The role of crosslinguistic influence in Greek child and adult learners' acquisition of English as a Foreign Language

跨语言影响的第二语言研究的焦点一直是句法-话语界面（Hulk & Muller, 2000）。Hulk和Muller（2000）观察到儿童较晚获得这种界面（以及它所调节的过程），据此提出了界面假说。根据这一假设，如果一个双语儿童的母语是模棱两可的（使用两种结构来表达语法-话语界面现象），但她的第二语言是明确的（使用一种结构来表达相同的现象），她会使用两种语言都显示的结构，即使在她的母语需要另一种结构的环境中。在这一背景下，主谓代词的虚显分布（属于句法-语篇界面的过程）成为一个富有成果的研究课题，引发了大量的研究。这表明，正如Hulk和Muller（2000）所预测的那样，双语儿童使用虚和显性代词的特点是受非虚主语语言（只使用显性主语代词）到虚主语语言（既使用显性主语代词又使用虚主语代词）的跨语言影响（如Argyri & Sorace, 2007; Paradis & Navarro, 2003; Serratrice, Sorace & Paoli, 2004）。事实上，在这些研究中，跨语言的影响导致双语参与者比单语者在他们的无主体语言中产生更多的显性主体。值得注意的是，上述一些研究强调语言优势可能在跨语言影响中发挥重要作用（Argyri & Sorace, 2007）。尽管具有开创性，但这些研究没有涉及大样本，因此无法提供可推广的结果。事实上，Argyri和Sorace（2007）对这一现象进行了为数不多的大规模调查，但包括8岁的参与者，因此无法揭示双语儿童的语言发展在其最初阶段是如何进行的。第二语言跨语影响的研究方法不同：该领域的研究人员没有关注结构问题，而是假设学习者的第一语言和第二语言之间的特征不匹配是跨语影响的原因（Tsimpli & Roussou, 1991）。Tsimpli和Roussou（1991）认为，要了解他们的语言允许空主语，儿童还必须了解主语本来要表达的数和人的特征必须印在动词上。在掌握了第一语言的特征之后，孩子们在学习第二语言或外语的特征（及其表达）不同时可能会遇到困难。在这种情况下，主语代词的分布再次成为研究的主要课题。在这个框架下产生的研究集中在学习非空主语语言（这些特征与名词相连）的母语为零主语语言（人称和数字特征与动词相连）的人身上。他们认为，特征不匹配的概念能够解释从空主语语言到非空主语语言的跨语言影响（例如，Prenza, 2014; White, 1985）。事实上，根据这些调查，学习者在非零主体语言中产生的零主体比单语使用者产生的零主体更多，而熟练程度被视为降低跨语言影响稳健性的重要因素（Prenza, 2014）。然而，这些研究包括小的、不可推广的样本，并专注于成人学习者；因此，与第二母语调查不同，它们不能阐明双语语言发展的过程。这两方面的研究尽管存在差异，但却是相似的。他们都认为跨语言影响在语言发展中是突出的，并且是由于语言的原因而产生的。确切的语言学原因各不相同——第二语言研究认为是结构歧义，而第二语言研究认为是结构歧义

The focus of 2L1 research on crosslinguistic influence has been the syntax-discourse interface (Hulk & Muller, 2000). Based on the observation that children acquire this interface (and thus, the processes it regulates) late, Hulk and Muller (2000) posited the Interface Hypothesis. According to the hypothesis, if a bilingual child's first language is ambiguous (using two structures to express a syntax-discourse interface phenomenon), but her second language is unambiguous (using one of the structures for the same phenomenon), she would use the structure that both languages manifest, even in contexts where her first language would require the other structure. Under this light, the distribution of null and overt subject pronouns (a process that pertains to the syntax-discourse interface) became a fruitful research topic that gave rise to numerous studies. These suggest that, as Hulk and Muller (2000) predicted, bilingual children's use of null and overt pronouns is characterised by crosslinguistic influence from the non-null-subject language (which uses only overt subject pronouns) to the null-subject language (which uses both overt and null subject pronouns) (e.g. Argyri & Sorace, 2007; Paradis & Navarro, 2003; Serratrice, Sorace & Paoli, 2004). Indeed, crosslinguistic influence caused bilingual participants in these studies to produce more overt subjects in their null-subject language than monolingual speakers of the language. Notably, some of the aforementioned studies emphasised that language dominance might play an important role in crosslinguistic influence (Argyri & Sorace, 2007). Albeit groundbreaking, these studies do not involve large samples and, therefore, cannot provide generalizable results. In fact, Argyri and Sorace (2007) conducted one of the few large(r) scale investigations of the phenomenon, but included eight-year-old participants and, consequently, cannot reveal how bilingual children's language development proceeds in its earliest stages.The approach of L2 research on crosslinguistic influence was different: instead of focusing on structural matters, researchers in the field postulated that feature-mismatch between the learner's first and second language is the cause of crosslinguistic influence (Tsimpli & Roussou, 1991). According to Tsimpli and Roussou (1991), to learn that their language allows null subjects, children must also learn that the number and person features that the subject would otherwise express must be imprinted on the verb. Having acquired their first language's features, children might face difficulties when learning a second or foreign language whose features (and their expression) differs. Under this light, the distribution of subject pronouns became, once again, the main topic of investigation. The studies that were produced under this framework focused on native speakers of a null-subject language (where person and number features are linked to the verb) who were learning a non-null-subject language (where these features are linked to the noun). They suggested that the idea of feature-mismatch was able to explain crosslinguistic influence from the null to the non-null-subject language (e.g. Prenza, 2014; White, 1985). Indeed, according to these investigations, learners produced more null subjects in their non-null-subject language than monolingual speakers of that language, while proficiency was seen as an important factor in decreasing the robustness of crosslinguistic influence (Prenza, 2014). However, these studies included small, non-generalisable samples and focused on adult learners; thus, not unlike 2L1 investigations, they cannot elucidate the process of bilingual language development. The two strands of research are similar despite their differences. Both agree that crosslinguistic influence is prominent in language development and occurs due to linguistic reasons. The precise linguistic reasons differ - with 2L1 research suggesting structural ambiguity, and L2 resea