It is not the case that Computational models trained on realistic child-directed corpora successfully acquire structure-dependent rules without innate grammatical constraints, as Pullum and Scholz demonstrate.
?Set your confidence on the premises below to see your aggregate.