The Weekly Newsletter of MIT Linguistics

Issue of Monday, August 2nd, 2021

Summer Syntax Square 8/03 - Suzana Fong (MIT)

Speaker: Suzana Fong (MIT)
Title: Pronouncing PRO in Wolof
Time: Tuesday, August 3, 2021 1pm—2pm EDT
Location: Zoom
Abstract: In Wolof, subject and object control clauses differ in whether or not the embedded controlled subject is pronounced. In subject control, the embedded subject is obligatorily null (1), while in object control, the embedded subject is obligatorily an overt pronoun (2).
(1) Subject control
Maymuna fas-na *(mu) jàng taalif b-i.
Maymuna try-na.3sg *(3gs.subj) read poem cm.sg-def
‘Maymuna tried to read the poem.’
(2) Object control
Dimbali-na-a a-b xale *(mu) jàng téere b-i.
help-na-1sg indef-cm.sg child *(3sg.subj) read book cm.sg-def
‘I helped a child read the book.’
Prima facie, the presence of the pronoun in (2) may suggest that this construction is not an instance of control (cf. English I helped a child (*she) read the book). Nonetheless, that (2) is indeed a case of control can be argued on the basis of the signature properties that characterize this type of construction (i.e. obligatory bound reading, de se interpretation, and sloppy reading under ellipsis).
The main questions that (1) vs. (2) give rise to are then: why do subject and object control in Wolof differ in the phonological realization of controlled PRO? Which control theory is compatible with such realization?
Answers to the first question may be provided by further differences between (1) and (2). Specifically, the absence of a pronounced subject in subject control (1) correlates with the obligatoriness of clitic climbing (Gowda & Wu 2020; Martinović 2021) and the prohibition of WH-resumption (i.e. the occurrence of a clitic in the position where a WH-phrase moves from in the embedded clause). Conversely, the obligatoriness of a pronounced subject in object control (2) correlates with the prohibition of clitic climbing and the obligatoriness of WH-resumption.
Based on these properties, I suggest that object control clauses project a ΣP which “impedes” movement (in a way that needs formalization). Assuming that obligatory control is derived by movement (Hornstein 1999 et seq), I model the pronounced PRO in (2) as the partial residue of movement (Van Urk 2018) that has been impeded (Lee 2003). Likewise, WH-resumption is modeled in the same way (i.e. as a residue of impeded movement), which accounts for why these properties dovetail in Wolof object control.
Subject control (1), on the other hand, would be an instance of restructuring (Wurmbrand 1998 et seq.), as already proposed by Gowda & Wu (2020) and Martinović (2021). This captures why subject control has the opposite properties as those showcased by object control.
Finally, I briefly examine existing control theories regarding how they treat the phonological realization of PRO and to what extent they are compatible with control phenomena in Wolof.