Remarks on Oku’s generalization: anti-agreement and subject ellipsis in Spanish and Japanese

Resumo

I focus on the so-called Oku’s generalization in order to show: (i) that a main division among at least two types of pro-drop languages is irreducible, i.e., radical vs. consistent pro-drop (pace DUGUINE, 2013), and (ii) that the distinction can be understood in terms of two types of ellipsis, namely phrasal DP-ellipsis (Japanese) and head pronoun ellipsis (Spanish). Such a distinction follows from a general model for ellipsis within and across languages and does not have to be stipulated for the particular distribution of null subjects in Japanese and Spanish. Thus, the difference between radical and consistent pro-drop languages does not follow from the deep vs. surface anaphora distinction, but from the timing of ellipsis in each language. In addition, I conjecture that the lack of DP-ellipsis in Spanish, and consistent pro-drop languages in general, is due to the agreement-Case system. In this respect, my analysis is committed to the so-called anti-agreement hypothesis, according to which the availability of DP-ellipsis is connected to absence of morphological agreement. Finally, even when I dispense with the deep vs. surface distinction as a way of deriving Oku’s generalization, I claim that null NP anaphora, a type of deep anaphora, is syntactically active at least in radical and partial pro-drop languages (BARBOSA, 2019).