On 12/1/2023 3:46 PM, olcott wrote:
> On 12/1/2023 2:14 PM, Jim Burns wrote:
>> On 12/1/2023 1:42 PM, olcott wrote:
>>> When
>>> provable from the axioms of L means
>>> true in L, and
>>> unprovable in L means
>>> untrue in L
>>> then
>>> incompleteness and undecidability
>>> cannot exist.
>>
>> When
>> the objects in the language of L
>> can be represented by
>> objects in the domain of L
>> then
>> one object represents
>> "x can't be proved in L"
>>
>> Consider
>> | "x can't be proved in L" can't be proved in L
Oops.
Better:
"preceded by its quotation can't be proved in L"
preceded by its quotation can't be proved in L.
Not a self-reference, but
a self-description.
>> If it's true,
>> it can't be proved, and
>> L is incomplete.
>
> I stipulate that
> this means that x is simply untrue in L.
1/2. You introduce a private meaning.
> This <is> the way that
> the entire body of analytic truth really works.
2/2. You claim (stipulate?) that
everyone is using your private meaning.
You think that stipulating is
your dragon-slaying sword.
Stipulating is almost always inappropriate.
But, yes, in those instances in which
it is appropriate, it is a dragon-slayer.
I can stipulate that ABC is a right triangle.
By doing that, I chalk the outline of
the conversation. _I_ am talking about
a right triangle. If you join me, _you_ are
talking about a right triangle. If you reject
my stipulation, you haven't joined me.
Stipulation slays non-right-triangle dragon.
I can't (I really, really shouldn't) stipulate
that the square of the hypotenuse of ABC is
equal to the sum of the squares of
the two remaining sides of ABC.
The goal should be to convince you of that.
A stipulation does no such thing.
Used in this way,
a stipulation does not slay dragons,
it stomps off the field, crying about how
the dragon unfairly didn't lay down and die.
> That math diverges from this is its error.
A finite sequence of claims with
no first-false claim
has no false claim.
Q in ⟨… P∨Q ¬P Q …⟩ is not first-false.
If you stipulate otherwise,
you'd be better off with 15-story kittens.
>> If it's false,
>> it can be proved,
>> but it's false! and
>> L is inconsistent.
>>
>> When
>> the objects in the language of L
>> can be represented by
>> objects in the domain of L
>> then
>> the choice is between
>> incomplete and inconsistent.
>
> False dichotomy.
Theorem.
Convincing everyone the choices are
incomplete or inconsistent
is what the proof of the theorem is for.
Ignoring a proof does not make it wrong.
> It is perfectly consistent to say that
> G is untrue in F.
> When unprovable means untrue
> then it does not mean incomplete.
Words have meanings.
Remove the words, and the meanings remain,
(silently now)
with the same nature they always had.
----
Consider the system ST [Boolos] with
empty set, adjunct, and extensionality.
| ∃x∀u: u∉x
| ∀x∀y∃z: ∀u(u∈z ⟺ u∈x ∨ u=y)
| ∀x∀y: x=y ⟺ ∀u(u∈x ⇔ u∈y)
I stipulate that ST is
what we're taking about, right now.
Reject my stipulation. Go ahead.
Now we two are aren't talking about anything.
No wins, no losses, but no progress.
I stipulate that
"z is the adjunct of y to x" means
z = x†y ⟺ ∀u(u∈z ⟺ u∈x ∋ u=y)
x†y = x∪{y}
x†y†z = (x†y)†z
"0 is the empty set" means
∀u: u∉0
"y is the successor of x" means
y = x⁺¹ ⟺ y = x†x
"x is the predecessor of y" means
x = y⁻¹ ⟺ y = x⁺¹
"x is less than y" means
x < y ⟺ x ∈ y
"x is a natural number" means
ℕ∋(x) ⟺
x = 0 ∨
(x⁻¹<x ∧ ∀u<x:(u=0 v u⁻¹<x))
"z is the ordered pair ⟨x,y⟩" means
z = ⟨x,y⟩ ⟺ z = 0†(0†x)†(0†x†y)
⟨x,y⟩ = {{x},{x,y}} [Kuratowski]
⟨x,y,z⟩ = ⟨⟨x,y⟩,z⟩
⟨x,…,y,z⟩ = ⟨⟨x,…,y⟩,z⟩
For ℕ∋(x) ℕ∋(y) ℕ∋(z)
"z is the sum of x and y" means
x + y = z ⟺
⟨ ⟨x,0,x⟩ ⟨x,0⁺¹,x⁺¹⟩ … ⟨x,y,z⟩ ⟩ exists
such that
for each of its splits Fᣔ<ᣔH
some ⟨x,i,j⟩‖⟨x,i⁺¹,j⁺¹⟩ is last‖first in F‖H
For ℕ∋(x) ℕ∋(y) ℕ∋(z)
"z is the product of x and y" means
x × y = z ⟺
⟨ ⟨x,0,0⟩ ⟨x,0⁺¹,0+x⟩ … ⟨x,y,z⟩ ⟩ exists
such that
for each of its splits Fᣔ<ᣔH
some ⟨x,i,j⟩‖⟨x,i⁺¹,j+x⟩ is last‖first in F‖H
For ℕ∋(n)
"z is f applied to x recursively n times" means
z = f⁽ⁿ⁾(x) ⟺
⟨ ⟨0,x⟩ ⟨0⁺¹,f(x)⟩ … ⟨n,z⟩ ⟩ exists
such that
for each of its splits Fᣔ<ᣔH
some ⟨i,y⟩‖⟨i⁺¹,f(y)⟩ is last‖first in F‖H
That is the usual natural number arithmetic,
given here as definitions in and theorems of ST
----
Apart from
| ∃x∀u: u∉x
| ∀x∀y∃z: ∀u(u∈z ⟺ u∈x ∨ u=y)
| ∀x∀y: x=y ⟺ ∀u(u∈x ⇔ u∈y)
|
all those stipulations are optional.
For the rest, they are
definiens ⟺ definiendum
Remove the definiens and
the definiendum remains.
Those stipulations are essential for humans
communicating to other humans
how to represent arithmetic in ST,
but their truth or falsity is unaffected
by those stipulations.
Those are dragon-slayer stipulations, but
only because rejecting them has no consequences.