Session 1D

Recent psycholinguistics models identify syntactic priming as a possible mechanism underlying implicit error-based learning of syntax in second (L2) language learners [1, 2]. They predict that 1) priming effects without lexical overlap will lead to long-lasting changes in learners’ knowledge and use of syntactic structures; 2) priming effects with lexical overlap will lead to short-term changes in language use because they rely on explicit memory; 3) the degree of priming will depend on speakers’ learning rate varying with individual differences in attention, motivation and proficiency, factors that are particularly relevant to second language learning [3, 4]. Using picture description tasks, we investigated English L2 French learners’ primed production of two syntactic alternations: active/passive (Study 1) and fronted/non-fronted temporal adverbial phrases (Study 2). In both studies, we manipulated between-subjects whether there was lexical overlap between prime and target sentences. We measured immediate priming (repeating a syntactic structure after a prime) and long-term priming (increase in target structures production in immediate post-tests without primes relative to pre-tests). We assessed attention, motivation and proficiency with questionnaires widely used in L2 research [e.g. 5, 6, 7]. Study 1: lexical overlap mediated immediate priming as the learners produced more passives after passive vs. active primes with overlap (N=48, 51.8%) than without (N=42, 8.3%). The learners showed long-term priming, producing more passives in the post-test than the pre-test. Surprisingly, this increase seemed greater following priming with overlap (33.6%) than without (19.0%). Learners’ magnitude of long-term priming related to: 1) their attention, defined as noticing the syntactic form of the stimuli [6], with (r=0.38, p< .05) and without overlap (r=0.48, p< .05); 2) their motivation to learn French with (r=0.33, p< .05) and without overlap (r=0.31, p< .05); 3) their proficiency (r=0.47, p< .05), with overlap only. Study 2: lexical overlap mediated immediate priming as the learners produced more fronted sentences after fronted vs. non-fronted primes with overlap (N=49, 49.5%) than without (N=45, 14.6%). The learners showed long-term priming, producing more fronted sentences in the post-test than in the pre-test. This increase seemed similar with (20%) and without overlap (22%). The magnitude of long-term priming was related to learners’ attention (r=-.31, p< .05) in the lexical overlap condition only. Learners’ priming effects were not related to their motivation or proficiency. In sum, learners were more likely to continue producing passives if they noticed this structure regardless of overlap condition. However, they were less likely to continue producing fronted sentences if they noticed this structure only in the overlap condition. Learners’ motivation to learn French and their proficiency only related to priming of passives: the former related to long-term priming of passives regardless of overlap condition and the latter to long-term priming of passives in the overlap condition. Thus, the degree of priming did vary with individual differences but such variation was not consistent across syntactic alternations. In line with implicit learning accounts of syntactic priming, the learners experienced long-term effects of syntactic priming on their production of passive and fronted sentences. Surprisingly, lexical overlap seemed to boost not only immediate but also long-term priming of passives. In overall, syntactic priming effects seem more likely to change according to individual differences in learner characteristics for priming of passives only which has important implications for models of priming. Further analyses will explore interactions between factors. [1] Chang, Dell, & Bock. (2006). Psychological Review, 113, 234-272. [2] Jackson. (2017). Second Language Research, 34, 539-552. [3] Takahashi. (2005). Applied Linguistics, 26, 90-120. [4] Ushioda. (2016). Language Teaching, 49, 564-577. [5] Bernolet, Hartsuiker, & Pickering. (2013). Cognition, 127, 287-306. [6] Brooks, & Kempe. (2013). Memory & Cognition, 41,281-296. [7] Dornyei, & Kormos. (2000). Language Teaching Research, 4, 275-300.

The form of anaphoric forms in discourse (null and overt pronominal subjects as well as NPs) is constrained by information-status factors (e.g., topic-continuity and topic-shift), the type of language (null vs non-null subject languages), or the number of activated antecedents, amongst other factors – cf. the bulk of literature in L2 English and L2 Spanish (Blackwell & Quesada, 2012; Filiaci, Sorace, & Carreiras, 2014; Lozano, 2018; Prentza, 2014; Ryan, 2015; among others). Anaphora Resolution (AR), a phenomenon at the syntax-discourse interface, has been shown to be problematic in L2 acquisition as learners (L2ers) are overexplicit (i.e., they redundantly use fuller anaphoric forms than is required) (Lozano, 2016; Ryan, 2015) and rarely attain native-like competence at the syntax-discourse interface even at very advanced levels (Sorace, 2011). Different SLA traditions have investigated the source of these deficits, which is still under debate. We take a corpus-based, theoretically-motivated and systematic approach to understand the factors constraining the development of AR in L2 discourse. We investigate the production of 3rd person anaphoric forms in subject position in L2 English and L2 Spanish from two written learner corpora, COREFL (Corpus of English as a Foreign Language) and CEDEL2 (Corpus Escrito del Español como L2) (www.learnercorpora.com), to analyse adult L1 Spanish-L2 English and L1 English-L2 Spanish L2ers’ written production across proficiency levels (A2-C2) plus two control groups of English and Spanish natives (N= 152 texts) based on a silent film-retell task (Chaplin video). We tagged the multiple factors constraining AR via a linguistically-informed and theoretically-motivated tagset based on previous work (Lozano, 2016; Quesada and Lozano, forthcoming). Results show the development of AR from A2 to C2 level in both L2 English and L2 Spanish learners. The production of anaphoric forms and their acquisition difficulties vary depending on the language combination. In particular, L2 English learners are less redundant in topic-continuity contexts than L2 Spanish learners, while all of them are more felicitous in topic-shift contexts, which shows that not all syntax-discourse properties of AR are equally vulnerable. L2 English learners produce more overt forms but less null pronouns in topic-continuity at beginner levels, whereas English natives show a higher production of null pronouns. However, the difference amongst groups is less marked than in L2 Spanish learners, as they start showing lower rates of null pronouns in topic-continuity contexts compared to their high production by Spanish natives. Additionally, L1 English-L2 Spanish learners show cross-linguistic influence in topic-continuity contexts as they only use null pronouns in topic-continuity coordinate contexts, which is the only context where null pronouns are produced in native English. By contrast, L1 Spanish-L2 English learners know the regulations of null forms in the L2 and do not transfer its use from their L1. Regarding the number of activated antecedents, there is an overall trend showing that L2ers and natives use fuller anaphoric forms as the number of activated antecedents increases. In short, the acquisition of anaphoric forms is constrained by general factors regardless of the language combination. However, the directionality in language and other linguistic factors determine the L2 developmental process and learners’ ultimate attainment. References: https://1drv.ms/b/s!ArpPoL6zgv90lPVfTFPnfvEzF42Hhg?e=NmIJXi

Beyond sensitivity of subject pronouns to syntactic cues (i.e., Position of the Antecedent Hypothesis, PAH; Carminati 2002), recent studies have also identified their sensitivity to discourse cues, such as the information status of the antecedent. In particular, they have identified an anti-focus effect that makes a focused antecedent less accessible in intrasentential contexts (de la Fuente 2015 for null pronouns in Spanish, Colonna et al. 2012 in German and French), an effect that is not attested in non-native populations (Patterson et al. 2017 in L2-German). However, to date, no studies have contrasted the effects of focus structures for null and overt pronouns, in null subject languages, nor investigated bilingual populations, characterised by having competing resources that may affect the integration of information in interface phenomena, as is the case with pronoun solving (Sorace 2016). Therefore, the aim of this research is twofold: 1) to analyse the impact of contrastive focus via cleft structures on the interpretation of null and overt anaphoric subject pronouns in bilingual Spanish in contact with Catalan, two languages that show microvariation in pronominal resolution patterns (Bel & García-Alcaraz 2018), and 2) to investigate whether early Spanish-Catalan bilinguals are sensitive to discourse-level cues as a function of language dominance. 
Three groups of Spanish-Catalan bilinguals (34 Catalan-dominants, 31 balanced bilinguals and 29 Spanish-dominants) and a group of Spanish monolinguals completed a two-alternative preference task (using a slider scale) in Spanish. Regarding internal linguistic factors, two conditions were manipulated: information structure (unmarked structures vs. subject-cleft constructions vs. object-cleft constructions) and pronoun type (null vs. overt). Participants were asked to choose between two antecedents to interpret a totally ambiguous subject pronoun in an intrasentential context as referring to a subject or an object antecedent, as shown in (1).
Fue a Laura a quien asustó María cuando entró a la habitación. (object focus, null pronoun)
¿Quién entró a la habitación?            
María / Laura (subject vs. object antecedent)
Results show a different impact of focus constructions on anaphora resolution depending on pronoun type and group. In monolingual Spanish and bilingual Spanish by Spanish-dominants, the anti-focus effect emerges only for overt pronouns when the object antecedent is clefted. Catalan-dominant bilinguals, however, show an anti-focus effect for both null pronouns (when the subject is clefted) and overt pronouns (when the object is clefted). Thus, in intrasentential contexts, focusing via it-cleft reduces the accessibility of the focused antecedent for subject pronouns (Colonna et al. 2012, de la Fuente 2015), being this effect stronger for overt pronouns. Finally, balanced bilinguals do not seem to be sensitive to focus constructions effects: no differences in the antecedents' choices of null nor overt pronouns are attested when marked and unmarked information structures are contrasted. Differences between groups will be discussed in relation to Sorace's (2016) view that anaphora resolution can be cognitively taxing for bilinguals, although it cannot be generalized to all bilingual populations. The possible influence of Catalan on bilingual Spanish will also be further explored.

Compulsory Non-Bare Nouns as Definites in L2 and L3 Mandarin Grammars
Unlike English, Mandarin is widely considered an article-less language where bare nouns can be used to express definiteness (Cheng and Sybesma, 1999). However, this is not the case across the board. In proper-hypernym definite contexts, bare nouns are not allowed in Mandarin and non-bare nouns, e.g. Demonstrative-Classifier-NPs (Dem-Cl-NPs), are compulsory, as in (1):
(1) Shangzhou wo  qu-le  Beijing. *Chengshi/Na-ge chengshi  hen da.
      last week   I   go-ASP Beijing             city/that-CL city        very big 
      Intended meaning: "Last week I went to Beijing. The city is big."
In Cantonese, another article-less language, Classifier-NPs like go lousi "CL teacher" are used to express definiteness, while bare nouns like lousi "teacher" must be indefinite and is therefore disallowed in definite contexts. In English, an article system is required for (in)definites, and bare nouns are underspecified for definiteness.
This paper reports on an empirical study examining whether or not L2 and L3 Mandarin speakers are able to acquire bare nouns as definites in Mandarin in general and the compulsory non-bare nouns as definites in certain contexts in particular. Investigating a less-examined language triad, it aims to contribute to the identification of the source of transfer in L3 acquisition (Rothman, 2015). The study involved 158 participants including native English speakers learning L2 Mandarin, L1 English L2 Cantonese speakers learning L3 Mandarin, L1 Cantonese L2 English speakers learning L3 Mandarin, and native Mandarin speakers, and learners of L2/L3 Mandarin at both initial and developmental stages were included. They completed an AJT task in which bare nouns, e.g. laoshi "teacher", or Dem-Cl-NPs, e.g. na-ge laoshi "that-CL teacher", appears in anaphoric, associational, and proper-hypernym definite contexts. 
Our data suggests that transfer comes from the previously-acquired language that is structurally similar to L3 Mandarin and occurs on a property-by-property basis. Evidence for both facilitative and detrimental transfer from Cantonese is noticed. Moreover, our investigation of discourse-related target properties suggests that they are not necessarily problematic for L3 learners.
References
Cheng, Lisa Lai-Shen, and Sybesma, Rint. (1999). Bare and not-so-bare nouns and the structure of NP. Linguistic Inquiry, 30, 509–542. 
Rothman, J. (2015). Linguistic and cognitive motivations for the Typological Primacy Model of third language transfer: Timing of acquisition and proficiency considered. Bilingualism: Language and Cognition, 18(2), 179–190.
Westergaard, M., Mitrofanova, N., Mykhaylyk, R. and Rodina, Y. (2016). Crosslinguistic influence in the acquisition of a third language: The Linguistic Proximity Model. International Journal of Bilingualism, OnlineFirst, DOI: 10.1177/1367006916648859