This article investigates the acquisition of two structures involving the syntax-discourse interface, locative inversion (LI) and there-constructions (TC), by advanced and near-native speakers of English whose L1s are European Portuguese and French. Its purpose is to test two hypotheses on the end-state of L2 acquisition at the syntax-discourse interface: the Interface Hypothesis (IH) and the L1+input Hypothesis (LIH). The former proposes that this interface is a locus permanent optionality due to processing inefficiencies related to bilingualism. The latter claims that syntax-discourse structures only generate persistent problems when they are different in the L1 and rare in the input. A battery of timed and untimed tasks tested the verbs and discourse contexts compatible with LI and TC. All L2 groups exhibited optionality regarding the phenomena tested at least in one type of task. Crucially, the control group did not display optionality in any tasks. These results indicate that, as the IH predicts, the syntax-discourse interface is an area of persistent optionality, even when the L1 is similar to the L2. They moreover suggest that optionality is gradient and influenced by, at least, four factors: construction frequency, the quantity and/or distance of the contextual information to be processed, L1-L2 (dis)similarity, and L2 proficiency.

中文翻译：

---

句法-话语界面二语习得的梯度选择：来自高级和接近母语的英语倒置的证据

本文研究了高级和接近母语的英语母语者（其 L1 为欧洲葡萄牙语和法语）对涉及句法-话语界面的两种结构的习得，即位置倒置 (LI) 和结构式 (TC)。其目的是检验句法-话语界面上关于二语习得结束状态的两个假设：界面假设（IH）和二语+输入假设（LIH）。前者提出，由于与双语相关的处理效率低下，该接口是一个永久可选的位置。后者声称句法话语结构只有在 L1 中不同并且在输入中很少见时才会产生持久性问题。一系列定时和非定时任务测试了与 LI 和 TC 兼容的动词和话语上下文。所有 L2 组都表现出关于至少在一种类型的任务中测试的现象的可选性。至关重要的是，对照组在任何任务中都没有表现出可选性。这些结果表明，正如 IH 所预测的那样，即使 L1 与 L2 相似，句法-话语接口也是一个持续可选的区域。此外，他们认为可选性是梯度的，并且至少受四个因素的影响：构建频率、要处理的上下文信息的数量和/或距离、L1-L2（不同）相似性和 L2 熟练程度。