In a recent publication, Michel and Löhr (2024) propose a predictive processing (PP) theory of copredication. They do so in an attempt to account for previously documented predicate ordering effects, whereby copredication acceptability is significantly impacted by the order of predicates. I review how their PP model is unfit for purpose, does not account for the empirically documented effects, does not provide novel experimental directions or conceptual innovations, and ultimately falls back on assumptions already presented in the copredication literature. Hence, the PP model fails empirically and theoretically.