24

I've stumbled upon the definition of exact sequence, particularly on Wikipedia, and noted the use of $\hookrightarrow$ to denote a monomorphism and $\twoheadrightarrow$ for epimorphisms.

I was wondering whether this notation was widely used, or if it is common to define a morphism in the general form and indicate its characteristics explicitly (e.g. "an epimorphism $f \colon X \to Y$").

Also, if epimorphisms and monomorphisms have their own special arrows, are isomorphisms notated by a special symbol as well, maybe a juxtaposition of $\hookrightarrow$ and $\twoheadrightarrow$?

Finally, are there other kinds of morphisms (or more generally, relations) that are usually notated by different arrows depending on the type of morphism, particularly in the context of category theory?

Thanks.

t.b.
  • 78,116
Abel
  • 1,792
  • 7
    Note that there are categories where an arrow that is both monic and epic may fail to be an isomorphism. For example, the map $x \to \exp 2 \pi x i$, $[0, 1) \to S^1$ is both monic and epic, but not an isomorphism in the category of topological spaces and continuous functions. – Zhen Lin Feb 02 '11 at 14:15
  • 2
    You're right. I probably should have said "bimorphisms", which I understand is common terminology for monic and epic morphisms. – Abel Feb 02 '11 at 15:56
  • 1
    This isn't obviously categorical (though I think it can be phrased as such) but often immersions of manifolds are notated with the same inclusion hook arrow except that the hook part comes down and crosses the main horizontal shaft. – Aaron Mazel-Gee Feb 02 '11 at 17:41

4 Answers4

24

Some people use those notations, some don't. Using $\hookrightarrow$ to mean that the map is a mono is not a great idea, in my opinion, and I much prefer $\rightarrowtail$ and use the former only to denote inclusions. Even when using that notation, I would say things like «consider the epimorphism $f:X\twoheadrightarrow Y$».

In some contexts (for example, when dealing with exact categories) one uses $\rightarrowtail$ and $\twoheadrightarrow$ to denote that the map is not only a mono or an epi, but that it has certain special properties (for example, that it is a split mono, a cofibration, or what not)

Denoting isomorphisms by mixing $\twoheadrightarrow$ and $\rightarrowtail$ is something I don't recall seeing.

You will find that there are no rules on notation, and that everyone uses pretty much whatever they like---the only important thing is that when you use what you like you make it clear to the reader.

Mariano Suárez-Álvarez
  • 135,076
  • Could your native language be significant? I noticed you used angular quotation marks. Ethnicity (not race) influences mathematical notation; for example I know Britons like $u$ for sequences while Americans prefer $a$, Britons like $\boldsymbol v$ for vectors while Americans prefer $\mathbf v$, $\mathrm d x$ over $dx$, etc. – gen-ℤ ready to perish Jul 27 '19 at 05:20
  • @gen-zreadytoperish or in Britain, I think, people are taught $y=mx+b$ whereas in Australia, people are taught $y=mx+c$. I just use $y=mx+n$. I suppose employing whatever kind of notation suits you best, as long as your audience knows what you mean when reading your work or papers :) – Mr Pie Mar 10 '20 at 09:08
15

I agree with everything Mariano says, especially about $\hookrightarrow$ and $\rightarrowtail$ only being for inclusions. I wanted to add that I usually denote an isomorphism by $\xrightarrow{\sim}$, which kind of evokes $\simeq$ and $\cong$. Another type of "morphism" that often gets a special arrow is a natural transformation, namely $\alpha:F\Rightarrow G$ for a natural transformation between the functors $F,G:\mathcal{A}\rightarrow\mathcal{B}$ (see here).

Zev Chonoles
  • 129,973
9

It may be worth noting that these arrows can mean different things in different contexts. For example, in model categories, one often uses either $\hookrightarrow$ or $\rightarrowtail$ to indicate that a map is a cofibration, and $\twoheadrightarrow$ to indicate a fibration. If a map is a weak eqivalence, we denote this by placing a $\sim$ over the relevant arrow.

This could potentially lead to ambiguity. For example, in the category of topological spaces, the standard model structure declares fibrations to be Serre fibrations. However, Serre fibrations need not be surjective (that is, an epimorphism in $\mathbf{Top}$). In many cases, one of the pairs cofibration/monomorphism or fibration/epimorphism will coincide, and in all cases, a weak equivalence will become an isomorphism in the homotopy category, so there is generally no confusion.

SL2
  • 4,321
3

There is a Unicode character 2916 (⤖) for "bijective mapping". My algebraic topology instructor, educated in Canada and New York, wasn't familiar with it when I employed it.

Mike
  • 223
  • 1
  • 11