On 9/3/2021 5:29 PM, WM wrote:
> Jim Burns schrieb
> am Freitag, 3. September 2021 um 21:27:39 UTC+2:
>> On 9/3/2021 9:22 AM, WM wrote:
>>> Before all fractions have been enumerated there, at least one
>>> fraction must have been enumerated in every unit interval.
>>
>> No.
>
> So all rational numbers in all intervals can be enumerated
> without having enumerated at least one rationale number
> in every interval?
Before all rationals have been enumerated, at least one
unit interval does not have any rationals enumerated in it.
>> Consider the much simpler hypothesis that your claim here
>> is wrong. Before all rationals have been enumerated there,
>> at least one unit interval has zero rationals enumerated in it.
>
> No, not "before" but after all rationals have been enumerated,
> one interval has no enumerated rationals in it.
> That is what you claim.
No, this is what I claim:
Before all rationals have been enumerated, at least one
unit interval does not have any rationals enumerated in it.
Your confusion is probably caused by assuming that I also
claim that there is a last natural number.
I don't claim that.
I claim...
For each k in N+, the set of positive naturals,
a steppable {1,...,k} exists in which,
for adjacent i,j, j = i+1.
For each p in Q+, the set of positive rationals,
j and k exist in N+ such that k*p = j.
( One way to define this multiplication '*' is geometrically.
( Construct similar right triangles K0J and U0P with sides
( d(0K) = k, d(0J) = j, d(0U) = 1, d(0P) = p.
( Because they're similar, p/1 = j/k, k*p = j.
I claim we have enumerated all the elements of Q+ with
the elements of N+. By that, I mean
I claim that index: Q+ --> N+ can be shown such that
for each p in Q+, a unique i in N+ exists, i = index(p),
index(p) = index(q) iff p = q,
and (extra credit, for k*k*j*j/rad(j))
for each k in N+,a unique p in Q+ exists, index(q) = k,
One consequence of these claims is that,
before all rationals have been enumerated, at least one
unit interval does not have any rationals enumerated in it.
"Before all rationals have been enumerated", there is
at least one rational p which has not been enumerated.
p has an index, i = index(p). i is FISONable.
A steppable set {0,...,i-1} exists in which, for all
adjacent j,k, k = j+1.
Let Q[i-1] be the set of rationals indexed by
an element of {0,...,i-1}.
Let r be the maximum rational in Q[i-1]
(Lemma: maximum r exists. [1])
r is in Q+
j,k exist in N+, r = j/k
The set { i =< j/k } is a subset of {0,...,j},
is also steppable,
is not empty,
has a last member m.
m =< j/k and m+1 > j/k.
Each rational in unit interval (m+1,m+2] is larger than
each rational indexed by one of {1,...,i-1}.
The unit interval (m+1,m+2] has no rationals in it
which are indexed before p, index(p) = i.
Therefore,
before all rationals have been enumerated, at least one
unit interval does not have any rationals enumerated in it.
[1]
The set Q[i-1] of rationals indexed by one of {1,...,i-1}
contains a maximum r.
Consider BestYet, a subset of {1,...,i-1}.
k is in BestYet iff
the rational indexed by k is larger than
each rational indexed by one of 1,...,k-1
Each subset of {1,...,i-1} contains a first and a last,
unless it's empty.
BestYet is a subset of {1,...,i-1}
1 is in BestYet. It's not empty.
BestYet contains a last. Call it m.
m is the index for some rational r, index(r) = m
Consider EvenBetter, a subset of {m+1,...,i-1} of
indexes of rationals following r that are larger than r.
EvenBetter is a subset of {1,...,i-1}.
Either it is empty or it contains a first and a last.
Suppose n is the first element of EvenBetter,
and index(s) = n.
s > r, everything between s and r is =< r,
everything before r is =< r
s is larger than each rational indexed by 1,...,n-1.
But then n = index(s) should be in BestYet.
But then, m = index(r) is NOT last in BestYet.
Contradiction.
Therefore, EvenBetter has no first.
EvenBetter is empty.
So. We have m = index(r).
m is last in BestYet, so r > each rational indexed in {1,...,m-1}
EvenBetter is empty, so r > each rational indexed in {m+1,...,i-1}
r > each rational indexed in {1,...,i-1} other than itself.
Therefore,
r is the maximum rational indexed by one of {1,...,i-1}.
>>> For this sake ℵo natural numbers are required.
>>> More are not available.
>>
>> I claim that all the integers Z = {...,-3,-2,-1,0,1,2,3,...}
>> can be enumerated by the naturals N = {0,1,2,3,...}.
>
> This error is harder to detect, but it is not an excuse for
> the blatant mistake that I have discovered.
The bijection between Z and N is simpler than the bijection
between Q+ and N+. Why would simplicity and clarity
make this "blatant mistake" *harder* to detect?
I think that this is because your "detection process"
requires the generation of large quantities of fog
to obscure what you are saying.
This is why my counter-argument is mostly clarification of
what a natural is, what a rational is, what an enumeration is,
the significance of truth-preservation.
The points I make are intended to be simple points.
It is their simplicity which opposes your obfuscation.