This definition of cofinality relies on the axiom of choice, as it uses the fact that every non-empty set of cardinal numbers has a least member. The cofinality of a partially ordered set A can alternatively be defined as the least ordinal x such that there is a function from x to A with cofinal image. This second definition makes sense without the axiom of choice. If the axiom of choice is assumed, as will be the case in the rest of this article, then the two definitions are equivalent.
If A admits a totally ordered cofinal subset, then we can find a subset B which is well-ordered and cofinal in A. Any subset of B is also well-ordered. Two cofinal subsets of B with minimal cardinality (i.e. their cardinality is the cofinality of B) need not be order isomorphic (for example if , then both and viewed as subsets of B have the countable cardinality of the cofinality of B but are not order isomorphic.) But cofinal subsets of B with minimal order type will be order isomorphic.
The cofinality of an ordinal ? is the smallest ordinal ? which is the order type of a cofinal subset of ?. The cofinality of a set of ordinals or any other well-ordered set is the cofinality of the order type of that set.
Thus for a limit ordinal ?, there exists a ?-indexed strictly increasing sequence with limit ?. For example, the cofinality of ?² is ?, because the sequence ?·m (where m ranges over the natural numbers) tends to ?²; but, more generally, any countable limit ordinal has cofinality ?. An uncountable limit ordinal may have either cofinality ? as does ?? or an uncountable cofinality.
The cofinality of 0 is 0. The cofinality of any successor ordinal is 1. The cofinality of any nonzero limit ordinal is an infinite regular cardinal.
A regular ordinal is an ordinal which is equal to its cofinality. A singular ordinal is any ordinal which is not regular.
Every regular ordinal is the initial ordinal of a cardinal. Any limit of regular ordinals is a limit of initial ordinals and thus is also initial but need not be regular. Assuming the axiom of choice, is regular for each ?. In this case, the ordinals 0, 1, , , and are regular, whereas 2, 3, , and ??·2 are initial ordinals which are not regular.
The cofinality of any ordinal ? is a regular ordinal, i.e. the cofinality of the cofinality of ? is the same as the cofinality of ?. So the cofinality operation is idempotent.
If ? is an infinite cardinal number, then cf(?) is the least cardinal such that there is an unbounded function from cf(?) to ?; cf(?) is also the cardinality of the smallest set of strictly smaller cardinals whose sum is ?; more precisely
That the set above is nonempty comes from the fact that
i.e. the disjoint union of ? singleton sets. This implies immediately that cf(?)
Using König's theorem, one can prove ? < ?cf(?) and ? < cf(2?) for any infinite cardinal ?.
The last inequality implies that the cofinality of the cardinality of the continuum must be uncountable. On the other hand,
the ordinal number ? being the first infinite ordinal, so that the cofinality of is card(?) = . (In particular, is singular.) Therefore,
(Compare to the continuum hypothesis, which states .)
Generalizing this argument, one can prove that for a limit ordinal ?
On the other hand, if the axiom of choice holds, then for a successor or zero ordinal ?