Terminology Question: non-Arrhenius

Useful nucleus

A temperature activated phenomenon/process, K, is said to be Arrhenius if dlog(K)/d(1/T) is constant where T is the absolute temperature.

Now suppose a process exhibits two constant slopes (m1,m2) when plotted versus (1/T), say m1 governs the low T behavior and m2 governs the high T behavior. Can one call this a non-Arrhenius process. Or is it necessary for the curvature of the plot to have a nonzero value in order for this process to be called non-Arrhenius.