It has long been assumed that consonant alternation of Nivkh consists of both spirantization and hardening. However, careful observation reveals that the two processes are not equivalent. Specifically, hardening is strictly bound to occur to transitive verbs, while no such restriction exists in spirantization. The traditional approach to this asymmetry is straightforward, namely, stipulating that phonological processes are bound to particular syntactic structures, e.g., hardening to VPs and spirantization to NPs. However,such a structure-specific stipulation is a powerful grammatical device and should be treated with caution. A preferable approach is possible if we posit the plosive-initial form as the underlying form of transitive verbs initiating with a fricative. As a consequence, we can eliminate hardening from the phonology of Nivkh and its absence in NPs can be explained.