|Abstract:||This paper focuses on three examples of aspect splits, wherein a particular Case, agreement form, or preposition is restricted, to or prohibited in, the perfective aspect. In Hindi and related languages, ergative Case is restricted to the perfective aspect. This split provides evidence for contextual (positional) faithfulness when its effect is viewed in combination with a second split in Nepali which expands the distribution of ergative Case into some imperfective contexts. The second example of an aspect split has nothing to do with ergativity. In Palauan, preposition insertion is prohibited in the perfective aspect. This can be accounted for with either DEP-perfective (P) or *P/perfective. This Palauan split is interesting because it involves an aspect driven choice between two possible 'repairs' to a phenomenon in clauses that resembles coda conditions or onset faithfulness in syllables. The third example of an aspect split, in Yucatec Maya and related languages, involves a prohibition against using one agreement series (Set A) in the perfective aspect. Because agreement is not present in the input to syntax, either a contextually restricted DEP constraint or markedness constraint will do, but the context can only be perfective (not imperfective). Thus, all three aspect splits can be analyzed as contextual faithfulness, but only one must be; and the context of an aspect split appears to be restricted to perfective (rather than imperfective).