Will Nediger recently presented a paper at GASLA 13 (Generative Approaches to Second Language Acquisition), held at the University of Indiana, Bloomington. This paper is based on a joint research between Will, Acrisio Pires, and Pedro Guijarro-Fuentes (Plymouth University/Universitat de les Illes Balears). The title and abstract of their presentation are given below.
An Experimental Study of the L2 Acquisition of Spanish Differential Object Marking
In Spanish, some direct objects are marked with the particle a, arguably depending on syntax-semantic properties of the object, the verb, and the subject, a phenomenon known as Differential Object Marking (DOM). We present an experimental study on the L2 acquisition of the syntax-semantics properties triggering DOM. The study sheds light on competing theories of L2 acquisition, including the Interpretability Hypothesis and the Feature Reassembly Hypothesis.
We tested native English speakers who are advanced adult L2 learners of Spanish and live in Spain, in contrast to an adult control group of native Spanish speakers. The study included three tasks: a grammaticality judgment task, an elicited production task (in which subjects could either insert a before an object, or insert nothing), and a discourse context-sentence matching task, in addition to proficiency tests and ethnographic survey. The main tasks tested semantic properties of the object, verb and subject that have been invoked as possible triggers or blockers of DOM.
Two of the conditions manipulated topicality and specificity. Traditionally, the main semantic features influencing DOM have been argued to be animacy and specificity (e.g. Torrego 1998). It has also been argued that apparent specificity effects are a reflex of topicality effects (Leonetti 2008). To account for both possibilities, in one condition the object was topicalized with clitic doubling (clitic left-dislocation, CLLD), yielding a specific reading , and in the other the object was topicalized but without clitic doubling (non-CLLD), thus interpretable as specific or non-specific .
Other conditions manipulated different semantic properties relevant to DOM, including object animacy  and affectedness , verb telicity , and subject agentivity , to investigate whether these semantic properties involved in DOM can be acquired by advanced L2 learners. Specificity and occurrence of ‘a’ was controlled across these conditions. These conditions allowed us to investigate the Interpretability Hypothesis (e.g. Tsimpli & Dimitrakopoulou 2007), according to which uninterpretable features absent in the L1 are a greater challenge for L2 learners than interpretable features.
The data collected so far suggest that advanced L2 speakers do have great difficulty acquiring semantic properties of DOM. They perform below native speakers in most conditions, including  to . However, L2 subjects, like native speakers, perform close to ceiling on two conditions blocking DOM: one in which the object is a bare noun (necessarily non-specific) and one in which the verb is existential (making the object necessarily non-specific).
Our results do not support the Interpretability Hypothesis, given the lack of a correlation between interpretability of features and success in acquisition. We explore the hypothesis that the piecemeal L2 acquisition supports the Feature Reassembly Hypothesis (Lardiere 2009), by which L2 acquisition involves reassembly of features in the L2, and the difficulty of acquisition is linked to the amount of reassembly required, and argue that the Feature Reassembly Hypothesis better captures the pattern of data observed. These results are consistent with and extend previous work on DOM acquisition, particularly Guijarro-Fuentes (2012). We also discuss the implications for the notion of “parameter” in a Minimalist framework.
 Clitic left-dislocation
A muchos estudiantes, ya los conocía
‘Many students, I already knew’
 Non-clitic left-dislocation
(A) muchos estudiantes, ya conocía
‘Many students, I already knew’
 [-/+animate, +specific object]
Busco una casa que tiene tres pisos vs. Busco a una vendedora que vende las frutas.
‘I am looking for a house with three floors’ ‘I seek a saleswoman who sells fruit.’
 [+affected, +animate, -specific object]
El accidente hirió (a) dos pasajeros
‘The accident injured two passengers
 [+telic verb], [+animate, -specific object]
Juan halló (a) una niña en una canasta en su porche
‘Juan found a baby in a basket on his porch’
 [+agent subject], [+animate, -specific object]
La enferma está llamando a una enfermera
‘The sick woman is calling for a nurse’