A rational function is called positive if all its Taylor coefficients are positive. Friedrichs-Lewy conjecture states the positivity of the rational function \begin{eqnarray*}\frac{1}{ (1-x)(1- y)+(1- y)(1-z)+(1-z)(1-x)} = \sum\limits_{ k,m,n\ge0} a_{k,m, n }x^k y^mz^n. \end{eqnarray*} The conjecture was first proved by G. Szego.

Let $P_n=\prod\limits_{i=1}^n(1-x_i)$, is it true that the following rational function is positive \begin{eqnarray*}\frac{1}{ \sum\limits_{i=1}^n\frac{P_n}{1-x_i}}=\sum\limits_{i_1,i_2\cdots, i_n\ge 0}a_{i_1,i_2\cdots, i_n }x_1^{i_1}x_2^{i_2}\cdots x_n^{i_n}\quad? \end{eqnarray*}

A reference:

Armin Straub, Positivity of Szegö's rational function, Advances in Applied Mathematics 41 Issue 2 (2008) pp 255–264, doi:10.1016/j.aam.2007.10.001, (Wayback Machine pdf)


1 Answer 1


The answer is yes.

This was already proved in Gabor Szegö's original paper from 1933:

G. Szegö, Über gewisse Potenzreihen mit lauter positiven Koeffizienten, Mathematische Zeitschrift, Volume 37, Number 1, 674-688, DOI: 10.1007/BF01474608

The result can be found in Paragraph 3 "Verallgemeinerungen". However, apparently the simplified proof method due to Armin Straub which you mention applies also to this generalization.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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