11
$\begingroup$

Consider the $n$-element subsets $\{a_1<a_2<\cdots <a_n\}$ of $\{1,\ldots ,2n\}$ satisfying $a_i\geq 2i$ for all $i=1,\ldots ,n$. The number of such subsets is given by $${2n\choose n}-{2n\choose n-1}=\frac{1}{n+1}{2n\choose n},$$ which is the $n$th Catalan Number.

I want to know if the $q$-Catalan number $$\frac{q^{n}}{[n+1]_q}{2n\choose n}_q={2n\choose n}_q-{2n\choose n-1}_q$$ counts some kind of special $n$-dimensional subspaces inside $\mathbb{F}_q^{2n}$? Note that ${2n \choose n}_q$ is the total number of $n$-dimensional subspaces of $\mathbb{F}_q^{2n}$ ($\mathbb{F}_q$ denotes finite field of order $q$).

$\endgroup$
5
  • $\begingroup$ A generalized question has been asked (originally) by Shapiro, and still waiting for answers. See here starting with page 16: users.math.msu.edu/users/sagan/Papers/Old/gfp.pdf $\endgroup$ Commented Feb 8, 2017 at 19:27
  • $\begingroup$ Incidentally, I think that the RHS of your identity should be $\binom{2n}{n}_q-q\binom{2n}{n-1}_q$. $\endgroup$ Commented Feb 8, 2017 at 20:35
  • 1
    $\begingroup$ @VladimirDotsenko Oops, thanks for pointing it out. Actually I want to keep the RHS as it is. I hope I have fixed it. $\endgroup$ Commented Feb 9, 2017 at 0:17
  • 1
    $\begingroup$ I have made a small edit: added the factor $q^n$ in the previous q-analog. Sorry for the typo before $\endgroup$ Commented Feb 9, 2017 at 0:27
  • $\begingroup$ Well, there are at least three meaningful q-analogs of the Catalan Numbers, and you could ask the same question for each of them. See Johann Cigler's answer to an old question of mine: mathoverflow.net/questions/89996/… $\endgroup$
    – Wolfgang
    Commented Feb 12, 2017 at 21:16

2 Answers 2

5
$\begingroup$

An answer to your question was given in ``Rank Polynomials" by Brandt, Dipper, James, and Lyle, published in Proc. London Math. Soc. (3) 98 (2009), 1-18. A special case of Theorem 2.6 in that paper answers your question.

$\endgroup$
1
  • $\begingroup$ This is great, thank you for the reference. $\endgroup$ Commented Feb 14, 2017 at 12:50
0
$\begingroup$

Wikipedia has that

$q$-binomial $ \binom{n}{k}_q$ counts the subspaces of dimension $k$ in the vector space $\mathbb{F}_q^n$

As a non-expert I would hope that with an appropriate group action I could find objects enumerated by

$$ \frac{q^n}{[n+1]_q} \binom{2n}{n}_q $$

and I would look through Richard Stanley's Enumerative Combinatorics and hope for the best.


There also seems to be a result from 2010

where $q$-Catalan numbers are shown to enumerate invariants related to the Hilbert scheme of $n$ points on $\mathbb{C}^2$ (which respect to an equivariant torus action).

There are clearly some vector spaces related to the $q$-catalan numbers Gorsky has $q,t$ Catalan numbers and just set $t = 1$ ( I am looking for the correct degeneration). I do not underestand why $(q,t)$-enumeration is so trendy, as one could have arbitrary deformations.

My double-use of the letter $q$ is a bit suspect since $q$ could be:

  • a prime number
  • a unit complex number

and this ambiguity persists in the finite fields / modular forms literature.

$\endgroup$
3
  • $\begingroup$ (Maybe main) reasons for the q,t-Catalan enumeration being so trendy are that it plays an important role in the (1) study of the diagonal coinvariants of the symmetric group with beautiful combinatorics and deep connections to symmetric function theory, commutative algebra, and Hilbert schemes of points in the plain, and ... $\endgroup$ Commented Feb 9, 2017 at 8:32
  • $\begingroup$ ... (2, closely related to (1)) representation theory of Cherednik algebras (which come with a natural bigrading from $V[ \mathbb{C} \oplus \mathbb{C}^* ]$ (here, one also sees that q,t-enumeration is special for the action of reflection groups on $\mathbb{C} \oplus \mathbb{C}^*$) and its connections to things like knot theory (see your link above). $\endgroup$ Commented Feb 9, 2017 at 8:32
  • $\begingroup$ Setting $t=1$ in the $q,t$-Catalan numbers gives different kind of $q$-Catalan numbers. $\endgroup$ Commented Feb 9, 2017 at 13:30

You must log in to answer this question.

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