Many researchers now believe that the representations and processes underlying syntactical development are specific to a "language faculty." If so, reference animacy would not be expected to influence acquisition of linguistic structures such as the passive sentence construction. Specifically, children should be comparably able to acquire the passive sentence construction from animate (A) patient instances (e.g., "the baby was picked up by the clown") and static inanimate (SI) patient instances (e.g., "the book was picked up by the clown"). After tests for comprehension of semantically reversible passives, 52 children aged 2.6 to 4.8 years repeated a passive sentence description of A-patient drawings; 51 children aged 2.10 to 5.3 years received descriptions of SI-patient drawings. In post-teaching, all described animate agent drawings, counterbalanced for animate and inanimate patients. Then 45 A-patient and 45 SI-patient children were retested for comprehension. Mean comprehension increased significantly only in A-patient children. Also, more A-patient children than SI-patient children "knew" word order relations for passives in post-teaching production tests. Findings indicate that the conceptual distinctions that children make between animate and inanimate beings can constrain language learning, and imply that language acquisition can not be fully understood in terms of the functioning of a purportedly language-specific faculty. (Author/MSE)