Input matters in L2 acquisition at the interfaces: The case of null objects in L2 European Portuguese

Activity: Talk or presentationOral presentation


In recent years, generative SLA research has started to examine the role of input in grammar acquisition (Rothman & Slabakova, 2017), a factor that was traditionally neglected in this field. According to some authors (e.g., Dominguez & Arche, 2014; Slabakova, 2015), properties at the syntax-discourse interface, which are proposed to be the main area of permanent difficulties in L2 acquisition by the Interface Hypothesis (Sorace & Filiaci, 2006; Sorace, 2011), only cause persistent difficulties to L2 learners when the evidence in the input is not frequent and transparent and the L1 and the L2 differ in the relevant respects.
The acquisition of null objects in European Portuguese (EP) is an appropriate testing ground for this hypothesis, because: (i) they are more productive in spoken/colloquial registers than in standard EP, where clitics are the unmarked pronominal option (Rinke, 2022); (ii) the possibility of definite null objects distinguishes EP from other Romance languages, namely Spanish; and (iii) EP null objects involve the syntax-discourse interface (null objects require a salient and immediately accessible antecedent in the discourse or in the situational context), as well as the syntax-semantics interface (null objects tend to be inanimate). These properties have been extensively investigated in L1 EP (e.g., Flores, Rinke & Sopata, 2020), but not in L2.
To fill this gap, this study investigates the acquisition of clitics and null objects in L2 EP. Participants were 25 L1 EP speakers and 30 L1 Spanish-L2 EP adult learners (10 intermediates, 10 advanced, 10 near-natives). We tested clitics and null objects using two tasks that have been proposed to tap primarily into implicit knowledge (Ellis, 2005): an elicited oral production task (EOPT) and a timed acceptability judgement task (TAJT). Both tasks crossed the variables accessibility of the antecedent (immediately vs. not immediately accessible) and animacy (animate vs. inanimate).
In the EOPT, all groups produced significantly more clitics than null objects across conditions and no significant effects of animacy and accessibility were found. In the TAJT, the native and the L2 groups accepted clitics significantly more than null objects in all conditions. Regarding null objects, the native group exhibits significant effects of animacy (acceptance of null objects is higher in the inanimate condition) and accessibility (acceptance of null objects is higher in the immediately accessible condition).
L2 groups do not display any animacy or accessibility effects, even at the near-native level. These results disconfirm the prediction of the Interface Hypothesis that, unlike properties at the syntax-discourse interface, grammar-internal properties are fully acquirable in an L2. Moreover, they show that, at least when the L1 and the L2 are different, learners may have permanent difficulties wrt interface properties when they are infrequent in the input. Finally, they indicate that, unlike what some studies suggest (e.g., Slabakova, 2015; Sorace, 2014), input factors do not selectively affect the syntax-discourse interface. Grammar-internal interfaces are also affected. Thus, convergence in L2 acquisition at the interfaces depends (at least in part) on the frequency and transparency of the evidence in the input.

Period1 Sept 2023
Event titleThe 32nd Conference of the European Second Language Association
Event typeConference
LocationBirmingham, United KingdomShow on map
Degree of RecognitionInternational