1
$\begingroup$

This is probably an incredibly stupid question, but I'm a novice to representation theory and finite field theory, as I've just been introduced to these concepts, so all I really need is confirmation of the following:

Let $S_n$ be the symmetric group of $n$ symbols and let $F=\mathbb{F}_q$ be a finite field of $q$ elements and characteristic $p>n$. Then $F$ is a splitting field for $S_n$ (that is to say every irreducible $FS_n$-module is absolutely irreducible). Furthermore, every irreducible representation of $S_n$ over $F$ is also irreducible over $\mathbb{C}$.

This seems almost tautological to me, but it would be nice to have confirmation just to make sure I haven't misunderstood thinks spectacularly.

As a further (potentially stupid) question, are these irreducible representations over $F$ the 'same' as those over $\mathbb{C}$, albeit with reduction modulo $p$. Do they have the same degree for instance? Again from the definitions, this would appear to me to be an emphatic yes, but again it would be nice to have confirmation.

Thank you.

$\endgroup$
4
  • 2
    $\begingroup$ The question as stated does nor make sense. How can can a reprentation over $\mathbb C$ be "irreducible over $F$"? What does make sense is something like: every representation $V$ of a symmetric group over $\mathbb C$ is in fact defined over $\mathbb Z$ (this is a rather non-trivial fact!), and then one can reduce it modulo $p$ to obtain a representation $V'$ over $F_p$, and then one can ask if $V'$ is irred. if $V$ is. $\endgroup$ May 11, 2013 at 23:10
  • $\begingroup$ After your edit: You interchanged the roles of $F$ and $\mathbb C$ but still it does not make sense: a representation of $S_n$ over a finite field is not, in any way, a representation of that group over $\mathbb C$. $\endgroup$ May 11, 2013 at 23:16
  • $\begingroup$ Oh dear, maybe I have misunderstood things. Or at least I've phrased things very poorly. The number of irreducible representations of a symmetric group corresponds to the number of distinct partitions of $n$, so there are 5 irreducible representations of $S_4$ over $\mathbb{C}$, for example. I know for a fact that there are also $5$ irreducible representations of $S_4$ over the field $F$. Is there a natural 1-1 correspondence between these representations? Do they have the same 'form', so-to-speak? Furthermore, does this hold true for all $n$? I hope this clears things up. $\endgroup$
    – Damien
    May 11, 2013 at 23:31
  • 1
    $\begingroup$ Regardless of the mess in your statements, I sort of understand your problem. The thing you have in mind is called the Specht module, which can be defined over $\mathbb{Z}$, parameterised by partitions of $n$ (bijective to ccl's of $S_n$). These modules are irreducible when extends to $\mathbb{C}$, but most of them are not irreducible anymore when extend to the modular splitting fields (when char$F$ divides $n!$). But they still have a nice property: if you quotient out this module by its radical, it is either irreducible or zero. $\endgroup$
    – Aaron
    May 12, 2013 at 14:16

1 Answer 1

6
$\begingroup$

One can show that for every $\def\bb{\mathbb}\bb CS_n$-module $V$ there is a $\bb ZS_n$-module $V'$ which is free as an abelian group and of rank equal to the dimension of $V$ as a vectr space, such that $V'\otimes_{\bb Z}\bb C\cong V$ as $\bb CS_n$-modules. This is a rather special property of $S_n$.

Now, if we start with $V$, construct $V'$ as above, then we can also construct $V''=V'\otimes_{\bb Z}\bb F_p$, and this is a representation of $\bb F_pS_n$. The passage from $V$ to $V''$ involves a choice, that of $V'$. When one understands what it entails, one can wonder if the mapping $V \leadsto V''$ gives a bijection between irreps of $\bb CS_n$ an irreps of $\bb F_pS_n$. It does, as long as $p>n$.

Other finite groups which have this property that its complex irreps are defined over $\bb Z$ are the Weyl groups (of which $S_n$ is an example) and surely others. For a general group $G$, a complex representation $V$ is always defined over the ring $\mathcal O$ of integers of some splitting field of $G$ (one can show that there is a splitting field which is cyclotomic) and then one can "reduce" $V$ modulo prime ideals of $\mathcal O$. The theory is extremely rich, complicated and deep. You can find this explain in the classic book by Curtis and Reiner, for example. The litle book of Serre has lots of information on the problem of rationality of representations, that is, of trying to restrict the field over which a representation is defined, but here we want considerably more; we want to restrict to a ring whose field of quotients is a field of definition of the representation. Already for cyclic groups this is a non-obvious problem.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .