Id: arxiv/2601.19614 Url: https://arxiv.org/abs/2601.19614 Tags: math.PR, math-ph Score: 0 Title: Derivatives of Gaussian multiplicative chaos Domain: math.PR Source: arxiv Authors: Antoine Jego Reviews: 0 Subject: math.PR Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-01-27 Page: https://delphi.science/r/arxiv/2601.19614 Consider a logarithmically-correlated Gaussian field $X$ in $d$ dimensions. For all $γ\in (-\sqrt{2d},\sqrt{2d})$, we show that the derivatives $\frac{\partial^k}{\partialγ^k} :e^{γX_ε}:$ of the regularised Gaussian multiplicative chaos $:e^{γX_ε}:$ converge as $ε\to 0$. By deriving optimal bounds on their growth as $k\to\infty$, we control the power expansion of $:e^{γX_ε}:$ about each $γ\in(-\sqrt{2d},\sqrt{2d})$. This yields an alternative approach to complex Gaussian multiplicative chaos in the whole subcritical regime, based entirely on real-valued quantities. One of our key technical…