Central limit theorem for random partitions under the Plancherel measure

Abstract

In this work, we obtain the central limit theorem for fluctuations of Young diagrams around their limit shape in the bulk of the "spectrum" of partitions of a large integer n (under the Plancherel measure). More specifically, we show that, under the suitable normalization (growing as the square root of log n), the corresponding random process converges, in the sense of finite dimensional distributions, to a Gaussian process with independent values. The proof uses heavily the determinantal structure of the correlation functions and is based on the application of the Costin-Lebowitz-Soshnikov central limit theorem. At the spectrum edges, the fluctuation asymptotics is expressed in terms of the largest members of the Airy ensemble; in particular, at the upper edge the limit distribution appears to be discrete (without any normalization). These results admit an elegant symmetric reformulation under the rotation of Young diagrams by 45 degrees, where the normalization no longer depends on the location of the spectrum point. We also discuss the link of our central limit theorem with an earlier result by S.V. Kerov on the convergence to a generalized Gaussian process.