1 Version Information

24.05.21: a lot more minor corrections (thanks to Joel Summerfield).

08.02.18: Corrections to the Levy Reflection Principle.
10.01.18: More corrections and some minor additions.
02.03.17: More corrections (thanks to Peter Neumann).
08.02.17: Added material up to L |= V = L and fixed more errors.
27.01.17: Many thanks to Peter Neumann for the numerous little errors he
spotted.
16.01.17: I have rewritten and restructured some of the introductory ma-
terial. At the moment these notes cover the course up to and including the
definition of the cumulative hierarchy.

2 How to use the Lecture Notes

These lecture notes start with a lot of technicalities the reason for which will
become clearer (I hope) later on. In the lectures, I will thus start with Section 3
and then skip to Section 7, go back to Section 5 and then proceed from Section
8 onwards. I will not cover Section 4 systematically but as the need arises.
If you find any mistakes, please contact me.

3 Background
We work in first order logic with equality and are concerned with theories with
one binary relation, written ∈. The language is called the Language of Set
Theory (LST).
Our objects are going to be called sets and denoted (generally) by small latin
letters.
For quantifiers, we use the abbreviations

∀x ∈ y ψ ≡ ∀x [x ∈ y → ψ]

and
∃x ∈ y ψ ≡ ∃x [x ∈ y ∧ ψ] .
Finally we single out a specific collection of formulae: we call a formula a
∆0 formula if every quantifier is bounded. Formally, we define the collection
of ∆0 formula (in the metatheory) by recursion: it is the smallest collection
of formulae that contains the atomic ones (x ∈ y and x = y) is closed under
the logical connectives ∧, ∨, →, ↔, ¬ and if ψ is ∆0 then so are ∀x ∈ y ψ and
∃x ∈ y ψ (y must be a variable).
Note that it might not make sense to talk about the ‘collection’ of ∆0 for-
mulae. This depends on your metatheory! But for any specific formula, it is
trivial to verify (in any reasonable metatheory) whether or not this is ∆0 .

1
3.1 Mathematics and Metamathematics
Note that we are, mostly, studying ∈-structures from the ‘outside’. Of course,
the question arises in which theory we do this, i.e. what our ‘metatheory’ is.
We are purposely vague about this. Any reasonable finitistic metatheory should
work. It is however important that we do not confuse the metatheory and the
theory. We will, for example, prove theorems within our theory (see for example
the section on Ordinals). We will also prove theorems which cannot even be
stated about in our theory: for example if we say ‘for every formula φ of LST
...’, this is a theorem in the metatheory.

4 Defined Notions
4.1 Classes
A class Cφ is a formula φ(t) with one free variable (here t). Instead of writing
φ(x) we write x ∈ Cφ .
We frequently do not give an explicit formula for a class but simply denote
it by a capital latin letter.
In the following we define various notions for sets. When these make sense
for classes, we will use the same notation. As an example of this, if Cφ is a class
given by the formula φ(t) we write
∀x ∈ Cφ ψ ≡ ∀x [x ∈ Cφ → ψ] ≡ ∀x [φ(x) → ψ] .

4.1.1 The Universe

We will write U for the class given by the formula t = t (so that trivially
∀x x ∈ U ).
Note however that the formula ∀x ∈ U ψ is not ∆0 : it is after all just an
abbreviation for the formula ∀x [x = x → ψ] and this is not ∆0 .

4.2 Defined Notions

To enhance readability of our formulae, we define abbreviations for formulae of
LST.
The formulae given are somewhat arbitrary, there are plenty others which
are equivalent. We try to give ∆0 -formulae - this is important later on when we
talk about absoluteness.
We will often define something like
z = {x, y} ≡ φ(x, y, z)
but then use {x, y} ∈ w to mean ∃t ∈ w [t = {x, y}] or more explicitly ∃t ∈
w φ(x, y, t). This is ∆0 provided that the definition of w = {x, y}, namely φ,
was ∆0 .
We will also write w ∈ {x, y}. This is harder to define abstractly, so we are
(sometimes) explicit about it.

2
4.3 Comprehension
For a formula φ
z = {y : φ(y)} ≡ ∀t [t ∈ z ↔ φ(t)] .

4.4 Subset
x ⊆ y ≡ ∀t ∈ x t ∈ y.

4.5 Emptyset

z = ∅ ≡∀t ∈ z [t 6= t]
t ∈ ∅ ≡t 6= t
∅ ∈ t ≡∃z ∈ t z = ∅.

4.6 Unordered Pair

z = {x, y} ≡ x ∈ z ∧ y ∈ z ∧ ∀t ∈ z [t = x ∨ t = y]
and
t ∈ {x, y} ≡ t = x ∨ t = y.
We use the shorthand
{x} ≡ {x, x} .

4.7 Ordered Pair

z = hx, yi ≡ z = {{x} , {x, y}} ≡∃t ∈ z [t = {x, x}] ∧

∃t ∈ z [t = {x, y}] ∧
∀t ∈ z [t = {x, x} ∨ t = {x, y}] .

Note that the first statement only makes sense if {x} and {x, y} are objects of
our model, whereas the full formula is defined independently of the existence of
{x} and {x, y}.

z is an ordered pair ≡ ∃a ∈ z∃b ∈ z∃x ∈ a∃y ∈ b z = hx, yi

We could of course simply write down the more obvious

z is an ordered pair ≡ ∃x∃y z = hx, yi

but this is not ∆0 and we prefer ∆0 formulae for reasons which will become
clear later (see Lemma 1).
We will frequently refer to the first and second coordinate of the ordered
pair. To do so we define

x = π1 (z) ≡ z is an ordered pair ∧ ∃b ∈ z∃y ∈ b [z = hx, yi]

3
and
y = π2 (z) ≡ z is an ordered pair ∧ ∃a ∈ z∃x ∈ a [z = hx, yi] .
We will also use the fact that

z = z 0 are ordered pairs ↔ ∃x, y [π1 (z) = x = π1 (z 0 ) ∧ π2 (z) = y = π2 (z 0 )]

4.8 Relations
We are only interested in binary relations, so simply use the word ‘relation’ for
binary relations.

r is a relation ≡ ∀t ∈ r t is an ordered pair.

r is a transitive relation ≡ r is a relation ∧∀u ∈ r∀v ∈ r [π2 (u) = π1 (v) → hπ1 (u), π2 (v)i ∈ z] .

It is somewhat unclear how the above is translated into a formula of LST (i.e.
how to eliminate the defined notions). Here is one way to do that:

r is a transitive relation ≡r is a relation ∧

∀u ∈ r∀v ∈ r∃x, y, z
[[x = π1 (u) ∧ y = π2 (u) ∧ y = π1 (v) ∧ z = π2 (v)] → ∃w ∈ r [w = hx, zi]] .

In this formula, not every quantifier is bounded, but we can write one down
where every quantifier is indeed bounded: replace the last bit by

∃a ∈ u∃x ∈ a∃b ∈ u∃y ∈ b∃c ∈ v∃z ∈ c [u = hx, yi ∧ v = hy, zi → ∃w ∈ r [w = hx, zi]] .

The more you do this, the more unwieldy the formula becomes. We will in the
future hence avoid these complicated formulae, but you should always check
that you can carry out this replacement.
We will need to consider relations on sets so we will define

x ∈ dom(r) ≡ r is a relation ∧ ∃z ∈ r [x = π1 (z)]

and
y ∈ ran(r) ≡ r is a relation ∧ ∃z ∈ r [y = π2 (z)]
as well as the classes (!)

dom(r) = {x : x ∈ dom(r)} ran(r) = {y : y ∈ ran(r)} .

Although these definitions seem self-referential, they are not: the string ‘x ∈
dom(r)’ is simply replaced by ‘r is a relation∧∃z ∈ r [x = π1 (z)]’. In particular,
if r is not a relation then dom(r) = ∅.
Note that we seem to have defined infinitely many classes here, which is of
course a bad thing. But in practice we will only ever use finitely many instances

4
of this or insist on r being a set and having sufficiently many axioms available
to show that this implies that dom(r) and ran(r) are sets.
So we can define

r is a relation on x ≡ r is a relation ∧ dom(r) ⊆ x ∧ ran(r) ⊆ x

and

r is a transitive relation on x ≡ r is a relation on x∧∀u, v, w ∈ x [hu, vi ∈ r ∧ hv, wi ∈ r → hu, wi ∈ r] .

Note that this is ∆0 (provided r and x are sets).

Similarly

r is a reflexive relation on x ≡ r is a relation on x ∧ ∀t ∈ x [ht, ti ∈ r]

r is an irreflexive relation on x ≡ r is a relation on x ∧ ∀t ∈ x [ht, ti 6∈ r] .

Of course, usually we write binary relations in infix notation, i.e. arb instead
of ha, bi ∈ r.

4.9 Orders
r is a partial strict order on x ≡ r is a transitive, irreflexive relation on x.

r is a total strict order on x ≡r is a transitive, irreflexive relation on x∧

∀u, v ∈ x [hu, vi ∈ r ∨ hv, ui ∈ r ∨ u = v] .

r is a strict well-order on x ≡r is a strict total order on x∧

∀p [p ⊆ x ∧ p 6= ∅ → ∃m ∈ p [∀t ∈ p ¬ [t r m]]]

Note that this latter definition is not a ∆0 -formula (and in fact cannot be
replaced by one).

4.10 Functions
f is a function ≡ f is a relation ∧∀x ∈ dom(f )∀y ∈ ran(f )∀y 0 ∈ ran(f ) [xf y ∧ xf y 0 → y = y 0 ]
We can replace this by a ∆0 formula (as it stands it is not since dom(f ) and
ran(f ) are defined notions and not variables).
If we have shown that f is a function and x ∈ dom(f ), we will write f (x)
for the unique y such that xf y.

5
4.11 Class functions
Note that the notion of being a function makes sense for classes. In this case
we talk about ‘class functions’. We spell out what this means formally (because
class functions are central to what we will be doing).
Definition 1. Suppose φ(t) is a formula of LST. We say that φ is a class
function if and only if

∀t [φ(t) → t is an ordered pair] ∧

∀x∀y∀y [φ(hx, yi) ∧ φ(hx, y 0 i) → y = y 0 ] .
0

Usually we denote class functions by capital latin letters and write F (x) = y
for φ(hx, yi) (where φ is the formula defining F ).
If A, B are classes, we say that F : A → B is a class function if and only if

F is a class function ∧
dom(F ) = A ∧ ran(F ) ⊆ B

4.12 Union
[
z= x ≡ [∀y ∈ x∀t ∈ y [t ∈ z]] ∧ [∀t ∈ z∃y ∈ x [t ∈ y ∧ y ∈ x]]
and
z = x ∪ y ≡ x ⊆ z ∧ y ⊆ z ∧ ∀t ∈ z [t ∈ x ∨ t ∈ y] .
Again, you might prefer the more natural
[
z= x ≡ ∀t [t ∈ z ↔ ∃y ∈ x t ∈ y]

but this is not ∆0 .

4.13 Powerset
z = P (x) ≡ ∀t [t ∈ z ↔ t ⊆ x] .
Although we won’t in fact prove it in this course (but I will frequently remark
on it), there is no ∆0 formula which we could use here.

4.14 Successor
z = x + 1 ≡ z = x ∪ {x, x} ≡ x ⊆ z ∧ x ∈ z ∧ ∀t ∈ z [t ∈ x ∨ t = x] .

4.15 Inductive Set

Ind(x) ≡ ∅ ∈ x ∧ ∀t ∈ x [t + 1 ∈ x] .

6
 5 Axiom Summary
Recall the definitions:

x ⊆ y ≡∀t ∈ x t ∈ y
z = ∅ ≡∀t ∈ z t 6= t
z = {x, y} ≡x ∈ z ∧ y ∈ z ∧ ∀t ∈ z [t = x ∨ t = y]
[
z= x ≡∀t ∈ z∃y ∈ x [t ∈ y] ∧ ∀y ∈ x∀t ∈ y [t ∈ z]
z = P (x) ≡∀t [t ⊆ x → t ∈ z] ∧ ∀t ∈ z [t ⊆ x]
z = S(x) ≡x ∈ z ∧ ∀t ∈ x [t ∈ z] ∧ ∀t ∈ z [t = x ∨ t ∈ x]

and then state (any free variables are implicitly universally quantified)
Extensionality
x⊆y∧y ⊆x→x=y

Separation For each formula φ(v1 , . . . , vn , vn+1 ) with all free variables shown

∀x∃y y = {z ∈ x : φ (a1 , . . . , an , z)} .

Emptyset
∃z z = ∅

Pairing
∀x∀y∃z z = {x, y}

Union [
∀x∃y y = x

Powerset
∀x∃y y = P (x)

Replacement For each formula φ(p1 , . . . , pn , t, u) (with all free variables displayed)

∀a1 , . . . , an ∀d
∀x ∈ d∃!y φ(a1 , . . . , an , x, y)
→
∃z z = {y : ∃x ∈ d φ(a1 , . . . , an , x, y)}

Infinity
∃z [∃x ∈ z x = ∅ ∧ ∀y ∈ z∃w ∈ z w = S(y)]

Foundation
∀x [x 6= ∅ → ∃y ∈ x [∀z ∈ x [z 6∈ y]]]

Choice We will eventually state the Well-ordering Principle which is equiva-

lent to the usual Axiom of Choice (see Part b, Set Theory).

7
 5.1 Axiom Systems
We will sometimes work in axiom systems which do not include all of the above
axioms. Standard abbreviations are
ˆ ZF− (sometimes denoted by ZF? ): Extensionality + Separation +
Emptyset + Pairing + Union + Powerset + Replacement + Infinity.
ˆ ZF: ZF− + Foundation.
ˆ ZFC: ZF + Choice.

6 Alternative Axiom Summary

If you would like to remove the defined notions above, you may end up with
something like the following:

Extensionality
∀x∀y [∀z [z ∈ x ↔ z ∈ y] → x = y]

Separation For each formula φ(p1 , . . . , pn , t) (with all free variables displayed)

∀a1 , . . . , an ∀x∃y∀z [z ∈ y ↔ z ∈ x ∧ φ (a1 , . . . , an , z)]

This is also called the Comprehension Scheme and is often written as

∀a1 , . . . , an ∀x∃y y = {z ∈ x : φ (a1 , . . . , an , z)} .

Emptyset
∃x∀y y 6∈ x
Note that Emptyset in fact follows form Separation with the formula
φ(t) ≡ t 6= t and the existence of any x (or by assuming that all models
are non-empty).
Pairing
∀x∀y∃z∀t [t ∈ z ↔ [t = x ∨ t = y]]
Note that this definition of z = {x, y} ≡ ∀t [t ∈ z ↔ [t = x ∨ t = y]] is not
∆0 , so we would have to prove manually that it is absolute for transitive
non-empty classes A ⊆ B (assuming ‘enough’ of ZF− ). With Comprehen-
sion this is equivalent to

∀x∀y∃z [x ∈ z ∧ y ∈ z]

Union
∀x∃y∀w [w ∈ y ↔ ∃z [w ∈ z ∧ z ∈ y]]
With Comprehension this is equivalent to

∀x∃y∀z∀w [w ∈ z ∧ z ∈ x → w ∈ y]

8
 Powerset
∀x∃y∀z [z ∈ y ↔ ∀w [w ∈ z → w ∈ x]]
With Comprehension this is equivalent to

∀x∃y∀z [∀w [w ∈ z → w ∈ x] → z ∈ y]

Replacement For each formula φ(p1 , . . . , pn , t, u) (with all free variables displayed)

∀a1 , . . . , an ∀d
∀x ∈ d∃!y φ(a1 , . . . , an , x, y)
→
∃z z = {y : ∃x ∈ d φ(a1 , . . . , an , x, y)}

Infinity h [ i
∃z ∃x ∈ z x = ∅ ∧ ∀y ∈ z∃w ∈ z w = {y, {y, y}}

Foundation
∀x [x 6= ∅ → ∃y ∈ x [∀z ∈ x [z 6∈ y]]]

Choice We will eventually state the Well-ordering Principle which is equiva-

lent to the usual Axiom of Choice (see Part b, Set Theory).

9
7 Relativization and Absoluteness
I will only define ‘relativization’ and ‘absoluteness’ for LST, but it can easily
be defined for any theory. The examples in lectures are intended to be given in
the relevant theories.
Definition 2. Given a class A and a formula φ of LST, the relativization of φ
to A is the formula φA where each quantifier is bounded by A. Formally, by
induction on the complexity of the formula we define
ˆ (x = y)A ≡ (x = y);
ˆ (x ∈ y)A ≡ (x ∈ y);
ˆ (¬φ)A ≡ ¬φA ;
ˆ (φ ∧ ψ)A ≡ (φA ∧ ψ A );
ˆ (φ ∨ ψ)A ≡ (φA ∨ ψ A );
ˆ (φ → ψ)A ≡ (φA → ψ A );
ˆ (φ ↔ ψ)A ≡ (φA ↔ ψ A );
ˆ (∃x φ)A ≡ ∃x ∈ A φA ≡ ∃x (x ∈ A ∧ φA );
ˆ (∀x φ)A ≡ ∀x ∈ A φA ≡ ∀x (x ∈ A → φA ).
If a1 , . . . , an ∈ A and φ free variables x1 , . . . , xn then we also write

A |= φ(a1 , . . . , an ) ≡ φ(a1 , . . . , an )A .

This is just your standard ‘interpretation of φ in the model A’, except that
of course A might not be a ‘model’ (it might not be a set but only a class) and
the above is purely syntactic.
Remark 1 (Expanding on the previous sentence). Intuitively, the relativization
of a formula φ to a class A is simply the interpretation of of φ in the model
(A, ∈). The problem we are facing is that if A is really a class and not a set (in
our meta-theory), then it does not make sense to talk about the model (A, ∈).
We cannot (or do not want to) use the semantic notions, so have to rely on
purely syntactical defintions.
If you are willing to work in a meta-theory in which you can show that ‘if
a theory is consistent, then it has a model’ this problem goes away. You will
assume that ZF− is consistent, and call its model (U, ∈) (so U really does exist
as an object of study - just like you usually assume in ordinary mathematics
that R really does exist as an object of study). You then work with a subobject
A of U and can interpret formulae in (A, ∈) in the classical model theoretic
sense.
If you do this, formally we will then be showing that ‘if ZF− is consistent,
then so is ZF and ZFC and ZFC + CH’. But of course, it is very likely that if

10
your meta-theory is strong enough to show that ‘if a theory is consistent, then
it has a model’, then it will include some version of Choice, so that somewhat
defeats the purpose.
If you are prepared to formally jump through the extra logical hoops, then
you can get away with a much weaker meta-theory (some finitistic meta-theory)
and prove the ‘stronger’ statement ‘if ZFC + CH is inconsistent, then so is
ZF− ’. For (most) practical purposes it is enough to think about the intuitive
definition, but keep these comments in the back of your mind.
Definition 3. Suppose A is a class and φ(x1 , . . . , xn ) is a formula of LST (with
all free variables shown) and Γ is a collection of sentences.
We say that A models φ (or A believes φ) in the context Γ, written A |=Γ φ
if and only if
Γ ` ∀a1 , . . . , an ∈ A φ(a1 , . . . , an )A .
If ∆ is a collection of formulae of LST, we say that A models ∆ (or A
believes in ∆) in the context Γ, written A |=Γ ∆ if and only if for each φ from
∆, A |=Γ φ.
Usually we do not specify Γ explicitly and take it to be some suitable sub-
collection of ZFC.
Definition 4. Suppose φ(x1 , . . . , xn ) is a formula of LST (with all free variables
shown), A, B are classes and Γ is a collection of sentences such that Γ ` A ⊆ B.
We say that φ is absolute for A, B in the context Γ if and only if
Γ ` ∀a1 , . . . , an ∈ A (φ(a1 , . . . , an )A ↔ φ(a1 , . . . , an )B ).
In this case we write A φ B.
Note that often we do not mention Γ. For us, it will always be a subset of
ZFC and we take it whatever subset we need in a proof.
Intuitively this says that the two models (A, ∈) and (B, ∈) have the same
believe about some formula φ whenever that makes sense (i.e. all free variables
are instantiated with elements of both A and B). Since formally, we will not
be working with ‘models’ (of ZF− or ZF or ZFC), we have to give a syntactic
definition which only relies on the existence of proofs (finite sequences of finite
strings of symbols which follow some easily checkable rules).
Recall that a class A is transitive if and only if ∀x ∈ A x ⊆ A.
Definition 5. Suppose φ is a formula of LST.
We say that φ is absolute (for transitive classes satisfying ∆) [in the context
Γ] if and only if for any (transitive) classes A, B (such that Γ ` A |= ∆ and
Γ ` B |= ∆) and Γ ` A ⊆ B, φ is absolute for A, B in the context Γ.
Lemma 1. ∆0 formulae are absolute for transitive classes.
Proof. Let A, B be transitive classes such that A ⊆ B.
We do this by induction of the complexity of the formula. By the definition
of relativization, the only interesting steps are the quantifiers. We do the exis-
tential case. The universal case is similar or can be deduced by the replacement
∀ ≡ ¬∃¬.

11
Case φ ≡ ∃x ∈ y ψ(x, y, x1 , . . . , xn ): Fix x1 , . . . , xn , y ∈ A.
First assume A |= ∃x ∈ y ψ(x, y, x1 , . . . , xn ), i.e. ∃x ∈ A (x ∈ y ∧
ψ(x, y, x1 , . . . , xn )A ). Find x ∈ A such that x ∈ y ∧ ψ(x, y, x1 , . . . , xn )A .
Since A ⊆ B we have x ∈ B and by absoluteness of ψ (inductive assumption)
ψ(x, y, x1 , . . . , xn )B . Thus B |= ∃x ∈ y ψ(x, y, x1 , . . . , xn ).
Conversely, assume that B |= ∃x ∈ y ψ(x, y, x1 , . . . , xn ), i.e. ∃x ∈ B (x ∈ y ∧
ψ(x, y, x1 , dots, xn )B ). Find x ∈ B such that x ∈ y ∈ A and ψ(x, y, x1 , . . . , xn )B .
By transitivity of A we have x ∈ A and by absoluteness of ψ (inductive assump-
tion) ψ(x, y, x1 , . . . , xn )A . Thus A |= ∃x ∈ y ψ(x, y, x1 , . . . , xn ).
Remark 2. Note that the above in fact shows that if ψ(x) is absolute for A ⊆ B
and A |= ∃x ψ then B |= ∃x ψ. It is only for the reverse direction that we need
transitivity and the fact that the quantifier is bounded.
Remark 3. This is a classic ‘meta-theorem’. View it as a ‘factory’ that produces
proofs: we will be interested (mostly) in specific applications of this to specific
formulae. So for example we will have reason to show that ‘pairing is absolute’,
i.e. that if (we can prove that) A ⊆ B are non-empty transitive classes, then
we can prove that ∀x, y, z ∈ A [A |= z = {x, y} ↔ B |= z = {x, y}] (this is a
sentence in LST). We could (and you should) go through these proofs by hand,
the above lemma is simply an abbreviating step.

8 The Ordinals
In this section we develop just enough of the theory of ordinals to prove the
results of the next section. You are strongly encouraged to consult the literature
for more results about ordinals.
Unless otherwise specified, we work in ZF− − Powerset (it may be fun to
figure out exactly which axioms are needed to prove the results below).
Definition 6.
On = {α : α is transitive ∧ ∈ is a well-order on α}
For the benefit of the reader we write out α ∈ On:
∀x ∈ α∀y ∈ x y ∈ α [α is transitive]
∧∀x ∈ α [x 6∈ x]
∧∀x, y, z ∈ α [x ∈ y ∧ y ∈ z → x ∈ z]
∧∀x, y ∈ α [x ∈ y ∨ y ∈ x ∨ x = y] [∈ is a strict total order on α]
∧∀x [x ⊆ α ∧ x 6= ∅ → ∃m ∈ x∀z ∈ m z 6∈ x] [∈ is well-founded on α]
Lemma 2.
ZF ` α ∈ On ↔ α is transitive and totally ordered by ∈
Thus if A, B are non-empty transitive classes satisfying (enough of ) ZF then
x ∈ On is absolute.

12
Lemma 3. 1. ∀α ∈ On α 6∈ α;
2. ∅ ∈ On;
3. ∀α ∈ On α + 1 := α ∪ {α} ∈ On;

4. ∀α, β ∈ On α ∩ β ∈ On;
5. ∀α, β ∈ On [α ⊆ β → α ∈ β ∨ α = β];
6. ∀α ∈ On α ⊆ On;

Proof. 1. Suppose α ∈ On and α ∈ α. Then by irreflexivity, α 6∈ α, a

contradiction.
2. ∅ is vacuously transitive and ∈ is vacuously a well-order on ∅.
3. Suppose α ∈ On.

α + 1 is transitive: Suppose x ∈ y ∈ α + 1. Either y ∈ α and then by

transitivity of α, x ∈ α or y = α and then x ∈ α. In any case x ∈ α ⊆ α+1.

∈ is irreflexive on α + 1: Suppose x ∈ α + 1. Then either x ∈ α so

that by assumption x 6∈ x or x = α and thus α 6∈ α by (1).

∈ is transitive on α + 1: Suppose x, y, z ∈ α + 1 and x ∈ y ∈ z. If

x, y, z ∈ α then we get x ∈ z from the assumption that α ∈ On. Next,
assume x = α. Then y 6= α by (1) so y ∈ α. By transitivity of α we have
α = x ∈ α contradicting (1). Similarly if y = α (use z 6= α). So, suppose
z = α. Then by transitivity of α we get x ∈ α = z.

∈ is a total order on α + 1: We need to verify trichotomy, so suppose

x, y ∈ α + 1. If x = y = α we are done. If x, y ∈ α we are done by the
assumption that α ∈ On. If (wlog) x ∈ α, y = α then we are also done.

∈ is well-founded on α + 1: Now suppose ∅ 6= x ⊆ α + 1. If x = {α}

then α is clearly the ∈-minimal element of x (as α 6∈ α by (1)). So assume
x ∩ α 6= ∅: then ∅ =
6 x ∩ α ⊆ α so x ∩ α has an ∈-minimal element m. Now
if t ∈ x then either t ∈ α so that t 6∈ m by construction of m or t = α. But
α = t ∈ x ∈ α gives α ∈ α by transitivity of α, a contradiction. Hence m
is ∈-minimal in x.

4. Suppose α, β ∈ On and x ∈ y ∈ α ∩ β. By transitivity of α and β

respectively we obtain that x ∈ α ∩ β. Now observe that the restriction
of any well-founded strict total order to a subset (or subclass) is a well-
founded strict total order.

13
 5. Suppose α, β ∈ On with α ⊆ β but α 6= β. Then β \ α is a non-empty
subset of β so has an ∈-minimal element x.
We claim x = α: Firstly if t ∈ x but t 6∈ α then by transitivity of β we
have t ∈ β \ α contradicting ∈-minimality of x. Thus x ⊆ α.
Secondly, if t ∈ α then by assumption (namely α ⊆ β) t ∈ β. So one of
x ∈ t or t ∈ x or t = x must be true. If x ∈ t then by transitivity of α we
have x ∈ α contradicting x ∈ β \ α. Similarly if t = x. Thus t ∈ x giving
α ⊆ x, as required.
6. Suppose α ∈ On and x ∈ α. If r ∈ t ∈ x then r ∈ x by transitivity of ∈
on α. Also, because α is transitive we have x ⊆ α so that ∈ restricts to a
well-founded strict total order on x. Hence x ∈ On.

Theorem 1. ∈ is a well-founded strict total order on On.

Theorem 2. If X is a non-empty subclass of On, then X has an ∈-minimal
element.
Proof. We show:

∈ is a strict total order on On: By the lemma, ∈ is irreflexive. If α, β, γ ∈

On and α ∈ β ∈ γ then by transitivity of γ we have α ∈ γ as required. For
totality, assume that α, β ∈ On. Let x = α ∩ β and note that by the lemma
x ∈ On and x ⊆ α and x ⊆ β. So by the lemma (x ∈ α or x = α) and (x ∈ β
or x = β). Unless x ∈ α and x ∈ β we have one of α ∈ β, β ∈ α or α = β. But
if we assume x ∈ α and x ∈ β. Then x ∈ α ∩ β = x contradicting the lemma.

∈ is well-founded on On: Suppose that x is a non-empty subset of On. Pick

α ∈ x and let y = α ∩ x. If y = ∅ then α is ∈-minimal in x. If on the other
hand y 6= ∅ then y is a non-empty subset of α so has an ∈-minimal element m.
By construction m ∈ x. If t ∈ x ∩ m then as m ∈ α and α is transitive we have
t ∈ α so t ∈ x ∩ α = y contradicting minimality of y. Thus x ∩ m = ∅ and m is
in fact ∈-minimal in x.
Note that this works independent of whether x is a class or a set.
Remark 4. Again, note that the first theorem is in fact a theorem of ZF− −
Powerset. The second theorem is a result in our meta-theory. It cannot even
be stated in the theory, since we don’t have classes in our theory.

Corollary 1 (Induction on On - another meta-theorem). Suppose that φ(t) is

a formula. Then

ZF− − Powerset ` [∀α ∈ On [∀β ∈ α φ(β)] → φ(α)] → ∀α ∈ On φ(α).

14
Definition 7. Suppose that α ∈ On.
α is a successor ≡ ∃β ∈ α α = β + 1
α is a limit ≡ α 6= ∅ ∧ α is not a successor
α is finite ≡ [α is a successor ∨ α = ∅] ∧ ∀t ∈ α [t = ∅ ∨ t is a successor]
x = ω ≡ x ∈ On ∧ Ind(x) ∧ ∀t ∈ x [t is finite]
Lim = {γ ∈ On : γ is a limit} .
Proof. We need to check that x = ω is in fact a definition, i.e. that if x and
y satsify the RHS then x = y: Note that x, y ∈ On gives wlog that x ∈ y (or
x = y and we are done). Then x is ‘finite’ so in particular a successor, say β + 1.
But then β ∈ x and as x is inductive, x = β + 1 ∈ x, a contradiction.
Lemma 4.
∀z [Ind(z) → ω ⊆ z]
Proof. Suppose z is an inductive set. By induction on the elements of ω we show
ω ⊆ z. Formally, suppose n ∈ ω \ z. As ∈ is well-founded on ω, we may assume
that n is ∈-minimal in ω \ z. As Ind(z) we cannot have n = ∅. Thus n = m + 1
for some m and by transitivity of ω we have m ∈ ω. By minimality of n, we
must have m ∈ z. But Ind(z) then gives n = m + 1 ∈ z, a contradiction.
Remark 5. Often ω is defined as the ‘smallest’ inductive set containing ∅, i.e.
x = ω ≡ Ind(x) ∧ ∀z [Ind(z) → x ⊆ z] and then ‘x is finiteq’ as x ∈ ω. The
advantage of this is that it doesn’t require the technology of the ordinals to
make sense. The disadvantage is that it is less clear that ω is absolute for non-
empty transitive classes satisfying enough of ZF (Foundation is crucial since
otherwise ‘being an ordinal’ might not be absoulte).
S S
Lemma 5. Suppose x ⊆ On and x is a set. Then x ∈ On and sup∈ x = x.
S S
Proof. If r ∈ t ∈ x then there is α ∈ x with r ∈ t ∈ α so r ∈ α ⊆ x since α
is transitive. S S
Next, if t ∈ x then t ∈ α for some α ∈ x so t ∈ On. Hence x ⊆ On and
S
so x is well-orderedS by ∈.
Now let α0 = x. If β ∈ x and t ∈ β then t ∈ α0 so that β ⊆ α0 . Thus α0
is an upper bound for x.
Finally, if α0 ∈ On is an upper bound for x and t ∈ α0 then find β ∈ x with
t ∈ β ⊆ α0 (since α0 is an upper bound for x). Thus α0 ⊆ α0 as required.
Lemma 6. Suppose α 6= ∅ is an ordinal.
α is a limit ordinal if and only if ∀β ∈ α β + 1 ∈ α.
Proof. Suppose α is a non-empty limit ordinal and that β ∈ α. By transitivity
of α we have β ⊆ α so β + 1 = β ∪ {β} ⊆ α. Hence we must have β + 1 ∈ α or
β + 1 = α. Since α is a limit (not a successor), we can’t have β + 1 = α (noting
that β is an ordinal) and hence β + 1 ∈ α as claimed.
Conversely, suppose α is not a limit. Because α 6= ∅, it is a successor, so
α = β + 1 for some ordinal β. Then β ∈ α and β + 1 6∈ α as required.

15
9 Recursion
A few more meta-theorems.
Theorem 3 (The informal Recursion Theorem). Suppose F is a class function
on U and that a ∈ U . Then there is a ‘unique’ class function G on On such
that:

1. G(0) = a;
2. ∀α ∈ On G(α + 1) = F (G(α));
S
3. ∀γ ∈ Lim G(γ) = {G(α) : α < γ}

We restate this more formally:

Definition 8. Suppose that F is a class function on U and that a ∈ U . We let

ψF,a (α, g) ≡α ∈ On ∧ g is a function on α + 1∧

g(0) = a∧
∀β ∈ α [g(β + 1) = F (g(β))] ∧
h [ i
∀γ ∈ Lim ∩ α + 1 g(γ) = {g(β) : β ∈ γ}

expressing that g is a function on α + 1 and that g satisfies the conditions above

on its domain. We let

GF,a ≡ {hα, yi : α ∈ On ∧ [∃g [ψF,a (α, g) ∧ hα, yi ∈ g]]} .

Theorem 4 (The Recursion Theorem). If ZF− − Powerset proves that F is

a class function on U and a ∈ U , then ZF− − Powerset proves that GF,a is a
class function on On and

GF,a (0) = a∧
∀β ∈ On GF,a (β + 1) = F (GF,a (β))∧
[
∀γ ∈ Lim G(γ) = {G(β) : β ∈ γ}

and ZF− − Powerset proves that if G, H are class functions satisfying the
displayed formula then ∀α ∈ On H(α) = G(α).
Even this is (technically speaking) not formal enough. For once, I give the
formal version of this theorem without the uniqueness bit (with defined notions
not yet eliminated):
Theorem 5 (The Formal Recursion Theorem). Suppose φ(t) is a formula with
free variable displayed. As always we write F = {x : φ(x)} for the class defined
by φ

16
 ZF− − Powerset `
[F is a class function on U ] →
∀a ∈ U
GF,a is a class function on On ∧
GF,a (0) = a∧
∀β ∈ On GF,a (β + 1) = F (GF,a (β))∧
[
∀γ ∈ Lim G(γ) = {G(β) : β ∈ γ}

As an exercise, you can try to remove all the defined notions from this
theorem.
Needless to say, for examination purposes the Informal Recursion Theorem
is sufficient (though for the proof you will likely have to define ψF,a and GF,a ).

Lemma 7. Suppose ZF− − Powerset proves that F is a class function on U

and that a ∈ U . ZF− − Powerset proves that for every α ∈ On there is a
unique g such that ψF,a (α, g).
Proof. Induction on On:
For α = 0, use g = {h0, ai} (which exists by Pairing) and uniqueness follows
from Extensionality.
For α = β + 1: Let g 0 be the unique function for β which exists by the
inductive hypothesis. Let g = g 0 ∪ {hα, g 0 (β)i} (this is a set by Union, Pairing
and inductive hypothesis). It is clear that g is a function on α + 1 which
satsifies the conditions (noting that α + 1 \ β + 1 = {α} so there are no new limit
ordinals). Uniqueness follwos once more by the requirement of the conditions
and Extensionality.
For γ ∈ Lim: For each β ∈ γ, let gβ be the unique function given by
the lemma for β and note that if β < β 0 then ĝ = gβ 0 |β+1 is a function on
β + 1 that satisfies ψF,a (β 0 , ĝ), so must equal gβ . Next note that
S {gβ : β ∈ γ} =
{g
S : ∃β ∈ γ ψ F,a (β, g)} is a set by Replacement . So g = {hγ, {gβ (β) : β ∈ γ}i}∪
{gβ : β ∈ γ} is a set by Union and Pairing. Now, by the note (that the gβ
agree on the common elements of their domains) g is a function on γ + 1 and it
is clear that g satisfies the conditions.
Proof of the Recursion Theorem. Note that we do two proofs at once (for ef-
ficiency reasons): one proof giving the theorem as stated and one giving the
theorem ‘up to some limit ordinal γ’. The latter only uses the previous lemma
for α < γ, whereas the former uses the lemma for all ordinals.
We start by demonstrating that G is indeed a class function on On (resp.
γ): it is clear from the definition of G that G is a relation from On to U . For
α ∈ On (resp. α < γ), we apply the previous lemma to see that there is g such
that ψF,a (α, g). Since α ∈ α + 1, there is y ∈ U with hα, yi ∈ g. Hence the
domain of G is all of On. Finally assume that there is α (resp. α < γ) and

17
y, y 0 ∈ U with hα, yi , hα, y 0 i ∈ G. Then we may take the witnessing g, g 0 and
note that by the previous lemma g = g 0 so that y = y 0 as required.
Finally we need to check that G satisfies the formula (up to γ). But this
follows directly from the definition of G and induction on On.

Remark 6. Let us note that if A, B are transitive non-empty classes satisfying

(enough of) ZF, On ⊆ A ⊆ B, F is absolute for A, B and a ∈ A is given by a
defined notion absolute for A, B then G is absolute for A, B.
This requires some careful checking (e.g. ψF,a is absolute for A, B) but is
essentially straightforward once we have that On is absolute for A, B.

There are some variations on recursion which we will use:

Theorem 6. Suppose A, B are classes and that F : A×B → B and H : A → B
are class functions. Then there is a unique class function G : A × On → B such
that:

ˆ G(x, 0) = H(x);

ˆ ∀α ∈ On G(x, α + 1) = F (x, G(x, α));

ˆ ∀γ ∈ Lim G(γ) = β<γ G(β).

S

Proof. As the proof of the Recursion Theorem, with ψ now being ψF,H (x, α, g)
where g(0) = a is replaced by g(0) = H(x) and

GF,H ≡ {hhx, αi , yi : α ∈ On ∧ x ∈ A ∧ [∃g [ψF,H (x, α, g) ∧ hα, yi ∈ g]]} .

Also, we may sometimes only define G up to some fixed ordinal α0 (typically

ω). The proof works as before, except that we insist on α + 1 < α0 throughout.
Theorem 7. Suppose A, B are classes and that F : A×B → B and H : A → B
are class functions and 0 < α0 ∈ On. Then there is a unique class function
G : A × α0 → B such that:
ˆ G(x, 0) = H(x);

ˆ ∀α ∈ On [α + 1 ∈ α0 → G(x, α + 1) = F (x, G(x, α))];

ˆ ∀γ ∈ Lim ∩ α0 G(γ) = β<γ G(β).

S

If α0 = ω then Replacement is not needed.

If A is a set (and Replacement holds or A is a singleton (finite set)) then
G is a set.

18
10 The Cumulative (von Neumann) Hierarchy
V
Definition 9. Let F (x) = P (x). Then ZF− proves that F is a class function
and we apply the Recursion Theorem with a = ∅ to obtain a class function V
on On such that (writing Vα for V (α))
ˆ V0 = ∅;

ˆ ∀α ∈ On Vα+1 = P (Vα );

ˆ ∀γ ∈ Lim Vγ = β<γ Vβ .
S
S
Abusing notation, we write V = α∈On Vα .
V is called the Cumulative (von Neumann) Hierarchy.
Lemma 8. ZF− proves that for all α ∈ On:

ˆ Vα is transitive;

ˆ Vα ⊆ Vα+1 ;

ˆ α ∈ Vα+1 ;

Hence V is a transitive non-empty class containing On and for α, β ∈ On we

have
α ⊆ β → Vα ⊆ Vβ .
Proof. We prove this by simultaneous induction on α:

Base Case: α = ∅: Vacuously V0 = ∅ is transitive and contained in V1 =

P (∅) = {∅}. By inspection 0 = ∅ ∈ V1 .

Successor Step: Suppose (1)-(3) hold for α ∈ On. We will show that they
hold for α + 1:

1. Let r ∈ t ∈ Vα+1 = P (Vα ). Then t ⊆ Vα (by construction) so r ∈ Vα ⊆

Vα+1 (by inductive hypothesis) as required.
2. Let t ∈ Vα+1 . Then t ⊆ Vα ⊆ Vα+1 (by inductive hypothesis) so that
t ∈ P (Vα+1 ) = Vα+1+1 .
3. By inductive hyothesis α ⊆ Vα ⊆ Vα+1 and α ∈ Vα+1 so that α +1 ⊆ Vα+1
giving α + 1 ∈ P (Vα+1 ) = Vα+1+1 .

19
Limit Step:
S Suppose γ ∈ Lim and (1)-(3) hold for α < γ. Recall that Vγ =
α<γ Vα .

1. If r ∈ t ∈ Vγ then find α < γ with r ∈ t ∈ Vα and by transitivity of Vα we

have r ∈ Vα ⊆ Vγ .
2. Suppose t ∈ Vγ . Then t ∈ Vα for some α < γ so by transitivity of Vα ,
t ⊆ Vα ⊆ Vγ , giving t ∈ P (Vγ ) = Vγ+1 .
3. For each α < γ we have α + 1 < γ (since γ is not a successor) and thus
α ∈ Vα+1 ⊆ Vγ . Hence γ ⊆ Vγ so γ ∈ Vγ+1 .
The ‘Hence’ now follows from (3) and induction.
We prove a little utitlity lemma (we only need this to show that V |=
Replacement but it makes the other axioms easier to check).
Lemma 9. ZF− proves that if z is a set and z ⊆ V then z ∈ V .
Proof. For t ∈ z let αt be the least ordinal with t ∈ Vαt (formally, we write
down a formula expressing this and check that this is a class function on z).
Then β = sup {αt + 1 : t ∈ z} is a set by Replacement (and some others which
we needed to establish that sets of ordinals have sups). By the previous lemma
z ⊆ Vβ so z ∈ Vβ+1 as required.
Theorem 8. For every φ ∈ ZF, V |= φ, i.e. ZF− ` φV .
In fact, if A is a transitive, non-empty class such that A |= Separation and
∀x ⊆ A∃z ∈ A x ⊆ z then A |= ZF− .
Proof. We prove this for V but remark that the proof also works for the A as
specified.

Extensionality: Extensionality follows from transitivity of V .

Separation: We do Separation next (since then the weaker versions of the

axioms stated in here imply the stronger versions with ↔ in place of →): so
let φ(t; v1 , . . . , vn ) be a formula of LST with all free variables shown and let
a1 , . . . , an , y ∈ V . Define

z = t : t ∈ V ∧ t ∈ y ∧ φ(t; a1 , . . . , an )V .


Then z is a set by Separation (in U ) and by the little lemma z ∈ V (in fact,
if y ∈ Vα then z ⊆ Vα so z ∈ Vα+1 avoids the lemma and hence the use of
Replacement).
We need to check that

V |= ∀t [t ∈ z ↔ t ∈ y ∧ φ(t; a1 , . . . , an )]

or more explicitly (using the definition of |=) that

ZF− ` ∀t ∈ V t ∈ z ↔ t ∈ y ∧ φ(t; a1 , . . . , an )V .
 

20
So let t ∈ V and assume t ∈ z. By definition of z (in U ) t ∈ y ∧ φ(t; a1 , . . . , an )V
as required. Now let t ∈ V and assume t ∈ y ∧ φ(t; a1 , . . . , an )V . Then by
definition of z (in U ) we have t ∈ z as required.

Replacement: We follow the strategy from above: given a formula φ we cook

up a formula ψ and apply Replacement with ψ in U to obtain some z. Then
we check z ∈ V and that V believes the right stuff about z. Since the format of
the axiom is a little more complicated, the proof is slightly messier:
So suppose that φ(x, t, d; v1 , . . . , vn ) is a formula of LST with all free variables
shown. Let a1 , . . . , an , d ∈ V and assume that V |= ∀x ∈ d∃!yφ(x, y, d). Since
V is transitive and d ∈ V this gives ZF− ` ∀x ∈ d∃!y ∈ V φ(x, y, d)V . So let
ψ(x, y, d) be
y ∈ V ∧ φ(x, y, d; a1 , . . . , an )V
and observe that we may apply Replacement in U to obtain c such that
∀x ∈ d∃y ∈ cψ(x, y, d). Let ĉ = c ∩ V ⊆ V (we are using Separation here - this
is unnecessary if we had used the stronger form of Replacement) so ĉ ∈ V .
Now we show that

V |= ∀x ∈ d∃y ∈ ĉφ(x, y, d; a1 , . . . , an ).

Assume x ∈ d and obtain y ∈ c (not ĉ!) such that ψ(x, y, d). But then y ∈ V ,
so y ∈ ĉ and φ(x, y, d, a1 , . . . , an )V .

Pairing: The formula defining x = {y, z} is ∆0 so absolute for V, U and if

y, z ∈ V then {y, z} ⊆ V so {y, z} ∈ V . (Again, strictly speaking we don’t need
the utitility lemma: we can choose α, β so that y ∈ Vα , z ∈ Vβ and wlog α ≤ β.
Then {y, z} ⊆ Vβ so {y, z} ∈ Vβ+1 .)
S
Union: Write it out in detail as an exercise. Essentially: the formula z = x
S absolute for transitive classes and by transitivity t ∈ x ∈ V → t ⊆ V so
is
S x ∈ V (and again, we could choose minimal αt ∈ On and then observe that
x ⊆ Vβ where β = sup {αt : t ∈ x}).

Powerset: Here we show that ∀x ∈ V ∃z ∈ V P (x) ∩ V ⊆ z: If x ∈ V , then

let α ∈ On (minimal or not) such that x ∈ Vα and let z = Vα + 1 ⊆ V so z ∈ V
(and again, the utility lemma is not really needed). If t ∈ P (x) ∩ V then t ∈ Vα
so t ⊆ Vα so t ∈ z.

Infinity: Exercise (go through the details which show that V |= Ind(ω) by
absoluteness and observe ω ∈ Vω+1 ).

Foundation: Let x ∈ V and assume x 6= ∅. So there is α ∈ On such that

x∩Vα 6= ∅. Consider {β ∈ α + 1 : x ∩ Vβ 6= ∅} and note that this is a non-empty
set of ordinals,
S so has an ∈-minimal element µ. Note that µ cannot be a limit
as x ∩ β<µ Vβ 6= ∅ means that there is β < µ with x ∩ Vβ 6= ∅ contradicting

21
minimality of µ. Also µ 6= ∅ as V∅ = ∅. So µ = η + 1 for some ordinal η. Pick
m ∈ x ∩ Vµ (which exists by construction of µ). Then m ⊆ Vη (by definition of
Vη+1 = Vµ ). So if t ∈ m ∩ x then t ∈ Vη ∩ x once again contradicting minimality
of µ. Thus m ∩ x = ∅ as required.
Theorem 9. If ZF is inconsistent, then so is ZF− .
Proof. Suppose there is are proofs P1 giving ZF ` ψ and P2 giving ZF ` ¬ψ
for a sentence ψ.
Note that for every axiom φ used in the proofs above, by the previous the-
orem we can write down proofs of ZF− ` φV and follow them with PV1 and
PV2 (i.e. where every line of Pi is relativized to V ) to obtain a ZF− ` ψ V and
ZF− ` ¬ψ V and thus ZF− is inconsistent.
The precise details of course depend on your formal proof system, but it will
be important that we have ZF− ` ∃x x ∈ V .

10.1 An alternative definition of V :

There is an alternative approach to defining V . For a set x, we define the
transitive closure of x as the smallest transitive set containing x and denote
it by trcl(x). That a transitive set containingS x exists follows from applying
the Recursion Theorem (on ω) with F = x and a = x and noting that
G(ω) ∪ {x} will then by transitive. We can then form the minimal one, just like
the alternative definition of ω.
Now, we can define a set x to be hereditarily well-founded if ∈ is well founded
on trcl(x) and let V = {x : x is hereditarily well-founded}. We can then show
that ∈ is well founded on every subset of V (so V |= Foundation) and also
that applying any axiom of ZF− to hereditarily well-founded sets gives new sets
which are hereditarily well-founded. This will then (with a bit of extra work)
show that V |= ZF.
We can recover the Vα by observing that in fact ∈ is well-founded on V and
recursively defining
 
 [ [ 
Vα = x ∈ V \ Vβ : x is ∈-minimal in V \ Vβ
 
β<α β<α

although it is not clear that the Vα thus defined are sets! See Kunen for details
on this approach.
It is a nice approach in that we explicitly construct the largest V which could
be well-founded (and transitive) and then check that it works. Our approach is
to ‘discover’ V and it is ‘pure luck’ that it does satisfy ZF.

11 Gödel’s Constructible Universe L

We will now work in ZF (and note the instances of Powerset we will use) to
define a ‘smaller’ universe L. Our surrounding universe is now V = {x : x = x}

22
(under ZF this is the same as V U as defined in the previous section, but I want
to emphasize that we assume Foundation).

11.1 The Definable Subsets

We want to define
 
there is a formula φ(v1 , . . . , vn , vn+1 ) and a1 , . . . , an ∈ x such that
Def (x) = y : .
∀t [t ∈ y ↔ t ∈ x ∧ (x, ∈) |= φ(a1 , . . . , an , t)]
Of course, we cannot do this formally since we cannot quantify over
formulae!
However, there is a workaround. We can internalize (x, ∈) |= φ(a1 , . . . , an )
within any class A which satisfies enough of ZF − Powerset. More precisely
there is a set F ∈ A (in fact F ⊆ ω), a class function val : A × ω × A → {0, 1}
and a function free : ω → ω <ω , free ∈ A, such that whenever φ(vk1 , . . . , vkn ) is
a formula of LST with all free variables shown then there is dφe ∈ F such that
ZF − Powerset ` {k1 , . . . , kn } = free(dφe)
ZF − Powerset `∀x ∈ A ∀a ∈ x<ω [val(x, dφe , a) = 1 ↔ [free(dφe) ⊆ dom(a) ∧ φ(a(k1 ), . . . , a(kn ))x ]] .
Using this, we can define
y ∈ Def (x) ≡ ∃m ∈ F ∃a ∈ xfree(m)\{1} [1 ∈ free(m) ∧ ∀t ∈ x [t ∈ y ↔ val(x, m, a ∪ {h1, ti}) = 1]] .
For later, we note that if A ⊆ B are non-empty transitive classes satisfying
enough of ZF − Powerset then F as well as val and hence y ∈ Def (x) are
absolute for A, B.
We can also see that y ∈ Def (x) → y ⊆ x so that if A satisfies additionally
Powerset (and a further instance of Separation) then in fact the above defines
the class function Def : A → A via
Def (x) = {y ∈ P (x) : y ∈ Def (x)} .
In fact, there is no need for Powerset: you can use Replacement instead.

11.1.1 The details of defining Def - Non-Examinable

There are a lot of different ways of defining val, F and free as well as dφe. We
present one of them, the details of which are not important.
We first note that we can define the usual arithmetic functions on ω (inter-
preted as N) by the Recursion Theorem and that these are absolute.
In the meta-theory, we then use a ‘nice’ Gödel numbering of the formu-
lae of LST (although this is only relevant at the very end - but it does help
understanding). Of course this does depend on our language, so we need to
fix it: The terms are v 0 . . .0 (or more formally we define recursively t0 = {v 0 },
tn+1 = {s0 : s ∈ tn }) and we code them by
(
2; t = v0
dte =
2 dse ; t = s0

23
So ‘terms’ are powers of 2 and we write v k instead of v 0 . . .0 (k 0 s) (for sanity
reasons) and we let T = 2k : k ∈ ω, k ≥ 1 .
Next, the atomic formulae are (for t, s terms, so we can think dte , dse ∈ T )

t=s
t∈s

coded by

dt = se =3dte 5dse 71
dt ∈ se =3dte 5dse 72

and we let A = 3t 5s 7k : t, s ∈ T, k ∈ {1, 2} .
Finally, the formulae are

φ; φ an atomic formula
¬φ; φ a formula
φ ∧ ψ; ψ, φ formulae
∀vk φ; vk a term, φ a formula

coded by

d¬φe =3dφe 73
dφ ∧ ψe =3dφe 5dψe 74
d∀vk φe =3dφe 5dvk e 75

and we let

F = A ∪ 3p 73 : p ∈ F ∪ 3p 5q 74 : p, q ∈ F ∪ 3p 5t 75 : p ∈ F, t ∈ T .
  

Of course, the definition of F doesn’t seem to make sense, so we should (by

recursion on ω) set

F0 =A
Fn+1 =F0 ∪ 3p 73 : p ∈ Fn ∪ 3p 5q 74 : p, q ∈ Fn ∪ 3p 5t 75 : p ∈ Fn , t ∈ T
  
[
F = Fn .
n∈ω

For convenience, we have chosen a minimal language. It is not difficult to

see how to deal with a more complicated language having all logical connectives
as well as existential quantifiers.
We note that T, A, F can be defined in a sufficiently large fragment of ZF −
Powerset and are absolute for transitive non-empty models of this fragment.
Now we define the function free on ω which takes values in ω <ω (finite
functions into ω, formally I think I want ω <ω = {f : F → ω : F finite ⊆ ω}) as

24
follows (by recursion on ω):
free(0) = {0}


 {0} ; n + 1 6∈ F
n + 1 ∈ F ∧ n + 1 = 3k 73

free(k);
free(n + 1) =


 free(k) ∪ free(l); n + 1 ∈ F ∧ n + 1 = 3k 5l 74
free(k) \ {l} ; n + 1 ∈ F ∧ n + 1 = 3k 5l 75 .


You should convince yourself that free gives {0} if the input is not (the code
for) a formula and otherwise the set of free variables in the formula.
(Note that I have made sure that 0 6∈ T so that 0 ∈ free(k) if and only if
k 6∈ F .)
We observe that free is absolute for non-empty transitive classes satisfying
enough of ZF − Powerset.
Finally, given x, we can define a function valx : ω × x<ω → {0, 1, 2} by
recursion on ω (here I interpret x<ω = {a : b → x : b finite ⊂ ω}). Essentially,
this is the model-theoretic satisfaction relation |= adapted to our representation
of φ by dφe. In the big case distinction, I have put the logical symbol which is
dealt with in square brackets before the n + 1.
valx (0, a) = 0


 0; n + 1 6∈ F

0; n + 1 ∈ F ∧ free(n + 1) 6⊆ dom(a)


 h i
k l

[=] n + 1 ∈ F ∧ free(n + 1) ⊆ dom(a) ∧ ∃k, l ∈ ω n + 1 = 32 52 71 ∧ a(k) 6= a(l)



 0;
 h i
k l

[=] n + 1 ∈ F ∧ free(n + 1) ⊆ dom(a) ∧ ∃k, l ∈ ω n + 1 = 32 52 71 ∧ a(k) = a(l)



 1;
 h i
k l

[∈] n + 1 ∈ F ∧ free(n + 1) ⊆ dom(a) ∧ ∃k, l ∈ ω n + 1 = 32 52 72 ∧ a(k) 6∈ a(l)



 0;
 h i
k l

[∈] n + 1 ∈ F ∧ free(n + 1) ⊆ dom(a) ∧ ∃k, l ∈ ω n + 1 = 32 52 72 ∧ a(k) ∈ a(l)



 1;

  
0; [¬] n + 1 ∈ F ∧ free(n + 1) ⊆ dom(a) ∧ ∃k, l ∈ ω n + 1 = 3k 73 ∧ valx (k, a) = 1


  
[¬] n + 1 ∈ F ∧ free(n + 1) ⊆ dom(a) ∧ ∃k, l ∈ ω n + 1 = 3k 73 ∧ valx (k, a) = 0

1;
valx (n + 1, a) = 


 0; [∧] n + 1 ∈ F ∧ free(n + 1) ⊆ dom(a) ∧ ∃k, l ∈ ω n + 1 = 3k 5l 74 ∧
[valx (k, a) = 0 ∨ valx (l, a) = 0]]




 
1; [∧] n + 1 ∈ F ∧ free(n + 1) ⊆ dom(a) ∧ ∃k, l ∈ ω n + 1 = 3k 5l 74 ∧





[valx (k, a) = 1 ∧ valx (l, a) = 1]]



 
[∀] n + 1 ∈ F ∧ free(n + 1) ⊆ dom(a) ∧ ∃k, l ∈ ω n + 1 = 3k 5l 75 ∧




 0;
 
∃â ∈ x<ω â|free(n+1)\{l} = a|free(n+1)\{l} ∧ l ∈ dom(â) → valx (k, â) = 0




 
1; [∀] n + 1 ∈ F ∧ free(n + 1) ⊆ dom(a) ∧ ∃k, l ∈ ω n + 1 = 3k 5l 75 ∧




  
∀â ∈ x<ω â|free(n+1)\{l} = a|free(n+1)\{l} ∧ l ∈ dom(â) → valx (k, â) = 1





0 otherwise - this case should not occur


Note that because x<ω is absolute (for transitive non-empty classes satisfying
enough of ZF − Powerset), valx is in fact absolute for these transitive non-
empty classes.

25
 For a formula φ(vk1 , . . . , vkn ) of LST with all free variables shown, and
a1 , . . . , an ∈ x we define

(x, ∈) |= φ(a1 , . . . , an ) ≡ valx (dφe , {hki , ai i : i = 1, . . . , n}) = 1.

11.2 A few definable sets

We work in any non-empty transitive class A satisfying enough of ZF−Powerset.
Lemma 10. Suppose x is a set.
1. ∅ ∈ Def (∅) hence Def (∅) = {∅}.
2. x ∈ Def (x).
3. For y, z ∈ x, {y, z} ∈ Def (x). More generally, finite subsets of x are
definable.
Proof. We give the relevant formulae and parameters and leave it to the reader
to verify that these define the appropriate sets:
1. φ(v1 ) ≡ v1 6= v1 and a = ∅;
2. φ(v1 ) ≡ v1 = v1 and a = ∅
3. φ(v1 , v2 , v3 ) ≡ v1 = v3 ∨ v2 = v3 and a = {h1, yi , h2, zi}; in general, we
use φ(v1 , . . . , vn , vn+1 ) ≡ v1 = vn+1 ∨ · · · ∨ vn = vn+1 to obtain sets with
(up to) n elements;
Remark 7. We will in fact spell out what really happens in the last of these
cases (when we take A = V ), just to convince you that you really don’t want
to do this (without a computer), but that it is not hard, just tedious. Note
that since val is absolute for non-empty transitive classes satisfying enough of
ZF − Powerset we get that (x, ∈) |= φ(a1 , . . . , an ) is absolute for these classes.
Also φx ≡ φ ≡ φA , so taking A = V does not lose generality.
First we work out dv1 = v3 ∨ v2 = v3 e. Note that for use, ∨ was only an
abbreviation so v1 = v3 ∨ v2 = v3 ≡ ¬ [¬ [v1 = v3 ] ∧ ¬ [v2 = v3 ]].
With our concrete definition,
1 3
n1 = dv1 = v3 e = 32 52 7

and 2 3
n2 = dv2 = v3 e = 32 52 7.
Then
n3 = d¬ [v1 = v3 ]e = 3n1 73
and
n4 = d¬ [v2 = v3 ]e = 3n2 73 .
Thus
n5 = d¬ [v1 = v3 ] ∧ ¬ [v2 = v3 ]e = 3n3 5n4 74

26
and finally
21 523 7 3 22 523 7 3
33 7 33 7 4
n = d¬ [¬ [v1 = v3 ] ∧ ¬ [v2 = v3 ]]e = 3n5 73 = 33 5 7
73 .

Clearly (well, by construction) n ∈ F and free(n) = {1, 2, 3}.

We now need to check that

t ∈ {y, z} ↔ val (x, n, a ∪ {h3, ti}) = 1.

So suppose that t = y. So, let us work out

val (x, n1 , a ∪ {h3, ti}) = 1

1 3
since n1 ∈ F , free(32 52 7) = {1, 3} ⊆ {1, 2, 3} = dom (a ∪ {h3, ti}) and k =
k l
1, l3 are such that n1 = 32 52 7 and in fact a(1) = y = t = a(3).
Thus
val (x, n3 , a ∪ {h3, ti}) = 0
and so
val (x, n5 , a ∪ {h3, ti}) = 0
giving
val (x, n, a ∪ {h3, ti}) = 1
as required (we left out the detailed checks for these last three claims).
The case t = z works similarly.
For the converse, if t 6= y and t 6= z then val (x, n1 , a ∪ {h3, ti}) = 0
as well as val (x, n2 , a ∪ {h3, ti}) = 0. Thus val (x, n3 , a ∪ {h3, ti}) = 1 as
well as val (x, n3 , a ∪ {h3, ti}) = 1 giving val (x, n5 , a ∪ {h3, ti}) = 1 so that
val x, n, a ∪ {h3, ti} = 0.

11.3 The definable sets L

Definition 10. By recursion on On we define a class function L : On → V
(strictly LV since our definition happens relativized to V )
ˆ L0 = ∅;
ˆ ∀α ∈ On Lα+1 = Def (Lα );
ˆ ∀γ ∈ Lim Lγ = β<γ Lβ
S
S
We also write L for the class α∈On Lα
Lemma 11. ZF− proves: Forall α ∈ On:
1. Lα ⊆ Vα ;
2. Lα ∈ Lα+1 ;

27
 3. Lα is transitive and Lα ⊆ Lα+1 ,
4. Lα ∩ On = α and α ∈ Lα+1
Proof. 1. By induction on α: base case and limit stages are trivial; for suc-
cessor steps note that Def (x) ⊆ P (x).
2. See Lemma 10.
3. By induction on α: again, the base case and the limit stages are straight-
forward. So suppose x ∈ Lα+1 .
If t ∈ x ∈ Lα+1 then x ⊆ Lα so t ∈ Lα ⊆ Lα+1 by inductive assumption.
Thus Lα+1 is transitive.
As Lα+1 is transitive x = {t : t ∈ Lα ∧ t ∈ x} ∈ Def (Lα+1 ) = Lα+2
giving ⊆.
4. Again, by induction on α: The non-trivial bit of the base case has been
covered in Lemma 10.
For the successor step: assume Lα ∩ On = α and α ∈ Lα+1 . Let β ∈
Lα+1 ∩ On. If α ∈ β ⊆ Lα then α ∈ Lα ∩ On = α, a contradiction. So
β = α or β ∈ α (as ∈ totally orders On) and hence β ∈ α ∪ {α} = α + 1.
Conversely, assume that β ∈ α + 1. Either β = α and then by inductive
assumption β = α ∈ Lα+1 or β ∈ α ∈ Lα+1 and by transitivity of Lα+1
we obtain β ∈ Lα+1 . Also φ(v1 , v2 ) ≡ v1 ∈ v2 ∨ v1 = v2 applied with
v2 = α shows that α + 1 ∈ Lα+2 .
For the limit: suppose γ ∈ Lim and the result is true for α < γ. If
β ∈ Lγ ∩ On then β ∈ Lα for some α < γ and hence β ∈ γ. Conversely, if
β ∈ γ then β < γ so β ∈ Lβ+1 ⊆ Lγ . But then φ(v1 ) ≡ v1 is an ordinal
shows that γ ∈ Def (Lγ ) as required.

11.4 ZF in L I
Theorem 10. For each φ ∈ ZF,
ZF ` φL
i.e. L |= ZF.
Proof of φL for the ‘easy’ axioms φ. We check the axioms in turn.

Extensionality: L is transitive, so satisfies ExtensionalityL (like V does).

Emptyset: ∅V ∈ L and z = ∅ is absolute, so ∅L = ∅V witnesses EmptysetL .

Pairing: Suppose x, y ∈ L and find α ∈ On such that x, y ∈ Lα+1 . By

V
an earlier lemma {x, y} ∈ Def (Lα+1 ) = Lα+2 ⊆ L and by absoluteness of
L V
z = {x, y}, {x, y} = {x, y} witnesses PairingL (for x, y).

28
 S V
Union: Suppose x ∈ L and find α ∈ On such that x ∈ Lα+1 . Let z = ( x) .
Note that if t ∈ y ∈ x ∈ Lα+1 then by transitivity of Lα+1 we have t ∈ Lα+1 so
that z ⊆ Lα+1 .
By definition of z,

t ∈ z ↔ [t ∈ Lα+1 ∧ ∃y ∈ x [t ∈ y]] .

The formula ∃y ∈ x [t ∈ y] is ∆0 so absolute and hence

h i
L
t ∈ z ↔ t ∈ Lα+1 ∧ [∃y ∈ x [t ∈ y]] α+1 .

But then
t ∈ z ↔ [t ∈ Lα+1 ∧ (Lα+1 , ∈) |= ∃y ∈ x [t ∈ y])] .
Thus we let φ(v1 , v2 ) ≡ ∃v3 ∈ v1 [v2 ∈ v3 ] (and a1 = x) to see that z ∈
Def Lα+1 = Lα+2 .
S L
Finally, by absoluteness of z = x, z = ( x) witnesses UnionL (for x).
S

Powerset: Suppose x ∈ L and find αx ∈ On such that x ∈ Lαx . Let z =

V
P (x) ∩ L, so that
t ∈ z ↔ t ⊆ x ∧ t ∈ L.
For each t ∈ z, let αt ∈ On be minimal such that t ∈ Lαt +1 and let α =
sup {αt + 1 : t ∈ z} ∪ {αx } ∈ On so that z ⊆ Lα and x ∈ Lα . Hence
Lα
t ∈ z ↔ t ⊆ x ∧ t ∈ Lα ↔ [t ⊆ x] ∧ t ∈ Lα

by absoluteness of t ⊆ x. Thus

t ∈ z ↔ t ∈ Lα ∧ (Lα , ∈) |= t ⊆ x

and hence φ(v1 , v2 ) ≡ v2 ⊆ v1 ≡ ∀v3 ∈ v2 [v3 ∈ v1 ] (and a1 = x) witnesses that

z ∈ Def (Lα ) = Lα+1 ⊆ L.
We now need observe that the above discussion shows
h i
L L
[∀t [t ∈ z ↔ t ⊆ x]] ≡ ∀t ∈ L t ∈ z ↔ [t ⊆ x]

L
so that indeed z = P (x) witnesses PowersetL (for x).

V
Foundation: If x ∈ L then x ∈ V so find m ∈ V such that [m is ∈-minimal in x] .
Then m ∈ x ∈ L gives m ∈ L and being ∈-minimal in x is absolute (∀y ∈
L
x [y 6∈ m] is ∆0 ), so [m is ∈-minimal in x] as required.

29
Infinity: By the Lemma ω V ∈ Lω+1 and φ(z) ≡ ∅ ∈ z ∧ ∀x ∈ z [x ∪ {x} ∈ z]
is absolute. Since φ(ω V )V we obtain φ(ω V )L as required (and we could note
that ω L = ω V ).
Remark 8. It is worth noting that the above proof shows that

ˆ Each Lα satsifies Extensionality and Foundation and if ω ∈ α then Lα

satisfies Infinity.
ˆ For every limit ordinal γ, Lγ satisfies Pairing and Union.

Proof attempt at instances of SeparationL . Suppose φ(v1 , . . . , vn , vn+1 ) is a for-

mula of LST with all free variables displayed and that a1 , . . . , an , x ∈ L. As
before we find α ∈ L such that a1 , . . . , an , x ∈ Lα . By Separation in V applied
with the formula φL , we obtain

z = t ∈ x : φ(a1 , . . . , an , t)L


and since z ⊆ x ⊆ Lα we obtain z ⊆ Lα .

So we may be tempted to show that z ∈ Def (Lα ) with the formula φ(v1 , . . . , vn , vn+1 )L ∧
t ∈ x. This encounters a problem: first the relativization L has to be rela-
tivized itself. Let’s write ψL (vn+2 ) for the formula
 defining L. If for example
φ ≡ ∀vn+2 ψ then φL ≡ ∀vn+2 ψL (vn+2 ) → ψ L , so that

L α h
Lα i
φL ≡ ∀vn+2 vn+2 ∈ Lα ∧ ψL (vn+2 )Lα ∧ ψ L .

At this stage we are in no position to prove that ψL is absolute for Lα , V

because Lα might not satisfy all of ZF− we need (even if we assume that
α is a limit ordinal). So, to avoid this, we try again, this time maybe with
φ(v1 , . . . , vn , vn+1 )∧t ∈ x since we automatically relativize to Lα+1 when check-
ing (Lα+1 , ∈) |= φ(a1 , . . . , an , t). But here we also encounter a problem. To
carry out this plan, we will need to show that

t ∈ z ↔ t ∈ x ∧ φ(a1 , . . . , an , t)Lα

and we only know that

t ∈ z ↔ t ∈ x ∧ φ(a1 , . . . , an , t)L .

Thus we need to get our hands on some ordinal γ ⊇ α such that φ(v1 , . . . , vn , vn+1 )
is absolute for Lγ , L. Provided we have this γ, we then have

t ∈ z ↔ t ∈ x ∧ φ(a1 , . . . , an , t)Lγ

so that z ∈ Def (Lγ ). This γ does exist as we will show in the next theorem,
the Levy Reflection Principle.
L
Then by construction [z = {t ∈ x : φ(a1 , . . . , an , t)}] and we are done.

30
The Levy Reflection Principle
We aim to show the following meta-theorem. First the informal, wordy version:
Theorem 11. For every formula φ of LST, if A : On → V is a non-decreasing
S
function of non-empty transitive sets with ∀γ ∈ Lim Aγ = β∈γ Aβ and
class S
B = α∈On Aα then there are arbitrarily large limit ordinals γ such that φ is
absolute for Aγ , B.
Or, more formally.
Theorem 12. Suppose A is a formula of LST with one free variable and that
φ(v1 , . . . , vn ) is some formula of LST.
We will assume that S A is a class function on On (see below) and then write
Aα = A(α) and B = α∈On Aα = {x : ∃α ∈ On x ∈ Aα }
ZF − Powerset proves:

A is a class function on On∧

A1 6= ∅∧
∀α, β ∈ On [α ⊆ β → Aα ⊆ Aβ ] ∧
∀α ∈ On Aα is transitive ∧
[
∀γ ∈ Lim Aγ = {Aβ : β ∈ γ}
→
∀α ∈ On ∃γ ∈ Lim [α ⊆ γ ∧ φ is absolute for Aγ , B] .

where of course ‘φ is absolute for Aγ , B’ means

∀a1 , . . . , an ∈ Aγ φ(a1 , . . . , an )Aγ ↔ φ(a1 , . . . , an )B

 

We first prove a lemma, the well known Tarski–Vaught criterion for abso-
luteness (from Model Theory). But first we need a definition in the meta-theory.
Definition 11. A list φ1 , . . . , φn of formulae of LST is subformula closed if and
only if every subformula of each φk appears in the list.
Note that whenever φ1 , . . . , φn is a list of formulae of LST, we can add all
the subformulae to obtain a subformula closed list φ1 , . . . , φn , . . . , φm .
Lemma 12 (Tarksi–Vaught criterion). Suppose φ1 , . . . , φn is a subformula closed
list of formulae of LST which do not contain the universal quantifier ∀ (but may
contain ∃) and that A ⊆ B are transitive, non-empty classes.
Then each of φ1 , . . . , φn is absolute for A, B provided that for every formula
φk (v1 , . . . , vm ) of the form ∃vm+1 φj (v1 , . . . , vn , vm+1 ) we have

∀a1 , . . . , am ∈ A φk (a1 , . . . , am )B → ∃am+1 ∈ A φj (a1 , . . . , an , am+1 )B .

 

Proof. By induction on the complexity of the (most complicated) formula. Atomic

formulae are trivially absolute. The logical connectives ∧, ¬, ∨, →, ↔ are trivial

31
by induction. Thus assume that the most complicated formula is φn (v1 , . . . , vm ) ≡
∃vm+1 φj (v1 , . . . , vm , vm+1 ) and that the statement is true for simpler formulae.
In particular, φj is absolute for A, B.
So let a1 , . . . , am ∈ A. Now observe

φn (a1 , . . . , am )A ≡ ∃am+1 ∈ A φj (a1 , . . . , am , am+1 )A .

Since φj is absolute

∃am+1 ∈ A φj (a1 , . . . , am , am+1 )A ↔ ∃am+1 ∈ A φj (a1 , . . . , am , am+1 )B .

   

Since A ⊆ B (for the → direction) and by the assumption (for the ←-direction)

∃am+1 ∈ A φj (a1 , . . . , am , am+1 )B ↔ φn (a1 , . . . , am )B .

 

Hence φn (a1 , . . . , am )A ↔ φn (a1 , . . . , am )B as required.

The Tarski–Vaught criterion can be understood as a ‘closure’ criterion: if
an existential formula φ is true in B then we can in fact find an element of A
witnessing φB .
To prove the Levy Reflection principle, we emulate the proof of the Skolem
Theorem, constructing the Skolem closure. We start with some Aα and ‘add
witnesses’ to all existential subformulae in B that are true in B. Of course,
adding these witnesses gives us new ‘parameters’ so that we have to consider
more instantiations of formulae.
As a concrete example, consider (in the language of fields) the statement
φ(x) ≡ ∃y y 2 = x. Then φ is not absolute for Q, C (because 2 ∈ Q and
∃y ∈ Q y 2 = 2 is false but ∃y ∈ C y 2 = 2 is true). We start for example by
adding all square √ roots of rationals and closing under
√ the field operations to get
a field F1 . Now 2 ∈ F1 , but ¬∃y ∈ F1 y 2 = 2. Thus we have to repeat:
form F2 by adding all square roots of elements of F1 and close under the field
operations. Again, there may (or may not) be (new) elements of F2 which don’t
have square
S roots in F2 , so we keep adding them to form the Fn . Finally, we
set F = n∈ω Fn . Now F is a field (easy to check that the increasing
√ union of
fields is a field) and if x ∈ F then x ∈ Fn for some n and hence x ∈ Fn+1 ⊆ F
witnesses ∃y ∈ F y 2 = x.
Of course, we may have to do this with lots of formulae at the same time
and we can’t simply add elements, but will have to go up in the Aα -hierarchy.
The principle of the proof is the same, though.
Proof of the Levy Reflection Principle. Let φ be a formula of LST. Replace all
universal quantifiers ∀x by logically equivalent ¬∃x¬ and let φ1 , . . . , φn be a
subformula closed list not mentioning ∀ and containing (the logical equivalent
of) φ. Write ik for the number of free variables of φk , k = 1, . . . , n.
k
Let α ∈ On. By recursion on m ∈ ω we define αm ∈ On, fm : Aiαkm → On.
Of course α0 = α + 1 (so that α0 ≥ 0).
k
Having defined αm , we define fm as follows. For a ∈ Aiαkm , writing aj = a(j),
if φk (v1 , . . . , vik ) ≡ ∃t ψ(v1 , . . . , vik , t) and φk (a1 , . . . , aik )B then fmk
(a) is the

32
least ordinal β such that ∃t ∈ Aβ ψ(a1 , . . . , aik , t)B and fm k
(a) = 0 otherwise.
k B
We need to argue that fm is a function: if φk (a1 , . . . , aik ) then pick a witness
t ∈ B such that ψ(a1 , . . . , aik , t)B and β̂ ∈ On such that t ∈ Aβ̂ then fm k
(a) =
n o
min β ∈ β̂ + 1 : ∃t ∈ Aβ ψ(a1 , . . . , aik )B .
We then set
 k
(a) : a ∈ Aik : k = 1, . . . , n .
 
αm+1 = sup αm + 1, sup fm

Finally, we claim that γ = sup {αm+1 : m ∈ ω} is as required.

Firstly since α < α + 1 = α0 < α1 < . . . we have that γ ∈ Lim and α ⊆ γ
and that each αm ∈ γ.
Next we check that Aγ , B satsify the Tarski–Vaught criterion. So let φl (v1 , . . . , vik )
be a formula of the form ∃t ψ(v1 , . . . , vik , t). Let a1 , . . . , ak ∈ Aγ and assume
φ(a1 , . . . , aik )B . Set aS= {hj, aj i : j = {1, . . . , ik }}. For each aj , j = 1, . . . , k
note that aj ∈ Aγ = β∈γ Aβ implies that there is β ∈ γ such that aj ∈ Aβ .
Since γ = supm αm , we can find αm such that β ⊆ αm so that aj ∈ Aαm .
Thus (by taking the maximum of such witnesses), there is m ∈ ω such that all
aj ∈ Aαm and so a ∈ Aiαkm . Since φ(a1 , . . . , aik )B by definition of fm k
(a) there
B k
is t ∈ Afm k (a) with ψ(a1 , . . . , ai , t) . Then fm (a) ≤ αm+1 ∈ γ so that in fact
k

t ∈ Aγ . Thus we have shown ∃t ∈ Aγ ψ(a1 , . . . , aik , t)B .

Hence by the Tarski–Vaught criterion, each φl is absolute for Aγ , B as
claimed.
k k
Remark
 9. Note that in the definition of fm (a), we cannot argue that fm (a) =
min β ∈ On : ∃t ∈ Aβ ψ(a1 , . . . , an , t)B since the RHS is a class and we cannot
use the well-order theorem of ∈ on On for infinitely many classes.
In constructing the Skolem closure, we only would need to add a witness
instead of going up the hierarchy. But we cannot choose (potentially) infinitely
many witnesses because we do not assume Choice (yet).
Remark 10. Note that the proof works if we substitute On by a limit ordinal
γ0 that is closed under countable suprema, i.e. such that for all α : ω → γ0 ; m 7→
αm we have sup {αm : m ∈ ω} ∈ γ0 .
Remark 11. I have been extremely explicit in the proof. You can shorten it
by quite a bit by leaving out the detailed checks, which are essentially easy.
Remark 12. Note that this proof is very much not uniform with respect to
the formulae and as such is a true meta-theory proof. You cannot ‘internalize’
it (using codes for formulae etc) and expect to get the ‘obvious’ result.

ZF in L II
We now check the two axiom schemes, Separation and Replacement in L.
We write down a clean version of the proof of Separation.

33
Proof of instances of SeparationL . Suppose φ(v1 , . . . , vn , vn+1 ) is a formula of
LST (with all free variables shown) and a1 , . . . , an , x ∈ L. Find α ∈ On such
that a1 , . . . , an , x ∈ Lα .
In V , apply separation with vn+1 ∈ vn+2 ∧ φL and v1 = a1 , . . . , vn =
an , vn+2 = x to obtain z ∈ V such that

∀t t ∈ z ↔ t ∈ x ∧ φ(a1 , . . . , an , t)L
 

(leaving out the trivial relativization to V ).

Apply the Levy Reflection Principle to φ and the hierarchy Lα to find γ ∈ On
such that α ⊆ γ and φ is absolute for Lγ , L. Thus

∀t t ∈ z ↔ t ∈ x ∧ φ(a1 , . . . , an , t)Lγ .
 

But φ(a1 , . . . , an , t)L

γ ↔ (Lγ , ∈) |= φ(a1 , . . . , an , t) so that z ∈ Def (Lγ ) =
Lγ+1 ⊆ L.
Finally since L is transitive, t ∈ x ∈ L → t ∈ L and t ∈ z ⊆ x → t ∈ L so
that we obtain

∀t ∈ L t ∈ z ↔ t ∈ x ∧ φ(a1 , . . . , an , t)L
 

which of course is equivalent to

L
[z = {t ∈ x : φ(a1 , . . . , an , t)}]

as required.
For Replacement we proceed similarly.
Proof of instances of ReplacementL . Suppose φ(v1 , . . . , vn , vn+1 , vn+2 ) is a for-
mula of LST (with all free variables shown) and a1 , . . . , an , d ∈ L. Assume
L
[∀x ∈ d ∃!y φ(a1 , . . . , an , x, y)] ≡ ∀x ∈ d ∃!y ∈ L φ(a1 , . . . , an , x, y)L

(since L is transitive x ∈ d is equivalent to x ∈ d ∧ x ∈ L).

Let ψ(v1 , . . . , vn , vn+1 , vn+2 ) ≡ vn+2 ∈ L ∧ ψ(v1 , . . . , vn , vn+1 , vn+2 )L .
We check that ψ codes a function on d in V , i.e.

∀x ∈ d ∃!y ψ(a1 , . . . , an , x, y) ≡ ∀x ∈ d ∃!y y ∈ L ∧ φ(a1 , . . . , an , x, y)L .

 

So assume x ∈ d. By assumption, we can find y ∈ L such that φ(a1 , . . . , an , x, y)L

so that y ∈ ∧φ(a1 , . . . , an , x, y)L , showing existence. Now if y, y 0 ∈ V such that
ψ(a1 , . . . , an , x, y). Then y, y 0 ∈ L and φ(a1 , . . . , an , x, y)L by definition of ψ so
L
that by assumption [y = y 0 ] ≡ y = y 0 .
Thus we can apply Replacement in V to obtain z ∈ V such that

∀y [y ∈ z ↔ ∃x ∈ d ψ(a1 , . . . , an , x, y)] ≡ ∀y y ∈ z ↔ ∃x ∈ d y ∈ L ∧ φ(a1 , . . . , an , x, y)L .

  

Note that therefore z ⊆ L. For each y ∈ z, let αy ∈ On be minimal such that

y ∈ Lαy (see remark) and let α ∈ On be such that a1 , . . . , an , d ∈ Lα . Set
α = sup {α} ∪ {αy : y ∈ z} ∈ On so that z ⊆ Lα and a1 , . . . , an , d ∈ Lα .

34
 Apply Levy’s Reflection Principle, to obtain γ ∈ Lim, α ⊆ γ such that φ is
absolute for Lγ , L. Then

y ∈ z ↔ y ∈ Lγ ∧ ∃x ∈ d φ(a1 , . . . , an , x, y)Lγ

so that z ∈ Def (Lγ ) = Lγ+1 .

By construction
L
[z = {y : ∃x ∈ d φ(a1 , . . . , an , x, y)}]

as required.

Remark 13. It is worth spelling out what exactly goes on in ‘let αy ∈ On be

minimal ...’. Formally, we consider the class

F = {hy, αi : y ∈ z ∧ α ∈ On ∧ y ∈ Lα ∧ [∀β ∈ On y ∈ Lβ → α ⊆ β]}

and the formula θ(v1 , v2 ) ≡ hv1 , v2 i ∈ F .

We show that this satisfies the condition for Replacement (in V ) with
d = z. So let y ∈ z ⊆ L and pick α̂ ∈ On such that y ∈ Lα̂ . Now note that
(by Separation in V ) and well-foundedness of ∈ on On (the ‘set’-version), α =
min {β ∈ α̂ ∪ {α̂} : y ∈ Lβ } is well-defined (since the set on the RHS contains α̂
and is thus non-empty). Since hy, αi ∈ F we obtain existence. Uniqueness on
the other hand is clear from the definition and the fact that ∈ is a total order
on On.
The same proof shows that F is a class function on z (and because z is a
set, it is in fact a function!). So by Replacement we can find z ∈ V such that

z = {F (y) : y ∈ z}

and note that F (y) = αy (as they were called in the proof).
Remark 14. We don’t need to explicitly evoke Levy’s Reflection Principle in
the proof of ReplacementL . Having obtained α such that z ⊆ Lα , we could
simply apply Separation in L to get

{y ∈ Lα : y ∈ z}

(where y ∈ z has to be replaced by an appropriate formula) defining z in V . Of

course, we still use the Levy Reflection Principle implicitly in the proof of the
appropriate instance of SeparationL .

12 Choice
We use a ‘very strong’ form of choice (of course, they are all equivalent under
ZF - but if your ‘base theory’ is slightly weaker then this form of choice typically
implies all the other commonly used ones):

35
Definition 12.
Choice ≡ ∀x ∃r r is a well-order on x
We will in fact prove an even stronger version about L, generally called
‘Global Choice’:
Definition 13. We say that a non-empty transitive class A satisfies ‘Global
Choice’ if and only if we can write down an explicit formula φ(x, y) with two
free variables such that φ is a well-order on A, i.e.

ZF `∀x ∈ A ¬φ(x, x) ∧
∀x, y, z ∈ A [φ(x, y) ∧ φ(y, z) → φ(x, z)] ∧
∀x, y ∈ A [φ(x, y) ∨ φ(y, x) ∨ x = y] ∧
∀x ∈ A [x 6= ∅ → [∃m ∈ x ∀y ∈ x ¬φ(y, m)]]

Lemma 13. ZF proves that Global Choice implies Choice.

Proof. Suppose φ(a, b) is as stated in Global Choice and fix x. By Separation,
we let
r = {ha, bi ∈ x × x : φ(a, b)}
and note that r ‘works’.
Remark 15. When applying this to L, we are of course implicitly using the
Reflection Principle (which we used to show Separation) to show that r ∈ L.
Assuming that < is a well-order on x, we proceed to write down a well-order
ˆ on x<ω = {f : n → x : n ∈ ω}:
<
Definition 14. Suppose < is a relation on x. We define

ˆ ≡f, g ∈ x<ω ∧
f <g
[dom(f ) ( dom(g)∨
[dom(f ) = dom(g) ∧ ∃m ∈ dom(f ) [f (m) < g(m) ∧ ∀k ∈ m f (k) = g(k)]]]

ˆ is a well-order on x<ω .
Lemma 14. If < is a well-order on x, then <
ˆ to
We use this to lift well-orders from x to Def (x). Essentially, we use <
well-order ω × x<ω and then use the smallest witness (n, a) ∈ ω × x<ω which
‘codes’ z ∈ Def (x) to be the ‘size’ of z.
Definition 15. Suppose < is a relation on x. We define

˜ ≡u, v ∈ Def (x)∧

u<v
∃nu ∈ F ∃au ∈ x<ω
[u = {t ∈ x : val(x, nu , au ) = 1} ∧
∀nv ∈ F ∀av ∈ x<ω [v = {t ∈ x : val(x, nv , av ) = 1} → [nu < nv ∨ [nu = nv ∧ au <a

ˆ v ]]]

˜ is a well-order on Def (x).

Lemma 15. If < is a well-order on x then <

36
Theorem 13. ZF proves that L satisfies Global Choice.
Proof. By Recursion on On we define the class function <α on On by
<∅ = ∅,
˜ α y}
<α+1 =<α ∪ {hx, yi : x ∈ Lα ∧ y ∈ Lα+1 \ Lα }∪{hx, yi : x, y ∈ Lα+1 \ Lα ∧ x<
and of course if γ ∈ Lim then
[
<γ = {<β : β ∈ γ}
It is easy to verify inductively that each <α is a well-order on Lα extending
every <β for β ∈ α.
Thus we can define
φ(x, y) ≡ x, y ∈ L ∧ ∀α ∈ On [x, y ∈ Lα → x <α y]
and again easily verify that this is a global well-order.

Corollary 2. If ZFC is inconsistent, then so is ZF.

13 V =L
To proceed we will prove a strong statement that L satisfies, namely that ‘V =
L’. Let us first formally define what this means.
Recall that on the one hand L and V were class functions on On and on
the other hand classes {x : ∃α ∈ On x ∈ Lα } and {x : ∃α ∈ On x ∈ Vα }. The
statement V = L then simply means that these two classes are equal, i.e.
∀x [∃α ∈ On x ∈ Lα ↔ ∃α ∈ On x ∈ Vα ] .
Since we assume Foundation (which was shown equivalent to ∀x x ∈ V ) this
is in fact equivalent to
∀x ∃α ∈ On x ∈ Lα .
For a non-empty class A that satisfies enough ZF − Powerset to define a class
function LA on OnA , we can relativize this to obtain
A
[V = L] ≡ ∀x ∈ A ∃α ∈ OnA x ∈ LA
α.

(Note that in fact we can relativize to any non-empty class A, but then L might
not in fact be a class function and we would have to write out carefully what
we mean by x ∈ Lα .)
That L satisfies V = L, i.e. that
L
ZF ` [V = L]
is remarkable. To appreciate this, assume for the moment that V 6= L. Then
V L (V defined in L) will in fact be L and thus different from ‘the real’ V . Thus
it could be that LL ( LV (or in fact LL ) LV ).
We thus need to delve into the definition of L to prove

37
Theorem 14. The class function Def is absolute for non-empty transitive
classes satisfying enough of ZF − Powerset.
Hence the class function L is absolute for non-empty transitive classes A
satisfying enough of ZF − Powerset in the sense that ∀α ∈ OnA LA α = Lα .

The proof is essentially straightforward: we verify that val and hence Def
are absolute and then absoluteness will ‘trickle up’ (through the Recursion The-
orem) to L.
But for absoluteness of val we need a Lemma: recall the quantifier case of
the definition of val:

val(x, d∀vk φe , a) = 1 ↔ ∀â ∈ x<ω . . .

What is written in . . . is absolute (can be expressed by a ∆0 -formula), but we

will also need that x<ω is absolute and exists as a set.
Let’s write

φ(f, x) ≡ ∃n ∈ On [n is finite ∧ f is a function from n to x]

and that
x<ω = {f : φ(f, x)} .
Lemma 16. Suppose A ⊆ B are non-empty transitive classes satisfying enough
of ZF − Powerset.
h
A
B i
ZF − Powerset ` ∀x ∈ A x<ω = x<ω .

Proof. Recall that

f ∈ x<ω ≡ φ(f, x) ≡ f is a function on ω ∧ ∃n ∈ ω dom(f ) ⊆ n.

φ is absolute for transitive classes satisfying enough of ZF − Powerset since it

can be expressed with a ∆0 -formula.
A B
Hence (x<ω ) ⊆ (x<ω ) .
So assume that the inclusion is strict. Let n ∈ ω be minimal such that

B
A
∃f ∈ x<ω \ x<ω ∧ dom(f ) ⊆ n.

Because the formula φ above is absolute, this means that f ∈ B \ A.

If n = ∅ then f = ∅ ∈ A, a contradiction.
So suppose n = m + 1 and note that by minimality of n, m ∈ dom(f ).
Again by minimality of n, f |m ∈ A. By transitivity of A and f (m) ∈ x ∈ A
and the fact that A satisfies Pairing and Union, we get m ∈ A and f =
f |m ∪ {hm, f (m)i} ∈ A, a contradiction.
Lemma 17. Suppose A is a non-empty transitive class satisfying enough of
ZF − Powerset.

ZF − Powerset ` ∀x ∈ A x<ω ∈ A.

38
Proof. All the notions used in this proof are absolute, and we will use this
silently. Also, we will not state which axioms we need exactly, but take care not
to use Powerset.
Note that x∅ = ∅ ∈ A. Also

x1 = {{h0, ti} : t ∈ x} ∈ A.

Next, if a, b ∈ A then
[
a×b= {{hr, ti : r ∈ a} : t ∈ b} ∈ A.

Assume that xn ∈ A. Then

xn+1 = {f ∪ {hn, ti} : hf, ti ∈ xn × x} ∈ A.

Thus by induction on n ∈ ω, ∀n ∈ ω xn ∈ A.
Finally [
x<ω = {xn : n ∈ ω} ∈ A.

Proof of Theorem. Suppose A is a non-empty transitive class satisfying enough

of ZF − Powerset. All ‘absolute’ below refer to ‘absolute for A, {x : x = x}’.
By absoluteness of ordinal addition and multiplication and existence and
absoluteness of x<ω , the sets T and F are absolute, as are free and val.
Hence by absoluteness of ω and again absoluteness of xω ,
  
<ω dom(a) = free(n) \ {0} ∧ 0 ∈ free(n)∧
Def (x) = z : ∃n ∈ F ∃a ∈ x
∀t ∈ x [t ∈ z ↔ val(x, n, a ∪ {h0, ti}) = 1)]

is absolute.
Finally, by the absoluteness of class functions defined by Recursion (on On)
(from absolute class functions F , here F = Def ), L is absolute.

Theorem 15.
L
ZF − Powerset ` [V = L] .
L
Proof. The statement [V = L] means

∀x ∈ L ∃α ∈ OnL x ∈ LL
α.

Since OnL ↔ On (being an ordinal is absolute for L, V ) and LL α = Lα (by

L
absoluteness of Def and induction on α), [V = L] is equivalent to

∀x ∈ L ∃α ∈ On x ∈ Lα

which is true by definition of L.

Corollary 3. If ZFC + V=L is inconsistent, then so is ZF.

39
14 Mostowski Collapse
Definition 16. Suppose that A is a set (in fact the proof also works for classes)
and R is a well-founded, set-like relation on A. Writing pred(x) = {y ∈ A : yRx}
we define by generalized recursion mos(x) = {mos(y) : y ∈ pred(x)}. We write
mos[A] for the image of A under mos.

Lemma 18. If R is well-founded and set-like on a set (or class) A, then

mos[A] = {mos(x) : x ∈ A} is transitive.
Proof. If r ∈ t ∈ mos[A] then find x ∈ A with mos(x) = t. Then r ∈ t = mos(x)
so there is y ∈ A such that yRx and mos(y) = r. Hence r ∈ mos[A].
Lemma 19. If R is well-founded, set-like and extensional (∀x, y ∈ A pred(x) =
pred(y) → x = y) on A then mos is injective on A and in fact an isomorphism
between (A, R) and (mos[A], ∈).
Proof. Suppose mos is not injective. Then the set (class)

{x ∈ A : ∃b ∈ A [b 6= x ∧ mos(b) = mos(x)]}

is non-empty so has an R-minimal element a. Fix b ∈ A with mos(b) = mos(a)

and b 6= a. If there is c ∈ A such that cRb then mos(c) ∈ mos(b) = mos(a)
so there is d ∈ A such that dRa and mos(c) = mos(d). But dRa so by R-
minimality of a, c = d giving cRa. Similarly, if there is d ∈ A such that dRa
then mos(d) ∈ mos(a) = mos(b) so there is c ∈ A such that cRb and mos(c) =
mos(d). Again, R-minimality of a gives d = c. Thus pred(b) = pred(a) and
since R is extensional b = a follows, a contradiction.
Clearly, mos is thus a bijection between A and mos[A] and by construction
∀a, b ∈ A [aRb ↔ mos(a) ∈ mos(b)].
Lemma 20. If R =∈ (which from ZF is well-founded, set-like and extensional)
and T ⊆ A is transitive, then mos|T = id|T .

Proof. If not, then set (class) {t ∈ T : mos(t) 6= t} is non-empty, so has an ∈-

minimal element m. Now mos(m) 6= m means that
ˆ either there is d ∈ mos(m) \ m. Find t ∈ A such that tRm (which means
t ∈ m as R =∈) and mos(t) = d. Then t ∈ m ∈ T so t ∈ T and thus by
R-minimality of m, we must have d = mos(t) = t ∈ m, a contradiction;
ˆ or there is t ∈ m \ mos(m). Again t ∈ m ∈ T gives t ∈ T and tRm so that
t = mos(t) ∈ mos(m), a contradiction.

40
15 Cardinals
We now assume ZFC but indicate the use of Choice and Powerset when they
are necessary for our proofs.
You should have seen most of this material in your first Set Theory course.
Specific attention should be paid to Lemma 24 and its proof which requires
neither Powerset nor Choice.
Definition 17. For x, y ∈ V we write x  y if and only if there is an injection
f : x → y.
We write x ≈ y if and only if there is a bijection f : x → y.
Lemma 21. 1.  is a reflexive, transitive relation on V (i.e. a pre-order);

2. ≈ is a symmetric, reflexive, transitive relation on V (i.e. an equivalence

relation)
Theorem 16 (Schröder-Bernstein). For every x, y ∈ V , x  y∧y  x → x ≈ y.
Lemma 22 (requires Choice). For x, y ∈ V : if f : x → y is a surjection then
y  x.
Proof. Well-order x and define g : y → x by g(t) = the least element of f −1 [t].

Theorem 17 (requires Choice). Every set is bijective with an ordinal. Hence

 is total (i.e. ∀x, y [x  y ∨ y  x].

Proof. Suppose x is a set. Well-order x by < and apply the Mostowski Collapse
(where the relation is the well-order on x - note that this is well-founded, set-like
and extensional) to obtain a t such that t is ∈-well-ordered and transitive, so
that t is an ordinal and mos : (x, <) → t is a bijection.
Definition 18 (requires Chioice).

|x| = the least α ∈ On such that α ≈ x

An ordinal α is a cardinal if and only if ∀β ∈ α β 6≈ α.

We write Card for the class of cardinals.

Lemma 23. 1. ∀x |x| is a cardinal;

2. ∀α ∈ On |α| ⊆ α;
3. ∀κ ∈ Card |κ| = κ;
4. every infinite cardinal is a limit ordinal;

Proof. 1. By minimality of |x|.

2. By the definition of |α| and the fact that id : α → α is a bijection.

41
 3. By the previous results |κ| ⊆ κ and |κ| ∈ κ being an immediate contra-
diction to the definition of cardinal.
4. If ω ⊆ α ∈ On then 
β + 1
 β∈ω
β 7→ β β ∈α\ω

0 β=α


is a bijection α + 1 → α so that α + 1 cannot be a cardinal.

Lemma 24. For all α ∈ On with α ≥ ω, |α| = |α × α|.

Proof. Clearly t 7→ ht, ti is an injection from α into α × α, so |α| ≤ |α × α|
For the other direction:
We well-order On×On by (t, s) < (t0 , s0 ) if and only if max {t, s} < max {t0 , s0 }
or max {t, s} = max {t0 , s0 } ∧ [t < t0 ∨ [t = t0 ∧ s < s0 ]].
Now, induct on α. For α = ω, write down an explicit injection (e.g. (n, m) 7→
2n 3m ).
So assume this is true below some ordinal α: If α is not a cardinal, then
α ≈ β for β < α and then clearly

α ≈ β ≈ β × β ≈ α × α.

So assume that α is a cardinal: Note that (α × α, <) is isomorphic to an

ordinal δ (by using the Mostowski Collapse). If α ∈ δ, then let hβ1 , β2 i ∈ α×α so
that α = mos (hβ1 , β2 i). Let β = max {β1 + 1, β2 + 1}. As α is a limit ordinal,
β ∈ α so that |β| = |β × β|. But then hβ1 , β2 i ∈ β × β so that α ∈ mos[β × β].
But |β × β| = |β| so that mos[β × β] ≤ |β| giving α ≤ |β|  β, a contradiction
to α being a cardinal. Hence δ ≤ α and thus |α × α| ≤ |α| = α as required.
In Part B ‘Set Theory’you may have seen a different proof which uses Choice
(in the form of Zorn’s Lemma) and Powerset to show this for all infinite sets
x. Of course with the well-ordering principle (our form of Choice) our proof
yields the same result: let f : x → α be a bijection and g : α × α → α be the
bijection from what we’ve just proven. Then f −1 ◦ g ◦ (f × f ) : x × x → x is the
required bijection.
Our proof has two advantages: it defines an explicit, absolute injection α ×
α → α (and in fact an explicit absolute bijection On × On → On by the class
version of the Mostowski collapse (absolute follows because Recursion preserves
absoluteness and < is absolute). Also we avoid the use of Powerset.
Definition 19. For cardinals α, β, define α ⊕ β = |α × {0} ∪ β × {1}| and
α ⊗ β = |α × β|.
Note that ⊕ and ⊗ do not require Choice as the lexicographic order is an
explicit well-order.
Lemma 25. 1. ⊕, ⊗ are weakly order preserving in both arguments;

42
 2. ⊕ is associative, commutative and for infinite α, β, α ⊕ β = max {α, β}.
3. ⊗ is associative and commutative and for infinite α, β, α⊗β = max {α, β}.
4. ⊗ distributes over ⊕;

5. ⊕ and ⊗ coincide with the usual (ordinal) addition and multiplication on

ω.
Definition
  20 (requires Choice and Powerset). For cardinals α, β, define
αβ = αβ .

Note that on the right hand side αβ is the set of all functions from β to α.
Lemma 26 (requires Choice and Powerset). Using + and . instead of ⊕ and
⊗ (i.e. cardinal arithmetic). For cardinals κ, λ, θ:
 θ
1. κλ.θ = κλ ;

2. κλ+θ = κλ .κθ ;
3. 2κ = |P (κ)|;
4. if 2 ≤ κ ≤ λ, ω ≤ λ then κλ = 2λ

Definition 21 (requires Choice and Powerset). Suppose I is a non-empty

set and f : I → Card; i 7→ κi a functions. We define
 
X [ 
κi =  κi × {i}
 
 
i∈I i∈I

and ( )
Y  [ 
κi =  g : I → κi : ∀i ∈ I g(i) ∈ κi  .
 
 
i∈I i∈I

Lemma 27 (requires Choice and Powerset). Under hypothesis as in the def-

inition:
P Q
1. if I = {0, 1} then i∈I κi = κ0 ⊕ κ1 and i∈I κi = κ0 ⊗ κ1 ;
P P
2. if κ is any cardinal then κ ⊗ i∈I κi = i∈I κ ⊗ κi ;
I → Card is constant with value κ (i.e. ∀i ∈ I κi = κ) and I ∈ Card
3. if f : Q
then i∈I κi = κI ;

Theorem 18 (König’s Inequality - requires Choice and Powerset). Suppose

Pg : I →QCard. Write κi = f (i) and λi = g(i).
I is a non-empty set and f,
If ∀i ∈ I κi < λi then i κi < i λi .

43
 P Q
Proof. Suppose that h : i κi → i λi is a function. We will show that h is
not a surjection.
For i ∈ I, let πi : i λi → λi ; x 7→ xi be the projection onto the ith
Q
coordinate.
For each i ∈ I, note that |πi ◦ h[κi × {i}]| ≤ κi < λi so that πi ◦ h cannot be
onto. Let ci ∈ λi \ πi ◦ h[κi × {i}]Qbe least (in P
the well-order of λi ).
Now let c = {hi, ci i : i ∈ I} ∈ i λi . If α ∈ i κi then πi ◦ h(α) 6= ci so that
h(α) 6= c. Thus h is not surjective.
Definition 22. For any ordinal α, define cf(α) to be the least ordinal β such
that there is an unbounded function f : β → α (i.e. f satisfies ∀γ ∈ α∃δ ∈
β γ ≤ f (δ)).
α is regular if and only if cf(α) = α. α is singular if it is not regular.
Note that the identity from α to itself is unbounded, so cf(α) is well-defined.
Lemma 28. For all α ∈ On:

1. cf(α) ≤ α;
2. cf(α) ∈ Card;
3. cf(α + 1) = 1 and for limit ordinals α, cf(α) ≥ ω;

4. if f : cf(α) → α is unbounded then there is a weakly increasing g : cf(α) →

α that is unbounded;
5. cf(cf(α)) = cf(α), so cf(α) is regular;
Proof. 1. α → α; β 7→ β is unbounded;

2. Suppose β < cf(α) and assume β ≈ cf(α). Let h be a witnessing bijection

and f : cf(α) → α unbounded. Then f ◦ h : β → α is unbounded,
contradicting minimality of cf(α).
3. α+1 is non-empty, so cf(α+1) 6= 0; but 0 7→ α defines an unbounded map
from 1 into α + 1. If on the other hand f : n + 1 → α is unbounded then
either f (n) is the maximal element of f [n+1] or f |n is also unbounded. In
the former case α = f (n) + 1, in the other cf(α) < n + 1. So by induction
on n, cf(α) 6= n for limit ordinals α.
4. This is clear for successor ordinals α, so assume that α is a limit and
hence that cf(α)S is a limit ordinal. Define g : cf(α) → On by g(β) =
supδ<β f (δ) = δ<β f (δ). Clearly g is weakly increasing. To see that g
maps into α, use minimality of cf(α): if g(β) ≥ α for some β ∈ cf(α) then
f |β is unbounded in in α: if η ∈ α ⊆ g(β) then there is δ < β with η ∈ f (δ)
a contradiction. Finally if η ∈ α then find β ∈ cf(α) with η ≤ f (β) and
note that as cf(α) is a limit ordinal, η ≤ f (β) ≤ g(β + 1) makes sense.

44
 5. clear if α is a successor from above. So assume α is a limit ordinal and
let h : cf(cf(α)) → cf(α) and f : cf(α) → α be unbounded. Without
loss of generality, h, f are weakly increasing. Define g : cf(cf(α)) → α by
g(β) = f (γ) where γ ∈ cf(α) is least such that h(β) ≤ γ.
To see that g is unbounded, let η ∈ α, find δ ∈ cf(α) such that η ≤ f (δ)
find β ∈ cf(cf(α)) such that f (δ) ≤ h(β) and check that g(β) ≥ f (h(β)) ≥
f (δ) ≥ η as required.

Theorem 19 (König’s Inequality). For every infinite cardinal κ, κ < cf(2κ ).

Proof. Let λ = 2κ and θ = cf(2κ ) and assume θ ≤ κ. Letting f : θ → λ

be unbounded and weakly increasing, and writing κα = f (α) and noting that
κα ∈ λ so that |κα | < λ, we have
[ X Y
λ= κα ≤ |κα | < λ = λθ = 2κ⊗θ ≤ 2κ⊗κ = 2κ = λ
α∈θ α∈θ α∈θ

a contradiction.

16 GCH
We will now show that ZF C + V L ` GCH where GCH is the formula

∀λ ∈ Card 2λ = λ+ .

We will state and prove a couple of theorems first which are essentially
refinements of what we have done before.

Theorem 20 (ZFC).

∀α ∈ On [α ≥ ω → |Lα | = |α|]

Proof. By induction on α. First by induction on n ∈ ω, we show that |Ln | = |Vn |

is finite so that |Vω | = |Lω | = ω establishing the base case.
Now, if |Lα | = |α| ≥ ℵ0 then observe that |ω × ALα × {Lα }| = ℵ0 ⊗|α|⊗1 =
|α| and that the map (n, σ, Lα ) 7→ {t ∈ Lα : val(n, σ ∪ {h0, ti} , Lα ) = T } is a
surjection from ω × ALα × {Lα } onto Lα+1 so that (using Choice) |Lα | ≤ |α| =
|α + 1|. On the other hand α + 1 ⊆ Lα+1 so that |α + 1| ≤ |L α+1 |. 
S S
The limit case is trivial: if γ ∈ Lim, |γ| =  alpha<γ α =  α<γ Lα  ≤
 

|Lγ |.
Theorem 21 (ZFC). If M is a transitive set which satisfies (enough of ) ZF −
Powerset + V = L then M = Lγ for γ = M ∩ On = min On \ M and γ is a
limit ordinal.

45
Proof. Firstly note that transitivity of M shows that min On \ M = M ∩ On.
Next, if α ∈ M then α + 1 = α ∪ {α} ∈ M by Pairing and Union and the fact
that α + 1 is absolute. Hence γ must be a limit.
Now, suppose x ∈ M . Since M |= V = L this means ∃δ ∈ M ∩OnM x ∈ LM δ .
Both On and Lδ are absolute for M, V , so we have in fact ∃δ ∈ γ x ∈ Lδ . Fixing
some such δ we see that x ∈ Lδ ⊆ Lγ .
Finally, assume x ∈ Lγ . As γ is a limit ordinal, there is δ ∈ γ with x ∈ Lδ .
But then δ ∈ M and Lδ = LM δ ⊆ M so that x ∈ M as required.

Theorem 22 (ZFC). For every regular uncountable cardinal κ, Lκ |= ZF −

Powerset + V = L (or enough thereof if you are nervous).

S seen earlier that for example x, y ∈ Lα → {x, y} ∈ Lα+1 ,

Proof. We have
x ∈ Lα → x ∈ Lα , ω ∈ Lω+1 etc so that our proofs that L satisfies each
of Union, Pairing, Infinity, Foundation, Extensionality go through. We
need to check Separation and Replacement.
For Separation: first modify the theorem and proof of Levy’s Reflection
principle as follows: replace On by κ (regular uncountable cardinal) and add
the condition ∀β ∈ κ |Aβ | < κ. Then each Fi (~a) < κ and there are < κ many
~a ∈ Anαi so that Gi (α) < κ by regularity. Since there are only finitely many i,
K(α) = max Gi (α) < κ. But then the sequence αn consists of countably (so
< κ) many ordinals < κ so by regularity γ = sup αn < κ. Now apply the proof
that L |= Separation to Lκ , noting that |Lα | = |α| < κ for α < κ as κ is a
cardinal.
For Replacement: we employ a similar strategy, essentially checking that
the set we produce still belongs to Lκ . If φ(x, y, ~v ) is the formula, ~a ∈ Lnκ
and Lκ believes that ∀x∀y, y 0 [φ(x, y, ~a) ∧ φ(x, y 0 , ~a) → y = y 0 ] and d ∈ Lκ then
firstly d ∈ Lα for some α < κ and hence |d| ≤ |α + 1| < κ. So, writing
αx for the least ordinal such that yx ∈ Lαx (where yx is the unique element
of Lλ such that φ(x, yx , ~a)) we see that αx < κ. Regularity of κ shows that
α = sup {αx : x ∈ d} < κ so that the z = yx : x ∈ d ∧ φ(x, yx , ~a)Lκ ⊆ Lα so
that z ∈ Lα+1 ⊆ Lλ .
For V = L, we essentially follow the proof that L |= V = L, i.e. use
absoluteness of the class function L (for Lκ , V - we now know that Lκ satisfies
enough of ZF − Powerset): recall that V = L ≡ ∀x existsα ∈ On x ∈ Lα , that
OnLκ = κ and that for α ∈ κ we have (Lα )Lκ = Lα . Thus (after relativizing
Lκ
S = L) ↔ ∀x ∈ Lκ ∃α ∈ κ x ∈ Lα which follows from
and using these facts) (V
the definition of Lκ = α∈κ Lα since κ is a limit ordinal (infinite cardinals are
limit ordinals).
Theorem 23 (Downward Löwenheim Skolem Theorem). If A is an infinite
subset of a set B then there is C such that A ⊆ C ⊆ B, |A| = |C| and for every
formula φ of LST, φ is absolute for C, B (i.e. C is an elementary submodel of
B, often written C  B).
Proof. See a book on model theory. For a rough idea (see remarks after the
proof), follow the proof of the Reflection Principle: Enumerate the formu-
las of LST by φi , i ∈ ω. Instead of having that Fi choose ordinals α, let

46
them choose (using Choice) witnessing elements b (i.e. well order B; if B |=
φi (b1 , . . . , bni ) = ∃xψ(x, b1 , . . . , bni ) we let Fi (b1 , . . . , bni ) be the least element
b of B such that ψ(b, b1 , . . . , bn )). Now, starting with A0 = A, let An+1 =
An ∪ {Fi (a S1 , . . . , ani ) : i ∈ ω ∧ a1 , . . . , ani ∈ An }. Note that |An+1 | = |An | and
that C = n∈ω An is as required.
Of course we cannot enumerate the ‘real’formulas inside our theory (and
specifically not with the ω of our theory), so we would need to internalize them,
i.e. we enumerate the Gödel numbers (in our theory) of formulas and then
things like B |= φ(b1 , . . . , bn ) is replaced by val(dφe , σ, B) for a suitable σ. We
need to convince ourselves once more that this works as intended.
Alternatively, we note that for our purposes later, we only ever consider a
specific (though not explicitly described) finite set of formulas, so we could just
use these.
Recall that Hκ = {x : |T C(x)| < κ} where T C(x) is the smalles transitive
set t with x ⊆ t.

Theorem 24. ZFC + V = L proves for infinite cardinals κ, Hκ = Lκ and

hence GCH.
Proof. The critical case is to show Lκ+ = Hκ+ (since we have already seen that
Lω = Hω and ‘limit cardinals’ work by unions):
First Lκ ⊆ Hκ : If x ∈ Lκ then x ∈ Lα for α < κ so that x ⊆ Lα (transitivity)
and thus |T C(x)| ≤ |Lα | = |α| < κ.
For Hκ+ ⊆ Lκ+ : Fix x ∈ Hκ+ and let t = T C({x}). Note that t = {x} ∪
T C(x) so that |t| ≤ κ. Using V = L, find a regular, uncountable λ such that
+
t ∈ Lλ (find an infinite α ∈ On such that t ∈ Lα and then note that λ = |α| is
regular and uncountable). Noting that Lλ satsifies (enough of) ZF−Powerset+
V = L apply the Downward Löwenheim Skolem Theorem to obtain s such that
t ⊆ s ⊆ Lλ , |s| = |t| and s satsifies (enough of) ZF − Powerset + V = L.
Then consider the Mostowski Collapse mos[s] of s. On the one hand t ⊆ s
is transitive so mos|t = id|t and in particular x = mos(x) ∈ mos[s]. On
the other hand mos[s] satisfies the same ∈-sentences as s (since mos is an ∈-
isomorphism) so that mos[s] satisfies (enough of) ZF − Powerset + V = L and
hence mos[s] = Lγ for γ = s ∩ On and x ∈ Lγ . But then

|γ| = |Lγ | = |mos[s]| ≤ |s| = |t| ≤ κ

so that γ < κ+ and x ∈ Lκ+ as required.

Finally, to see that this implies GCH, consider any infinite cardinal λ. If
A ∈ P (λ) (the powerset being taken in L of course) then T C(A) ⊆ A ∪ λ (as
the RHS is transitive) so that |T C(A)| ≤ λ. Hence A ∈ Hλ+ . Thus P (λ) ⊆
Hλ+ = Lλ+ and thus |P (λ)| ≤ |Lλ+ | ≤ λ+ . On the other hand by Cantor’s
Theorem |P (λ)| ≥ λ+ so that |P (λ)| = λ+ as required.

47