A Universal Fixed Point: The Cauchy Distribution as A Unifying Principle in Modern Probability Theory

This article re-examines the Cauchy distribution, a probability law famously regarded as "pathological" in classical probability theory due to its lack of a defined mean and higher-order moments. We argue that this perspective is incomplete and that the Cauchy distribution, rather than being an anomaly, occupies a unique and fundamental position as a universal fixed point that consistently reconciles four distinct and major frameworks of modern probability theory: classical (tensor), free, Boolean, and monotone. The central thesis is that the Cauchy distribution is not merely a "bridge" but a canonical, invariant object whose properties remain structurally unchanged across these seemingly disparate notions of independence.

Our investigation leverages the analytical structure of linearizing transforms associated with each convolution type, including the classical Fourier transform, the free R-transform, the Boolean K-transform, and the monotone reciprocal Cauchy transform. We demonstrate that for the Cauchy distribution, these transforms take on exceptionally simple forms—either constant or affine—a property that serves as the root cause of its universal stability. This analytical simplicity leads to a central finding: the Cauchy family is strictly 1-stable and closed under all four convolution types with an identical parameter addition law, a consistency unmatched by any other known probability distribution.

Building on this, we establish a series of universal convergence theorems that position the Cauchy distribution as a canonical attractor for a vast class of heavy-tailed distributions. This challenges the Gaussian distribution's traditional role as the sole central limit, advocating for a paradigm shift in how we model phenomena where classical independence assumptions fail. The paper's most profound contribution is the rigorous extension of the Bercovici-Pata bijection to include Boolean and monotone frameworks, proving that the Cauchy distribution is the unique fixed point of these maps. This fixed-point property confirms the distribution's invariant nature, demonstrating that its essential structure is preserved across these four probabilistic "languages," thereby revealing a deep intellectual unity underlying the field. We further prove that all of these scalar results extend verbatim to the operator-valued (amalgamated) setting, providing a direct link to quantum probability theory and random matrix theory.

The work provides a unified framework for the analysis of probability measures that lack finite moments, a class of laws increasingly relevant to a range of applications, from financial mathematics to network theory and quantum physics. By establishing the Cauchy distribution as a fundamental organizing principle, our findings offer a new perspective that transcends classical probability's limitations and provides a foundation for the study of complex systems in the twenty-first century.