Sarah Bernolet

Shared syntactic representations in bilinguals: Evidence for the role of word-order repetition

Studies on syntactic priming strongly suggest that bilinguals can store a single integrated representation of constructions that are similar in both languages (e.g., Spanish and English passives; R. J. Hartsuiker, M. J. Pickering, & E. Veltkamp, 2004). However, they may store 2 separate representations of constructions that involve different word orders (e.g., German and English passives; H. Loebell & K. Bock, 2003). In 5 experiments, the authors investigated within--and between--languages priming of Dutch, English, and German relative clauses. The authors found priming within Dutch (Experiment 1) and within English as a 2nd language (Experiments 2 and 4). An important finding is that priming occurred from Dutch to German (Experiment 5), which both have verb-final relative clauses; but it did not occur between Dutch and English (Experiments 3 and 4), which differ in relative-clause word order. The results suggest that word-order repetition is needed for the construction of integrated syntactic representations.

Persistence of emphasis in language production: A cross-linguistic approach

This study investigates the way in which speakers determine which aspects of an utterance to emphasize and how this affects the form of utterances. To do this, we ask whether the binding between emphasis and thematic roles persists between utterances. In one within-language (Dutch-Dutch) and three cross-linguistic (Dutch-English) structural priming experiments, we measured persistence effects for four different Dutch transitives (actives, PP-initial passives, PP-medial passives, and PP-final passives). Whereas English allows only one passive (PP-final passive), Dutch allows three different variants with the same functional assignment, but different constituent structures. Additionally, the degree of emphasis on the agent differs significantly between the PP-initial passive and the other passives (Experiment 1). Experiment 2 showed persistence of actives, PP-medial, and PP-final passives in Dutch, but no priming between passives with different constituent structures. Experiments 3 and 4, however, showed that both PP-medial and PP-final passives prime the use of English passives. Experiment 5 confirmed that the emphasis on thematic roles persists: the proportion of passives in the PP-initial passive condition fell midway between the proportions produced in the active and PP-medial passive condition.

Does verb bias modulate syntactic priming

In a corpus analysis of spontaneous speech Jaeger and Snider (2007) found that the strength of structural priming is correlated with verb alternation bias. This finding is consistent with an implicit learning account of syntactic priming: because the implicit learning model implemented by Chang (2002), Chang, Dell, and Bock (2006), and Chang, Dell, Bock, and Griffin (2000) uses probabilistic information about different verb-structure combinations to predict the form of sentences, it predicts that primes exert stronger priming when they are less expected, given the syntactic preference of their main verb. We tested this claim experimentally by comparing the strength of double-object dative priming (DO) and prepositional object dative priming (PO) between dative verbs with differing syntactic preferences in a syntactic priming experiment. The syntactic preferences of the prime and target verbs were first measured in a picture description experiment. Consistent with an implicit learning account, the results showed a verb-specific effect of inverse preference: the strength of DO-priming was modulated by the alternation bias of the dative verbs that were used in the primes.

From language-specific to shared syntactic representations: The influence of second language proficiency on syntactic sharing in bilinguals

Studies on cross-linguistic syntactic priming suggest that bilinguals can share syntactic representations across languages (e.g., Hartsuiker, Pickering, & Veltkamp, 2004). But how are these representations established in late learners of a second language? Specifically, are representations of syntactic structures in a second language (L2) immediately collapsed with similar structures of the first language (L1), or are they initially represented separately? In order to investigate this, we primed the use of English genitives with Dutch (Experiment 1) and English (Experiment 2) genitives (e.g., het hemd van de jongen/the shirt of the boy vs. de jongen zijn hemd/the boys shirt) in late Dutch-English bilinguals with varying levels of proficiency in English (their L2). The head nouns of prime and target constructions either had the same meaning (hemd/shirt - shirt) or a different meaning (duim/thumb - shirt), in order to test whether the use of both genitives was generalized across nouns. Experiment 1 found stronger between-language priming for more than less proficient bilinguals in both conditions, thus suggesting a shift from language-specific to shared syntactic representations. Experiment 2 suggests that these early, language-specific syntactic representations might be item-specific: Less proficient bilinguals showed much weaker priming when the heads of prime and target constructions had different meanings than when they were repeated.

Effects of phonological feedback on the selection of syntax: evidence from between-language syntactic priming

Research on word production in bilinguals has often shown an advantage for cognate words. According to some accounts, this cognate effect is caused by feedback from a level that represents information about phonemes (or graphemes) to a level concerned with the word. In order to investigate whether phonological feedback influences the selection of words and syntactic constructions in late bilinguals, we investigated syntactic priming between Dutch and English genitive constructions (e.g., the fork of the girl vs. the girls fork). The head nouns of prime and target constructions were always translation equivalents. Half of these were Dutch–English cognates with a large phonological overlap (e.g., vork–fork), the other half were non-cognates that had very few phonemes in common (e.g., eend–duck). Cognate status boosted between-language syntactic priming. Further analyses showed a continuous effect of phonological overlap for cognates and non-cognates, indicating that this boost was at least partly caused by feedback from the translation equivalents’ shared phonemes.

The development of shared syntax in second language learning

According to Hartsuiker et al.s (2004) shared-syntax account bilinguals share syntactic representations across languages whenever these representations are similar enough. But how does such a system develop in the course of second language (L2) learning? We will review recent work on cross-linguistic structural priming, which considered priming in early second language learners and late second language learners as a function of proficiency. We will then sketch our account of L2 syntactic acquisition. We assume an early phase in which the learner relies on transfer from L1 and imitation, followed by phases in which language- and item-specific syntactic representations are added and in which such representations become increasingly abstract. We argue that structural priming effects in L2 (and between L1 and L2) depend on the structure of this developing network but also on explicit memory processes. We speculate that these memory processes might aid the formation of new representations.

Syntactic priming in bilingual patients with parallel and differential aphasia

Background: Syntactic priming is the phenomenon by which the production or processing of a sentence is facilitated when that sentence is preceded by a sentence with a similar syntactic structure. Previous research has shown that this phenomenon also occurs across languages, i.e., hearing a sentence in one language can facilitate the production of a sentence with the same structure in another language. This suggests that syntactic representations are shared across languages. Aims: The aim of the current study is to investigate this cross-lingual syntactic priming in patients with bilingual aphasia. To address this aim, we asked the following three research questions: (1) do patients with bilingual aphasia show priming effects within and across languages? (2) do these priming effects differ from the priming effects observed in control participants? and (3) does the pattern of priming effects interact with the type of aphasia? Methods & Procedures: We tested two groups of patients: one group had similar impairments in both languages (parallel aphasia); in the other group, the impairments were larger in one of the languages (differential aphasia). We investigated syntactic priming within and across languages by means of a dialogue experiment. Outcomes & results: We found significant cross-lingual priming effects in both patient groups as well as in a control group. In addition, the effect size of both patient groups was similar to that of the control group. Conclusions: These findings support models that incorporate shared syntactic representations across languages, and are in favour of a non-localised account of differential aphasia in bilingual aphasia.

Syntactic representations in late learners of a second language : a learning trajectory

Several studies have shown that syntactic structures can be primed between the different languages of a bilingual. Bilingual production models put forward by Hartsuiker, Pickering and Veltkamp (2004) and Pickering and Hartuisker (2008) therefor assume that bilinguals share syntactic structures between languages as much as possible. In this paper, we discuss a model for the development of these shared syntactic structures in late learners of a second language (Hartsuiker & Bernolet, 2016). More specifically, we discuss evidence for three central claims of the model, namely that 1) L2 syntactic representations move from being item-specific to being more abstract, 2) L2 representations become more and more integrated with existing L1 representations, and 3) L1 influences on syntactic processing and production in the L2 occur in early and late phases of L2 syntactic development. Acquiring the syntax of a language is a very complex and multi-faceted process: A learner needs to learn the meaning of lexical items, form syntactic categories, and distill the complete set of rules and processes that govern the structure of sentences from the input in order to learn how syntax reflects the relation between the different elements in a sentence. Children learn their native syntax (or syntaxes, if they are native bilinguals) quite rapidly, and without explicit instruction: In a year’s time, their utterances evolve from twoword phrases like “Bye-bye daddy” and “Doggy sit” to complex sentences as “Can I have some milk, please, mommy?”. For late learners of a second language, the acquisition of syntax is a stumbling block: It requires much effort and attention, and often some kind of explicit instruction. The learning context is in some cases very different for native syntax learning and the late acquisition of second language (L2) syntax: While native syntax is picked up implicitly from interactions with speakers in the child’s direct environment, second language syntax can be formally taught when children learn a foreign language at school. Another important difference between native syntax learning and late learning of second language syntax is that in the latter case, the syntax of the first language (L1) is already fully acquired. General knowledge of syntactic concepts and systems might aid second language learning, but it could also hinder learners. The knowledge of L1 syntax might lead to ‘educated guesses’ for the comprehension and production of L2 syntax, which might require more conscious thought and effort than the ‘trial-and-error-processing’ that characterizes the acquisition of L1 syntax. Additionally, late learners might not as feel free as children do to produce ill-formed sentences in the process of learning. How do syntactic representations in the second language develop in these late learners and what is the relation between the L1 and L2 syntax? We try to answer these questions by presenting a theoretical model for the acquisition of L2 syntax in late learners of a second language (Hartsuiker & Bernolet, 2016) and discussing the different studies that led to the formulation of this model. A model of the trajectory of late L2 syntactic acquisition The production model we proposed in Hartsuiker and Bernolet (2015) is an elaboration on the bilingual production models of Hartsuiker, Pickering and Veltkamp (2004) and Pickering and Hartsuiker (2008). Hartsuiker et al. (2004) discovered that syntactic structures can be primed between both languages of a bilingual: Spanish–English bilinguals were more likely to describe pictures in English (L2) using passives when they had just processed a Spanish (L1) passive prime sentence (El camión es perseguido por el taxi [The truck is being chased by the taxi]), than when they had processed Spanish actives (El taxi persigue el camión [The taxi chases the truck]) or intransitive sentences (El taxi acelera [The taxi accelerates]). Hartsuiker et al. (2004) concluded that Spanish–English bilinguals use the same syntactic rules when comprehending and producing Spanish or English passives and activate the same memory representations from both languages. Hence they proposed a lexical-syntactic model for bilingual sentence production in which syntactic information is shared between languages as much as possible. In their model, which is based on Pickering and Branigan’s (1998) model of lexico-syntactic representations, information about syntactic structures is represented in combinatorial nodes, which are connected to lemma nodes in a single integrated lexicon. Thus, the lemmas of all Spanish and English transitive verbs are connected to a shared categorical node indicating that these verbs can be combined with two noun phrases in order to form a passive sentence. In the last few years, numerous studies have shown between-language syntactic priming using a variety of language pairs, constructions, and paradigms (Bernolet, Hartsuiker & Pickering, 2012; Cai, Pickering, Yan & Branigan, 2011; Kantola & Van Gompel, 2011; Salamoura & Williams, 2006, 2007; Schoonbaert, Hartsuiker & Pickering, 2007; Shin & Christianson, 2009; 2012), suggesting that syntactic representations are indeed shared between languages as much as possible. Pickering and Hartsuiker (2008) made the additional assumption that, according to a lexicalist model with fully shared syntax, priming between languages should be of comparable magnitude to priming within a language, because both languages would use one and the same node. In our 2015 paper we argued that this is probably not true for all bilinguals (Hartsuiker & Bernolet, 2015). Hartsuiker et al.’s (2004) and Pickering and Hartsuiker’s (2008) models might capture important aspects of syntactic representations in native bilinguals, who simultaneously developed syntactic representations in both of their languages, and for syntactic representations in late learners of a second languages that have reached a very high level of L2 proficiency. These bilinguals can be expected to show abstract structural priming of comparable magnitude withinand across-languages. For bilinguals who did not yet reach this level of L2 proficiency, we expect between-language priming to be weaker than within-language priming, because their L2 syntactic representations are still under development. In order to be able to make predictions about the L2 syntactic output that can be generated by L2 learners and about the strength of within and between-language priming through the development of L2 syntax, we proposed a developmental model with 5 consecutive stages. In our model, learning is determined by two principles, namely representational specificity and economy: The learning system aims to capture all relevant differences between linguistic representation, but , at the same time, it tries to minimize the number of representations by sharing what can be shared. We assume that explicit memory processes play a large role in the early phases of acquisition, and that longer term learning is caused by Hebbian learning or by the implicit learning of syntactic structures (Chang, Dell, & Bock, 2006). The different stages of our developmental model for the representation of L2 syntax are depicted in Figure 1. The model shows the learning trajectories for the English s-genitive (s-gen: the boy’s doll) and the English post-modified noun phrase (N RC: the ball that is red), structures that can both be combined with the nouns pop [doll] and bal [ball]. For simplicity’s sake, the model only shows two lexical nodes and two combinatorial nodes, while leaving out the other nodes to which these nodes are connected (e.g., the conceptual nodes, the language node, the word category nodes, etc.). The representations of the English of-genitive and the English pre-modified noun phrase are also left out, though we assume that these develop together with their syntactic alternatives, the English s-genitive and post-modified noun phrase. Given that the model represents the stages of late L2 syntactic acquisition, the L1 syntax is already fully represented in the first panel of the figure, which represents the initial stage of acquisition for both structures. We assume that, in this initial stage, L2 acquisition begins with learning of lexical representations without firm connections to syntactic information. A Dutch learner of English as an L2 might have learned simple nouns like ‘doll’ and ‘boy’, but might not know which English constructions can be used to express that the doll is owned by the boy (the doll of the boy/the boy’s doll). If this learner wants to express a possessive relation between both nouns, she can only rely on her knowlegde of L1 genitive structures to do the job. Transferring the L1 syntax can lead to transfer errors such as “It’s the doll from the boy (Het is de pop van de jongen)” and “It’s the boy his doll (Het is de jongen zijn pop)”, but also to correct productions “It’s the doll of the boy (Het is de pop van de jongen)”. Another strategy that learners can use at this stage, is to imitate structures that are produced by more proficient speakers. Production is in this case rather reproductive than creative: A question like “Shall I wear the red dress or the blue one?” is likely to elicit parallel responses like “the blue one” and “the red dress”, but no creative response like “the red one”. We assume that these imitations are based on the retrieval of the example sentences from explicit memory. Hence, these repeated or minimally edited L2 structures only appear quickly after the learner perceived an example sentence. To summarize, in this initial stage of the acquisition of L2 syntax production is characterized by L1 transfer and primed L2 structures in immediate conditions when there is a high lexical overlap between consecutive sentences. Comprehension of L2 syntax is also guided by L1 syntactic preferences at this stage, as no L2 syntactic representations are available yet. Figure 4. Devel

Syntactic levels, lexicalism, and ellipsis: The jury is still out

Structural priming data are sometimes compatible with several theoretical views, as shown here for three key theoretical claims. One reason is that prime sentences affect multiple representational levels driving syntactic choice. Additionally, priming is affected by further cognitive functions (e.g., memory). We therefore see priming as a useful tool for the investigation of linguistic representation but not the only tool.