The central thesis of this book is that an inanimate subject signals to language learners that it is likely to be derived, and that the main predicate it is adjacent to is what I have called a displacing predicate (one that does not underlyingly select an external argument, or “semantic subject”). There are four ingredients needed to support this thesis, and in the preceding chapters I have provided three of these ingredients: (1) languages universally prefer subjects to be animate and allow inanimate subjects if they are derived or displaced (Chapter 3); (2) adults are sensitive to NP animacy in their parsing of sentences and preferentially treat inanimate subjects as derived (Chapter 4); (3) the animate–inanimate distinction is available to children's conceptual apparatus and children, like adults, treat inanimate subjects as derived in experimental settings (Chapter 5). The fourth ingredient is to extend the experimental demonstrations of Chapter 5 to language acquisition “in the wild.” That is, we must show that the subject animacy cues that appear to guide children's guesses about sentence structure (and predicate categorization) in experimental settings are plausibly used by children in the natural course of language acquisition.

We will do this in two ways. The first is to demonstrate that these cues are available to children in the language they hear around them. To do this I will present data on patterns of subject animacy in input to children with the different kinds of predicates we have focused on here, showing that subjects are inanimate asymmetrically more often with displacing predicates than with non-displacing predicates. One could think of this as a computational-level analysis: it shows that, in principle, the cues are present in the input data for children to use.