This paper focuses on the long-standing problem in Bantu syntax of why some objects lose the ability to be realized as object markers (OMs) in the passive. The standard answer to this question since the work of Gary and Keenan (1977) is that the passive and object marker require the same property (e.g., a grammatical relation or a particular case) and if this property is usurped by the passive, it will not be available for an OM. This paper suggests an alternative approach, which maintains that OMs are blocked in the passive, not because the passive and object marker compete for the same scarce resource, but rather because the addition of the passive morpheme adds a structural level that blocks the access of the OMs to case. This modification makes it possible to account for the fact that several standard OMs can absorb case in an active construction, while access to this property is blocked for all standard OMs in a passive construction. This modification has the added benefit of allowing the theory to account for the odd behavior of SiSwati in a simple way and to capture the parallelism between the behavior of Runyambo and SiSwati. (MDM)