0

If there exists no injection $X \rightarrow Y$ does that mean that there exists a surjection $Y \rightarrow X$ ?

If there exists no surjection $X \rightarrow Y$ does that mean that there exists an injection $Y \rightarrow X$ ?

In other words, are two sets always comparable by size ?

  • 1
    I want to add more duplicates, but apparently I've reached the limit... so you're going to have to search for them yourself. – Asaf Karagila Jan 18 '18 at 21:24

1 Answers1

6

EDIT: you've made a mistake in how you phrase comparability of sets. What you've written is false: take e.g. $X=\{1,2\}, Y=\{1\}$ for the first version and $X=\{1\}, Y=\{1, 2\}$ for the second version.

The right statement of comparability is:

For any sets $X, Y$, either there is an injection from $X$ to $Y$ or there is an injection from $Y$ to $X$.

And (at least, in the presence of choice), we can also use surjections instead: for any sets $X, Y$, there is an injection from $X$ to $Y$ iff there is surjection from $Y$ to $X$. This means we can equivalently restate comparability as:

For any sets $X, Y$, either there is a surjection from $X$ to $Y$ or there is a surjection from $Y$ to $X$.

But in general, in the absence of choice we can't alternate between injections and surjections this way: the axiom of choice is in fact equivalent to the statement "for any sets $X, Y$, there is an injection from $X$ to $Y$ iff there is surjection from $Y$ to $X$."


This is exactly what the axiom of choice is all about! (Or at least, it's one of the things; AC is prolific.)

The comparability of all sets with respect to cardinality is a consequence of the axiom of choice. In fact, it's equivalent to it (over the usual axioms for set theory without choice, ZF): ZF proves that the axiom of choice holds iff cardinalities are always comparable. Meanwhile, there are reasonable set theories (like ZF+AD) in which the answer is "no."

In general, choice is assumed unless otherwise stated - so the context-free answer to your question would be "yes."


There are a couple different ways to prove trichotomy from choice:

  • Use Zorn's lemma, taking as the relevant poset the set of all partial injections from one set to the other, and think about what a maximal element is.

  • Use the well-ordering theorem, and show that any two well-orderings are comparable.

To prove choice from trichotomy is actually quite simple, and uses a neat fact: that for any set $X$, there is a well-orderable set $Y$ which does not inject into $X$. (This is Hartogs' theorem.) Now note that trichotomy then implies that every set injects into a well-orderable set, hence is well-orderable (why?), so we have the well-ordering theorem; and proving choice from the well-ordering theorem is easy.

Noah Schweber
  • 245,398