I am trying to fill in the details for part (ii) of Theorem 5.15 in Jech's Set Theory:

Theorem 5.15 If GCH holds and $\kappa$ and $\lambda$ are infinite cardinals, then

(i) If $\kappa \leq \lambda$, then $\kappa^\lambda = \lambda^+$.

(ii) If $\mathrm{cf}(\kappa) \leq \lambda < \kappa$, then $\kappa^\lambda = \kappa^+$.

(iii) If $\lambda < \mathrm{cf}(\kappa)$, then $\kappa^\lambda = \kappa$.

In the proof of (ii) he just states that it follows from these two lemmas:

Lemma 5.7 If $|A| = \kappa \geq \lambda$, then the set $[A]^\lambda$ has cardinality $\kappa^\lambda.$

(Here $[A]^\lambda = \{X \subset A : |X| = \lambda\}.$)

Lemma 5.8 If $\lambda$ is an infinite cardinal and $\kappa_i > 0$ for each $i < \lambda$, then

$$\sum_{i<\lambda}{\kappa_i} = \lambda\cdot\sup_{i<\lambda}{\kappa_i}.$$

I am struggling to construct an explicit proof using these results and a few facts about cardinal arithmetic such as absorption for infinite cardinals and cardinalities of power sets.

Since, assuming GCH, $\kappa^+ = 2^\kappa = |\mathcal{P}(A)|$ for some $A$ with cardinality $\kappa$, I thought I could come up with some sequence $\{\mu_i : i < \mathrm{cf}(\kappa)\}$ such that

$$\kappa^+ = 2^\kappa = |\mathcal{P}(A)| = \sum_{i < \mathrm{cf}(\kappa)} |[A]^{\mu_i}| = \sum_{i < \mathrm{cf}(\kappa)} \kappa^{\mu_i} = \mathrm{cf}(\kappa)\cdot\sup_{i<\mathrm{cf}(\kappa)}\kappa^{\mu_i} = \mathrm{cf}(\kappa)\cdot\kappa^\lambda = \kappa^\lambda$$

Is this approach a good idea? How would I go about finding the appropriate $\mu_i$? I suppose it will have to make use of the assumption that $\mathrm{cf}(\kappa) \leq \lambda < \kappa$, but I don't see how.

I'd much appreciate any hints on how to proceed (either with my suggestion or another way).


You fell into a trap that also caught me when I was just starting to read mathematical books. When you see a reference to (5.7) it does not mean Lemma 5.7, but rather the displayed formula whose tag is 5.7, which in this case is the formula:

$$\kappa\leq\kappa^\lambda\leq 2^\kappa.$$ Similarly, (5.8) refers to the inequality $$\kappa<\kappa^\lambda\quad\text{ if }\lambda\geq\operatorname{cf}\kappa.$$

And indeed these are the inequalities needed to prove (i) and (ii).

| cite | improve this answer | |
  • $\begingroup$ Oh! That makes a lot more of sense. Thanks! I think it was more confusing as the line immediately above cites Lemma 5.6, then it cites (5.7) and (5.8). $\endgroup$ – ryan221b Dec 4 '18 at 16:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.