J Charles Alderson
Lancaster University, UK
Università di Verona, Italy
University of Hawai'i at Manoa, USA
University of Edinburgh, UK
Diagnosing strengths and weaknesses in S/FL reading: What do SLA and Testing have to offer?
J Charles Alderson
Language test developers and language testing researchers have traditionally been concerned with developing and researching high-stakes proficiency tests, with more emphasis on testing reading, writing and listening than on tests of speaking. SLA researchers, on the other hand, have concentrated on devising small-scale tasks which will elicit spoken and written performances to be analysed in detail. It can be argued that SLA researchers are more interested in understanding language ability and development at the micro level, and that language testers have more concern for establishing reliable and valid measures of overall proficiency. Language teachers have rather been forgotten and left to their own devices to assess their learners’ strengths and weaknesses as well as their progress. However, language testing researchers’ recent interest in diagnostic testing claims to redress this imbalance. But diagnostic testing requires a better understanding of language abilities at a less general level than hitherto, and this has posed challenges to testers to define their diagnostic constructs both theoretically and operationally.
The aim of the SLATE group (Second Language Acquisition and Testing in Europe - see www.slate.eu.org) is to bring together language testing researchers and researchers of second language acquisition to create synergies to our mutual benefit. In theory, SLA should be able to offer insights into the construct of reading in a S/FL, and testing ought to be able to deliver the results of tests of S/FL reading proficiency abilities which should be based on theoretical insights and which can inform such theory. But does this happen in practice? In this paper I ask whether there is such a thing as language-testing-informed SLA, and / or SLA-informed language testing. I will report on progress in three inter-related research projects, co-funded by the UK's Economic and Social Research Council, The Finnish Academy of Sciences, The PISA Project of the OECD, and Pearson Assessment.
Beyond canonical order: the acquisition of marked word orders in Italian as a second language
Learners can go a long way in making themselves understood by mapping conceptual arguments such as agent, event and patient straight onto canonical word order. However, in order to maximise communicative intentions they need to go beyond it. This paper will discuss such notions as canonical order, functional assignment and marked word orders within the framework of Manfred Pienemann’s Processability Theory. This SLA theory seems particularly suited to do so, having recently expanded its scope in several directions. Most important among these are, first, the addition a discourse-pragmatically motivated syntactic component (Pienemann, Di Biase & Kawaguchi 2005) to the orginal morpho-syntactic one (Pienemann 1998), and secondly, its typological validation on new languages.Italian is a headmarking, prodrop language located towards the less configurational end of the typological spectrum, characterised by a rich morphology and a flexible syntax which is highly sensitive to pragmatic and discourse choices. For these typological characteristics Italian seems particularly suited for discussing the notion of canonical word order, and then showing how learners acquire the skills to free up its rigidity and learn to topicalise and focalise different elements of their message in an unequivocal (that is, grammatically accurate) manner for the listener.
Drawing on work on Italian L2 done chiefly by and with Di Biase on declarative sentences (cf. e.g. Di Biase & Kawaguchi 2002; Di Biase 2007; Bettoni, Di Biase & Nuzzo 2009) and Ginelli on interrogatives (cf. e.g. Ginelli 2010; Bettoni & Ginelli in press), the paper will show the advantages of reconceptualising the 1998 and 2005 main tenets of Processability Theory (cf. Bettoni & Di Biase, in press; Di Biase & Kawaguchi, in press) in accordance with recent progress in its two feeder disciplines dealing with language production (i.e., Levelt’s Model) and linguistic knowledge (i.e., Bresnan’s (2001) Lexical Functional Grammar). With any L2, the main task for the learner is the acquisition of its lexicon, which is by far the greatest source of diversity, not only linguistically but also culturally. This entails learning for every new word the full bundle of its features, including, at the lemma level (Levelt, Roelofs & Meyer 1999), all the information needed to encode it grammatically and combine it with other compatible lemmas. If the sentence in which words are placed is declarative, simple, active, affirmative, minimally presuppositional, and pragmatically neutral, the speaker’s choices are limited, and encoding operations and their structural outcomes mostly obligatory. So when grammatical encoding begins misunderstandings are few even if words exhibit minimal diacritic features and values, and are rigidly organised in constituent structure according to the canonical order of the target language. But in order to enhance expressiveness and impose their own perspective on events, states and evaluations, speakers may wish to make marked choices at the discourse-pragmatic level. A second source of great diversity is the way in which different languages can handle – or prefer to handle – the consequences of these choices, and the means they use to do so. As grammatical encoding proceeds, thanks to the gradual building up of functional structure, lemmas acquire richer features (inflectional morphology), and constituents can be placed in a freer linear order. All this entails moving from default solutions, simpler to process, to nondefault ones, cognitively more costly to the learner but easier on the listener.
Understanding instructed SLA: Constructs, contexts, and consequences
John M. Norris
Constructs are generally understood to be idealized and unobservable concepts, or groups of inter-related concepts (often referred to as models), used to explain particularly important phenomena in various domains of scientific inquiry. Within instructed Second Language Acquisition (ISLA), constructs provide the basic constitutive elements underlying theories of how languages are, or should be, taught and learned under educational circumstances. “Focus-on-Form”, “socio-pragmatic competence”, “implicit linguistic knowledge”, “academic speaking ability”, and “the ZPD” are examples of some of the many kinds of constructs that play a role in guiding how we think, talk about, and research language education in more or less theoretically informed ways. While familiar, constructs are also tricky--they emerge out of particular intellectual milieus for particular reasons, they are themselves socially constructed, and they may be de-constructed by those who use them in both research and applied instruction. In this talk, I will consider how ISLA constructs are developed and put into practice by individuals as well as research and professional communities. In particular, I will highlight the complicating challenges of constructs that are intended to inform the “I” in ISLA, including the defining role played by contexts of instruction, and the eventual consequences for teaching and learning that derive from ISLA research. By exploring several well-known examples, I will demonstrate how constructs may either help to illuminate or further confuse our understandings of instructed learning, as they are defined, operationalized, analyzed, communicated, and interpreted by researchers and educational practitioners. Finally, I will propose a modest agenda for the ISLA research community, with the goal of improving the validity of interpretations that derive from our constructs and, by extension, the meaningfulness of our work for language instruction.
SLA as bilingualism: or, it’s time to see the forest for the trees
Focusing on recent studies of language and cognition in very advanced L2 speakers, I will address the general question of whether it is still productive to regard SLA as a separate research field rather than a type of bilingual development.
Since Coppieters’s 1987 pioneering paper on near-nativeness, research on ultimate attainment has moved a long way from early descriptive studies towards theoretically motivated and testable predictions about age effects on L2 final outcomes. Much of this progress has been made possible by two factors. On the one hand, the scope of research has broadened from the exclusive use of L1 monolinguals as the reference point for L2 speakers to a comparison between adult L2 speakers and other bilingual groups (i.e. child bilingual L1 acquirers, early and late consecutive child L2 learners, L1 speakers undergoing attrition). On the other hand, the increasing interdisciplinarity of research on bilingual language development has allowed the expansion of research scope and the integration of theoretical insights and methodological tools from different fields - not just linguistics but also neuroscience, cognitive psychology, evolutionary biology.
Despite these developments, many SLA researchers still engage in inward-looking ideological debates on which linguistic theory is best or whether a cognitive or a sociocultural perspective is the most promising, I propose that a big tent approach to SLA within the study of bilingualism may be much more insightful than research confined (and disseminated only) within the boundaries of a narrowly-defined SLA field.
Call for papers