox if 𝑤 is
𝛿
orthodox, we posit
24.4.1. Definition of the enumerator:

e === {(𝑛, 𝑥)|𝑛 ∈ 𝜔 ∧ 𝑥 ∈ 𝑛

U}
𝛿

24.4.2. Theorem (e is orthdox)

76
24.5 The enumeration postulates 24 CHOICE, POWER, POTENCY AND COUNTABILITY

Proof: As 𝑛
U, for 𝑛 ∈ 𝜔 is orthodox. □
𝛿

24.4.3. Theorem (The functionality of e)

⊢M ∀𝑥∀𝑦∀𝑧( (𝑥, 𝑦) ∈ e ∧ (𝑥, 𝑧) ∈ e → 𝑦 = 𝑧)

Proof: Obvious, given Definitions 24.3.2 and 24.4.1 and Theorem 24.4.2. □

24.4.4. Theorem
⊢M e‘𝑛 ≎ 𝑥 ↔ (𝑛, 𝑥) ∈ e ↔ 𝑛 ∈ 𝜔 ∧ 𝑥 ∈ 𝑛
(U)
𝛿

Proof: On account of Definitions 24.2.1 and 24.4.1, and Theorem 24.4.3. □

24.4.5. Theorem
e is a bijection from 𝜔 to U.

Proof: Given §24.1, as the orders of 𝜔 and Ω match, and for any constant 𝑎, 𝑎𝜂Ω, as all
sets are finite positive von Neumann ordinals according to the meta language. □

24.5 The enumeration postulates

For any ordinal 𝛼:
24.5.1. Postulate:

Π𝑎Π𝑏 constant(𝑎) & constant(𝑏) ⇒
  
(Ξ, 𝛼) ⊩ ∀𝑛(𝑛 ∈ 𝜔 → ∃=𝑛 𝑥(𝑥 ◁ 𝑎) ∧ ∃=𝑛 𝑦(𝑦 ◁ 𝑏) → 𝑎 = 𝑏 )

24.5.2. Postulate:
  
Π𝑎, 𝑏, 𝑐, ((Ξ, 𝛼) ⊩ ∀𝑛 𝑛 ∈ 𝜔 → ( ⟨𝑎, 𝑏⟩ ∈ e & ∃=𝑛 𝑥(𝑥 ◁ 𝑏) ↔
  
⟨{𝑣|𝑣 ∈ 𝑎 ∨ 𝑣 = 𝑎}, 𝑐⟩ ∈ e & ∃ =(𝑛+1)
𝑥(𝑥 ◁ 𝑐) )

24.5.3. Postulate:
(Ξ, 𝛼) ⊩ ∀𝑛(𝑛 ∈ 𝜔 → ∃𝑦(⟨𝑛, 𝑦⟩ ∈ e))

24.5.4. Postulate:
(Ξ, 𝛼) ⊩ ∀𝑦∃𝑛(𝑛 ∈ 𝜔 ∧ ⟨𝑛, 𝑦⟩ ∈ e)

24.5.5. Postulate:
(Ξ, 𝛼) ⊩ ∀𝑛∀𝑛′∀𝑦(⟨𝑛, 𝑦⟩ ∈ e ∧ ⟨𝑛′, 𝑦⟩ ∈ e → 𝑛 = 𝑛′)

24.5.6. Postulate:
(Ξ, 𝛼) ⊩ ∀𝑛∀𝑦∀𝑧(⟨𝑛, 𝑦⟩ ∈ e ∧ ⟨𝑛, 𝑦⟩ ∈ e → 𝑦 = 𝑧)

77
24.6 Absolute countability 24 CHOICE, POWER, POTENCY AND COUNTABILITY

Some consequences of the enumeration postulates:

24.5.7. Theorem:
⊢M ∃=0 𝑥(𝑥 ◁ L )

24.5.8. Theorem:

Π𝑎Π𝑏 constant(𝑎) & constant(𝑏) ⇒

⊢M ∀𝑛(𝑛 ∈ 𝜔 → ∃=𝑛 𝑥(𝑥 ◁ 𝑎) ∧ ∃=𝑛 𝑦(𝑦 ◁ 𝑏) → 𝑎 = 𝑏 )

  

24.5.9. Theorem:
M
h  i
Π𝑏 constant(𝑏) ⇒ ⊢ ⟨∅, 𝑏⟩ ∈ e ↔ ∃ 𝑥(𝑥 ◁ 𝑏)
=0

24.5.10. Theorem:
Π𝑎, 𝑏, 𝑐, (⊢M ∀𝑛 𝑛 ∈ 𝜔 → ( ⟨𝑛, 𝑏⟩ ∈ e & ∃=𝑛 𝑥(𝑥 ◁ 𝑏) ↔
  

  
⟨{𝑣|𝑣 ∈ 𝑛 ∨ 𝑣 = 𝑛}, 𝑐⟩ ∈ e & ∃=(𝑛+1) 𝑥(𝑥 ◁ 𝑐) ) )

24.5.11. Theorem:
⊢M ∀𝑦∃𝑛(𝑛 ∈ 𝜔 ∧ ⟨𝑛, 𝑦⟩ ∈ e)

Proof: As all sets are finite von Neumann ordinals of the meta language, and 𝜔 has the
same order as Ω. □

24.6 Absolutely all sets are countable

If some set is uncountable, some set of subsets of 𝜔 is uncountable. We have earlier in-
troduced the power set P (a) = {𝑥|𝑥 ⊂ a}, and will first consider its import on the question.
Thereon we consider the potency set of a set a as given by
24.6.1. Definition: P(a) = {𝑥|𝑥 = {𝑦|𝑦 ∈ 𝑥 ∧ 𝑦 ∈ a}}.
The potency set construction is very important in §25. Here the preoccupation is with
showing that neither power sets nor potency sets generate uncountable sets.

24.6.1 e restricted to P ({𝑥|𝑥 ∈ 𝜔})

e restricted to the power set of {𝑥|𝑥 ∈ 𝜔} is
e| P ({𝑥|𝑥∈𝜔}) = {(𝑥, 𝑦)|(𝑥, 𝑦) ∈ e ∧ 𝑦 ∈ P ({𝑥|𝑥 ∈ 𝜔})}, (24.6.2)
which has 𝜔 as domain and P ({𝑥|𝑥 ∈ 𝜔}) as range. Given Definitions 24.3.2 and 24.4.1,
equation 24.6.2 may be equivalently stated as

e| P ({𝑥|𝑥∈𝜔}) = {(𝑥, 𝑦)|𝑥 ∈ 𝜔 ∧ 𝑦 ∈ 𝑥

U ∧ 𝑦 ∈ P ({𝑥|𝑥 ∈ 𝜔})}. (24.6.3)
𝛿

78
24.6 Absolute countability 24 CHOICE, POWER, POTENCY AND COUNTABILITY

24.6.4. Theorem

⊢M ∀𝑢∀𝑣∀𝑤((𝑢, 𝑣) ∈ e| P ({𝑥|𝑥∈𝜔}) ∧ (𝑢, 𝑤) ∈ e| P ({𝑥|𝑥∈𝜔}) → 𝑣 = 𝑤).

Proof: Obvious, from the built up of e| P ({𝑥|𝑥∈𝜔}) with orthodox function 𝑥

□
𝛿
.

To attempt Cantor’s proof by contradiction for uncountability, assume that e| P ({𝑥|𝑥∈𝜔})

surjects from 𝜔 to P ({𝑥|𝑥 ∈ 𝜔}) and posit
24.6.5. Definition:
S = {𝑥|𝑥 ∈ 𝜔 ∧ 𝑥 ∉ e| P ({𝑥|𝑥∈𝜔}) ‘𝑥}.

24.6.6. Theorem For an 𝑚 ∈ 𝜔,

⊢M (𝑚, S) ∈ e| P ({𝑥|𝑥∈𝜔}) .

Proof: A consequence of Equation 24.6.3 and alethic comprehension is

⊢M (𝑚, S) ∈ e| P ({𝑥|𝑥∈𝜔}) ↔ T 𝑚 ∈ 𝜔 ∧ S ∈ 𝑚
U ∧ S ∈ P ({𝑥|𝑥 ∈ 𝜔}) .
𝛿

Let 𝑚 ∈ 𝜔 be the natural number such that ⊢M S ∈ 𝑚 U, so ⊢M 𝑚 ∈ 𝜔∧S ∈ 𝑚 U. But besides,

𝛿 𝛿
⊢M S ∈ P ({𝑥|𝑥 ∈ 𝜔}), as ⊢M S ⊂ {𝑥|𝑥 ∈ 𝜔}. So ⊢M 𝑚 ∈ 𝜔 ∧ S ∈ 𝑚 U ∧ S ∈ P ({𝑥|𝑥 ∈ 𝜔}). Thus,
𝛿
on account of inference mode 9.1.2.1, ⊢M T 𝑚 ∈ 𝜔 ∧ S ∈ 𝑚 U ∧ S ∈ P ({𝑥|𝑥 ∈ 𝜔}) . Finish
𝛿
by using the maxim mode 9.2.5. □

24.6.7. Theorem There is an 𝑚 ∈ 𝜔 such that ⊢M e| P ({𝑥|𝑥∈𝜔}) ‘𝑚 ≎ S.

Proof: Invoke Theorems 24.6.4 and 24.6.6, and Definition 24.2.1. □

From Definition 24.6.5 and alethic comprehension,

⊢M 𝑚 ∈ S ↔ T 𝑚 ∈ 𝜔 ∧ 𝑚 ∉ e| P ({𝑥|𝑥∈𝜔}) ‘𝑚 . (24.6.8)

Given Definition 24.2.3, euro and 𝑥

U
𝛿

⊢M 𝑚 ∈ S ↔ T 𝑚 ∈ 𝜔 ∧ ∀𝑦(e| P ({𝑥|𝑥∈𝜔}) ‘𝑚 ≎ 𝑦 → 𝑚 ∉ 𝑦) . (24.6.9)

Given Theorem 24.6.7, and the fact that there is only one 𝑚 ∈ 𝜔 such that S ∈ 𝑚
U, for
𝛿
the appropriate 𝑚, ∀𝑦(e| P ({𝑥|𝑥∈𝜔}) ‘𝑚 ≎ 𝑦 ↔ 𝑦 = S). So that

⊢M 𝑚 ∈ S ↔ T 𝑚 ∈ 𝜔 ∧ 𝑚 ∉ S) . (24.6.10)

But it was assumed that 𝑚 ∈ 𝜔, which is an orthodox statement, so that

79
24.6 Absolute countability 24 CHOICE, POWER, POTENCY AND COUNTABILITY

⊢M 𝑚 ∈ S ↔ T 𝑚 ∉ S) . (24.6.11)

As

⊢ T 𝑚 ∉ S) → 𝑚 ∉ S (24.6.12)

and

⊢ 𝑚 ∈ S → T 𝑚 ∈ S) , (24.6.13)

it follows that

⊢𝑚∉S (24.6.14)

and

⊢T 𝑚∉S →T 𝑚∈S . (24.6.15)

But

⊢M T 𝑚 ∈ S → ¬T 𝑚 ∉ S , (24.6.16)

so that

⊢ T 𝑚 ∉ S → ¬T 𝑚 ∉ S , (24.6.17)

and consequently

⊢ ¬T 𝑚 ∉ S . (24.6.18)

But an instance of the inference mode 9.1.2.11 is

⊢ ¬T 𝑚 ∉ S ⇒ ⊢ 𝑚 ∈ S, (24.6.19)

so that

⊢ 𝑚 ∈ S. (24.6.20)

A joining of equations 24.6.14 and 24.6.20 results in

80
24.6 Absolute countability 24 CHOICE, POWER, POTENCY AND COUNTABILITY

⊢ 𝑚 ∈ S & ⊢ 𝑚 ∉ S. (24.6.21)

But this merely amounts to an argumentum ad paradoxo, and it has not been proven that
e| P ({𝑥|𝑥∈𝜔}) is not a function with domain 𝜔 which is onto its range P ({𝑥|𝑥 ∈ 𝜔}).

24.6.2 e restricted to P(𝜔)

The potency set of 𝜔 is
P(𝜔) = {𝑥|𝑥 = {𝑦|𝑦 ∈ 𝑥 ∧ 𝑦 ∈ 𝜔}}. (24.6.22)

e restricted to the potency set of 𝜔 is

e| P(𝜔) = {(𝑥, 𝑦)|(𝑥, 𝑦) ∈ e ∧ 𝑦 ∈ P(𝜔)}, (24.6.23)

which has 𝜔 as domain and P(𝜔) as its range. Given Definition 24.4.1, equation 24.6.23
may be equivalently stated as

e| P(𝜔) = {(𝑥, 𝑦)|𝑥 ∈ 𝜔 ∧ 𝑦 ∈ 𝑥

(U) ∧ 𝑦 ∈ P(𝜔)}, (24.6.24)
𝛿

24.6.25. Fact e, P(𝜔) and e| P(𝜔) are orthodox.

Proof: e is orthodox given Theorem 24.4.2, P(𝜔) on account of the theory of identity,
and e| P(𝜔) is orthodox because e and P(𝜔) are orthodox. □

24.6.26. Fact
⊢M ∀𝑥∀𝑦∀𝑧((𝑥, 𝑦) ∈ e| P(𝜔) ∧ (𝑥, 𝑧) ∈ e| P(𝜔) → 𝑦 = 𝑧).

Proof: As e is functional. □

24.6.27. Assumption Orthodox function e| P(𝜔) surjects from 𝜔 to P(𝜔):

∀𝑤(𝑤 ∈ P(𝜔) → ∃𝑣(𝑣 ∈ 𝜔 ∧ e| P(𝜔) ‘𝑣 ≎ 𝑤)).

24.6.28. Definition:
S = {𝑥|𝑥 ∈ 𝜔 ∧ 𝑥 ∉ e| P(𝜔) ‘𝑥}.

24.6.29. Assumption S = {𝑦|𝑦 ∈ 𝜔 ∧ 𝑦 ∈ S}: S ∈ P(𝜔).

24.6.30. Assumption S is orthodox.

Proof: From Assumption 24.6.29, Axiom 23.8 and Theorem 23.4. □

24.6.31. Assumption An 𝑚 ∈ 𝜔 is such that e| P(𝜔) ‘𝑚 ≎ S.

81
24.6 Absolute countability 24 CHOICE, POWER, POTENCY AND COUNTABILITY

Proof: From Assumption 24.6.27. □

24.6.32. Assumption ⊢M ∀𝑥(𝑥 ∈ S ↔ 𝑥 ∈ 𝜔 ∧ 𝑥 ∉ e| P(𝜔) ‘𝑥).

Proof: Given Definition 24.6.28 and the fact that S is orthodox. □

24.6.33. Assumption ⊢M ∀𝑥(𝑥 ∈ S ↔ 𝑥 ∈ 𝜔 ∧ ∀𝑦(e| P(𝜔) ‘𝑥 ≎ 𝑦 → 𝑥 ∉ 𝑦).

Proof: On account of Definition 24.2.3 and Assumption 24.6.32. □

24.6.34. Assumption ⊢M (𝑚 ∈ S → 𝑚 ∉ S).

Proof: It was agreed in Assumption 24.6.31 that for an 𝑚 ∈ 𝜔, e| P(𝜔) ‘𝑚 ≎ S. □

24.6.35. Assumption 𝑚 ∉ 𝑆 → ∃𝑦(e| P(𝜔) ‘𝑚 ≎ 𝑦 ∧ 𝑚 ∈ 𝑦).

Proof: From Assumption 24.6.33, the agreement of Assumption 24.6.31. □

24.6.36. Theorem For functional 𝑓 :

If ⊢M ∃𝑦( 𝑓 ‘𝑎 ≎ 𝑦 ∧ 𝑎 ∈ 𝑦) and ⊢M 𝑓 ‘𝑎 ≎ 𝑐, then ⊢M 𝑎 ∈ 𝑐.

Proof: Because ⊢M [(𝑎, 𝑦) ∈ 𝑓 ∧ (𝑎, 𝑐) ∈ 𝑓 ] → 𝑦 = 𝑐, as 𝑓 is functional, and because

⊢M ((𝑑, 𝑒) ∈ 𝑓 ↔ 𝑓 ‘𝑑 ≎ 𝑒) if ⊢M ( 𝑓 is functional). □

24.6.37. Assumption ⊢M (𝑚 ∉ S → 𝑚 ∈ S).

Proof: Appeal to Assumption 24.6.35 and Theorem 24.6.36. □

24.6.38. Assumption ⊢M 𝑚 ∈ S ∧ 𝑚 ∉ S

Proof: From Assumptions 24.6.34 and 24.6.37. □

The contradiction in the maximal context of Assumption 24.6.38 is false, so it follow that
a previous assumption is to be discarded. We do that by stating the following
24.6.39. Theorem Assumption 24.6.29 is false, and ⊢M S ≠ {𝑥|𝑥 ∈ S ∧ 𝑥 ∈ 𝜔}.

Proof: The discussion in §24.6.2. □

82
 25 AND THE THEORIES OF VONSETS

25 £ and the theories of vonsets

If people do not believe that mathematics is simple, it is

only because they do not realize how complicated life is.
John von Neumann

25.1. Definition: Set theory £ is £ plus Axioms 23.6, 23.8, 23.10.

Recall Definitions 23.3 and 23.12.
25.2. Definition: £H&R(D) is £ plus H(D) plus R(D), with D as in Definition 25.4.6.
Let NBGC + TA be Neumann-Bernays-Gödel set theory with Global Choice and Tarski’s
Axiom. An interpretation of NBGC + TA is developed in £HR(D) below.
Natural weakenings and extensions of NBGC + TA are as well taken to be theories of
vonsets. Needless to say, but all vonsets are sets, though some sets are not vonsets.
The term “natural” in the previous paragraph is left undefined, as investigations should
not be restrained. So we here disregard philosophical quandaries related to the fact that
the term “vonset” may have different meanings, whatever that is, in natural extensions
of NBG which are not consistent with each other, such as NBG + the Axiom of choice,
and NBG + the Axiom of determinacy.

25.1 The potency vonset

We saw in §20 that power sets as classically defined are mathematically useless, as they
are paradoxical lest of a non-paradoxical universal set.
Potency vonsets are potency sets, as all vonsets are sets.
The notion of potency set was introduced in Definition 24.6.1:

P(𝑎) === {𝑥|𝑥 = {𝑦|𝑦 ∈ 𝑥 ∧ 𝑦 ∈ 𝑎}}.

25.1.1. Theorem: The potency vonset of a vonset 𝑎 contains precisely 𝑎’s hypovonsets,
in the sense of Definition 23.5.

Proof: Use Axioms 23.10 and 25.4.10 and Theorem 23.9.entail that vonsets are heritors,
and from Axioms 23.6 and 23.8. □

25.1.2. Theorem:

P(𝑎) is orthodox, and all of its members are hereditarily heritors.

Proof: P(𝑎) is orthodox by the logic of identity. Its members, if any, are heritors on
account of Axiom 23.8, and are hereditarily heritors given Axiom 23.10. □

83
25.2 The Grothendieck vonset of w relative to v 25 AND THE THEORIES OF VONSETS

25.1.3. Theorem:
P(𝑎) is empty if 𝑎 is not an heritor.

Proof: Appeal to Axiom 23.8. □

25.1.4. Theorem:
⊢M ∀𝑥(𝑥 ∈ P(𝑎) ↔ H(𝑥) ∧ H(𝑎) ∧ 𝑥 ⊂ 𝑎).

Proof: Appeal to Theorem 23.9 and Definition 24.6.1. □

25.2 The Grothendieck vonset of w relative to v

25.2.1. Definition: Let

G(𝑣, 𝑤, 𝑣 0 , 𝑣 1 , ) ===∀𝑦 𝑤 ∈ 𝑦 ∧ ∀𝑧 𝑧 ∈ 𝑦 → (𝑧 ∈ P(𝑣 1 ) ∧ P(𝑧) ∈ P(𝑣 1 ) ∧ P(𝑧) ∈ 𝑦) ∧
 

∀𝑧 𝑧 ∈ P(𝑦) ∧ 𝑧 ∉ 𝑦 → ∃ 𝑓 [ 𝑓 ∈ 𝑣 ∧ Bijection( 𝑓 )∧
(∀𝑥 0 )(𝑥0 ∈ 𝑦 → ∃𝑥 1 (𝑥 1 ∈ 𝑧 ∧ (𝑥 0 , 𝑥1 ) ∈ 𝑓 ))


(∀𝑥 1 )(𝑥1 ∈ 𝑧 → ∃𝑥 0 (𝑥 0 ∈ 𝑦 ∧ (𝑥 0 , 𝑥1 ) ∈ 𝑓 ))] → 𝑣 0 ∈ 𝑦

Use Theorem 19.1.5 to obtain the manifestation set with parameters G(𝑣, 𝑤),
25.2.2. Theorem The Grothendieck of w relative to v:
M

⊢ ∀𝑢∀𝑤 𝑢 ∈ G(v, w) ↔ T T ∀𝑦 𝑤 ∈ 𝑦 ∧ ∀𝑧 𝑧 ∈ 𝑦 →


(𝑧 ∈ P(G(v, w)) ∧ P(𝑧) ∈ P(G(v, w)) ∧ P(𝑧) ∈ 𝑦) ∧



∀𝑧 𝑧 ∈ P(𝑦) ∧ 𝑧 ∉ 𝑦 → ∃ 𝑓 [ 𝑓 ∈ 𝑣 ∧ Bijection( 𝑓 )∧
(∀𝑥0 )(𝑥 0 ∈ 𝑦 → ∃𝑥 1 (𝑥 1 ∈ 𝑧 ∧ (𝑥0 , 𝑥1 ) ∈ 𝑓 ))



(∀𝑥1 )(𝑥 1 ∈ 𝑧 → ∃𝑥 0 (𝑥0 ∈ 𝑦 ∧ (𝑥0 , 𝑥1 ) ∈ 𝑓 ))] → 𝑢 ∈ 𝑦

25.2.3. Theorem: G(𝑣, 𝑤) is orthodox for orthodox v and w, so that

⊢M ∀𝑢∀𝑤 𝑢 ∈ G(v, w) ↔∀𝑦 𝑤 ∈ 𝑦 ∧ ∀𝑧 𝑧 ∈ 𝑦 →
 

(𝑧 ∈ P(G(v, w)) ∧ P(𝑧) ∈ P(G(v, w)) ∧ P(𝑧) ∈ 𝑦) ∧



∀𝑧 𝑧 ∈ P(𝑦) ∧ 𝑧 ∉ 𝑦 → ∃ 𝑓 [ 𝑓 ∈ 𝑣 ∧ Bijection( 𝑓 )∧
(∀𝑥 0 )(𝑥0 ∈ 𝑦 → ∃𝑥 1 (𝑥 1 ∈ 𝑧 ∧ (𝑥 0 , 𝑥1 ) ∈ 𝑓 ))



(∀𝑥 1 )(𝑥1 ∈ 𝑧 → ∃𝑥 0 (𝑥 0 ∈ 𝑦 ∧ (𝑥 0 , 𝑥1 ) ∈ 𝑓 ))] → 𝑢 ∈ 𝑦

Proof: As in the proof that 𝜔 is orthodox, of Theorem 13.2.3 on page 43, noting that
P(G(v, w)) is an orthodox heritor by cause of Theorem 25.1.2. □

25.2.4. Remark: For appropriate v and w, Theorem 25.2.3 amounts to Tarski’s axiom,
which states that all sets are members of a Grothendieck-universe. Tarski-Grothendieck
set theory is usually presented as ZFC + Tarski’s axiom.

84
25.3 Capture 25 AND THE THEORIES OF VONSETS

25.3 Capture
In this section we presuppose that the sets and conditions invoked are orthodox.
25.3.1. Definition Capture with B from 𝑤:

C(B, 𝑤) === {𝑥|∃𝑦(𝑦 ∈ 𝑤 ∧ ∀𝑧((𝑥, 𝑦) 𝑦𝑧 ∈ B ↔ 𝑦 = 𝑧))}

25.3.2. Theorem: Capture is equivalent with replacement.

Proof: i) If a vonset is obtained from capture with B from 𝑤, it can be obtained from
replacement by using the functional condition ∀𝑧((𝑥, 𝑦) 𝑦𝑧 ∈ B ↔ 𝑦 = 𝑧). ii) If a vonset is
obtained from replacement by functional B so that ∀𝑥∀𝑦∀𝑧((𝑥, 𝑦) ∈ B ∧ (𝑥, 𝑧) ∈ B → 𝑦 = 𝑧),
it can be obtained from capture by using the condition as in Definition 25.3.1. □

25.3.3. Theorem: Capture, as replacement, entails specification.

Proof: Use the functional B′ === {(𝑥, 𝑦)|𝑥 ∈ B ∧ 𝑥 = 𝑦)} as capture vonset relative to a
vonset 𝑎, and observe that the existence of the vonset {𝑥|𝑥 ∈ 𝑎 ∧ B(𝑥)} is justified by
capture and extensionality, which holds for V and D below, as per Theorem 25.4.13. □

25.4 V and D
25.4.1. Definition of the drift of 𝑢:

𝔇(𝑢) = {𝑤|𝑤 ∈ 𝑢 ∨ ∀𝑣 𝑢 ∈ 𝑣 ∧ E = {(𝑥𝑖 , 𝑥 𝑗 )|(𝑥𝑖 , 𝑥 𝑗 ) ∈ 𝑢 2 ∧ 𝑥𝑖 ∈ 𝑥 𝑗 } ∈ 𝑣 ∧
∀𝑥𝑖 (𝑥𝑖 ∈ 𝑣 → {𝑦|𝑦 ∈ 𝑢 ∧ 𝑦 ∉ 𝑥𝑖 } ∈ 𝑣) ∧ ∀𝑥𝑖 ∀𝑥 𝑗 (𝑥𝑖 ∈ 𝑣 ∧ 𝑥 𝑗 ∈ 𝑣 → 𝑥𝑖 ∩ 𝑥 𝑗 ∈ 𝑣) ∧
∀𝑥𝑖 (𝑥𝑖 ∈ 𝑣 → dom(𝑥𝑖 ) = {𝑦|∃𝑥((𝑦, 𝑥) ∈ 𝑥𝑖 )} ∈ 𝑣) ∧
∀𝑥𝑖 (𝑥𝑖 ∈ 𝑣 → {𝑦|∃𝑥 𝑗 , 𝑥 𝑘 (𝑦 = (𝑥 𝑗 , 𝑥 𝑘 ) ∧ 𝑥 𝑗 ∈ 𝑥𝑖 ∧ 𝑣 𝑘 ∈ 𝑣 1 )} ∈ 𝑣) ∧
∀𝑥𝑖 (𝑥𝑖 ∈ 𝑣 → {𝑦|∃𝑥 𝑗 ∃𝑥 𝑘 ∃𝑥 𝑙 (𝑦 = (𝑥 𝑗 , 𝑥 𝑘 , 𝑥 𝑙 ) ∧ (𝑥 𝑘 , 𝑥 𝑙 , 𝑥 𝑗 ) ∈ 𝑥𝑖 } ∈ 𝑣) ∧


∀𝑥𝑖 (𝑥𝑖 ∈ 𝑣 → {𝑦|∃𝑥 𝑗 ∃𝑥 𝑘 ∃𝑥 𝑙 (𝑦 = (𝑥 𝑗 , 𝑥 𝑘 , 𝑥 𝑙 ) ∧ (𝑥 𝑗 , 𝑥 𝑙 , 𝑥 𝑘 ) ∈ 𝑥𝑖 } ∈ 𝑣) → 𝑤 ∈ 𝑣 }

25.4.2. Definition of V(𝑣 0 , 𝑣 1 ):

Ø
V(𝑣 0 , 𝑣 1 ) = ∀𝑣 (𝜔 ∈ 𝑣 ∧ ∀𝑤 ∈ 𝑣∀𝑥 ∈ 𝑣 : {𝑤, 𝑥} ∈ 𝑣 ∧ ∀𝑤 ∈ 𝑣 : 𝑤∈𝑣∧
∀𝑤 𝑣 : P(𝑤) = {𝑥|𝑥 = {𝑦|𝑦 ∈ 𝑥 ∧ 𝑦 ∈ 𝑤}} ∈ 𝑣 ∧
∈
∀𝑤 𝑣 : (𝑤) = {𝑥|(𝑥 ∈ 𝑤 ∧ ∀𝑦(𝑦 ∈ 𝑤 → 𝑥 ⊴ 𝑦)} ∈ 𝑣 ∧
∈
𝛿

∀𝑤 𝑣 : G(𝔇(𝑣 1 ), 𝑤) ∈ 𝑣 ∧
∈
∀𝑤 𝑣∀B ∈ 𝔇(𝑣 1 ) : C(B, 𝑤) = {𝑥|∃𝑦(𝑦 ∈ 𝑤 ∧ ∀𝑧((𝑥, 𝑦) 𝑦𝑧 ∈ B ↔ 𝑦 = 𝑧))} ∈ 𝑣)
∈


→ 𝑣0 ∈ 𝑣

85
25.4 V and D 25 AND THE THEORIES OF VONSETS

25.4.3. Definition of V via manifestation from Definition 25.4.2:

⊢M ∀𝑢 𝑢 ∈ V ↔T T ∀𝑣 [𝜔 ∈ 𝑣 ∧ ∀𝑤 ∈ 𝑣∀𝑥 ∈ 𝑣 : {𝑤, 𝑥} ∈ 𝑣 ∧ ∀𝑤 ∈ 𝑣 :
h  Ø
𝑤∈𝑣∧
∀𝑤𝑣 : P(𝑤) = {𝑥|𝑥 = {𝑦|𝑦 ∈ 𝑥 ∧ 𝑦 ∈ 𝑤}} ∈ 𝑣 ∧
∈
∀𝑤 ∈
𝑣 : (𝑤) = {𝑥|(𝑥 ∈ 𝑤 ∧ ∀𝑦(𝑦 ∈ 𝑤 → 𝑥 ⊴ 𝑦)} ∈ 𝑣 ∧
𝛿

∀𝑤𝑣 : G(𝔇( V), 𝑤) ∈ 𝑣 ∧

∈
∀𝑤𝑣∀B ∈ 𝔇( V) : C(B, 𝑤) = {𝑥|∃𝑦(𝑦 ∈ 𝑤 ∧ ∀𝑧((𝑥, 𝑦) 𝑦𝑧 ∈ B ↔ 𝑦 = 𝑧))} ∈ 𝑣]
∈
i
→𝑢∈𝑣

As V is orthodox on account of Theorems 23.4 and 25.4.11,

25.4.4. Theorem:
⊢M ∀𝑢 𝑢 ∈ V ↔ ∀𝑣 [𝜔 ∈ 𝑣 ∧ ∀𝑤 ∈ 𝑣∀𝑥 ∈ 𝑣 : {𝑤, 𝑥} ∈ 𝑣 ∧ ∀𝑤 ∈ 𝑣 :
h  Ø
𝑤∈𝑣∧
∀𝑤 𝑣 : P(𝑤) = {𝑥|𝑥 = {𝑦|𝑦 ∈ 𝑥 ∧ 𝑦 ∈ 𝑤}} ∈ 𝑣 ∧
∈
∀𝑤 ∈
𝑣 : (𝑤) = {𝑥|(𝑥 ∈ 𝑤 ∧ ∀𝑦(𝑦 ∈ 𝑤 → 𝑥 ⊴ 𝑦)} ∈ 𝑣 ∧
𝛿

∀𝑤 𝑣 : G(𝔇( V), 𝑤) ∈ 𝑣 ∧
∈
∀𝑤 𝑣∀B ∈ 𝔇( V) : C(B, 𝑤) = {𝑥|∃𝑦(𝑦 ∈ 𝑤 ∧ ∀𝑧((𝑥, 𝑦) 𝑦𝑧 ∈ B ↔ 𝑦 = 𝑧))} ∈ 𝑣]
∈
i
→𝑢∈𝑣

25.4.5. The drift equation: D =𝔇(V).

25.4.6. Definition of the drift of all classes:

D = {𝑤|𝑤 ∈ V ∨ ∀𝑣 V ∈ 𝑣 ∧ E = {(𝑥𝑖 , 𝑥 𝑗 )|(𝑥𝑖 , 𝑥 𝑗 ) ∈ V2 ∧ 𝑥𝑖 ∈ 𝑥 𝑗 } ∈ 𝑣 ∧
∀𝑥𝑖 (𝑥𝑖 ∈ 𝑣 → {𝑢|𝑢 ∈ V ∧ 𝑢 ∉ 𝑥𝑖 } ∈ 𝑣) ∧ ∀𝑥𝑖 ∀𝑥 𝑗 (𝑥𝑖 ∈ 𝑣 ∧ 𝑥 𝑗 ∈ 𝑣 → 𝑥𝑖 ∩ 𝑥 𝑗 ∈ 𝑣) ∧
∀𝑥𝑖 (𝑥𝑖 ∈ 𝑣 → dom(𝑥𝑖 ) = {𝑤|∃𝑥((𝑤, 𝑥) ∈ 𝑥𝑖 )} ∈ 𝑣) ∧
∀𝑥𝑖 (𝑥𝑖 ∈ 𝑣 → {𝑢|∃𝑥 𝑗 , 𝑥 𝑘 (𝑢 = (𝑥 𝑗 , 𝑥 𝑘 ) ∧ 𝑥 𝑗 ∈ 𝑥𝑖 ∧ 𝑣 𝑘 ∈ V)} ∈ 𝑣) ∧
∀𝑥𝑖 (𝑥𝑖 ∈ 𝑣 → {𝑢|∃𝑥 𝑗 ∃𝑥 𝑘 ∃𝑥 𝑙 (𝑢 = (𝑥 𝑗 , 𝑥 𝑘 , 𝑥 𝑙 ) ∧ (𝑥 𝑘 , 𝑥 𝑙 , 𝑥 𝑗 ) ∈ 𝑥𝑖 } ∈ 𝑣) ∧


∀𝑥𝑖 (𝑥𝑖 ∈ 𝑣 → {𝑢|∃𝑥 𝑗 ∃𝑥 𝑘 ∃𝑥 𝑙 (𝑢 = (𝑥 𝑗 , 𝑥 𝑘 , 𝑥 𝑙 ) ∧ (𝑥 𝑗 , 𝑥 𝑙 , 𝑥 𝑘 ) ∈ 𝑥𝑖 } ∈ 𝑣) → 𝑤 ∈ 𝑣 }

25.4.7. Fact: ⊢M V ⊂ D.
25.4.8. Theorem The definition of V, with recourse to D:
M
h  Ø
⊢ ∀𝑢 𝑢 ∈ V ↔ ∀𝑣 [𝜔 ∈ 𝑣 ∧ ∀𝑤 ∈ 𝑣∀𝑥 ∈ 𝑣 : {𝑤, 𝑥} ∈ 𝑣 ∧ ∀𝑤 ∈ 𝑣 : 𝑤∈𝑣∧
∀𝑤 𝑣 : P(𝑤) = {𝑥|𝑥 = {𝑦|𝑦 ∈ 𝑥 ∧ 𝑦 ∈ 𝑤}} ∈ 𝑣 ∧
∈
∀𝑤 ∈
𝑣 : (𝑤) = {𝑥|(𝑥 ∈ 𝑤 ∧ ∀𝑦(𝑦 ∈ 𝑤 → 𝑥 ⊴ 𝑦)} ∈ 𝑣 ∧
𝛿

∀𝑤 𝑣 : G( D, 𝑤) ∈ 𝑣 ∧
∈
∀𝑤 𝑣∀B ∈ D : C(B, 𝑤) = {𝑥|∃𝑦(𝑦 ∈ 𝑤 ∧ ∀𝑧((𝑥, 𝑦) 𝑦𝑧 ∈ B ↔ 𝑦 = 𝑧))} ∈ 𝑣]
∈
i
→𝑢∈𝑣

86
25.4 V and D 25 AND THE THEORIES OF VONSETS

25.4.9. Definition: V is the class of all vonsets.

25.4.10. Axiom: H( D).
25.4.11. Theorem: H( V).

Proof: On account of Axioms 23.10 and 25.4.10. □

25.4.12. Corollary: V and D are orthodox.

Proof: Use Axiom 25.4.10, Theorem 25.4.11 and Theorem 23.4. □

25.4.13. Theorem: Co-extensional members of V ∪ D are identical.

Proof: Use Axioms 23.10 and 25.4.10, Theorem 25.4.11 and, finally, Theorem 23.7. □

25.4.14. Theorem: V = {𝑥|𝑥 ∈ V ∧ 𝑥 ∈ D}.

Proof: As H( V) and H( D) on account of Axiom 25.4.10 and Theorem 25.4.11, appeals

to Theorems 23.9 and Fact 25.4.7 suffice to finish the proof. □

25.4.15. Axiom The drift is wellfounded:

R(D)

25.4.16. Theorem All classes are wellfounded.

Proof: Invoke the result of Exercise 23.13. □

25.4.17. Theorem All vonsets are wellfounded.

Proof: Given Fact 25.4.7, a vonset in V is as well a class member of D. So the vonset is
wellfounded on account of Theorem 25.4.16. □

25.4.18. Theorem D is not a class.

Proof: If D were a class, it would on account of Definition 25.4.6 follow that D∈D, which
contradicts Axiom 25.4.15. □

25.4.19. Remark: Instead of postulating Axiom 25.4.15, one may obtain a suitable reg-
ular class V* of all regular vonsets by taking it to be the class of all elements of a potency
set of an ordinal in V. That invokes the consistency proof of ZFC with regularity given the
consistency of ZFC− = ZFC without regularity, by (Kunen 1980, chapter 3), or a similar
relative consistency proof. Given Kunen’s result, however, and the relative consistency
results obtained earlier by (Skolem 1923) and (Neumann 1929), we know that we can
safely posit Axiom 25.4.15.

87
25.5 Primitive theorems for classes 25 AND THE THEORIES OF VONSETS

25.4.20. Theorem: V is not a vonset.

Proof: Appeal to Definition 25.4.9 and Theorem 25.4.15. □

25.5 Primitive theorems for classes

We leave is as an exercise to prove the following from Definition 25.4.6.
25.5.1. Theorem V is a class:
V ∈ D.

25.5.2. Theorem Membership class:

E = {(𝑥, 𝑦)|𝑥 ∈ V ∧ 𝑦 ∈ V ∧ 𝑥 ∈ 𝑦} ∈ D.

25.5.3. Theorem Intersection class:

∀A ∈ D∀B ∈ D∃C ∈ D∀𝑥(𝑥 ∈ C ↔ 𝑥 ∈ A ∧ 𝑥 ∈ B).

25.5.4. Theorem Complement class:

∀A ∈ D∃B ∈ D∀𝑥(𝑥 ∈ B ↔ 𝑥 ∉ A).

25.5.5. Theorem Domain class:

∀A ∈ D∃B ∈ D∀𝑥(𝑥 ∈ B ↔ ∃𝑦((𝑥, 𝑦) ∈ A)).

25.5.6. Theorem Product by V class:

∀A ∈ D∃B ∈ D∀𝑥(𝑥 ∈ B ↔ ∃𝑦∃𝑧(𝑥 = (𝑦, 𝑧) ∧ 𝑦 ∈ A ∧ 𝑧 ∈ V)).

25.5.7. Theorem Circular permutation class:

∀A ∈ D∃B ∈ D∀𝑥∀𝑦∀𝑧((𝑥, 𝑦, 𝑧) ∈ B ↔ (𝑦, 𝑧, 𝑥) ∈ A).

25.5.8. Theorem Transposition class:

∀A ∈ D∃B ∈ D∀𝑥∀𝑦∀𝑧((𝑥, 𝑦, 𝑧) ∈ B ↔ (𝑥, 𝑧, 𝑦) ∈ A).

88
25.6 The Tuple-lemmas 25 AND THE THEORIES OF VONSETS

25.6 The Tuple-lemmas

25.6.1. Lemma:

∀A ∈ D∃B1 ∈ D∀𝑥∀𝑦∀𝑧((𝑥, 𝑦, 𝑧) ∈ B1 ↔ (𝑥, 𝑦) ∈ A ∧ 𝑧 ∈ V).

25.6.2. Lemma:

∀A ∈ D∃B2 ∈ D∀𝑥∀𝑦∀𝑧((𝑥, 𝑧, 𝑦) ∈ B2 ↔ (𝑥, 𝑦) ∈ A ∧ 𝑧 ∈ V).

25.6.3. Lemma:

∀A ∈ D∃B3 ∈ D∀𝑥∀𝑦∀𝑧((𝑧, 𝑥, 𝑦) ∈ B3 ↔ (𝑥, 𝑦) ∈ A ∧ 𝑧 ∈ V).

25.6.4. Lemma:
∀A ∈ D∃B4 ∈ D∀𝑥∀𝑦((𝑦, 𝑥) ∈ B4 ↔ (𝑥, 𝑦) ∈ A).

Proof: Use Theorem 25.5.6 to get B1 , Theorem 25.5.8 on B1 to get B2 , Theorem 25.5.7
on B1 to get B3 , and use Theorem 25.5.7 on B2 , plus Theorem 25.5.5, to get B4 . □

25.7 The class existence theorem

25.8 The expansion lemma
25.9 Proof that V is orthodox
25.10 Proof that all members of V are orthodox
As H by Axiom ... This is done already.

25.11 Global well ordering

Useful explanation of
Global well ordering given global choice.

89
25.11 Global well ordering 25 AND THE THEORIES OF VONSETS

90
 26 SPACE FOR LIBRATIONIST CATEGORY THEORY?

26 Space for librationist category theory?

La filosofia è scritta in questo grandissimo libro, che contin-

uamente ci sta aperto innanzi agli occhi (io dico l’Universo),
ma non si può intendere, se prima non il sapere a intender la
lingua, e conoscer i caratteri ne quali è scritto. Egli è scritto
in lingua matematica, e i caratteri son triangoli, cerchi ed
altre figure geometriche, senza i quali mezzi è impossibile
intenderne umanamente parola; senza questi è un aggirarsi
vanamente per un oscuro labirinto.

Galilei

The author has learned that set theories as NBGC + TA are considered ideal for category
theory, and wants to investigate whether that can be done in the librationist framework
set up for mentioned set theories in §25.

91
26 SPACE FOR LIBRATIONIST CATEGORY THEORY?

92
 27 THE THEORY OF VANSETS NF IN H(W)

27 The theory of vansets NF in £H(W)

The analogy between the myth of mathematics and the myth of physics is, in some
additional and perhaps fortuitous ways, strikingly close. Consider, for example,
the crisis which was precipitated in the foundations of mathematics, at the turn of
the century, by the discovery of Russell’s paradox and other antinomies of set the-
ory. These contradictions had to be obviated by unintuitive, ad hoc devices; our
mathematical myth-making became deliberate and evident to all. But, what, of
physics? An antinomy arose between the undular and the corpuscular accounts
of light; and if this was not as out-and-out a contradiction as Russell’s paradox,
I suspect that the reason is that physics is not as out-and-out as mathematics.

Willard van Orman Quine, in (Quine 1961, pp. 18–19)

We give an account of Willard van Quine’s set theory New Foundations, of (Quine 1937),
via the axiomatization offered by (Hailperin 1944, p. 10), which is adapted here:
𝑃0 : ∃𝛽∀𝑥(𝑥 ∈ 𝛽 ↔ ∃𝑦(𝑥 ∈ 𝑦 ∧ 𝑥 ∉ 𝑦)
𝑃1 : ∀𝑢∀𝑣∃𝛽∀𝑥(𝑥 ∈ 𝛽 ↔ (𝑥 ∉ 𝑢 ∨ 𝑥 ∉ 𝑣))
𝑃2 : ∀𝛼∃𝛽∀𝑥∀𝑦(({𝑥}, {𝑦}) ∈ 𝛽 ↔ (𝑥, 𝑦) ∈ 𝛼)
𝑃3 : ∀𝛼∃𝛽∀𝑥∀𝑦∀𝑧((𝑥, 𝑦, 𝑧) ∈ 𝛽 ↔ (𝑥, 𝑦) ∈ 𝛼)
𝑃4 : ∀𝛼∃𝛽∀𝑥∀𝑦∀𝑧((𝑥, 𝑧, 𝑦) ∈ 𝛽 ↔ (𝑥, 𝑦) ∈ 𝛼)
𝑃5 : ∀𝛼∃𝛽∀𝑥∀𝑦((𝑦, 𝑥) ∈ 𝛽 ↔ 𝑥 ∈ 𝛼)
𝑃6 : ∀𝛼∃𝛽(𝑥 ∈ 𝛽 ↔ ∀𝑢((𝑢, {𝑥}) ∈ 𝛼))
𝑃7 : ∀𝛼∃𝛽∀𝑥∀𝑦((𝑦, 𝑥) ∈ 𝛽 ↔ (𝑥, 𝑦) ∈ 𝛼)
𝑃8 : ∃𝛽∀𝑥(𝑥 ∈ 𝛽 ↔ ∃𝑦(𝑥 = {𝑦}))
𝑃9 : ∃𝛽∀𝑥∀𝑦(({𝑥}, 𝑦) ∈ 𝛽 ↔ 𝑥 ∈ 𝑦)
Notice that 𝑃0 was not included in (Hailperin 1944, p. 10).
U was reserved for the full universal set {𝑥|𝑥 = 𝑥} of £. In the previous section V was
reserved for the class of all vonsets, as defined via manifestation there.
W, with associated mnemonic device die Welt, is reserved the Quinean vanset of all
vansets, as defined via manifestation below in this section.

93
 27 THE THEORY OF VANSETS NF IN H(W)

27.1. Definition:

W(𝑣 0 , 𝑣 1 ) === ∀𝑣 {𝑥|∃𝑦(𝑥 ∈ 𝑦 ∧ 𝑥 ∉ 𝑦)} ∈ 𝑣 ∧
∀𝑤∀𝑥(𝑤 ∈ 𝑣 ∧ 𝑥 ∈ 𝑣 → {𝑦 ∈ 𝑣 1 | (𝑦 ∉ 𝑤 ∨ 𝑦 ∉ 𝑥)} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {({𝑥}, {𝑦} ∈ 𝑣 1 |(𝑥, 𝑦) ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {(𝑥, 𝑦, 𝑧) ∈ 𝑣 1 |(𝑥, 𝑦) ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {(𝑥, 𝑧, 𝑦) ∈ 𝑣 1 |(𝑥, 𝑦) ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {(𝑦, 𝑥) ∈ 𝑣 1 |(𝑥, 𝑦) ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {𝑥 ∈ 𝑣 1 |∀𝑦(𝑦 ∈ 𝑣 1 → (𝑦, {𝑥}) ∈ 𝑤)} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {(𝑦, 𝑥) ∈ 𝑣 1 |𝑥 ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {𝑥 ∈ 𝑣 1 |∃𝑦(𝑦 ∈ 𝑣 1 ∧ 𝑥 = {𝑦})} ∈ 𝑣) ∧


∀𝑤(𝑤 ∈ 𝑣 → {({𝑥}, 𝑦) ∈ 𝑣 1 |𝑥 ∈ 𝑦} ∈ 𝑣 ]

→ 𝑣0 ∈ 𝑣

Use Definitions 27.1 and 19.1.2 to obtain

27.2. Theorem:

∀𝑢(𝑢 ∈ W ↔ T T ∀𝑣 {𝑥|∃𝑦(𝑥 ∈ 𝑦 ∧ 𝑥 ∉ 𝑦)} ∈ 𝑣 ∧
∀𝑤∀𝑥(𝑤 ∈ 𝑣 ∧ 𝑥 ∈ 𝑣 → {𝑦 ∈ W |(𝑦 ∉ 𝑤 ∨ 𝑦 ∉ 𝑥)} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {({𝑥}, {𝑦} ∈ W |(𝑥, 𝑦) ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {(𝑥, 𝑦, 𝑧) ∈ W |(𝑥, 𝑦) ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {(𝑥, 𝑧, 𝑦) ∈ W |(𝑥, 𝑦) ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {(𝑦, 𝑥) ∈ W |(𝑥, 𝑦) ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {𝑥 ∈ W |∀𝑦(𝑦 ∈ W → (𝑦, {𝑥}) ∈ 𝑤)} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {(𝑦, 𝑥) ∈ W |𝑥 ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {𝑥 ∈ W |∃𝑦(𝑦 ∈ W ∧ 𝑥 = {𝑦})} ∈ 𝑣) ∧

∀𝑤(𝑤 ∈ 𝑣 → {({𝑥}, 𝑦) ∈ W |𝑥 ∈ 𝑦} ∈ 𝑣)

→𝑢∈𝑣 )

27.3. Theorem: W is orthodox.

Proof: Adapt the the proof of Theorem 13.2.3. □

94
 27 THE THEORY OF VANSETS NF IN H(W)

27.4. Theorem:

∀𝑢(𝑢 ∈ W ↔ ∀𝑣 {𝑥|∃𝑦(𝑥 ∈ 𝑦 ∧ 𝑥 ∉ 𝑦)} ∈ 𝑣 ∧
∀𝑤∀𝑥(𝑤 ∈ 𝑣 ∧ 𝑥 ∈ 𝑣 → {𝑦 ∈ W |(𝑦 ∉ 𝑤 ∨ 𝑦 ∉ 𝑥)} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {({𝑥}, {𝑦} ∈ W |(𝑥, 𝑦) ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {(𝑥, 𝑦, 𝑧) ∈ W |(𝑥, 𝑦) ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {(𝑥, 𝑧, 𝑦) ∈ W |(𝑥, 𝑦) ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {(𝑦, 𝑥) ∈ W |(𝑥, 𝑦) ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {𝑥 ∈ W |∀𝑦(𝑦 ∈ W → (𝑦, {𝑥}) ∈ 𝑤)} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {(𝑦, 𝑥) ∈ W |𝑥 ∈ 𝑤} ∈ 𝑣) ∧
∀𝑤(𝑤 ∈ 𝑣 → {𝑥 ∈ W |∃𝑦(𝑦 ∈ W ∧ 𝑥 = {𝑦})} ∈ 𝑣) ∧

∀𝑤(𝑤 ∈ 𝑣 → {({𝑥}, 𝑦) ∈ W |𝑥 ∈ 𝑦} ∈ 𝑣)

→𝑢∈𝑣 )

Proof: A consequence of Theorem 27.2 as W is orthodox, given Theorem 27.3. □

27.5. Axiom: H( W)
27.6. Theorem: Co-extensional sets in W are identical.

Proof: Use Axiom 23.10 and Theorem 23.7. □

The proper identity for W is of course given by

27.7. Definition:
W
𝑎 = 𝑏 === ∀𝑣(𝑣 ∈ W → (𝑎 ∈ 𝑣 → 𝑏 ∈ 𝑣).

By Axiom 27.5, Theorem 27.6 and Theorem 27.4 combined with the results of (Hailperin
1944), it follows that £W accounts for Quine’s set theory NF.

95
27 THE THEORY OF VANSETS NF IN H(W)

96
 28 𝐴 IS TRUE JUST IF 𝐴 STATES THE TRUTH

28 𝐴 is true just if 𝐴 states the truth

La logique est l’hygiène des mathématiques.

André Weil

The following perspective upon the semantics is useful for some purposes.
28.1. Definition: The closure ordinal Ϙ is the truth.
28.2. Definition: The way of sentence A is [𝛿 : 𝛿 ⪯ Ϙ & (Ξ, 𝛿) ⊩ A].
28.3. Definition: A states the supremum of its way.
28.4. Definition: A expresses its way.
28.5. Definition: A is true just if A states the truth.
28.6. Definition: 𝐴 is false just if ¬𝐴 is true.
28.7. Definition: The way of 𝐴 ∧ 𝐵 is the way of 𝐴 ∩ the way of 𝐵.
28.8. Definition: The way of ¬𝐴 is Ϙ minus the way of 𝐴.
Here the sentence A is true should be interpreted as ⊢ T A , and the sentence A states
the truth as equivalent with ⊢ A.
Moreover, “just if” is here to be interpreted via the bidirectional entailment in

⊢ T A ⇔ ⊢ A.

It is a fact that
⊢ ¬T A ⇔ ⊢ ¬A,
so, consequently,
⊢M T A ⇔ ⊢M A.

The connectives are not truth-functional in librationism, but they are way-functional,
and can be accounted for by following classical interdefinability connections as in any
Boolean algebra: The way of the negation ¬A of A, is truth minus the way of A, and the
way of the conjunction A ∧ B is the intersection of the way of A and the way of B. The
ways of sentences built up from other connectives follow from their definitions in terms
¬ and ∧.
According to librationism, a true paradoxical sentence L and its true companion sentence
¬L complement each other. For the way of L, as defined in Definition 28.2, is in such a
case a set of ordinals with Ϙ as least upper bound, whereas as well the way of ¬L is a set
of ordinals with Ϙ as least upper bound; moreover, the ways of L and ¬L do not overlap.
Thus, by the Definition 28.4, L does not express the same as what ¬L expresses, for L
and ¬L have different ways.

97
28 𝐴 IS TRUE JUST IF 𝐴 STATES THE TRUTH

98
REFERENCES REFERENCES

References
Beeson, M. (1985). Foundations of Constructive Mathematics. Springer.
Bjørdal, F. A. (1998). “Towards a Foundation for Type-Free Reasoning”. The Logica Year-
book 1997. Ed. by Timothy Childers. FILOSOFIA, by the Academy of Sciences of the
Czech Republic, pp. 259–273.
– (2005). “There are Only Countably Many Objects”. The Logica Yearbook 2004. Ed. by
M. Bilkova and L. Behounek. FILOSOFIA, Prague, pp. 47–58.
– (2006). “Minimalistic Librationism: An Adequate, Acceptable, Consistent and Contra-
dictory Foundation”. The Logica Yearbook 2005. Ed. by M. Bilkova and L. Behounek.
FILOSOFIA, Prague, pp. 39–50.
– (2008). “«2+2=4» er misvisande”. Enhet i mangfold - Festskrift for Johan Arnt Myrstad.
Ed. by A. Leirfalll and T. Sandmel, pp. 55–66. isbn: 978-82-303-1179-0.
– (2011). “Considerations Contra Cantorianism”. The Logica Yearbook 2010. Ed. by M.
Pelis and V. Puncochar. Coll. Publ., London, pp. 43–52.
– (2012). “Librationist Closures of the Paradoxes”. Log. Log. Philos. 21.4, pp. 323–361.
– (2015). “On the Type-Free Paracoherent Foundation of Mathematics with the Sedate
Extension of Classical Logic by Librationist Set Theory £, and Specifically on why £
is Neither Inconsistent nor Contradictory nor Paraconsistent”. In New Directions in
Paraconsistent Logic. Proceedings from the 5th World Congress on Paraconsistency.
Kolkata, India, February 2014. Ed. by Mihir Chakraborty Jean-Yves Beziau and Soma
Dutta. Springer Proceedings in Mathematics & Statistics 152, pp. 509–515. isbn: 978-
81-322-2717-5.
Burgess, J. P. (1986). “The Truth is Never Simple”. J. Symb. Log. 51.3, pp. 663–681.
Cantini, A. (1996). Logical Frameworks for Truth and Abstraction: An Axiomatic Study.
Elsevier.
Church, A. (1976). “Comparison of Russell’s Resolution of the Semantical Antinomies
with that of Tarski”. J. Symb. Log. 41.4, pp. 747–760.
Forster, T. (2019). “Quine’s New Foundations”. The Stanford Encyclopedia of Philosophy.
Ed. by Edward N. Zalta. Summer 2019. Metaphysics Research Lab, Stanford University.
Fraenkel, A. A. and Y. Bar-Hillel (1958). Foundations of Set Theory. 1st edition. North-
Holland Publishing Company.
– (1973). Foundations of Set Theory. 2nd edition. Atlantic Highlands, NJ, USA: Elsevier.
Friedman, H. (1973). “The Consistency of Classical Set Theory Relative to a Set Theory
with Intuitionistic Logic”. J. Symb. Log. 38.2, pp. 315–319.
Friedman, H. and M. Sheard (1987). “An Axiomatic Approach to Self-Referential Truth”.
Ann. Pure Appl. Log. 33, pp. 1–21.
Gilmore, P. C. (1974). “The Consistency of Partial Set Theory Without Extensionality”. In:
Jech, pp. 147–153.
Gödel, K. (1931). “Über Formal Unentscheidbare Sätze der Principia Mathematica und
Verwandter Systeme I”. Monatshefte für Mathematik 38.1, pp. 173–198.
Grim, P. (1991). The Incomplete Universe: Totality, Knowledge, and Truth. Cambridge,
Massachusetts: The MIT Press.

99
REFERENCES REFERENCES

Hachtman, S. (2019). “Determinacy and Monotone Inductive Definitions”. Isr. J. Math.

230, pp. 71–96.
Hailperin, T. (1944). “A Set of Axioms for Logic”. Journal of Symbolic Logic 9.1, pp. 1–19.
doi: 10.2307/2267307.
Halbach, Volker (1994). “A System of Complete and Consistent Truth”. Notre Dame Jour-
nal of Formal Logic 35.3, pp. 311–327. doi: 10.1305/ndjfl/1040511340.
Herzberger, H. (1980). “Notes on Periodicity”. Circulated manuscript.
Hunter, G. (1971). Metalogic: An Introduction to the Metatheory of Standard First Order
Logic. Berkeley: University of California Press.
Jaskowski, S. (1948). “Rachunek Zdań dla Systemow Dedukcyjnych Sprzecznych”. Studia
Societatis Scientiarum Torunensis 1.5, pp. 57–77.
– (1999). “A Propositional Calculus for Inconsistent Deductive Systems”. Log. Log. Phi-
los. 7. Translation of (Jaskowski 1948), pp. 35–56.
Kunen, K. (1980). Set Theory - An Introduction to Independence Proofs. First edition.
Amsterdam: North-Holland Publishing Company. isbn: 0720422000.
Leitgeb, H. (2007). “What Theories of Truth Should Be Like (but Cannot Be)”. Philosophy
Compass 2.2, pp. 276–290. doi: 10.1111/j.1747-9991.2007.00070.x.
McGee, V. (1985). “How Truthlike can a Predicate be? A Negative Result”. J. Philos. Log.
14.4, pp. 399–410.
Montague, R. (1963). “Syntactic Treatment of Modality, with Corollaries on Reflection
Principles and Finite Axiomatizeability”. Acta Philosophica Fennica 16, pp. 153–167.
Neumann, J. von (1929). “Über eine Widerspruchfreiheitsfrage in der axiomatischen
Mengenlehre”. J. fur Reine Angew. Math. 160, pp. 227–241.
Priest, G., F. Berto, and Z. Weber (2022). Dialetheism. SEP. url: https : / / plato .
stanford.edu/archives/fall2022/entries/dialetheism/.
Quine, W. V. (1937). “New Foundations for Mathematical Logic”. American Mathematical
Monthly 44, pp. 70–80.
– (1961). “On what there is”. From a Logical point of view. Harvard University Press.
Chap. 1, pp. 1–19.
Ramsey, F. (1925). “The Foundations of Mathematics”. Proc. London Math. Soc. 25,
pp. 338–384.
Regis, E. (1988). Who got Einstein’s office? Eccentricity and Genius at the Institute for
Advanced Study. Simon & Schuster.
Scott, D. (1961). “More on the Axiom of Extensionality”. In: Bar-Hillel1962, pp. 115-131.
– (1974). “Axiomatizing Set Theory”. In: Jech, pp. 207-214.
Shapiro, S, ed. (1985). Intensional Mathematics. Elsevier.
Sheard, M. (2003). “Truth, Provability, and Naive Criteria”. In Principles of Truth. Ed. by
V. Halbach and L. Horsten, pp. 169–181.
Skolem, T. (1923). “Axiomatized set theory”. From Frege to Gödel. Ed. by J. van Hei-
jenoort. Trans. by S. Bauer-Mengelberg, pp. 291–301.
Smorynski, C. (1977). “The incompleteness theorems”. Handbook of Mathematical Logic.
Ed. by J. Barwise. North Holland, pp. 821–865.
Turner, R. (1990). Truth and Modality for Knowledge Representation. MIT Press.
Visser, A. (1989). “Handbook of Philosophical Logic”. Ed. by Dov Gabbay et al. 1st edition.
Vol. 4. Springer. Chap. Semantics and the Liar Paradox, pp. 617–706.

100
REFERENCES REFERENCES

Welch, P. D. (2011). “Weak Systems of Determinacy and Arithmetical Quasi-inductive

Definitions”. J. Symb. Log. 76, pp. 418–436.

101