1
$\begingroup$

I have only seen the notation $f : A \rightarrow B$ used for discussing functions. Can this notation be used for relations that are not functions?

For example, $\tan : \mathbb R \rightarrow \mathbb R$ appears to define a relation that is not left-total (for example $\frac \pi 2\in \mathbb R$ is not mapped to anything in codomain $\mathbb R$) and hence not a function. Would it be incorrect to express this relation, that is not a function, using this notation?

Edit: To borrow a great example from Arthur's answer below, consider the notation $\le\; : \mathbb R \rightarrow \mathbb R$. This seems to me to unambiguously express a relation that, while left-total is certainly not functional (e.g. $(3, 3), (3, 4) \in \;\le$) and hence not a function. If it's unambiguous, surely this notation can be used to express this relation that is not a function...

I guess I could express my question more clearly as: is there anything in the expression $f:A\rightarrow B$ that asserts that $f$ is not merely a relation but specifically a function?

$\endgroup$
4
$\begingroup$

There is a category of relations having sets as objects and binary relations as arrows.

In that context the notation: $$R:A\to B$$ where $R\subseteq A\times B$, is fine.


In my view (somewhat coloured by categories) the expression $f:A\to B$ on its own is not enough to state that we are dealing with a function (or a relation).

I do recognize it as an arrow in a category and the context must make clear which category.

So my answer to your last question is: "no".

$\endgroup$
0
2
$\begingroup$

An entirely correct way would be $\tan\subseteq \Bbb R\times\Bbb R$. I think $\tan: \Bbb R\to\Bbb R$ would be considered technically incorrect but in many circumstances within allowable abuse of notation, mostly because $\tan$ is still a partial function defined on most of the given domain. Writing something like $\leq{}:\Bbb R\to\Bbb R$ is probably taking it too far.

$\endgroup$
3
  • 1
    $\begingroup$ Rather then "function-like" would "Partial function" be more useful? $\endgroup$ Aug 24 '18 at 7:21
  • $\begingroup$ @QthePlatypus You're right. I fixed it up. $\endgroup$
    – Arthur
    Aug 24 '18 at 7:22
  • $\begingroup$ Hmm. I think the $\tan$ you are referring to (when you say it is a subset of $\mathbb R \times \mathbb R$) is not just function-like, but a function in all senses, because $\frac \pi 2$, for example, is (usually implicitly) not part of the domain. In my example I explicitly said that the domain was the entire $\mathbb R$... My example is not the relation we generally understand by $\tan$ stated on its own. $\endgroup$ Aug 24 '18 at 7:23

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.