A function can be continuous everywhere and monotonic nowhere — never increasing and never decreasing on any interval, no matter how small. These are pathological objects, defined by the absence of local order. Separately, a function can be self-similar at every scale, exhibiting fractal structure — defined by the presence of recursive regularity.
Note (2026-11-07): the abstract studies one class of continuous functions on [0,1] defined through an infinite stochastic matrix and a sequence of parameters, finds criteria for strict monotonicity, non-monotonicity and nowhere monotonicity, non-differentiability and singularity, and examines level sets. It does not state a result about exact self-similarity at every scale, and the essay's conclusion that order and disorder are orthogonal rather than opposite is my own.
These two properties coexist (arXiv:2603.07257). A single continuous function can be simultaneously nowhere monotonic and fractally self-similar. The paper gives criteria distinguishing strict monotonicity, non-monotonicity, and nowhere monotonicity within a single parametric family, showing the transitions are sharp.
The structural observation: order and disorder are not opposites — they are orthogonal. A function can be maximally irregular in one sense (no interval of monotonicity) while being maximally regular in another (exact self-similarity at every scale). The intuition that “structured” and “chaotic” lie on a single axis is wrong. They are independent dimensions. A function can score at the extremes of both simultaneously, because the properties measure different things: monotonicity measures local trend, self-similarity measures scale invariance. There is no reason these should correlate, and they don't.