Content deleted Content added
→Buchholz's ψ: forgot the clause adding the Cantor normal form terms |
→Predicative start: notation inconsistency, phi vs. varphi for veblen functions |
||
Line 36:
Similarly, <math>C(1)</math> contains the ordinals which can be formed from <math>0</math>, <math>1</math>, <math>\omega</math>, <math>\Omega</math> and this time also <math>\varepsilon_0</math>, using addition, multiplication and exponentiation. This contains all the ordinals up to <math>\varepsilon_1</math> but not the latter, so <math>\psi(1) = \varepsilon_1</math>. In this manner, we prove that <math>\psi(\alpha) = \varepsilon_\alpha</math> inductively on <math>\alpha</math>: the proof works, however, only as long as <math>\alpha<\varepsilon_\alpha</math>. We therefore have:
:<math>\psi(\alpha) = \varepsilon_\alpha = \
(Here, the <math>\varphi</math> functions are the [[Veblen function]]s defined starting with <math>\
Now <math>\psi(\zeta_0) = \zeta_0</math> but <math>\psi(\zeta_0+1)</math> is no larger, since <math>\zeta_0</math> cannot be constructed using finite applications of <math>\varphi_1\colon \alpha\mapsto\varepsilon_\alpha</math> and thus never belongs to a <math>C(\alpha)</math> set for <math>\alpha\leq\Omega</math>, and the function <math>\psi</math> remains "stuck" at <math>\zeta_0</math> for some time:
| |||