Could you verify my proof by transfinite induction of: $\langle a_\xi : \xi< \alpha\rangle =\langle b_\xi : \xi< \alpha\rangle$?

56 Views Asked by At

I'm am trying to comprehend the proof on page 22 of Thomas Jech's Set theory and work out my first ever proof by transfinite induction. The problem is this:

Le $G$ be a function (on $V$), then there can only be one $\alpha$-sequence satisfying $$(\forall\xi <\alpha)\; a_\xi = G\big(\langle a_\eta : \eta<\xi\rangle \big ) \tag{$\star$}\label{moi}.$$

(I actually have no idea what $V$ is supposed to be, here.)

The book says that if $\langle a_\xi : \xi< \alpha\rangle$ and $\langle b_\xi : \xi< \alpha\rangle$ are two $\alpha$-sequences satisfying \eqref{moi}, then one shows $a_\xi = b_\xi$ by induction on $\xi$.

Now for a basis I say that $\langle a_\eta:\eta<0\rangle=\emptyset=\langle b_\eta:\eta<0\rangle$ and so $a_0=G(\emptyset )= b_0$.
Question: Is this valid? I don't know wether or not $\emptyset$ is in the domain of $G$. Especially sice $G$ is so generally (vaguely) defined.

Now if we assume that $a_\xi=b_\xi$ for all $\xi\leq \zeta$, for some $\zeta$, then $$ \langle a_\eta :\eta <\zeta+1\rangle=\langle a_\eta:\eta\leq \zeta\rangle=\langle b_\eta:\eta\leq \zeta \rangle=\langle b_\eta: \eta< \zeta+1\rangle\implies a_{\zeta+1}=b_{\zeta+1}.$$ Question: Is this correct, at all?

Lastly, if we have $a_\xi=b_\xi $ for all $\xi<\zeta$, for some $\zeta$$$\langle a_\eta:\eta<\zeta\rangle=\langle b_\eta:\eta<\zeta\rangle\implies a_\zeta=b_\zeta.$$

Is this a valid proof?

1

There are 1 best solutions below

0
On BEST ANSWER

As Asaf said in the comments, $V$ is the class of all sets, so in particular $\varnothing$ is in the domain of $G$. Your argument is correct, but you don’t actually need to split it into an initial case, a successor case, and a limit case. Suppose that $\eta<\alpha$, and $a_\xi=b_\xi$ for each $\xi<\eta$. Then

$$a_\eta=G\big(\langle a_\xi:\xi<\eta\rangle\big)=G\big(\langle b_\xi:\xi<\eta\rangle\big)=b_\eta\;.\tag{1}$$

The hypothesis that $a_\xi=b_\xi$ for each $\xi<\eta$ is vacuously true for $\eta=0$, so the initial case comes for free, so to speak, and the calculation $(1)$ is valid for both limit and successor $\eta$.