This page is intended to hold proofs I'm working on of theorems related to Surreal numbers . I was curious about the topic and noticed a lack of online versions of proofs of theorems and statements in the article itself. For now the work is just a personal indulgence, but if it turns out to be interesting I may see about transferring it to Wikibooks as a supplemental article for people who read the Surreal numbers article and want some additional detail.
To review the basic definitions, a surreal number X is an object that consists of a (possibly empty) pair of collections of other surreal numbers called "X left" and "X right", denoted as XL and XR respectively. The normal notation is to say that X = {XL | XR }, and as we proceed it will be shown that the two sides are essentially sets of lower and upper bounds for X.
We also define the relation
≤
{\displaystyle \leq }
between two surreal numbers as:
Comparison Rule
For a surreal number x = { XL | XR } and y = { YL | YR } it holds that x ≤ y if and only if y is less than or equal to no member of XL , and no member of YR is less than or equal to x .
Additionally, if
x
≤
y
{\displaystyle x\leq y}
but
y
≰
x
{\displaystyle y\not \leq x}
then we say that
x
<
y
{\displaystyle x<y\,}
.
We restrict the construction of surreal numbers using the
≤
{\displaystyle \leq }
relation as follows:
Construction Rule
If L and R are two sets of surreal numbers and no member of R is less than or equal to any member of L then { L | R } is a surreal number.
Finally, we define a base surreal number, 0 = { | }, with empty left and right boundaries. With 0 defined, we now have a basis from which to inductively construct and prove theorems about surreal numbers.
Properties of
≤
{\displaystyle \leq }
edit
Our first priority is to show that
≤
{\displaystyle \leq }
will be able to help form an ordering on the surreal numbers analogous to the like relation on the real numbers. To begin, we will need to prove that
≤
{\displaystyle \leq }
is a total preorder on the surreal numbers, meaning that it is both total and transitive.
Since many of the proofs here will be by induction, we'll start with a simple basic lemma:
Lemma: 0
≤
{\displaystyle \leq }
0
edit
Proof: Since 0L and 0R are both empty, it follows that
∄
x
∈
0
L
s
.
t
.
0
≤
x
{\displaystyle \not \exists x\in 0_{L}\ s.t.\ 0\leq x}
and
∄
x
∈
0
R
s
.
t
.
x
≤
0
{\displaystyle \not \exists x\in 0_{R}\ s.t.\ x\leq 0}
. Therefore by definition
0
≤
0
◻
{\displaystyle 0\leq 0\;_{\Box }}
With the above proven, we can now proceed to start showing certain general properties of
≤
{\displaystyle \leq }
by induction.
Theorem: For all surreal numbers X,
X
≤
X
{\displaystyle X\leq X}
Proof by induction:
(Base) We proved above that
0
≤
0
{\displaystyle 0\leq 0}
(Induction) Let surreal number X be given and assume that for all
a
∈
X
L
∪
X
R
,
a
≤
a
{\displaystyle a\in X_{L}\cup X_{R}\;,a\leq a}
. Then we wish to show that
X
≤
X
{\displaystyle X\leq X}
.
Assume for the moment that
∃
a
∈
X
L
s
.
t
.
X
≤
a
{\displaystyle \exists a\in X_{L}\;s.t.\;X\leq a}
. Then by definition
∄
b
∈
X
L
s
.
t
.
a
≤
b
{\displaystyle \not \exists b\in X_{L}\;s.t.\;a\leq b}
. But by the inductive assumption
a
∈
X
L
{\displaystyle a\in X_{L}}
and
a
≤
a
{\displaystyle a\leq a}
. Therefore by contradiction
∄
a
∈
X
L
s
.
t
.
X
≤
a
{\displaystyle \not \exists a\in X_{L}\;s.t.\;X\leq a}
.
Now assume that
∃
a
∈
X
R
s
.
t
.
a
≤
X
{\displaystyle \exists a\in X_{R}\;s.t.\;a\leq X}
. Then by definition
∄
b
∈
X
R
s
.
t
.
b
≤
a
{\displaystyle \not \exists b\in X_{R}\;s.t.\;b\leq a}
. But by the inductive assumption
a
∈
X
R
{\displaystyle a\in X_{R}}
and
a
≤
a
{\displaystyle a\leq a}
. Therefore by contradiction
∄
a
∈
X
R
s
.
t
.
a
≤
X
{\displaystyle \not \exists a\in X_{R}\;s.t.\;a\leq X}
.
Thus combining 1) and 2) above we derive that
∄
a
∈
X
L
s
.
t
.
X
≤
a
{\displaystyle \not \exists a\in X_{L}\;s.t.\;X\leq a}
and
∄
a
∈
X
R
s
.
t
.
a
≤
X
{\displaystyle \not \exists a\in X_{R}\;s.t.\;a\leq X}
. So by definition
X
≤
X
{\displaystyle X\leq X}
.
◻
{\displaystyle _{\Box }}
Now that we've proven reflexivity, we'll proceed with some intermediary results.
Left side < X and X < right side
edit
Theorem For all
a
∈
X
L
a
≤
X
{\displaystyle a\in X_{L}\;a\leq X}
Proof by induction:
(Base case)
0
L
{\displaystyle 0_{L}\;}
and
0
R
{\displaystyle 0_{R}\;}
are empty so the theorem is vacuously true.
(Induction) Assume that for all
a
∈
X
L
{\displaystyle a\in X_{L}}
that
∀
b
∈
a
L
b
≤
a
{\displaystyle \forall b\in a_{L}\;b\leq a}
.
Let
a
∈
X
L
{\displaystyle a\in X_{L}}
. We want to show that
a
≤
X
{\displaystyle a\leq X}
.
Assume
∃
b
∈
X
R
b
≤
a
{\displaystyle \exists b\in X_{R}\;b\leq a}
. But since
b
∈
X
R
{\displaystyle b\in X_{R}}
and
a
∈
X
L
{\displaystyle a\in X_{L}}
by definition of right/left
b
≰
a
{\displaystyle b\not \leq a}
. So by contradiction
∄
b
∈
X
R
b
≤
a
{\displaystyle \not \exists b\in X_{R}\;b\leq a}
.
Now assume
∃
b
∈
a
L
X
≤
b
{\displaystyle \exists b\in a_{L}\;X\leq b}
. Then by the inductive assumption
b
≤
a
{\displaystyle b\leq a}
. But since
X
≤
b
{\displaystyle X\leq b}
it follows
∄
y
∈
X
L
b
≤
y
{\displaystyle \not \exists y\in X_{L}\;b\leq y}
, and so since
a
∈
X
L
{\displaystyle a\in X_{L}}
, we find that
b
≰
a
{\displaystyle b\not \leq a}
. Thus by contradiction
∄
b
∈
a
L
X
≤
b
{\displaystyle \not \exists b\in a_{L}\;X\leq b}
.
Therefore since
∄
b
∈
X
R
b
≤
a
{\displaystyle \not \exists b\in X_{R}\;b\leq a}
and
∄
b
∈
a
L
X
≤
b
{\displaystyle \not \exists b\in a_{L}\;X\leq b}
by definition
a
≤
X
{\displaystyle a\leq X}
.
◻
{\displaystyle _{\Box }}
Theorem For all
a
∈
X
R
X
≤
a
{\displaystyle a\in X_{R}\;X\leq a}
Proof by induction:
(Base case)
0
L
{\displaystyle 0_{L}\;}
and
0
R
{\displaystyle 0_{R}\;}
are empty so the theorem is vacuously true.
(Induction) Assume that for all
a
∈
X
R
{\displaystyle a\in X_{R}}
that
∀
b
∈
a
R
a
≤
b
{\displaystyle \forall b\in a_{R}\;a\leq b}
Let
a
∈
X
R
{\displaystyle a\in X_{R}}
. We want to show that
X
≤
a
{\displaystyle X\leq a}
.
Assume
∃
b
∈
X
L
a
≤
b
{\displaystyle \exists b\in X_{L}\;a\leq b}
. But since
a
∈
X
R
{\displaystyle a\in X_{R}}
and
b
∈
X
L
{\displaystyle b\in X_{L}}
by definition of right/left
a
≰
b
{\displaystyle a\not \leq b}
. So by contradiction
∄
b
∈
X
L
a
≤
b
{\displaystyle \not \exists b\in X_{L}\;a\leq b}
.
Now assume
∃
b
∈
a
R
b
≤
X
{\displaystyle \exists b\in a_{R}\;b\leq X}
. Then by the inductive assumption
a
≤
b
{\displaystyle a\leq b}
. But since
b
≤
X
{\displaystyle b\leq X}
it follows
∄
y
∈
X
R
y
≤
b
{\displaystyle \not \exists y\in X_{R}\;y\leq b}
, and so since
a
∈
X
R
{\displaystyle a\in X_{R}}
, we find that
a
≰
b
{\displaystyle a\not \leq b}
. Thus by contradiction
∄
b
∈
a
R
b
≤
X
{\displaystyle \not \exists b\in a_{R}\;b\leq X}
.
Therefore since
∄
b
∈
X
L
a
≤
b
{\displaystyle \not \exists b\in X_{L}\;a\leq b}
and
∄
b
∈
a
R
b
≤
X
{\displaystyle \not \exists b\in a_{R}\;b\leq X}
by definition
X
≤
a
{\displaystyle X\leq a}
.
◻
{\displaystyle _{\Box }}
Thus the above theorems show that, as expected, members of
X
L
{\displaystyle X_{L}\;}
are lower bounds of
X
{\displaystyle X\;}
, while members of
X
R
{\displaystyle X_{R}\;}
are upper bounds of
X
{\displaystyle X\;}
.
Lemma For all surreal numbers
X
,
X
∉
X
L
{\displaystyle X,X\not \in X_{L}}
and
X
∉
X
R
{\displaystyle X\not \in X_{R}}
Proof: Assume
X
∈
X
L
{\displaystyle X\in X_{L}}
. Then since
X
≤
X
{\displaystyle X\leq X}
it follows that
∃
a
∈
X
L
X
≤
a
{\displaystyle \exists a\in X_{L}\;X\leq a}
. But since
X
≤
X
{\displaystyle X\leq X}
, it follows by definition that
∄
a
∈
X
L
X
≤
a
{\displaystyle \not \exists a\in X_{L}\;X\leq a}
. So by contradiction
X
∉
X
L
{\displaystyle X\not \in X_{L}}
Now assume
X
∈
X
R
{\displaystyle X\in X_{R}}
. Then since
X
≤
X
{\displaystyle X\leq X}
it follows that
∃
a
∈
X
R
a
≤
X
{\displaystyle \exists a\in X_{R}\;a\leq X}
. But since
X
≤
X
{\displaystyle X\leq X}
, it follows by definition that
∄
a
∈
X
R
a
≤
X
{\displaystyle \not \exists a\in X_{R}\;a\leq X}
. So by contradiction
X
∉
X
R
{\displaystyle X\not \in X_{R}}
.
◻
{\displaystyle _{\Box }}
In fact, we can show that everything on the left is
<
X
{\displaystyle <X}
and everything on the right is
>
X
{\displaystyle >X}
.
Theorem For all
a
∈
X
L
a
<
X
{\displaystyle a\in X_{L}\;a<X}
Proof: Let
a
∈
X
L
{\displaystyle a\in X_{L}}
and assume
X
≤
a
{\displaystyle X\leq a}
. Then
∄
y
∈
X
L
a
≤
y
{\displaystyle \not \exists y\in X_{L}\;a\leq y}
. But
a
∈
X
L
{\displaystyle a\in X_{L}}
and
a
≤
a
{\displaystyle a\leq a}
so by contradiction
X
≰
a
{\displaystyle X\not \leq a}
. Therefore since
a
≤
X
{\displaystyle a\leq X}
and
X
≰
a
{\displaystyle X\not \leq a}
by definition
a
<
X
{\displaystyle a<X\;}
.
◻
{\displaystyle _{\Box }}
Theorem For all
a
∈
X
R
X
<
a
{\displaystyle a\in X_{R}\;X<a}
Proof: Let
a
∈
X
R
{\displaystyle a\in X_{R}}
and assume
a
≤
X
{\displaystyle a\leq X}
. Then
∄
y
∈
X
R
X
≤
a
{\displaystyle \not \exists y\in X_{R}\;X\leq a}
. But
a
∈
X
R
{\displaystyle a\in X_{R}}
and
a
≤
a
{\displaystyle a\leq a}
so by contradiction
a
≰
X
{\displaystyle a\not \leq X}
. Therefore since
X
≤
a
{\displaystyle X\leq a}
and
a
≰
X
{\displaystyle a\not \leq X}
by definition
X
<
a
{\displaystyle X<a\;}
.
◻
{\displaystyle _{\Box }}
Thus everything included in the left side can be considered a strict lowerbound less than X, and everything in the right side is a strict upperbound on X. Now to show totality.
Theorem For all surreals
X
{\displaystyle X\;}
and
Y
{\displaystyle Y\;}
either
X
≤
Y
{\displaystyle X\leq Y}
or
Y
≤
X
{\displaystyle Y\leq X}
Proof by induction:
(Base)
0
≤
0
{\displaystyle 0\leq 0}
(Induction) Assume that for all
a
,
b
∈
X
L
∪
X
R
{\displaystyle a,b\in X_{L}\cup X_{R}}
that either
a
≤
b
{\displaystyle a\leq b}
or
b
≤
a
{\displaystyle b\leq a}
. Then let
a
∈
X
L
∪
X
R
∪
{
X
}
{\displaystyle a\in X_{L}\cup X_{R}\cup \{X\}}
. We want to show that either
a
≤
X
{\displaystyle a\leq X}
or
X
≤
a
{\displaystyle X\leq a}
.
1) Assume
a
=
X
{\displaystyle a=X\;}
. Then since
≤
{\displaystyle \leq }
is reflexive
a
≤
a
{\displaystyle a\leq a}
.
2) Assume
a
∈
X
L
{\displaystyle a\in X_{L}}
. Then
a
≤
X
{\displaystyle a\leq X}
3) Assume
a
∈
X
R
{\displaystyle a\in X_{R}}
. Then
X
≤
a
{\displaystyle X\leq a}
.
◻
{\displaystyle _{\Box }}
An immediate corollary is that for all surreal numbers x and y if x is not equivalent to y, then either x < y or y < x.
With totality in hand, we'll now show that
≤
{\displaystyle \leq }
is transitive.
Theorem: If
X
≤
Y
{\displaystyle X\leq Y}
and
Y
≤
Z
{\displaystyle Y\leq Z}
then
X
≤
Z
{\displaystyle X\leq Z}
Proof by induction:
(Base)
0
≤
0
{\displaystyle 0\leq 0}
(Induction)
Assume that
∀
a
,
b
,
c
∈
X
L
∪
X
R
{\displaystyle \forall a,b,c\in X_{L}\cup X_{R}}
that if
a
≤
b
{\displaystyle a\leq b}
and
b
≤
c
{\displaystyle b\leq c}
then
a
≤
c
{\displaystyle a\leq c}
. Then we want to show the same transitive property holds for
a
,
b
∈
X
L
∪
X
R
{\displaystyle a,b\in X_{L}\cup X_{R}}
and
X
{\displaystyle X\;}
. There are three possible cases to consider to show transitivity.
1) Assume
a
≤
b
{\displaystyle a\leq b}
and
b
≤
X
{\displaystyle b\leq X}
. Since
b
≤
X
{\displaystyle b\leq X}
and
b
∈
X
L
∪
X
R
{\displaystyle b\in X_{L}\cup X_{R}}
it follows that
b
∈
X
L
{\displaystyle b\in X_{L}}
because
∀
y
∈
X
R
y
≰
X
{\displaystyle \forall y\in X_{R}y\not \leq X}
. Because
a
≤
b
{\displaystyle a\leq b}
and
∄
y
∈
X
R
y
≤
b
{\displaystyle \not \exists y\in X_{R}\;y\leq b}
,
a
∈
X
L
{\displaystyle a\in X_{L}}
. And therefore
a
≤
X
{\displaystyle a\leq X}
.
2) Assume
X
≤
a
{\displaystyle X\leq a}
and
a
≤
b
{\displaystyle a\leq b}
. Since
X
≤
a
{\displaystyle X\leq a}
it follows that
a
∈
X
R
{\displaystyle a\in X_{R}}
because
∀
y
∈
X
L
X
≰
y
{\displaystyle \forall y\in X_{L}X\not \leq y}
. Because
a
≤
b
{\displaystyle a\leq b}
and
∄
y
∈
X
L
a
≤
y
{\displaystyle \not \exists y\in X_{L}\;a\leq y}
,
b
∈
X
R
{\displaystyle b\in X_{R}}
. And therefore
X
≤
b
{\displaystyle X\leq b}
.
3) Assume
a
≤
X
{\displaystyle a\leq X}
and
X
≤
b
{\displaystyle X\leq b}
. Since
a
≤
X
{\displaystyle a\leq X}
it follows that
a
∈
X
L
{\displaystyle a\in X_{L}}
. And since
X
≤
b
{\displaystyle X\leq b}
it follows that
b
∈
X
R
{\displaystyle b\in X_{R}}
. Because
≤
{\displaystyle \leq }
is total, either
a
≤
b
{\displaystyle a\leq b}
or
b
≤
a
{\displaystyle b\leq a}
. But by the definition of construction since
b
∈
X
R
{\displaystyle b\in X_{R}}
and
a
∈
X
L
{\displaystyle a\in X_{L}}
it follows that
b
≰
a
{\displaystyle b\not \leq a}
. Thus
a
≤
b
{\displaystyle a\leq b}
.
◻
{\displaystyle _{\Box }}
We have now established that
≤
{\displaystyle \leq }
is a total preorder on the surreal numbers. We'll see momentarily that it is, in fact, not antisymmetric, but first let's construct a few surreal numbers to help see how they operate.
Constructing the first few surreal numbers
edit
The initial base surreal number will be
0
=
{
|
}
{\displaystyle 0=\{\;|\;\}}
, which we've dealt with above.
Now we begin constructing the surreal numbers inductively. At each step of the induction, we'll define
S
0
=
{
0
}
{\displaystyle S_{0}=\{0\}\;}
and
S
n
+
1
{\displaystyle S_{n+1}\;}
to be the set of surreals directly constructed from subsets of
S
n
{\displaystyle S_{n}\;}
.
So
S
1
=
{
{
0
|
}
,
{
|
0
}
}
{\displaystyle S_{1}=\{\{0|\},\{|0\}\}\;}
. Note that
{
0
|
0
}
{\displaystyle \{0|0\}\;}
is not a valid surreal number because
0
≤
0
{\displaystyle 0\leq 0\;}
.
Since
0
∈
{
|
0
}
R
{\displaystyle 0\in \{|0\}_{R}}
it follows that
{
|
0
}
<
0
{\displaystyle \{|0\}<0\;}
. Likewise since
0
∈
{
0
|
}
L
{\displaystyle 0\in \{0|\}_{L}}
it follows that
0
<
{
0
|
}
{\displaystyle 0<\{0|\}\;}
. Thus these three numbers are an ordered sequence which we will label as
−
1
=
{
|
0
}
{\displaystyle -1=\{|0\}\;}
and
1
=
{
0
|
}
{\displaystyle 1=\{0|\}\;}
. With that labelling we can write this simply as
−
1
<
0
<
1
{\displaystyle -1<0<1\;}
.
So far so good. Let's move on to
S
2
{\displaystyle S_{2}\;}
. One of the surreal numbers generated is
{
|
−
1
}
<
−
1
{\displaystyle \{|-1\}<-1\;}
. Similarly, we can generate
{
|
−
1
,
0
}
<
−
1
{\displaystyle \{|-1,0\}<-1\;}
. Notice though that
{
|
−
1
}
≤
{
|
−
1
,
0
}
{\displaystyle \{|-1\}\leq \{|-1,0\}}
and
{
|
−
1
,
0
}
≤
{
|
−
1
}
{\displaystyle \{|-1,0\}\leq \{|-1\}}
. Thus we can see that
≤
{\displaystyle \leq }
is not a total order but is rather a total preorder . To adjust things so we can work with a total order, we'll use the equivalence relation implied by the total preorder:
a
==
b
{\displaystyle a==b\;}
if and only if
a
≤
b
{\displaystyle a\leq b}
and
b
≤
a
{\displaystyle b\leq a}
and work with the equivalence classes implied by that relation. We write
[
X
]
{\displaystyle [X]\;}
to mean the equivalence class of all surreal numbers equivalent to X. For example, within
S
1
{\displaystyle S_{1}\;}
we have that
[
0
]
{\displaystyle [0]\;}
includes
0
,
{
−
1
|
}
,
{
−
1
|
1
}
,
{
|
1
}
{\displaystyle 0,\{-1|\},\{-1|1\},\{|1\}\;}
.
Completing
S
1
{\displaystyle S_{1}\;}
we have seven distinct equivalence classes, labelled -2, -1, -1/2, 0, 1/2, 1 and 2 respectively:
[-2] contains {|-1} (which we'll call -2) == {|-1,0} == {|-1,1} == {-1,0,1} <
[-1] contains -1 == {|0,1} <
[-1/2] contains {-1|0} (which we'll call -1/2) == {-1|0,1} <
[0] contains 0 == {-1|} == {-1|1} == {|1} <
[1/2] contains {0|1} (which we'll call 1/2) == {-1,0|1} <
[1] contains 1 == {-1,0|} <
[2] contains {1|} (which we'll call 2) == {0,1|} == {-1,1|} == {-1,0,1|}
At this point it would be useful to have a way to simplify how to tell when two surreal numbers are equivalent. The following theorems should prove handy in that regard.
Eliminating excess lower and upper bounds
edit
Theorem For a surreal number
X
{\displaystyle X\;}
, if
b
∈
X
L
{\displaystyle b\in X_{L}}
and
a
≤
b
{\displaystyle a\leq b}
then
X
==
{
X
L
∪
{
a
}
|
X
R
}
{\displaystyle X==\{X_{L}\cup \{a\}|X_{R}\}}
Proof
Let a, b and X be given such that
b
∈
X
L
{\displaystyle b\in X_{L}}
and
a
≤
b
{\displaystyle a\leq b}
, and let
Y
=
{
X
L
∪
{
a
}
|
X
R
}
{\displaystyle Y=\{X_{L}\cup \{a\}|X_{R}\}}
. We want to show
X
≤
Y
{\displaystyle X\leq Y}
and
Y
≤
X
{\displaystyle Y\leq X}
.
1a)Let
c
∈
Y
R
{\displaystyle c\in Y_{R}}
be given. Since
Y
R
=
X
R
{\displaystyle Y_{R}=X_{R}\;}
it follows that
c
∈
X
R
{\displaystyle c\in X_{R}}
and thus
c
≰
X
{\displaystyle c\not \leq X}
.
1b)Let
c
∈
X
L
{\displaystyle c\in X_{L}}
be given. Then
c
∈
X
L
∪
{
a
}
=
Y
L
{\displaystyle c\in X_{L}\cup \{a\}=Y_{L}}
, so
Y
≰
c
{\displaystyle Y\not \leq c}
. Therefore by 1a and 1b,
X
≤
Y
{\displaystyle X\leq Y}
.
2a)Let
c
∈
X
R
{\displaystyle c\in X_{R}}
be given. Since
X
R
=
Y
R
{\displaystyle X_{R}=Y_{R}\;}
,
c
∈
Y
R
{\displaystyle c\in Y_{R}}
and therefore
c
≰
Y
{\displaystyle c\not \leq Y}
.
2b)Let
c
∈
Y
L
{\displaystyle c\in Y_{L}}
be given. Then either
c
∈
X
L
{\displaystyle c\in X_{L}\;}
or
c
=
a
{\displaystyle c=a\;}
. If
c
∈
X
L
{\displaystyle c\in X_{L}}
then
X
≰
c
{\displaystyle X\not \leq c}
. If
c
=
a
{\displaystyle c=a\;}
then by our initial assumption
c
≤
b
{\displaystyle c\leq b}
and
b
∈
X
L
{\displaystyle b\in X_{L}}
. Assume
X
≤
c
{\displaystyle X\leq c}
. Then since
c
≤
b
{\displaystyle c\leq b}
by transitivity
X
≤
b
{\displaystyle X\leq b}
. But because
b
∈
X
L
{\displaystyle b\in X_{L}}
it follows that
X
≰
b
{\displaystyle X\not \leq b}
. So therefore by contradiction
X
≰
c
{\displaystyle X\not \leq c}
. And so by 2a and 2b
Y
≤
X
{\displaystyle Y\leq X}
.
◻
{\displaystyle _{\Box }}
The upshot of the above theorem is that, when it comes to equivalence, you can ignore lower bounds less than the maximal lower bound. So for example
{
0
|
1
}
==
{
−
1
,
0
|
1
}
==
{
−
2
,
0
|
1
}
==
{
−
2
,
1
,
0
|
1
}
{\displaystyle \{0|1\}==\{-1,0|1\}==\{-2,0|1\}==\{-2,1,0|1\}\;}
. Since the right hand sides are equal only the maximal lower bound 0 matters for purposes of equivalence.
A similar theorem exists for upper bounds.
Theorem For a surreal number
X
{\displaystyle X\;}
, if
b
∈
X
R
{\displaystyle b\in X_{R}}
and
b
≤
a
{\displaystyle b\leq a}
then
X
==
{
X
L
|
X
R
∪
{
a
}
}
{\displaystyle X==\{X_{L}|X_{R}\cup \{a\}\}}
Proof
Let a, b and X be given such that
b
∈
X
R
{\displaystyle b\in X_{R}}
and
b
≤
a
{\displaystyle b\leq a}
, and let
Y
=
{
X
L
|
X
R
∪
{
a
}
}
{\displaystyle Y=\{X_{L}|X_{R}\cup \{a\}\}}
. We want to show
X
≤
Y
{\displaystyle X\leq Y}
and
Y
≤
X
{\displaystyle Y\leq X}
.
1a)Let
c
∈
Y
L
{\displaystyle c\in Y_{L}}
be given. Since
Y
L
=
X
L
{\displaystyle Y_{L}=X_{L}\;}
it follows that
c
∈
X
L
{\displaystyle c\in X_{L}}
and thus
X
≰
c
{\displaystyle X\not \leq c}
.
1b)Let
c
∈
X
R
{\displaystyle c\in X_{R}}
be given. Then
c
∈
X
R
∪
{
a
}
=
Y
R
{\displaystyle c\in X_{R}\cup \{a\}=Y_{R}}
, so
c
≰
Y
{\displaystyle c\not \leq Y}
. Therefore by 1a and 1b,
Y
≤
X
{\displaystyle Y\leq X}
.
2a)Let
c
∈
X
L
{\displaystyle c\in X_{L}}
be given. Since
X
L
=
Y
L
{\displaystyle X_{L}=Y_{L}\;}
,
c
∈
Y
L
{\displaystyle c\in Y_{L}}
and therefore
Y
≰
c
{\displaystyle Y\not \leq c}
.
2b)Let
c
∈
Y
R
{\displaystyle c\in Y_{R}}
be given. Then either
c
∈
X
R
{\displaystyle c\in X_{R}\;}
or
c
=
a
{\displaystyle c=a\;}
. If
c
∈
X
R
{\displaystyle c\in X_{R}}
then
c
≰
X
{\displaystyle c\not \leq X}
. If
c
=
a
{\displaystyle c=a\;}
then by our initial assumption
b
≤
c
{\displaystyle b\leq c}
and
b
∈
X
R
{\displaystyle b\in X_{R}}
. Assume
c
≤
X
{\displaystyle c\leq X}
. Then since
b
≤
c
{\displaystyle b\leq c}
by transitivity
b
≤
X
{\displaystyle b\leq X}
. But because
b
∈
X
R
{\displaystyle b\in X_{R}}
it follows that
b
≰
X
{\displaystyle b\not \leq X}
. So therefore by contradiction
c
≰
X
{\displaystyle c\not \leq X}
. And so by 2a and 2b
X
≤
Y
{\displaystyle X\leq Y}
.
◻
{\displaystyle _{\Box }}
Therefore, if maximal lower bounds exist or minimal upper bounds exist, you can ignore anything but those when determining equivalence.
Labelling surreal numbers
edit
Above we talked above about labelling certain surreal numbers to correspond to various real numbers. We need to choose these labels such that the total ordering of the equivalence classes is isomorphic to the ordering of the reals. We also, as will be discussed later, want to choose our labels so that we can imbed the surreal numbers with the arithmetic operations of addition, negation and multiplication, and do it in such a way that it is naturally isomorphic to the corresponding operations in the reals. So if we perform an operation on two real numbers, the same operation on the corresponding surreal numbers gives the expected result.
We already defined 0 as a base case. Inductively, we can generate all the positive integer surreals by saying the successor to the number labelled n is
{
n
|
}
{\displaystyle \{n|\}\;}
. So
1
=
{
0
|
}
,
2
=
{
1
|
}
{\displaystyle 1=\{0|\},2=\{1|\}\;}
and so on.
We also have labelled
{
0
|
1
}
=
1
/
2
{\displaystyle \{0|1\}=1/2\;}
. With that plus the integers, once we develop operations for addition, negation and multiplication we will be able to identify equivalence classes of surreals with any real number of the form
m
/
2
n
{\displaystyle m/2^{n}}
for any integers m and n. So let's get to it....
Addition, Negation and Multiplication
edit
Conway defines addition, negation and multiplication as follows:
Addition
x
+
y
=
{
X
L
+
y
∪
x
+
Y
L
|
X
R
+
y
∪
x
+
Y
R
}
{\displaystyle x+y=\{X_{L}+y\cup x+Y_{L}|X_{R}+y\cup x+Y_{R}\}}
where for a set A and number x
A
+
x
=
{
a
+
x
|
a
∈
A
}
{\displaystyle A+x=\{a+x|a\in A\}}
and
x
+
A
=
{
x
+
a
|
a
∈
A
}
{\displaystyle x+A=\{x+a|a\in A\}}
Negation
−
x
=
{
−
X
R
|
−
X
L
}
{\displaystyle -x=\{-X_{R}|-X_{L}\}\;}
where for a set
A
,
−
A
=
{
−
a
|
a
∈
A
}
{\displaystyle A,-A=\{-a|a\in A\}}
Multiplication
x
y
=
{
(
X
L
y
+
x
Y
L
−
X
L
Y
L
)
∪
(
X
R
y
+
x
Y
R
−
X
R
Y
R
)
|
(
X
L
y
+
x
Y
R
−
X
L
Y
R
)
∪
(
X
R
y
+
x
Y
L
−
X
R
Y
L
)
}
{\displaystyle xy=\{(X_{L}y+xY_{L}-X_{L}Y_{L})\cup (X_{R}y+xY_{R}-X_{R}Y_{R})|(X_{L}y+xY_{R}-X_{L}Y_{R})\cup (X_{R}y+xY_{L}-X_{R}Y_{L})\}}
where
X
Y
=
{
x
y
|
x
∈
X
,
y
∈
Y
}
,
X
y
=
X
{
y
}
,
x
Y
=
{
x
}
Y
{\displaystyle XY=\{xy|x\in X,y\in Y\},Xy=X\{y\},xY=\{x\}Y}
The intent is that these operations should, together with
≤
{\displaystyle \leq }
, form an ordered field in the surreal numbers.
0 is the additive identity
edit
Theorem
x
+
0
=
0
+
x
=
x
{\displaystyle x+0=0+x=x\;}
Proof by induction:
(base)
0
+
0
=
{
0
L
+
0
∪
0
+
0
L
|
0
R
+
0
∪
0
+
0
R
}
=
{
{
a
+
0
|
a
∈
0
L
}
∪
{
0
+
a
|
a
∈
0
L
}
|
{
a
+
0
|
a
∈
0
R
}
∪
{
0
+
a
|
a
∈
0
R
}
}
{\displaystyle {\begin{aligned}0+0&=\{0_{L}+0\cup 0+0_{L}|0_{R}+0\cup 0+0_{R}\}\\&=\{\{a+0|a\in 0_{L}\}\cup \{0+a|a\in 0_{L}\}|\{a+0|a\in 0_{R}\}\cup \{0+a|a\in 0_{R}\}\}\\\end{aligned}}}
And since both
0
L
{\displaystyle 0_{L}\;}
and
0
R
{\displaystyle 0_{R}\;}
are empty,
0
+
0
=
{
|
}
=
0
{\displaystyle {\begin{aligned}0+0&=\{\;|\;\}=0\end{aligned}}}
(induction) Let surreal number x be given and assume that for all
z
∈
X
L
∪
X
R
0
+
z
=
z
+
0
=
z
{\displaystyle z\in X_{L}\cup X_{R}\;0+z=z+0=z}
. We want to show that
x
+
0
=
0
+
x
=
x
{\displaystyle x+0=0+x=x\;}
Since both
0
L
{\displaystyle 0_{L}\;}
and
0
R
{\displaystyle 0_{R}\;}
are empty, it follows that
x
+
0
=
{
X
L
+
0
∪
x
+
0
L
|
X
R
+
0
∪
x
+
0
R
}
=
{
X
L
+
0
∪
{
x
+
a
|
a
∈
0
L
}
|
X
R
+
0
∪
{
x
+
a
|
a
∈
0
R
}
}
=
{
X
L
+
0
|
X
R
+
0
}
=
{
{
a
+
0
|
a
∈
X
L
}
|
{
a
+
0
|
a
∈
X
R
}
}
{\displaystyle {\begin{aligned}x+0&=\{X_{L}+0\cup x+0_{L}|X_{R}+0\cup x+0_{R}\}\\&=\{X_{L}+0\cup \{x+a|a\in 0_{L}\}|X_{R}+0\cup \{x+a|a\in 0_{R}\}\}\\&=\{X_{L}+0|X_{R}+0\}\\&=\{\{a+0|a\in X_{L}\}|\{a+0|a\in X_{R}\}\}\\\end{aligned}}}
By our inductive assumption above
a
+
0
=
0
+
a
=
a
{\displaystyle a+0=0+a=a\;}
so
x
+
0
=
{
{
0
+
a
|
a
∈
X
L
}
|
{
0
+
a
|
a
∈
X
R
}
}
=
{
{
a
|
a
∈
X
L
}
|
{
a
|
a
∈
X
R
}
}
=
{
X
L
|
X
R
}
=
x
{\displaystyle {\begin{aligned}x+0&=\{\{0+a|a\in X_{L}\}|\{0+a|a\in X_{R}\}\}\\&=\{\{a|a\in X_{L}\}|\{a|a\in X_{R}\}\}\\&=\{X_{L}|X_{R}\}=x\end{aligned}}}
and since
0
L
{\displaystyle 0_{L}\;}
and
0
R
{\displaystyle 0_{R}\;}
are empty
x
+
0
=
{
0
+
X
L
|
0
+
X
R
}
=
{
{
a
+
x
|
a
∈
0
L
}
∪
0
+
X
L
|
{
a
+
x
|
a
∈
0
R
}
∪
0
+
X
R
}
=
{
0
L
+
x
∪
0
+
X
L
|
0
R
+
x
∪
0
+
X
R
}
=
0
+
x
◻
{\displaystyle {\begin{aligned}x+0&=\{0+X_{L}|0+X_{R}\}\\&=\{\{a+x|a\in 0_{L}\}\cup 0+X_{L}|\{a+x|a\in 0_{R}\}\cup 0+X_{R}\}\\&=\{0_{L}+x\cup 0+X_{L}|0_{R}+x\cup 0+X_{R}\}\\&=0+x\\\end{aligned}}_{\Box }}
Commutativity of addition
edit
Theorem
x
+
y
=
y
+
x
{\displaystyle x+y=y+x\;}
for all surreal numbers x and y
Proof by induction:
Let surreal numbers x and y be given.
(Base)
x
+
0
=
0
+
x
=
x
{\displaystyle x+0=0+x=x\;}
and
y
+
0
=
0
+
y
=
y
{\displaystyle y+0=0+y=y\;}
(Induction) Assume that for all
z
∈
X
L
∪
X
R
∪
Y
L
∪
Y
R
{\displaystyle z\in X_{L}\cup X_{R}\cup Y_{L}\cup Y_{R}\;}
that
x
+
z
=
z
+
x
{\displaystyle x+z=z+x\;}
and
y
+
z
=
z
+
y
{\displaystyle y+z=z+y\;}
. Then we want to show that
x
+
y
=
y
+
x
{\displaystyle x+y=y+x\;}
x
+
y
=
{
X
L
+
y
∪
x
+
Y
L
|
X
R
+
y
∪
x
+
Y
R
}
=
{
x
+
Y
L
∪
X
L
+
y
|
x
+
Y
R
∪
X
R
+
y
}
=
{
{
x
+
a
|
a
∈
Y
L
}
∪
{
a
+
y
|
a
∈
X
L
}
|
{
x
+
a
|
a
∈
Y
R
}
∪
{
a
+
y
|
a
∈
X
R
}
}
{\displaystyle {\begin{aligned}x+y&=\{X_{L}+y\cup x+Y_{L}|X_{R}+y\cup x+Y_{R}\}\\&=\{x+Y_{L}\cup X_{L}+y|x+Y_{R}\cup X_{R}+y\}\\&=\{\{x+a|a\in Y_{L}\}\cup \{a+y|a\in X_{L}\}|\{x+a|a\in Y_{R}\}\cup \{a+y|a\in X_{R}\}\}\\\end{aligned}}}
so by our inductive assumption,
x
+
y
=
{
{
a
+
x
|
a
∈
Y
L
}
∪
{
y
+
a
|
a
∈
X
L
}
|
{
a
+
x
|
a
∈
Y
R
}
∪
{
y
+
a
|
a
∈
X
R
}
}
=
{
Y
L
+
x
∪
y
+
X
L
|
Y
R
+
x
∪
y
+
X
R
}
=
y
+
x
◻
{\displaystyle {\begin{aligned}x+y&=\{\{a+x|a\in Y_{L}\}\cup \{y+a|a\in X_{L}\}|\{a+x|a\in Y_{R}\}\cup \{y+a|a\in X_{R}\}\}\\&=\{Y_{L}+x\cup y+X_{L}|Y_{R}+x\cup y+X_{R}\}\\&=y+x\;_{\Box }\end{aligned}}}
Associativity of addition
edit
Theorem For all surreal numbers
x
,
y
,
z
x
+
(
y
+
z
)
=
(
x
+
y
)
+
z
{\displaystyle x,y,z\;x+(y+z)=(x+y)+z}
Proof: Let surreal numbers x,y and z be given
(base) Since addition is commutative,
0
+
(
0
+
0
)
=
(
0
+
0
)
+
0
{\displaystyle 0+(0+0)=(0+0)+0\;}
(induction) Assume for all
a
∈
X
L
∪
X
R
∪
Y
L
∪
Y
R
∪
Z
L
∪
Z
R
{\displaystyle a\in X_{L}\cup X_{R}\cup Y_{L}\cup Y_{R}\cup Z_{L}\cup Z_{R}}
that:
a
+
(
x
+
y
)
=
(
a
+
x
)
+
y
{\displaystyle a+(x+y)=(a+x)+y}
a
+
(
y
+
z
)
=
(
a
+
y
)
+
z
{\displaystyle a+(y+z)=(a+y)+z}
a
+
(
x
+
z
)
=
(
a
+
x
)
+
z
{\displaystyle a+(x+z)=(a+x)+z}
x
+
(
a
+
y
)
=
(
x
+
a
)
+
y
{\displaystyle x+(a+y)=(x+a)+y}
x
+
(
a
+
z
)
=
(
x
+
a
)
+
z
{\displaystyle x+(a+z)=(x+a)+z}
y
+
(
a
+
z
)
=
(
y
+
a
)
+
z
{\displaystyle y+(a+z)=(y+a)+z}
x
+
(
y
+
a
)
=
(
x
+
y
)
+
a
{\displaystyle x+(y+a)=(x+y)+a}
x
+
(
z
+
a
)
=
(
x
+
z
)
+
a
{\displaystyle x+(z+a)=(x+z)+a}
y
+
(
z
+
a
)
=
(
y
+
z
)
+
a
{\displaystyle y+(z+a)=(y+z)+a}
Then we want to show that
x
+
(
y
+
z
)
=
(
x
+
y
)
+
z
{\displaystyle x+(y+z)=(x+y)+z\;}
x
+
(
y
+
z
)
=
{
X
L
+
(
y
+
z
)
∪
x
+
(
y
+
z
)
L
|
X
R
+
(
y
+
z
)
∪
x
+
(
y
+
z
)
R
}
=
{
{
a
+
(
y
+
z
)
|
a
∈
X
L
}
∪
{
x
+
a
|
a
∈
(
y
+
z
)
L
}
|
{
a
+
(
y
+
z
)
|
a
∈
X
R
}
∪
{
x
+
a
|
a
∈
(
y
+
z
)
R
}
}
=
{
{
a
+
(
y
+
z
)
|
a
∈
X
L
}
∪
{
x
+
a
|
a
∈
{
Y
L
+
z
∪
y
+
Z
L
}
}
|
{
a
+
(
y
+
z
)
|
a
∈
X
R
}
∪
{
x
+
a
|
a
∈
{
Y
R
+
z
∪
y
+
Z
R
}
}
}
=
{
{
a
+
(
y
+
z
)
|
a
∈
X
L
}
∪
{
x
+
a
|
a
∈
Y
L
+
z
}
∪
{
x
+
a
|
a
∈
y
+
Z
L
}
|
{
a
+
(
y
+
z
)
|
a
∈
X
R
}
∪
{
x
+
a
|
a
∈
Y
R
+
z
}
∪
{
x
+
a
|
a
∈
y
+
Z
R
}
}
=
{
{
a
+
(
y
+
z
)
|
a
∈
X
L
}
∪
{
x
+
(
a
+
z
)
|
a
∈
Y
L
}
∪
{
x
+
(
y
+
a
)
|
a
∈
Z
L
}
|
{
a
+
(
y
+
z
)
|
a
∈
X
R
}
∪
{
x
+
(
a
+
z
)
|
a
∈
Y
R
}
∪
{
x
+
(
y
+
a
)
|
a
∈
Z
R
}
}
{\displaystyle {\begin{aligned}x+(y+z)&=\{X_{L}+(y+z)\cup x+(y+z)_{L}|X_{R}+(y+z)\cup x+(y+z)_{R}\}\\&=\{\{a+(y+z)|a\in X_{L}\}\cup \{x+a|a\in (y+z)_{L}\}|\{a+(y+z)|a\in X_{R}\}\cup \{x+a|a\in (y+z)_{R}\}\}\\&=\{\{a+(y+z)|a\in X_{L}\}\cup \{x+a|a\in \{Y_{L}+z\cup y+Z_{L}\}\}|\{a+(y+z)|a\in X_{R}\}\cup \{x+a|a\in \{Y_{R}+z\cup y+Z_{R}\}\}\}\\&=\{\{a+(y+z)|a\in X_{L}\}\cup \{x+a|a\in Y_{L}+z\}\cup \{x+a|a\in y+Z_{L}\}|\{a+(y+z)|a\in X_{R}\}\cup \{x+a|a\in Y_{R}+z\}\cup \{x+a|a\in y+Z_{R}\}\}\\&=\{\{a+(y+z)|a\in X_{L}\}\cup \{x+(a+z)|a\in Y_{L}\}\cup \{x+(y+a)|a\in Z_{L}\}|\{a+(y+z)|a\in X_{R}\}\cup \{x+(a+z)|a\in Y_{R}\}\cup \{x+(y+a)|a\in Z_{R}\}\}\\\end{aligned}}}
And by our inductive assumptions, the above equals
…
=
{
{
(
a
+
y
)
+
z
|
a
∈
X
L
}
∪
{
(
x
+
a
)
+
z
|
a
∈
Y
L
}
∪
{
(
x
+
y
)
+
a
|
a
∈
Z
L
}
|
{
(
a
+
y
)
+
z
|
a
∈
X
R
}
∪
{
(
x
+
a
)
+
z
|
a
∈
Y
R
}
∪
{
(
x
+
y
)
+
a
|
a
∈
Z
R
}
}
=
{
{
a
+
z
|
a
∈
X
L
+
y
}
∪
{
a
+
z
|
a
∈
x
+
Y
L
}
∪
{
(
x
+
y
)
+
a
|
a
∈
Z
L
}
|
{
a
+
z
|
a
∈
X
R
+
y
}
∪
{
a
+
z
|
a
∈
x
+
Y
R
}
∪
{
(
x
+
y
)
+
a
|
a
∈
Z
R
}
}
=
{
{
a
+
z
|
a
∈
X
L
+
y
∪
x
+
Y
L
}
∪
{
(
x
+
y
)
+
a
|
a
∈
Z
L
}
|
{
a
+
z
|
a
∈
X
R
+
y
∪
x
+
Y
R
}
∪
{
(
x
+
y
)
+
a
|
a
∈
Z
R
}
=
{
{
a
+
z
|
a
∈
(
x
+
y
)
L
}
∪
{
(
x
+
y
)
+
a
|
a
∈
Z
L
}
|
{
a
+
z
|
a
∈
(
x
+
y
)
R
}
∪
{
(
x
+
y
)
+
a
|
a
∈
Z
R
}
}
=
{
(
x
+
y
)
L
+
z
∪
(
x
+
y
)
+
Z
L
|
(
x
+
y
)
R
+
z
∪
(
x
+
y
)
+
Z
R
}
=
(
x
+
y
)
+
z
◻
{\displaystyle {\begin{aligned}\dots &=\{\{(a+y)+z|a\in X_{L}\}\cup \{(x+a)+z|a\in Y_{L}\}\cup \{(x+y)+a|a\in Z_{L}\}|\{(a+y)+z|a\in X_{R}\}\cup \{(x+a)+z|a\in Y_{R}\}\cup \{(x+y)+a|a\in Z_{R}\}\}\\&=\{\{a+z|a\in X_{L}+y\}\cup \{a+z|a\in x+Y_{L}\}\cup \{(x+y)+a|a\in Z_{L}\}|\{a+z|a\in X_{R}+y\}\cup \{a+z|a\in x+Y_{R}\}\cup \{(x+y)+a|a\in Z_{R}\}\}\\&=\{\{a+z|a\in X_{L}+y\cup x+Y_{L}\}\cup \{(x+y)+a|a\in Z_{L}\}|\{a+z|a\in X_{R}+y\cup x+Y_{R}\}\cup \{(x+y)+a|a\in Z_{R}\}\\&=\{\{a+z|a\in (x+y)_{L}\}\cup \{(x+y)+a|a\in Z_{L}\}|\{a+z|a\in (x+y)_{R}\}\cup \{(x+y)+a|a\in Z_{R}\}\}\\&=\{(x+y)_{L}+z\cup (x+y)+Z_{L}|(x+y)_{R}+z\cup (x+y)+Z_{R}\}\\&=(x+y)+z\\_{\Box }\end{aligned}}}
Addition is order preserving
edit
Theorem: For all surreal numbers x, y and z if
x
≤
y
{\displaystyle x\leq y\;}
then
x
+
z
≤
y
+
z
{\displaystyle x+z\leq y+z\;}
Proof by induction:
(base)
0
≤
0
{\displaystyle 0\leq 0\;}
and
0
+
0
≤
0
+
0
{\displaystyle 0+0\leq 0+0\;}
.
(induction) Let surreal numbers x, y and z be given and assume that for all
a
,
b
∈
X
L
∪
X
R
∪
Y
L
∪
Y
R
∪
Z
L
∪
Z
R
{\displaystyle a,b\in X_{L}\cup X_{R}\cup Y_{L}\cup Y_{R}\cup Z_{L}\cup Z_{R}\;}
that if
a
≤
b
{\displaystyle a\leq b\;}
then
a
+
z
≤
b
+
z
{\displaystyle a+z\leq b+z\;}
. Then we want to show that if
x
≤
y
{\displaystyle x\leq y}
then
x
+
z
≤
y
+
z
{\displaystyle x+z\leq y+z}
.
1) First assume that
x
≤
y
{\displaystyle x\leq y}
and let
a
∈
(
y
+
z
)
R
{\displaystyle a\in (y+z)_{R}\;}
be given such that
a
≤
x
+
z
{\displaystyle a\leq x+z}
. Then
a
∈
Y
R
+
z
∪
y
+
Z
R
{\displaystyle a\in Y_{R}+z\cup y+Z_{R}}
so either
∃
b
∈
Y
R
a
=
b
+
z
{\displaystyle \exists b\in Y_{R}\;a=b+z}
or
∃
b
∈
Z
R
a
=
y
+
b
{\displaystyle \exists b\in Z_{R}\;a=y+b}
.
a) Let
b
∈
Y
R
{\displaystyle b\in Y_{R}\;}
be given such that
a
=
b
+
z
{\displaystyle a=b+z\;}
. Then
b
+
z
≤
x
+
z
{\displaystyle b+z\leq x+z}
. Since
b
∈
Y
R
{\displaystyle b\in Y_{R}}
, that implies that
y
≤
b
{\displaystyle y\leq b}
so
x
≤
y
≤
b
{\displaystyle x\leq y\leq b}
. And therefore since
x
≤
b
{\displaystyle x\leq b}
by our inductive assumption it follows that
x
+
z
≤
b
+
z
{\displaystyle x+z\leq b+z}
, and thus
x
+
z
==
b
+
z
{\displaystyle x+z==b+z\;}
. ...
We'll just need to prove two quick lemmas to proceed...
Lemma:
−
(
−
x
)
=
x
{\displaystyle -(-x)=x\;}
Proof by induction:
(Base) -(-0) = -0 = 0
(Induction) Let surreal number x be given and assume that for all
a
∈
X
L
∪
X
R
−
(
−
a
)
=
a
{\displaystyle a\in X_{L}\cup X_{R}\;-(-a)=a}
−
(
−
x
)
=
{
−
(
−
X
)
R
|
−
(
−
X
)
L
}
=
{
{
−
a
|
a
∈
(
−
X
)
R
}
|
{
−
a
|
a
∈
(
−
X
)
L
}
}
=
{
{
−
a
|
a
∈
−
X
L
}
|
{
−
a
|
a
∈
−
X
R
}
}
=
{
{
−
(
−
a
)
|
a
∈
X
L
}
|
{
−
(
−
a
)
|
a
∈
X
R
}
}
=
{
{
a
|
a
∈
X
L
}
|
{
a
|
a
∈
X
R
}
}
=
{
X
L
|
X
R
}
=
X
◻
{\displaystyle {\begin{aligned}-(-x)&=\{-(-X)_{R}|-(-X)_{L}\}\\&=\{\{-a|a\in (-X)_{R}\}|\{-a|a\in (-X)_{L}\}\}\\&=\{\{-a|a\in -X_{L}\}|\{-a|a\in -X_{R}\}\}\\&=\{\{-(-a)|a\in X_{L}\}|\{-(-a)|a\in X_{R}\}\}\\&=\{\{a|a\in X_{L}\}|\{a|a\in X_{R}\}\}\\&=\{X_{L}|X_{R}\}=X_{\Box }\end{aligned}}}
Lemma: If
x
≤
y
{\displaystyle x\leq y}
then
−
y
≤
−
x
{\displaystyle -y\leq -x}
Proof by induction:
(Base)
0
≤
0
{\displaystyle 0\leq 0}
and
−
0
≤
−
0
{\displaystyle -0\leq -0}
(Induction)
Let x and y be given such that
x
≤
y
{\displaystyle x\leq y}
and assume that for all
a
∈
X
L
∪
X
R
{\displaystyle a\in X_{L}\cup X_{R}}
that if
a
≤
y
{\displaystyle a\leq y}
then
−
y
≤
−
a
{\displaystyle -y\leq -a}
. We want to show that
−
y
≤
−
x
{\displaystyle -y\leq -x}
.
1)Let
a
∈
(
−
X
)
R
{\displaystyle a\in (-X)_{R}}
be given such that
a
≤
−
y
{\displaystyle a\leq -y}
. Then
−
a
∈
X
L
{\displaystyle -a\in X_{L}}
so
−
a
≤
x
≤
y
{\displaystyle -a\leq x\leq y}
. So by our inductive assumption
−
y
≤
a
{\displaystyle -y\leq a}
.
Theorem For all surreal numbers x, -x is also a surreal number
Proof by induction
(base) -0 = 0
(induction) Let x be given and assume for all
a
∈
X
L
∪
X
R
∄
y
∈
(
−
A
)
L
z
∈
(
−
A
)
R
z
≤
y
{\displaystyle a\in X_{L}\cup X_{R}\;\not \exists y\in (-A)_{L}z\in (-A)_{R}\;z\leq y}
. Then let
a
∈
(
−
X
)
L
b
∈
(
−
X
)
R
{\displaystyle a\in (-X)_{L}\;b\in (-X)_{R}}
be given.
It follows from the definition of negation that
a
∈
−
X
R
{\displaystyle a\in -X_{R}}
and
b
∈
−
X
L
{\displaystyle b\in -X_{L}}
. So
−
a
∈
X
R
{\displaystyle -a\in X_{R}}
and
−
b
∈
X
L
{\displaystyle -b\in X_{L}}
and therefore
−
b
≤
−
a
{\displaystyle -b\leq -a}
.
Theorem:
x
+
(
−
x
)
==
0
{\displaystyle x+(-x)==0\;}
Proof by induction:
(Base)
0
+
(
−
0
)
=
(
−
0
)
=
{
|
}
=
0
{\displaystyle 0+(-0)=(-0)=\{|\}=0\;}
(Induction) Assume that for all
z
∈
X
L
∪
X
R
∪
−
X
L
∪
−
X
R
z
+
(
−
z
)
==
0
{\displaystyle z\in X_{L}\cup X_{R}\cup -X_{L}\cup -X_{R}\;z+(-z)==0}
. Then we want to show that
x
+
(
−
x
)
==
0
{\displaystyle x+(-x)==0\;}
, ie. that
(
x
+
(
−
x
)
)
≤
0
{\displaystyle (x+(-x))\leq 0}
and
0
≤
(
x
+
(
−
x
)
)
{\displaystyle 0\leq (x+(-x))}
.
x
+
(
−
x
)
=
{
X
L
+
(
−
x
)
∪
x
+
(
−
x
)
L
|
X
R
+
(
−
x
)
∪
x
+
(
−
x
)
R
}
=
{
X
L
+
(
−
x
)
∪
x
+
(
−
X
R
)
|
X
R
+
(
−
x
)
∪
x
+
(
−
X
L
)
}
{\displaystyle {\begin{aligned}x+(-x)&=\{X_{L}+(-x)\cup x+(-x)_{L}|X_{R}+(-x)\cup x+(-x)_{R}\}\\&=\{X_{L}+(-x)\cup x+(-X_{R})|X_{R}+(-x)\cup x+(-X_{L})\}\\\end{aligned}}}
1) Assume
∃
a
∈
X
R
a
+
(
−
x
)
≤
0
{\displaystyle \exists a\in X_{R}\;a+(-x)\leq 0}
. Then by our inductive assumption
(
−
a
+
a
)
+
(
−
x
)
≤
−
a
+
0
{\displaystyle (-a+a)+(-x)\leq -a+0}
, so
0
+
(
−
x
)
≤
−
a
{\displaystyle 0+(-x)\leq -a}
and hence
−
x
≤
−
a
{\displaystyle -x\leq -a}
. But since
a
∈
X
R
{\displaystyle a\in X_{R}}
, it follows that
−
a
∈
−
X
L
{\displaystyle -a\in -X_{L}}
and therefore
−
x
≰
−
a
{\displaystyle -x\not \leq -a}
. So by contradiction
∄
a
∈
X
R
a
+
(
−
x
)
≤
0
{\displaystyle \not \exists a\in X_{R}\;a+(-x)\leq 0}
.
2) Assume
∃
a
∈
−
X
L
x
+
a
≤
0
{\displaystyle \exists a\in -X_{L}\;x+a\leq 0}
. Then by our inductive assumption
x
+
a
+
(
−
a
)
≤
−
a
{\displaystyle x+a+(-a)\leq -a}
and so
x
≤
−
a
{\displaystyle x\leq -a}
. However since
a
∈
−
X
L
{\displaystyle a\in -X_{L}}
it follows that
−
a
∈
X
L
{\displaystyle -a\in X_{L}}
since
−
(
−
a
)
=
a
{\displaystyle -(-a)=a}
, and thus
x
≰
−
a
{\displaystyle x\not \leq -a}
. Therefore by contradiction
∄
a
∈
−
X
L
x
+
a
≤
0
{\displaystyle \not \exists a\in -X_{L}\;x+a\leq 0}
.
So by 1) and 2) above
∄
a
∈
(
x
+
(
−
x
)
)
R
a
≤
0
{\displaystyle \not \exists a\in (x+(-x))_{R}\;a\leq 0}
. And since
0
L
{\displaystyle 0_{L}\;}
is empty, we have that
∄
a
∈
0
L
(
x
+
(
−
x
)
)
≤
a
{\displaystyle \not \exists a\in 0_{L}\;(x+(-x))\leq a}
, and therefore
(
x
+
(
−
x
)
)
≤
0
{\displaystyle (x+(-x))\leq 0}
.
3) Assume
∃
a
∈
X
L
0
≤
a
+
(
−
x
)
{\displaystyle \exists a\in X_{L}\;0\leq a+(-x)}
. Then by our inductive assumption
−
a
+
0
≤
(
−
a
+
a
)
+
(
−
x
)
{\displaystyle -a+0\leq (-a+a)+(-x)}
, so
−
a
≤
0
+
(
−
x
)
{\displaystyle -a\leq 0+(-x)}
and hence
−
a
≤
−
x
{\displaystyle -a\leq -x}
. But since
a
∈
X
L
{\displaystyle a\in X_{L}}
, it follows that
−
a
∈
−
X
R
{\displaystyle -a\in -X_{R}}
and therefore
−
a
≰
−
x
{\displaystyle -a\not \leq -x}
. So by contradiction
∄
a
∈
X
L
0
≤
a
+
(
−
x
)
{\displaystyle \not \exists a\in X_{L}\;0\leq a+(-x)}
.
4) Assume
∃
a
∈
−
X
R
0
≤
x
+
a
{\displaystyle \exists a\in -X_{R}\;0\leq x+a}
. Then by our inductive assumption
−
a
≤
x
+
a
+
(
−
a
)
{\displaystyle -a\leq x+a+(-a)}
and so
−
a
≤
x
{\displaystyle -a\leq x}
. However since
a
∈
−
X
R
{\displaystyle a\in -X_{R}}
it follows that
−
a
∈
X
R
{\displaystyle -a\in X_{R}}
since
−
(
−
a
)
=
a
{\displaystyle -(-a)=a\;}
, and thus
−
a
≰
x
{\displaystyle -a\not \leq x}
. Therefore by contradiction
∄
a
∈
−
X
R
0
≤
x
+
a
{\displaystyle \not \exists a\in -X_{R}\;0\leq x+a}
.
So by 3) and 4) above
∄
a
∈
(
x
+
(
−
x
)
)
L
0
≤
a
{\displaystyle \not \exists a\in (x+(-x))_{L}\;0\leq a}
. And since
0
R
{\displaystyle 0_{R}\;}
is empty, we have that
∄
a
∈
0
R
a
≤
(
x
+
(
−
x
)
)
{\displaystyle \not \exists a\in 0_{R}\;a\leq (x+(-x))}
, and therefore
0
≤
(
x
+
(
−
x
)
)
{\displaystyle 0\leq (x+(-x))}
and hence
0
==
(
x
+
(
−
x
)
)
◻
{\displaystyle 0==(x+(-x))\;_{\Box }}
.
Theorem For all surreal numbers
x
{\displaystyle x\;}
and
y
{\displaystyle y\;}
x
+
y
{\displaystyle x+y\;}
is also a surreal number
Proof by induction:
(base)
0
+
0
=
0
{\displaystyle 0+0=0\;}
, which is a surreal number.
(induction) Assume that for all
a
,
b
∈
X
L
∪
X
R
∪
Y
L
∪
Y
R
{\displaystyle a,b\in X_{L}\cup X_{R}\cup Y_{L}\cup Y_{R}}
that
a
+
b
,
a
+
x
,
a
+
y
{\displaystyle a+b,a+x,a+y\;}
are all surreal numbers. Then we want to show that
x
+
y
{\displaystyle x+y\;}
is also a surreal number.
x
+
y
=
{
X
L
+
y
∪
x
+
Y
L
|
X
R
+
y
∪
x
+
Y
R
}
{\displaystyle {\begin{aligned}x+y&=\{X_{L}+y\cup x+Y_{L}|X_{R}+y\cup x+Y_{R}\}\end{aligned}}}
1) Assume
∃
a
∈
X
R
b
∈
X
L
a
+
y
≤
b
+
y
{\displaystyle \exists a\in X_{R}b\in X_{L}\;a+y\leq b+y}
...
Distributivity of multiplication over addition
edit
1 is the multiplicative identity
edit