Power-law distributions based on exponential distributions: Latent scaling, spurious Zipf's law, and fractal rabbits

Abstract: The different between the inverse power function and the negative exponential function is significant. The former suggests a complex distribution, while the latter indicates a simple distribution. However, the association of the power-law distribution with the exponential distribution has been seldom researched. Using mathematical derivation and numerical experiments, I reveal that a power-law distribution can be created through averaging an exponential distribution. For the distributions defined in a 1-dimension space, the scaling exponent is 1; while for those defined in a 2-dimension space, the scaling exponent is 2. The findings of this study are as follows. First, the exponential distributions suggest a hidden scaling, but the scaling exponents suggest a Euclidean dimension. Second, special power-law distributions can be derived from exponential distributions, but they differ from the typical power-law distribution. Third, it is the real power-law distribution that can be related with fractal dimension. This study discloses the inherent relationship between simplicity and complexity. In practice, maybe the result presented in this paper can be employed to distinguish the real power laws from spurious power laws (e.g., the fake Zipf distribution).