Further to Blog T2, I conceived third thought experiment.
Let's assume there exists a Set Theory System (STS) that is made of ω layers as described below, where y is a positive integer while a, b, c and z are non-negative integers:-
- Base layer, STS(ω) (where STS(0) = ZFC): Ca(n) = C0,a(n) = C0,0,a(n) = ... (this was defined in Blog T2)
- First layer, STS(ω,ω): Cb,a(n) = C0,b,a(n) = C0,0,b,a(n) = ...
- Second layer, STS(ω,ω,ω): Cc,b,a(n) = C0,c,b,a(n) = C0,0,c,b,a(n) = ...
- (y-1)th layer, STS(ω◄ω): Cz◄y(n) = Cz,z,...,z,z(n), with y of z's.
In short, all lower layers are subsets of immediate upper layer; in other words, any layer is much more powerful and more robust than all layers below it:
- {STS(ω)} = {STS(0), STS(1), STS(2), ...}
- {STS(ω,ω)} =
- {STS(0,0), STS(0,1), STS(0,2), ...} = {STS(ω)}; plus
- {STS(1,0), STS(1,1), STS(1,2), ...}; plus
- {STS(2,0), STS(2,1), STS(2,2), ...}; plus
- {STS(3,ω)}, {STS(4,ω)} and so on.
- This shall continue for STS for 3 and more ω's.
- STS(0) is the ω-order set theory system that defines Rayo's number (which is at first order); I assumed it is ZFC plus extensions of ZFC.
- STS(1) is next level ω-order set theoty system that includes STS(0).
- This definition shall continue for the whole STS(ω◄ω) universe.
To put these into perspective: (≍ denotes 'is similar to')
- STS(0) to STS(1) ≍ STS(1) to STS(2) ≍ ... ≍ STS(1,0) to STS(1,1) ≍ STS(1,1) to STS(1,2) ≍ ... ≍ PA to ZFC;
- STS(ω) to STS(1,0) ≍ STS(1,ω) to STS(2,0) ≍ ... ≍ STS(ω,ω) to STS(1,0,0) ≍ ... ≍ STS(ω,ω.ω) to STS(1,0,0,0) ≍ ... ≍ Robinson Arithmatic to ZFC.
Growth rate comparison:
C0(n) << C1(n) << ... << Cn(n) <<< C1,0(n) << C1,1(n) << ... << Cn,n(n) <<< C1,0,0(n) << C1,0,1(n) ...
Since this is in uncomputable set theory realm, it's inappropriate to compare it with recursive computable realm, but still it resembles {φ(1,0), φ(1,0,0), φ(1,0,0,0), ...} of SVO, e.g. a sequence of functions can be defined as {C1,0(n), C1,0,0(n), C1,0,0,0(n), ...}. Hence I call it SSTS (Small Set Theory System).
Assuming both are well defined, I am confident that Cn◄n(n) is comparable to Bowers' n-System (that defines Utter Oblivion) as descibed by Ytosk.