Kunen's exercise on antitransitive relations

152 Views Asked by At

I am working in an exercise from Professor K. Kunen's 2011 book. Let me write some definitions.

Given a relarion $R$, one defines the relation $R^*$ as follows:

$$aR^*b \leftrightarrow \exists n\in \omega \exists s \in A^{n+1} (a=s(0)R\dots Rs(n)=b).$$

It is clear that $R^*$ is the least transitive relation contaning $R$.

A relation $R$ is antitransitive iff:

$$\forall xyz(xRy\wedge yRz\rightarrow x\not R z)$$

The exercise is:

a) Prove that if $R$ is well founded defined on a set $A$ and every element has finite rank, there exists a unique antitransitive relation $H$ such that $H^*=R^*$.

b) Show that there is a countable well order that is conter example for this if we remove the finite rank hipothesis.

I managed to prove the existence of a) and showed that the uniqueness statement is in fact false: Let $A=4$, $R=<$, $H=\{(0, 1), (1, 2), (2, 3)\}$ and $\tilde H=\{(0, 1), (1, 2), (2, 3), (0, 3)\}$. Is the uniqueness really false or am I missing something?

For $b)$ I am having trouble finding a coumterexample. It looks like no ordinal up to $\omega.\omega+\omega$ works. Any help is appreciated.

Edit: exercise I.9.47.