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

It is well known that the ordinary and exponential generating functions of a sequence of numbers are related by an integral transform (the Borel transformation).

Does there exist a combinatorial theory of integral transforms? The example above indicates that something might be going on "behind the scenes". Has anyone been able to formulate a precise combinatorial explanation of this phenomenon? If not this one, might other types of integral transforms have combinatorial interpretations?

Thanks for your interest!

share|cite|improve this question
I am not sure that I fully understand the question, but could the following book: G.P. Egorychev, Integral Representation and the Computation of the Combinatorial Sums, AMS, 1984 (Zbl 0453.05001) be relevant? There, the author goes in direction opposite to what you are seemingly interested in: starting from a combinatorial identity, to find its interpretation in terms of integral operators. – Pasha Zusmanovich Jun 20 '11 at 7:55
up vote 4 down vote accepted

I claim that there is a construction here similar to the one in "Lagrange inversion for species" by Gessel and Labelle, which can explain the general picture.

To every species $S:\mathcal{B}\to \mathcal{B}$ there should correspond a "labelled" version $L(S):\mathcal B\to \mathcal B$ whose exponential generating function is the ordinary generating function of $F$. The construction should start by looking at the span $$\mathcal B \leftarrow \mathcal B\times\mathcal B\rightarrow \mathcal B$$ with maps $p,q$ and introducing a "kernel" on $\mathcal B\times \mathcal B$ in a manner that $L$ can be defined from $$F(\mathcal B)\leftarrow F(\mathcal B\times \mathcal B)\rightarrow F(\mathcal B)$$ by a pull-push formalism.

This kernel $\kappa (X,Y)$ on $F(\mathcal B^2)$ should be a species of 2 sorts, but since labelling is the inverse of $\sum \frac{x^ny^n}{n!}$ it will actually be a virtual species (formal difference of species, so $\kappa$ is a categorification of $e^{-xy}$). Then one has to check that $$L=p _{\ast}(\kappa q^{\ast})$$ or something analogous to it holds. I.e. I suspect that there is a categorification of the Borel (also called Laplace-Carson?) transform which is a natural operation on species, and I believe that the appearances of these integral transforms arise from decategorifications of properties of spans of groupoids.

Unfortunately, I don't know where to find a treatment of this. I was looking a while back for a treatment of the (closely related) Laplace transform along similar lines but couldn't find any references. Since I can't work out details right now, perhaps someone who knows this can elaborate?

Also, there is a section on differential-integral equations with species in "Combinatorial species and tree-like structures" by F. Bergeron, G. Labelle and P. Leroux, which is worth looking at.

share|cite|improve this answer

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.