MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Here is a power series, which looks a bit like a Hypergeometric function series, but I don't think that it is. Has anyone any idea what it is? Here $n,p,r$ are integers with $n\ge 0$ and $p\ge r\ge 0$:

$$ f_{n,p,r}(x)\,=\,\sum_{s=0}^p \frac{x^s}{s!}\ \frac{p!\,(2n+p+s+2)!\,(n+r+s+2)!}{(p-s)!\,(2n+r+s+3)!\,(n+s+2)!} $$

Originally this occurred as a $q$-factorial series, but if the $q=1$ case given here could be recognised, it would be a big help.

Thanks for the answer below. A bit before I declare this answered (if I remember how to do that!) I would like to sneak in another question. Is there any sensible way to sum this in the case $x=-1$? The answer is likely zero if $r<p$ and reasonably simple if $r=p$, but how to prove that?

share|cite|improve this question
Isn't it hypergeometric? Maybe I'm messed up, but it looks like a multiple of ${}_3F_3(-p, n+r+3, 2n+p+3; n+3, n+r+2, 2n+r+4; -x)$. (This is from taking the ratio of the $s+1$ and $s$ terms and factoring.) Where did this function come from? It seems plausible that one could say more than just that it's hypergeometric, but I'm not sure what. – Henry Cohn Nov 6 '13 at 13:01
Thanks, that is a more parameter version of the Hypergeometric than I am used to! It comes from looking at the eigenvalues of the Hodge Laplacian in the 1-forms of the noncommutative sphere (guess q=1 is just commutative sphere). The sum, evaluated at x=-1, is used in finding the normalisation of the eigen-1-forms in the Hodge inner product. I am hoping that knowing the classical analogue will help in the quantum case. – Edwin Beggs Nov 6 '13 at 13:40
up vote 7 down vote accepted

Henry is almost right; the sum can be written as $$\frac { \left( 2n+p+2 \right) !\, \left( n+r+2 \right)!} { \left( 2n+r+3 \right) !\, \left( n+2 \right) !} {}_{3}F_{2}\left({{2n+p+3,n+r+3,-p\,\,\,}\atop {2n+r+4,n+3}}\Bigm |-x\right)$$ For $x=-1$ (no need to assume $r\le p$) the sum can be evaluated by Saalschütz's theorem (also called the Pfaff-Saalschütz theorem), giving $$ \left( -1 \right) ^{p}\frac { \left( 2n+p+2 \right) !\, \left( n+r +2 \right) !\,r!\, \left( n+p \right) !}{ \left( 2+n+p \right) !\, \left( 3+2n+r+p \right) !\, \left( r-p \right) !\,n!}. $$

Saalschütz's theorem is probably the most commonly used hypergeometric summation formula after Vandermonde's theorem (and it has a $q$-analogue). Zeilberger's algorithm can be used to prove any hypergeometric summation formula, but it won't tell you what it is. (Though Saalschütz's theorem can be found in the book $A=B$ in section 3.5, and section 3.6 has some information on how to identify hypergeometric summation formulas.)

share|cite|improve this answer

According to Mathematica, for $r=p$, the sum for $x=-1$ gives $$ \frac{(n+1) \left(-\frac{1}{4}\right)^p \left(\frac{1}{2} (2 n+3)\right)! p! (2 n+p+2)!}{(2 n+3)! (n+p+1) \left(\frac{1}{2} (2 n+2 p+3)\right)!} $$ This formula can can be proven using Zeilberger’s algorithm [D. Zeilberger, The method of creative telescoping, J. Symbolic Comput. 11 (1991) 195–204], implemented as a Mathematica package [P. Paule, M. Schorn, A Mathematica version of Zeilberger’s algorithm for proving binomial coefficient identities, J. Symbolic Comput. 20 (5–6) (1995) 673–698], which establishes recurrence relations for $f_{n,p,p}$.

share|cite|improve this answer
The canonical reference for this sort of thing is Petkovsek-Zeilberger A=B. – Igor Rivin Nov 6 '13 at 15:13
Thanks! I have been comprehensively answered on all points. Thanks to all who contributed! – Edwin Beggs Nov 6 '13 at 16:12
Don't forget that Wilf is also a co-author of A=B. – Ira Gessel Nov 7 '13 at 13:31

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.