Sylviane Granger

Learner English on Computer

Computer learner corpus (CLC) research is still in its infancy. With roots both in corpus linguistics and second language acquisition (SLA) studies, it uses the methods and tools of corpus linguistics to gain better insights into authentic learner language. This volume is intended to open up a new field of research to a wider audience and includes articles on all aspects of computer learner corpus research, from the initial stage of corpus compilation to the final stage of pedagogical application. Although the main focus of the book is English as a Foreign Language, the approach could easily be applied to other foreign languages. The book is divided into three main parts. Part 1 is a general outline of learner corpus design and analysis. In the first paper, I first discuss the place of learner corpus research within corpus linguistics, SLA and ELT, and then outline the basic principles underlying learner corpus compilation and computer-aided linguistic analysis of learner data. In the second paper, Fanny Meunier presents the major linguistic software tools, focusing in particular on their relevance for interlanguage studies. In Part 2, the principles outlined in the first part of the book are put into practice through a series of case studies examining various aspects of learner lexis, discourse and grammar.The first three papers analyse learner lexis from a variety of angles: high frequency words (Hakan Ringbom), adjective intensification (Gunter Lorenz) and recurrent word combinations (Sylvie De Cock et al.). The following three papers focus on discourse phenomena: connectors (Bengt Altenberg and Marie Tapper), questions (Tuija Virtanen) and features of reader/writer visibility (Stephanie Petch-Tyson). The research in the last two papers is carried out on tagged corpora and highlights aspects of learner grammar on the basis of the frequencies of individual tags (Sylviane Granger and Paul Rayson) and tag sequences (Jan Aarts and Sylviane Granger). Part 3 shows how CLC-based studies can help improve pedagogical tools: EFL grammars (Doug Biber and Randi Reppen), dictionaries (Patrick Gillard and Adam Gadsby), writing textbooks (Przemyslaw Kaszubski) and electronic tools (John Milton). Implications for classroom methodology are highlighted in the last paper (Sylviane Granger and Chris Tribble). Interest in learner corpora is growing fast, with both SLA specialists and ELT practicioners beginning to recognize their theoretical and practical value. There is no doubt in my mind that we are on the verge of a learner corpus boom. This volume will serve its purpose if it helps would-be learner corpus builders and analysts to get started in a field of research which is certain to make an important contribution to foreign language learning and teaching.

Computer-aided Error Analysis

There is no doubt that Error Analysis (EA) has represented a major step forward in the development of SLA research, but equally, it is also true that it has failed to fulfill all its promises (see below). In this article we aim to demonstrate that recognizing the limitations of EA does not necessarily spell its death. We propose instead that it should be reinvented in the form of Computer-aided Error Analysis (CEA), a new type of computer corpus annotation.

Computer Learner Corpus Research: Current Status and Future Prospects

Despite a mere decade of existence, the field of computer learner corpus (CLC) research has been the focus of so much active international work that it seems worth taking a retrospective look at the research accomplished to date and considering the prospects for future research in both Second Language Acquisition (SLA) studies and Foreign Language Teaching (FLT) that emerge. One of the main distinguishing features of computer learner corpora – and indeed one of their main strengths – is that they can be used by specialists from both these fields and thus constitute a possible point of contact between them. The first three sections of this chapter are devoted to a brief overview of the main aspects of CLC research: data collection, methodological approaches, learner corpus typology, and size and representativeness. Sections 4 and 5 review the tangible results of CLC research in the fields of SLA and FLT.

Formulaic language in learner corpora

Formulaic language is at the heart of corpus linguistic research, and learner corpus research (LCR) is no exception. As multiword units of all kinds (e.g., collocations, phrasal verbs, speech formulae) are notoriously difficult for learners, and corpus linguistic techniques are an extremely powerful way of exploring them, they were an obvious area for investigation by researchers from the very early days of LCR. In the first part of this article, the focus is on the types of learner corpus data investigated and the most popular method used to analyze them. The second section describes the types of word sequences analyzed in learner corpora and the methodologies used to extract them. In the rest of the article, we summarize some of the main findings of LCR studies of the learner phrasicon, distinguishing between co-occurrence and recurrence. Particular emphasis is also placed on the relationship between learners’ use of formulaic sequences and transfer from the learners first language. The article concludes with some proposals for future research in the field.

A Model for Choked Flow-through Cracks With Inlet Subcooling

The leaks through steam-generator cracks are the subject of research carried out in cooperation between EDF and UCL. A model to predict the mass flow rate has been developed and has been successfully validated. The purpose of the paper is to present this model and to show some comparisons between the results and the presently available data. The model takes into account the persistence of some metastable liquid in the crack and the special flow pattern which appears in such particular geometry. Although the model involves the use of several correlations (friction, heat transfer, ... ), no adjustment of parameters against the data has been needed, neither in the single-phase of the flow, nor in the two-phase part.

The Learner Corpus: A Revolution in Applied Linguistics

A report on ICLE (the International Corpus of Learner English), a project that collects written work from, and analyses the usage of, advanced adult EFL learners

The use of collocations by intermediate vs. advanced non-native writers: A bigram-based study.

Abstract Phraseological competence, the use of (semi-)prefabricated expressions in language, is a major component of second language acquisition. Recent research focused on lexical bundles, i.e. recurrent contiguous strings of words, has highlighted quantitative and qualitative differences between native and non-native speaker use of these strings. Few studies, however, have investigated the development of phraseological competence as a function of degree of proficiency in L2. Relying on a methodology used by Durrant and Schmitt (2009: IRAL 47, 157–177) to compare native and non-native speakers, the present study identifies significant differences in the way in which intermediate and advanced learners use collocations. In particular, the intermediate learners tend to overuse high frequency collocations (such as hard work) and underuse lower-frequency, but strongly associated, collocations (such as immortal souls). The concluding section addresses the limits of the study and points to possible applications in foreign language teaching and automated scoring.

Categorising spelling errors to assess L2 writing

Based on a corpus of 223 argumentative essays written by English as a foreign language learners, this study shows that spelling errors, whether detected manually or automatically, are a reliable predictor of the quality of L2 texts and that reliability is further improved by sub-categorising errors. However, the benefit derived from sub-categorisation is much lower in the case of errors automatically detected by means of the Microsoft Word 2007 spell checker, a situation which results from Words limited success in detecting and correcting some specific categories of L2 learner errors.

Automated Retrieval of Passives from Native and Learner Corpora Precision and Recall

The aim of this article is to assess how successfully annotated corpora can be analyzed using automated retrieval methods. The perspective adopted is that of a corpus linguist, not that of a computational linguist or software developer. The investigation is part of the International Corpus of Learner English (ICLE) project, whose aim is to identify differences in the written English of advanced learners and native speakers (Granger 1993, 1996). The ICLE is a computerized learner corpus of argumentative writing by EFL (English as a foreign language) learners from fourteen different mother tongue backgrounds and currently contains approximately one and a half million words. The LOCNESS corpus (Louvain Corpus of Native English Essays) contains comparable writing from native English writers. One of the priorities of the project is to look for ways of achieving automatic profiling of learner texts, the underlying idea being that each learner variety is characterized by its own matrix of lexical, grammatical, and discursive features. Within this framework, linguistic annotation tools have a crucial role to play, and so it is paradoxical that comparatively few corpus linguists are making use of the good linguistic annotation software that is now widely available. Most analyses are still based on raw corpora in conjunction with text retrieval software. Reluctance

Error annotation systems

and says only that this part of the learner utterance is unidiomatic, confl ating an implicit target hypothesis with an error tag (the annotator is only able to know that this expression is unidiomatic if he or she knows a more idiomatic expression). Different target hypotheses are not equivalent; a target hypothesis directly infl uences the following analysis. The Falko corpus consistently has two target hypotheses – the fi rst one deals with clear grammatical errors and the second one also corrects stylistic problems. The need for such an approach becomes clear in (11). The learner utterance in (11) contains a spelling error . The two occurrences of dependance have to be replaced by dependence . From a more abstract perspective, the whole phrase Dependence on gambling sounds unidiomatic if we take into account that the learner wants to refer to a specifi c kind of addiction. Similarly, dependence on drugs appears to be a marked expression as opposed to drug addiction . An annotation that wants to take this into consideration has to separate the description into the annotation of the spelling error and the annotation of the stylistic error in order not to lose one of the pieces of information. Example (12) illustrates this. The examples in this section show how important the step of formulating a target hypothesis is – the subsequent error classifi cation critically depends on this fi rst step. In order to operationalise the fi rst step of the error annotation , one can give guidelines for the formulation of target hypotheses, in addition to the guidelines for assigning error tags, which also need to be evaluated with regard to consistency (see Section 2.6 ). The problem of unclear error identifi cation has been discussed since the beginning of EA. Milton and Chowdhury ( 1994 ) have already suggested that sometimes multiple analyses should be coded in a learner corpus. If (11) Dependance on gambling is something like dependance on drugs (...) (ICLE-CZ-PRAG-0013.3) (12) LU Dependance on gambling TH 1 Dependence on gambling TH 2 Gambling addiction (10) LU it sleeps inside everyone from the start of being TH 1 it sleeps inside everyone since birth TH 2 it sleeps inside everyone from the beginning TH 3 it sleeps inside everyone UNIDIOMATIC 9781107041196c07_p135-158.indd 145 6/11/2015 1:48:09 PM LÜDELING AND HIRSCHMANN 146 the target hypothesis is left implicit or there is only one error analysis , the user is given an error annotation without knowing against which form the utterance was evaluated. In early corpora (pre-multi-layer, pre-XML) it was technically impossible to show the error exponent because errors could only be marked on one token. In corpora that use an XML format it is possible to mark spans, and target hypotheses are sometimes given in the XML mark-up. Only in standoff architectures, however, is it possible to give several competing target hypotheses. Examples of learner corpora with consistent and well-documented (multiple) target hypotheses are the Falko corpus, the trilingual MERLIN corpus (Wisniewski et al. 2013 ) or the Czech as a Second Language corpus (Rosen et al. 2014 ).