EUROSLA Yearbook - Volume 8, Issue 1, 2008

Recent psycholinguistic studies provide compelling evidence for the claim that both languages are active when second language (L2) learners and bilinguals process information in one language alone. The parallel activation of the two languages occurs even when individuals are performing highly practiced tasks such as reading, listening, and speaking, and even when they are highly proficient in both languages. The presence of cross-language activity in the absence of random errors, particularly for those who are highly proficient in the L2, suggests that a mechanism of cognitive control is in place to guide the selection of the intended language. The focus of current research is to understand the basis of this cognitive mechanism, how it varies as a function of individual differences in cognitive resources, and what consequences it holds for cognition more generally. In this paper we consider whether L2 learners and bilinguals are able to exploit cues to language status that might allow them to focus their attention on languagerelevant attributes of processing or to effectively inhibit information related to the language not in use as a means to control language selection. The results of the present study suggest that it is possible to create a functional language cue for planning the L2 and bias language selection.

The study compares pragmalinguistic knowledge of the L1 in non-routine situations in L2 users and L2 learners, aiming to see to what extent pragmalinguistic behaviours in the L1 are influenced by exposure to the L2. The first part of the paper discusses the nature of bilingual language knowledge and transfer in multicompetence. In the experiment, refusals to untypical requests were elicited by means of a discourse completion test (DCT). The L2 users in this study appear to be “more native than the native speaker” (Cook, Iarossi, Stellakis, & Tokumaru 2003) in their choice of semantic formulas, as compared to monolingual speakers of their L1. Since pragmatic transfer is ruled out by the baseline data from native speakers of English, an alternative explanation is proposed, based on the L2 users’ increased sensitivity to interactional demands (Genesee, Tucker & Lambert 1975), more varied experience at social interaction and changed perception of their L1.

The present article provides an overview of some recent research in bilingual first language acquisition with special reference to Romance languages. It addresses two language contact phenomena, language mixing at the lexical level and crosslinguistic influence at the syntactic level. For both contact phenomena, there is evidence that they are unrelated to language dominance (as measured in terms of MLU). Language mixing is negatively correlated with the number of utterances per minute, a measure for language fluency. Cross-linguistic influence at the syntactic level is due to computational complexity which is caused for example by the invasive interplay of pragmatics and syntax. This kind of interplay will be discussed on the basis of the presence or absence of the null-subject property, comparing Italian and German, and by studying the presence or absence of object clitics, comparing French and German. Since cross-linguistic influence affects bilingual individuals to different degrees (sometimes referred to as individual variation), we discuss the assumption that the degree to which the influence manifests itself differs as a function of fluency (measured in words produced per minute). The more fluent the child is in the language with the computationally complex analysis, the less the effect of cross-linguistic influence. In sum, the language which exhibits influence is determined by computational complexity as defined, for example, by the invasive interplay between pragmatics and syntax. The degree of manifestation of the influence depends on fluency in the computationally complex language.

The question of whether adult native speakers of Chinese, a language that does not morphosyntactically represent tense, are able to acquire tense in English has been a topic of great interest in part because it allows us to examine whether there is a critical period for features that are not instantiated in the native language (Hawkins 2001; Hawkins & Liszka 2003; Lardiere 1998a, 2003).While most previous studies have focused on production data, the present study examines the semantics of tense, investigating whether or not learners’ interpretations are sensitive to temporal distinctions. Native speakers of Mandarin are compared with native speakers of Japanese and Korean, languages which both morphosyntactically encode tense. Results of an interpretation task targeting the present progressive and past progressive in English show that by advanced levels of proficiency the three groups of learners performsimilarly. The results provide evidence that tense is fully acquirable in L2 acquisition regardless of the properties of the native language.

This paper reports the results of an experimental study on the resolution of intra-sentential anaphora in Italian, by native Italian speakers and near-native Italian speakers whose L1 is Croatian. In a picture-selection task, the two groups of speakers had to identify the antecedents of third person null and overt subject pronouns in ambiguous forward and backward anaphora sentences and their unambiguous counterparts. In all contexts under investigation, near-natives expressed native-like antecedent preferences, indicating that they have acquired not only the syntactic, but also the discourse-pragmatic constraints on the realisation of pronominal subjects in Italian. Given that Croatian and Italian do not differ with respect to the licensing and distribution of pronominal forms, the results provide additional evidence for an account attributing the well-attested tendency of bilingual speakers of a null subject language and a non-null subject language to overuse and misinterpret overt subjects in the null subject language to crosslinguistic influence.

Much recent research in SLA is guided by the hypothesis of L2 interface vulnerability (see Sorace 2005). This study contributes to this general project by examining the acquisition of two classes of subjunctive complement clauses in L2 Spanish: subjunctive complements of volitional predicates (purely syntactic) and subjunctive vs. indicative complements with negated epistemic matrix predicates, where the mood distinction is discourse dependent (thus involving the syntax–discourse interface). We provide an analysis of the volitional subjunctive in English and Spanish, suggesting that English learners of L2 Spanish need to access the functional projection Mood P and an uninterpretable modal feature on the Force head available to them from their formal English register grammar, and simultaneously must unacquire the structure of English for-to clauses. For negated epistemic predicates, our analysis maintains that they need to revalue the modal feature on the Force head from uninterpretable to interpretable, within the L2 grammar.With others (e.g. Borgonovo & Prévost 2003; Borgonovo, Bruhn de Garavito & Prévost 2005) and in line with Sorace’s (2000, 2003, 2005) notion of interface vulnerability, we maintain that the latter case is more difficult for L2 learners, which is borne out in the data we present. However, the data also show that the indicative/subjunctive distinction with negated epistemics can be acquired by advanced stages of acquisition, questioning the notion of obligatory residual optionality for all properties which require the integration of syntactic and discourse information.

In the literature, L2 learners of English are reported to have problems accepting grammatical unaccusative sentences like the cup broke yesterday although they accept the passivized form like the cup was broken yesterday (e.g. Yip 1995; Ju 2000). They also produce ungrammatical sentences like “If the World War III is happened, . . .”. It has been suggested that these phenomena are universal in the development of L2 English because they occur in the English of L2 learners with different L1s (cf. Ju 2000). In this paper, we present an empirical study investigating whether these non-target phenomena also occur in L2 Chinese. It is found that English speakers of Chinese at different proficiency levels fully accept the NP-V unaccusative construction. Optionality is observed in English speakers’ non-target passivized unaccusative verbs of change of state, which are argued to be due to L1 transfer when English speakers incorrectly treat Chinese change-of-state verbs as causative. Overall, our results suggest that the overpassivization phenomenon should not be considered universal in the L2 acquisition of unaccusativity despite their prevalence in L2 English.

Finiteness in adult Dutch is expressed by use of formal means of morpho-syntax, i.e. with both inflectional morphology and word order (verb-second). These formal means are linked to a projection of the functional category FIN as in:FINSpecFIN FINFIN VPSpecVP VPAt the initial stage of acquisition, there is no productive use of functional elements nor do learner languages have verb-second. This is due to the fact that initially learner languages are systems in which the functional category FIN is absent. Properties of validation and anchoring are expressed by lexical means: Expression of illocutionary force occurs with modal predicates, contextual anchoring is expressed with deictic elements. Instantiation of the functional projection FIN is caused by the acquisition of the auxiliary verb heb/heeft (have/has). As a functional category, AUX provides a position for functional elements to express modal or – in the default case – non-modal illocutionary force. With the instantiation of AUX, the initial, i.e. specifier position serves as a position for elements with topic function, i.e. for elements used to establish contextual anchoring of the utterance. The acquisition of AUX is the driving force in the development from the lexical to the functional stage. However, the acquisition of AUX does not come as a deus ex machina. Structures with AUX are learned for reasons of information processing. Utterances with an auxiliary are typically used as a topicalisation device. Due to the fact that structures with AUX are learned as a topicalisation device, it seems possible (1) to provide a functional explanation for why auxiliaries, inversion and subject pronouns are learned simultaneously and (2) to account for the fact that the acquisition process occurs similarly in both L1- and L2-learner languages.

Researchers argue that variable use of inflectional morphology demonstrates a lack of underlying syntactic representation in second language learners. Others disagree and argue for a full syntactic representation: stating that variable use of morphology during early the stages of learning is due to the difficulty of mapping an underlying representation onto surface form. Twelve beginning L2 learners, whose L1 was Japanese or Korean (East Asians), and nine L2 learners, whose L1 was various Indo-European languages (Non Asians) participated in an elicited imitation experiment, and additional twelve East Asians participated in a grammaticality judgment experiment. The targets included mono-clausal declarative sentences and wh-questions with the auxiliaries be, can, and do. In imitation, the prevalent error was auxiliary omission. Learners fronted whwords successfully. They also inverted auxiliaries successfully when they included them. Along with imitation, the results of grammaticality judgment support the argument that beginning L2 learners have difficulty producing the surface forms of the auxiliaries and tense morphology, not with forming the correct syntactic representation for questions.

This study investigates whether intermediate Japanese learners of English (JLE) show any variability in sensitivity to the overuse of 3rd person singular (3sg) -s and if they do, what the causes may be. The results of the experiment indicate that JLE do exhibit variability: JLE showed sensitivity to ungrammaticality caused by a discrepancy in person features between subjects and verbs. In addition, they were sensitive to number feature disagreement when the plurality of subjects was expressed syntactically, namely, by using the conjunction and (e.g., Tim and Paul), by the demonstrative these and by a numeral quantifier (e.g., these two secretaries). However, they were not sensitive to such disagreement when subjects were marked only by plural -s (e.g., The chefs). Based on these results, we suggest that the failure of JLE to use 3sg -s may not lie in the difficulty of subject–verb agreement, but in the detection of the number feature of sentential subjects. We suggest that intermediate JLE have problems both with the number feature at the lexicon/syntax level and with its morphological mapping at the level of morphology.

Two recent hypotheses which support the theory of full access to Universal Grammar have been proposed in order to account for variant data supplied by L2 learners. The Prosodic Transfer Hypothesis (Goad, White & Steele 2003) suggests that non-target-like behaviour by L2 learners is partially due to the differences in prosody between the L1 and L2 and the ensuing prosodic constraints; whilst the Missing Surface Inflection Hypothesis (Prévost & White 2000) proposes that problems are due to the learners’ variability in mapping abstract syntactic features onto morphological forms. This paper discusses a study of Japanese native speakers acquiring L3 German adjectival inflection in light of these two hypotheses. Data are provided from a written gap-filling task and from two oral production tasks. The results indicate stronger support for the MSIH.

The contribution of Spanish to the field of SLA continues to grow (Lafford & Salaberry 2003; Montrul 2004), and the need for good L2 Spanish datasets is becoming increasingly evident. In this paper we introduce a newly created database titled Spanish Learner Language Oral Corpus (SPLLOC), describing the rationale underlying the database design and methodology used for its construction. This project applying CHILDES tools to L2 Spanish follows successful creation of a collection of French L2 oral corpora (Rule et al. 2003), already available at www.flloc.soton.ac.uk. Creating a successful oral corpus is costly and available corpora are often built somewhat opportunistically from available material rather than designed in a balanced way to facilitate SLA research. The SPLLOC database has been designed to fill the existing gap in Spanish L2 resources and also to support a focused research agenda investigating learner development with respect to the verb phrase, clitic pronouns, and word order, from an interface perspective.