Parameter setting is either precipitous (Gibson & Wexler, 1994) or it is gradual in response to input frequency (Yang, 2002, 2004). In this study, we compare these models against the empirical domain of subject and (direct) object drop in Mandarin. We conducted a corpus-based study of the speech of 47 Mandarin-speaking children aged 1;2–6;5, and their caregivers, from the CHILDES database. The results show that before age 1;8 all the children used null subjects and null objects in a target-like fashion, which reveals that the parameter that governs null topics is set from very early on, even if the presence of disambiguating evidence for [+Null Topic] patterns is low. Besides, children's ba constructions, which require an overt object, reliably included this object from the first occurrence although its frequency was scarce in the input. Our results indicate that the setting of certain parameters occurred early independently of the input.

Few studies have assessed children’s comprehension of the null infinitival subject (PRO) in Obligatory Control complements when the matrix object is implicit. A study on European Portuguese showed that children have poor comprehension of control complements of dizer “tell” with an implicit object controller, displaying high rates of subject control with this verb (Gamas, 2016). Moreover, Mateu (2016) found that both Spanish and English-speaking children performed at adult level with subject control complements of promise-type verbs when the matrix object is implicit. Our experiment tested children’s comprehension of control with prometer “promise” and dizer “tell” with overt and implicit objects. The results show that, in the case of sentences with implicit objects, children obtained significantly higher rates of object control with dizer than with prometer, despite high rates of (non-adult) subject control with dizer. Furthermore, children obeyed the locality of Obligatory Control and were sensitive to the properties of the main verb.