In human sentence production, there is a tendency for animate referents (e.g. person, animal) to come earlier than inanimate ones (e.g. objects). This is because they are more conceptually accessible - more easily retrieved from memory (Bock & Warren 1985). Although Bock & Warren's English study suggests that conceptual accessibility does not affect word-order, English word-order is `fixed'- thus it would be arguable whether it would be the same case for the language which has relatively free-word order or not (e.g. Prat-Sala & Branigan, 1999 in Spanish).
In order to find out about it, using Japanese is particularly useful - it uses case-marking to determine the syntactic functions, in other words, as long as there is case-marking, the word order is freely changeable. Thus if the results in Japanese are the same as the ones in English, it would be much stronger evidence that conceptual accessibility influences grammatical function, but not word-order.
In my recent project, I conducted a sentence recall task using 6 different sentence types:
Bock and Warren's model predicts that 2 sets (1 and 2, 3 and 4) should show animacy effects, but not 5 and 6. However, if animacy affects the recall of 5 and 6, this would be evidence for animacy effects on word-order and hence evidence against Bock & Warren's hypothesis.