Since its discovery 65 years ago, the solid-state dynamic nuclear polarization effect has been rationalized in terms of mixing of the Zeeman energy levels. Before becoming aware of this explanation, Erb, Motchane and Uebersfeld proposed a dynamical equation to make sense of their experiments. Here we provide a formal justification of their phenomenological equation. The result is a different way of thinking about the solid-state effect, with novel implications for the effect in liquids.