This book presents a new theory of grammatical categories - the Universal Spine Hypothesis - and reinforces generative notions of Universal Grammar while accommodating insights from linguistic typology.
EDITORS: Plonsky, Luke and Schierloh, Maren TITLE: Selected Proceedings of the 2009 Second Language Research Forum SUBTITLE: Diverse Contributions to SLA SERIES TITLE: Cascadilla Proceedings Project PUBLISHER: Cascadilla Press YEAR: 2011
Christos Pliatsikas, School of Psychology, University of Birmingham, UK
This volume presents a selection of the papers that were presented at the 2009 Second Language Research Forum (SLRF). It contains a diverse array of methodologies and theoretical frameworks that investigate a variety of issues in second language (L2) research. A brief summary of each of the papers is presented in this review, followed by a general evaluation of the volume.
The volume starts with a paper by Jennifer Baker and Margaret Quesada, who investigate how L2 learners of Spanish make use of adverbials in order to interpret and select Spanish preterit and imperfect. Adverbials appear important to Spanish learners with an L1 such as English, which does not provide morphological marking of this aspectual distinction. The authors compared English speaking intermediate and advanced L2 learners of Spanish to native speakers of Spanish in an offline task. This included 10 cloze passages in which the participants had to select between the imperfect or preterit form of the target verbs. Half of the passages included adverbials, whereas in the other half, the context would bias towards the selection of preterit/imperfect. Baker and Quesada revealed that L2 learners were more influenced by adverbials in their selection of aspect, compared to native speakers, and that this trend was more profound in the case of intermediate learners; when adverbials were absent, they were least likely to select the expected form. The authors attribute this finding to classroom L2 instruction, and propose the limitation/elimination of adverbials in order for students to develop their sense of aspect in L2.
Kimberly Geeslin and Aarnes Gudmestad explore variation in subject expression in L2 Spanish via a sociolinguistic approach. Subject in Spanish can be expressed in various ways, such as null subject or various forms of overt subject. The authors investigated the forms chosen by L2 learners in sentences with ambiguous verbs or contexts (whether the referent switches), as well as whether their performance follows discourse-level variables, such as Referent Cohesiveness (i.e. distance and function of the previous mention of a referent) and Perseveration (i.e. continuity of the form across mentions of the referent). They interviewed native and non-native speakers of Spanish, and monitored for the subject forms that were used (e.g. pronouns, null pronouns, NPs, etc.). The results revealed that increased distance from the original mention affects the choice of the subject form of the referent, and also that the type of the chosen subject (e.g. null vs. overt) perseveres and follows up on the original type of subject. Importantly, these effects were common in native and non-native speakers of Spanish, proving that non-native speakers are equally responsive to discourse-level variables.
Turning to the domain of syntactic processing, Masahiro Hara investigates L2 processing of syntactic gaps, and how this is mediated by computational load. He tested advanced L2 learners of Japanese, a language that permits syntactic scrambling, with Korean or English as an L1. In a self-paced reading (SPR) probe task, Hara presented sentences that were either canonically ordered (SOV) or scrambled (OSV), with the latter posing a gap at the canonical position of the object. These sentences were followed by a probe word and reaction times (RTs) were collected. This task revealed that all groups reactivated the displaced NP driven by the syntactic gap. Additionally, Hara used another SPR task, where he monitored RTs in sentences that were either canonically ordered or included short or long scrambling. In this task, Korean learners showed evidence of processing the syntactic gap in the short-scrambling condition only, whereas English learners did not provide any evidence for gap processing. Based on these findings, Hara linked L2 performance to cognitive resource-limitations, since Koreans processed the gap in the condition with a moderate computational demand; on the other hand, Hara suggested that the absence of gap effects on English L2 speakers points to L1 transfer, as English does not feature scrambling.
Nam-Sook Jeong focuses on meaning negotiation in L2 via a computer-mediated setting. He tested 24 Korean L2 learners of English, which he split according to their proficiency (e.g. higher/lower), and created participant pairs of two types, namely homogenous (e.g. higher/higher or lower/lower proficiency), and heterogeneous (e.g. higher/lower). Participant pairs were instructed to conduct a conversation on MSN Messenger text-chat in English, where three tasks were performed in each of 11 sessions: a jigsaw task, a decision-making task and an open-ended task. The topics of the chat were pre-defined (e.g. marriage, cleaning, Christmas). The content of the electronic discussions was coded according to standard models of Meaning Negotiation. The results suggested that the type of task affected the quantity of meaning negotiation, as the jigsaw task produced greater production than the other two. Additionally, the two types of pairs did not differ in terms of the amount of meaning negotiation they produced, albeit in qualitative terms. Overall, the homogenous group produced more effective collaborative learning.
Kitaek Kim investigates processing of the copula ''be'' by Korean L2 learners of English. ''Be'' has been shown to be overgenerated by L2 learners (e.g. She is go home), and this has been suggested to be a topic marker or a functional category (such as tense). Kim hypothesized that, in the case of Korean learners, the overgenerated ''be'' starts off as a topic marker, which develops into a functional category as a function of proficiency level in L2. Kim studied 23 beginner students of English divided into three proficiency groups. The participants were given 20 topics and were asked to produce a writing sample in English within 15 minutes. The collected scripts revealed that the participants of lowest proficiency produced the overgenerated ''be'' more frequently compared to the other groups, and also as a topic marker. This group was identified as being in a topic-prominent (TopP) stage influenced by L1 typology. The most proficient group produced the overgeneralised ''be'' as a verbal inflection, and was identified as being in a subject-prominent (SP) stage closer to L2 typology. Kim proposed a continuum that describes the development of ''be'' in learners' interlanguage that goes from TopP to SP stages.
Jihye Lee investigates the acquisition of direction-giving in Korean L2. She tested 30 learners of Korean, of three proficiency levels, and 6 native Korean speakers. The participants saw several maps and were auditorily instructed to give directions to a specific point on each map. They had 1.5 minutes for each map and their responses were recorded. The data suggested that beginning and intermediate learners rely on bare imperatives for their directions, whereas advanced learners preferred bi-clausal imperatives. Additionally, advanced learners and native speakers were found to use more mitigation in their directions, such as external mitigation (i.e. reason) or syntactic mitigation (i.e. if-clauses). Finally, advanced learners were found to be native-like in their use of proper speech levels according to the recipient of their directions (e.g. friend vs. someone superior), while all learners were found to overuse polite and honorific suffixes compared to native speakers.
Eleonora Luzi starts off in a constructional framework to investigate whether SLA takes place through the acquisition, strengthening, and exposure to language constructions. She tested processing of Italian left-dislocation, where an element is moved to the beginning of a phrase in order to be topicalised. Left-dislocation can involve the subject or the direct object, where agreement with the verb is compulsory, and the indirect object or any other circumstantial, where agreement is not required. Luzi wanted to investigate whether the various types of left dislocation are acquired in a successive fashion. She tested 64 L2 learners of Italian divided into two proficiency groups, B1 (low) and C1 (high). Discourse data were collected from both groups. These extracts revealed that both groups produced all types of dislocations, apart from B1, which did not produce indirect object ones. Therefore, the construction complexity did not seem to influence the learners' performance. However, the fact that any incorrect dislocations were absent in the C1 data, suggests, according to Luzi, a continuum in the development of this construction, from the ''rough'' grammar stage, which involves constructualisation, to the ''fine'' grammar stage, with regular production of constructions.
A. Kate Miller is concerned with the constraints that underlie L2 syntactic processing, and whether these are due to gaps in L2 grammar or to L1/L2 processing differences. Miller conducted a cross-modal priming experiment in which sentences in French were visually presented to intermediate and advanced American L2 learners.nThe sentences included a structural gap created by the displacement of a constituent further up the sentence. An auditory probe that either matched or failed to match the antecedent (creating a 2x2 design) was presented either at the site of the gap or at an earlier control position. The task for the participants was to make an animacy decision on the probe. If L2 processing is structure-based, then it should permit the reactivation of the antecedent at the gap position, and facilitate the response to the matching probe, compared to the non-matching one. This was not proven the case for any of the L2 groups, but the advanced learners showed faster reaction to the probe in the gap position compared to the control position. Miller concluded that advanced learners were more sensitive to structure, and that they are potentially able to construct full syntactic representations in L2.
Lisa Pierce and Tania Ionin are interested in the acquisition of articles in L2 English, especially in cases where the L1 of speakers lacks articles. In these cases, L2 learners tend to omit articles or substitute them inappropriately. Recent theories have attributed these errors to constraints imposed by the L1 prosodic system. The authors tested two groups of L2 learners with L1s that do not have articles but that do differ in their prosodic system, namely, Korean (16) and Mandarin Chinese (14) speakers. They administered two tasks; first, an acceptability judgment task with sentence pairs, where the second sentence included an NP with either an article present (correct) or missing (incorrect). Both groups were proven poor at detecting missing articles, with the Koreans being more accurate. The participants also did a transcription of spoken sentences including definite and indefinite articles in various sentence positions. Both groups demonstrated errors in the form of omissions or article substitutions, with the Koreans performing better. The authors concluded that article acquisition is independent of proficiency level, and that the Koreans’ advantage may be attributed to the Korean rhythmic system, which facilitates perception of articles in L2 English.
Claire Renaud investigates feature selection in the acquisition of L2 French. She tested American learners of French at three proficiency levels, as well as native speakers of French, in two online SPR tasks. In both tasks, a context sentence preceded the target sentence, which was segmented into six parts, and the participants had to respond as to whether the target sentence was a good follow-up to the context sentence. The target sentences contained auxiliary verbs that either agreed in person and number to the subject (grammatical), or agreed in person only (ungrammatical). In the second task, the ungrammaticality was demonstrated by sentences containing feminine participles in a masculine context (and vice versa), which were compared to grammatical sentences. RTs were collected from the auxiliary-containing segment and the following ones. The results revealed that learners had good knowledge of the grammatical forms of auxiliaries, even at low proficiency levels. On the other hand, learners of lower proficiency were more accepting of feminine forms in a masculine context, suggesting a failure in their interlanguage grammar. Advanced learners were native-like in this task. In terms of RTs, all groups revealed similar effects of ungrammaticality. Based on these findings, Renaud suggested that an L2 parser guides the selection of the relevant features for the analysis of forms even before lexical encoding.
Natsue Sugaya investigated L2 processing of Japanese inflection on the premises of dual system theories (rule learning vs. associative learning). Japanese is an agglutinating language with a complex inflectional system, including two types of regular verbs, as well as irregular verbs. Sugaya tested 39 Mongolian and 2 Korean L2 learners of Japanese of various proficiency levels in two offline cloze tests where verbs were missing from short sentences. In the first task (nonce verb test), the participants had to choose among three candidates, which were regularly inflected nonce verbs in Japanese, in four inflectional forms. In the second task (real verb test) the participants were given the uninflected form of a real verb and had to fill the sentence gap with the correct inflection. Sugaya showed that the more proficient participants were more successful in applying morphological rules to real and nonce verbs. Also, it was found that the nonce verbs that resembled real ones in form were responded to more accurately, giving evidence for associative learning. Sugaya concluded that learning of complex inflectional systems in L2 requires both rule learning and exemplar-based learning.
Malena Weitze, Jeremiah McGhee and C. Ray Graham examine the issue of L1 effects in L2 acquisition, especially the acquisition of grammatical morphemes in L2. They tested 760 learners of English of several L1 backgrounds in an elicited imitation task. The participants were given sets of short auditory sentences which they had to repeat within a given time. The sentences included morphemes that have been shown to be acquired either early (e.g. -ing, -s (plural) and articles) or late (e.g. regular past -ed and 3rd person singular -s), and the participants were scored according to how accurately they produced the forms. The results showed that L2 learners with an L1 that resembles English in morphological structure (e.g. Spanish, Portuguese) were more accurate with early-acquired forms than late ones, whereas Japanese and Korean learners of English showed the opposite pattern. The authors suggested that L1 affects the order of acquisition of grammatical morphemes; however, their results were inconclusive because Chinese learners performed differently than the speakers of morphologically similar languages (e.g Japanese and Korean).
German Zarate-Sandez looks at processing of syllabic structure by L2 learners of English. More specifically, he tested phonological processing of diphthongs (e.g. /i/ in the word ''avión'') and hiatuses (e.g. ae in the word ''traer'') in L2 Spanish in order to investigate whether native and non-native speakers of Spanish follow the same strategies during syllabification of words with vowel complexes. He tested 167 English learners of Spanish in three proficiency levels (i.e. elementary, intermediate, advanced), and 22 native speakers of Spanish as the control group. The task consisted of a questionnaire with 36 words containing vowel sequences, and 17 words were L1-L2 cognates (e.g. ''historia''). The participants had to cut each word into syllables, and they were scored according to the number of the diphthongs they recognised. The results suggested that L2 learners tended to treat hiatuses in Spanish as diphthongs, contrary to native speakers, and this effect was largely driven by the influence of the cognate stimuli. The proficiency level did not appear to interact with their performance.
This is a highly informative volume, containing papers on a range of experimental fields from both established and emerging researchers; there are studies on sociolinguistic aspects of L2 (Geeslin and Gudmestad, Jeong, Lee), morphological (Sugaya, Weitze et al.), morphosyntactic (Baker and Quesada, Kim, Renaud) and phonological (Zarate-Sandez) processing, as well as syntactic aspects of L2 processing (Hara, Miller, Luzi, Pierce and Ionin). The presented experiments were conducted with several online (e.g. self-paced reading) and offline (e.g. cloze tests) tasks.
Of particular interest is the paper by Hara, who dismissed earlier suggestions that L2 learners do not process abstract syntactic structures, and attributed the absence of native-like effects to increased computational demands in L2 syntactic processing. This suggestion was in accordance with the findings by Miller, who showed evidence that L2 learners, at least advanced ones, resort to native-like structure-based syntactic processing. Additionally, and quite interestingly, several studies in this volume appear to lower the importance of proficiency level in areas such as the L2 processing of phonology in Spanish (Zarate-Sandez), auxiliary verbs in French (Renaud), and article acquisition in English (Pierce and Ionin), whereas in more syntax-related experiments proficiency does appear important (Baker and Quesada, Luzi, Miller). This distinction suggests that advanced proficiency level is not necessary for efficient L2 processing, but may be beneficial for specific language domains, especially syntax.
As such, the volume provides a good summary of the directions of ongoing research in second language acquisition (SLA). The volume would benefit from grouping papers into sections according to the subfield to which they are relevant instead of the current, alphabetical order. Several papers are highly relevant to each other, so it would be good if they were presented together. This would make the volume more coherent and assist the readers in comparing and contrasting relevant papers.
This volume is highly recommended to researchers in the fields of SLA and psycholinguistics, as well as to students or any other parties that require an up-to-date knowledge of issues in current SLA research. The diversity of this volume guarantees that researchers from a variety of different fields in psycholinguistics will find some interesting experimental findings and suggestions, as well as use them as a starting point for future research.
ABOUT THE REVIEWER
ABOUT THE REVIEWER:
Christos Pliatsikas received his PhD from the Department of Clinical
Language Sciences, University of Reading. His research interests are in the
area of psycholinguistics and neurolinguistics. He has used behavioural and
neuroimaging (fMRI) methods for the study of online morphological and
syntactic processing, especially by late second language learners. He is
currently employed as a Research Fellow by the School of Psychology,
University of Birmingham