« first day (531 days earlier)      last day (967 days later) » 

12:53 PM
@user21820 You may find this thing I came up with recently to be amusing (though it seems I wasn't the first to come up with it)

Let $i,k,n$ denote natural numbers and $F_0$ be the set of $\mathbb N\mapsto\mathbb N$ functions and $F_{k+1}$ be the set of $F_k\mapsto F_k$ functions. Let $f_i$ denote an element of $F_i$.

Define $\operatorname{Iterate}_k$ as an element of $F_k$ as follows:
Let $\operatorname{Iterate}_0=n\mapsto n+1$ be the successor function.
For $k\ne0$, let $\operatorname{Iterate}_k=f_{k-1}\mapsto f_{k-2}\mapsto\dots\mapsto f_1\mapsto f_0\mapsto n=f_{k-1}^n(f_{k-2})\dots(f_
$$\huge\overline{\underline{\color{red}\star TGIF\color{blue}\star}}$$
@amWhy
 
 
1 hour later…
2:04 PM
@SimplyBeautifulArt Shouldn't your mapping arrows be in reverse?
And I guess that's about the same as the finite levels of the FGH?
Anyway, good to hear from you after a 'long' while. =)
 
@SimplyBeautifulArt Yay!!!
 
@user21820 it ought to go to ε0
@user21820 I'm basically using this notation
 
@SimplyBeautifulArt No you appear to have the sequence backwards. Why are you mapping whatever to "n"?
It should be "n ↦ ...", shouldn't it?
@SimplyBeautifulArt And yea I missed that you didn't iterate the inner ones.
 
@user21820 no?
$\operatorname{Iterate}_k$ maps an element of $F_{k-1}$ to another element of $F_{k-1}$
 
Whatever it is, there seems to be a type error.
 
2:17 PM
Particularly it maps $f_{k-1}$ to $f_{k-2}\mapsto\dots$
I still don't see the issue :/
 
Hmm..
Iterate[1] = f[0] ↦ n according to your definition.
That seems to not be of the type you claimed.
 
No, it ought to be Iterate[1] = f[0] ↦ (n ↦ f[0]ⁿ(n))
n ↦ f[0]ⁿ(n) is an N ↦ N function.
 
Currently, as written, it does not make any sense. Perhaps you should rewrite what you actually mean...
@SimplyBeautifulArt In particular, you did not write this in your first message.
 
"For $k\ne0$, let $\operatorname{Iterate}_k=f_{k-1}\mapsto f_{k-2}\mapsto\dots\mapsto f_1\mapsto f_0\mapsto n=f_{k-1}^n(f_{k-2})\dots(f_1)(f_0)(n)$"
oh wait
mb, the last "=" should be "↦"
For $k\ne0$, let $\operatorname{Iterate}_k=f_{k-1}\mapsto f_{k-2}\mapsto\dots\mapsto f_1\mapsto f_0\mapsto n\mapsto f_{k-1}^n(f_{k-2})\dots(f_1)(f_0)(n)$
 
That... makes a lot of difference...
 
2:22 PM
mb
 
=P
 
@SimplyBeautifulArt Anyway yea I forgot a lot about the FGH already.
 
But it's probably better to just do the proof theory directly.
For each finite k you can prove totality of your function.
 
2:28 PM
Well this gives exact values for the normal FGH for ordinals less than ε0 :P
 
So clearly it's below ε0.
 
(supposedly I claim)
 
You just need k times of the induction, each time using a higher induction.
So PA will prove that each Iterate[k] is total for each k of the form 1+...+1.
But PA cannot prove ∀k∈nat ( Iterate[k] is total ).
 
 
2 hours later…
4:32 PM
 
@SimplyBeautifulArt: Just yesterday there was a related question:
0
A: $(\forall n, {\rm T} \vdash P(n)) \Rightarrow ({\rm T} \vdash \forall n, P(n))$ Why is it false?

user21820Actually there is a much simpler 'reason': $ \def\nn{\mathbb{N}} $ (1) "$\forall n \in \nn\ ( T \vdash P(n) )$" expands to "$\forall n \in \nn\ \exists q \in Strings\ ( \text{$q$ is a proof over $T$ of $P(n)$} )$". (2) "$T \vdash \forall n ( P(n) )$" expands to "$\exists q \in Strings\ (...

 

« first day (531 days earlier)      last day (967 days later) »