I have been working on applying Complexity Theory and Activity Theory to measuring / modelling second language development in online language learning (proficiency development, grammar, corrective feedback) and now to concept-based instruction. My secondary research interest is the research and development of online language courses. I am coordinating a project whose aim is the telling of the story and the stories of the German immigrants to Waterloo Region in Ontario, Canada.
Skills and Expertise
Academic WritingLearningE-LearningOnline LearningLanguage TeachingApplied LinguisticsTechnology Enhanced LearningEnglish LanguageForeign Language LearningSecond Language AcquisitionLanguage LearningSketchingArtificial IntelligenceTranslationSpeech and Language ProcessingCognitive LinguisticsNatural Language ProcessingWritingCorpus LinguisticsDiscourse StudiesResearch And DevelopmentTranslation StudiesSLAEditingLinguisticsText LinguisticsProject EvaluationInstructional DesignForeign LanguagesSyntaxGrammarMorphologyComplexity TheoryCorpus CompilationReplicationActivity TheoryComplexity ScienceParsingDigital GamesComplex Adaptive SystemsGermanIntelligent Tutoring SystemsComputer Assisted Language LearningMaterials DesignSociocultural Theory of Learning
Jan 1996 - Jul 2001
Field of study
- Language Engineering
Nov 1985 - Jul 1990
Pädagogische Hochschule Leipzig
Field of study
- Teaching Degree for German and Russian Languages and Literatures
Research Item (47)
Processes of learner-‐computer interaction are complex because a number of actors—learners, instructors, and L1 speakers—and components—computational hardware and software—participate in them and there are a community and multiple components that influence them, e.g., other learning materials, linguistic artefacts, and educational institutions, in their environment. Language-‐learning processes are complex because they involve many internal and environmental variables and components, such as proficiency, aptitude, motivation, and (online) learning environments and materials. These variables are not stable; they interact with one another and are therefore subject to change. To capture the interaction and interdependency of actors and components and their variables better, we describe language-‐learning processes as dynamic systems. Hence we can say that in the dynamic change of such a system, its variables co-‐adapt continuously. It is very important at this stage to reiterate that CAS are, essentially, complex processes. In other words, when we say CAS or just system, we mean the learning— not the learner, we mean the second-‐language development (SLD)—not a structure of acquired and applied knowledge, we mean the learner-‐computer interaction—not the software nor the computer, and we mean the online gaming—not the digital game. How did CAS come into Applied Linguistics and CALL? Since the late 1980s, we witnessed a proliferation of research approaches, concepts, and metaphors of complexity well beyond mathematics and the natural sciences, from where they originated. Books like Gleick's (1987) Chaos: Making a New Science popularized research on complex and (ostensibly) chaotic systems and made it accessible also for scholars in the social sciences and humanities. Over the three decades since, Complexity Theory, Dynamic Systems Theory, and Chaos Theory—related theories that discuss complex processes with a slightly different
‘Sometimes maligned for its allegedly behaviorist connotations but critical for success in many fields from music to sport to mathematics and language learning, practice is undergoing something of a revival in the applied linguistics literature’ (Long & Richards 2007, p. xi). This research timeline provides a systematic overview of the contributions of computer-assisted language learning (CALL) to the role, nature, and development of individual practice in language learning. We focus on written language practice in Tutorial CALL, corrective feedback and language awareness-raising in Intelligent CALL (ICALL), and individualization of the learning process through tailoring of learning sequences and contingent guidance.
The application of techniques from Artificial Intelligence (AI) to CALL has commonly been referred to as Intelligent CALL (ICALL). ICALL is only slightly older than the CALICO Journal, and this paper looks back at a quarter century of published research mainly in North America and by North American scholars. This 'inventory-taking' will provide a basis for establishing the place, context, and direction of ICALL in the CALICO Journal, CALICO organization, and CALL research and development in general. As a point of reference, I will use a list of desiderata for ICALL (Oxford, 1993) from the outside perspective of Second Language Acquisition (SLA).
The choice of grammatical framework in ICALL – the branch of CALL that applies artificial intelligence techniques – has important implications for both research and development. Matthews (1993) argued for one ‘that potentially meshes with SLA (second language acquisition)’ (p. 5) and sketches three criteria that facilitate the crucial decision of selecting a grammatical framework for an ICALL system: computational effectiveness, linguistic perspicuity and acquisitional perspicuity. We will use Matthews' three adequacy criteria to review recent research in construction grammar and propose its application in ICALL, particularly in projects which involve the building of a student model. Such a student modeling project – Mocha – will be sketched briefly to provide a concrete context for the conceptualisation and implementation of construction grammar. This grammatical framework has the potential to help overcome some challenges in ICALL and to facilitate a more thorough analysis of learner language in context and thus improve our knowledge about language learning processes.
The first issue of the 32nd volume of CALICO Journal marks a new beginning. This issue is the first available exclusively through Equinox Publishing (http://www.equinoxpub.com/home/journals/calico). The CALICO Journal was first published by CALICO in 1983 – which makes it the publication on CALL with the longest pedigree. The Journal appeared online as well as print form and then migrated to online only in 2007 (issue 25.1). Since 2011, it has been produced and published on the Open Journal System (OJS). It became apparent to us that the 'in-house' publication by CALICO would not be sustainable over the long term and we began to explore other options in early 2013. The resulting partnership with Equinox Publishing was sealed the following summer. During the transition period over the last couple of months, the 31-year archive of the Journal has been transferred to the OJS at Equinox. The databases with authors, reviewers, subscribers, and registered readers are also fully functional on the new OJS server. We are very grateful to colleagues at Equinox for the commitment to the CALICO Journal, their immense work during the transition, and their professional support. We have had every indication that the members of CALICO, our authors and readers, and the scholarly CALL community at large will reap substantial benefits from this new beginning. The Journal will continue to contain high-quality research articles on CALL and reviews of learning technologies and relevant books, which will appear three times a year in January, May, and September. Reviews will remain open access at http://www.equinoxpub.com/journals/index.php/CALICO; and research articles will become open access after a three-year embargo period. With the help of Equinox, this research and the reviews will reach a much
Currently there is a push toward offering more language courses online because they can provide students with new forms of social and learning interaction, widen their access to education, and offer an individualized learning experience in large classes. Little research exists examining how students transition between online and on-campus language courses and what effect this has on students’ academic success. We analyzed student data from language courses with online and on-campus counterparts, to better understand the extent to which online learning enables students to meet their intended learning outcomes. Our goals were to establish how the medium of learning (online vs. classroom) impacts students’ academic success, to identify patterns in the students’ transitioning between online and on-campus courses, and to gather evidence-based information about students’ course choices and their decisions about online vs. on-campus. In addition to the comprehensive statistical analysis of learner data, we conducted a qualitative analysis of language biography surveys and semi-structured interviews with students currently enrolled in the German program. While the statistical data provided a birds-eye view of student trajectories over 10 years, the surveys and interviews gave us in-depth information about individual learning trajectories and students’ curricular decisions.
Recent research in digital game-based language learning has been encouraging, yet it would benefit from research methods that focus on the gaming processes and second-language development (Larsen-Freeman, 2015) rather than learner/player reflection or individuals’ beliefs about the validity of gameplay. This has proven challenging as research methods are needed, which provide insight into the gameplay experiences and its many factors. Having the gameplay experience occur extramurally is desirable, but makes the direct observation of the gamers/learners’ activities by a researcher difficult. For this reason, we suggest approaching digital game-based language learning through complex adaptive systems research (Larsen-Freeman & Cameron, 2008a) and employing Dörnyei’s (2014) retrodictive qualitative modeling to capture the complex synchronic and diachronic variability of the learners/players and their individual nonlinear gaming trajectories with requisite data density and over a considerable period of time. This paper draws on a study examining language learners playing the online role-playing game World of Warcraft over four months. We will focus on the data collection in this observational study and the methods of analysis of a complex adaptive system, which helped to better understand the role of extramural digital gaming for the purpose of second-language development.
Complexity-theoretical approaches in Applied Linguistics are relatively new, but they hold great promise as an integrative (meta-)theory, provide new ways of hypothesizing about and conceptualizing the complex phenomena of language use and (second) language development, and also require different data gathering and analytical methods. This chapter sketches first the main tenets of a theory of complex adaptive systems (CAS) as it applies to technology-mediated language learning. The investigation of complex adaptive systems in computer-assisted language learning (CALL) is in its infancy; a selection of representative CAS studies in CALL will be discussed. As will become apparent, the main challenge for future research in this area is the design and application of robust commensurate research methods. General facets of a CAS methodology in CALL will be outlined and a general direction for future investigation will be given.
Question - If I meant to ask students about inflectional morphemes, and articles, How can I attribute errors to over-generalization or avoidance strategy?
re avoidance: The easiest might be to give them two tasks: one in which they can easily avoid certain constructions, one in which they cannot. An unconstrained text production task (Write X words on topic Y.) and a written translation (a short text or text passage) can both be embedded in a meaningful context and should give you the right data.
Over-generalization is more difficult. I would set up a stimulated recall scenario. Students do a suitable language-learning task. Immediately or very soon after the researcher goes through the task output with them. Asking questions for example on verb morphology and article use (referring to both accurate and inaccurate examples), such as why did you use this ending? why did you not use an article here? Both the text production and the stimulated recall are recorded, coded, and analyzed.
Question - Do you happen to know some useful publications concerned with "Error analysis", introduced by S. Pit. Corder?
You might want to have a look at the following papers by Corder, Selinker, and Schachter. There is, of course, a wealth of others. Particularly interesting might be
Interlanguage: Forty Years Later
Editors: ZhaoHong Han and Elaine Tarone
Amsterdam: John Benjamins
Corder, S.P. (1974). The Significance of Learners' Errors. In J. C. Richards (Ed.), Error Analysis: Perspectives on Second Language Acquisition (pp. 19-27). London: Longman.
Corder, S.P. (1974). Error Analysis. In J. P. B. Allen & P. Corder (Eds.), The Edinburgh Course in Applied Linguistics. Volume 3 - Techniques in Applied Linguistics (pp. 122-131). London: Oxford University Press.
Corder, S.P. (1975). Error Analysis, Interlanguage and Second Language Acquisition (Survey Article). Language Teaching and Linguistics. Abstracts, 8, 201-218.
Corder, S.P. (1981). Error Analysis and Interlanguage. Oxford: Oxford University Press.
Schachter, J. (1974). An Error in Error Analysis. Language Learning, 27, 205-214.
Selinker, L. (1974). Interlanguage. In J. C. Richards (Ed.), Error Analysis : Perspectives on Second Language Acquisition (pp. 31-54). London: Longman.
Selinker, L. (1992). Rediscovering Interlanguage. London: Longman.
With this thematic issue on replication studies in CALL, we would like to draw attention to the importance of looking back both in CALL research and the development of learning technology. Replicating CALL research and evaluating (commercialized) language-learning tools, software, systems, and environments afford engagement with past findings, outcomes, and results and, through this engagement, yield new insight into current and future language learning in technology-rich contexts. Caws and Heift (in press) argue that, although research and evaluation can be distinguished, in CALL, this distinction is not always clear cut. For example, researchers rely on the evaluation of learning outcomes, task designs, facets of complex learning processes in their studies; and software evaluators ground their reviews in research findings and often apply methods akin to those in applied-linguistics research. In this editorial , we will first introduce the theme of this issue – replication studies – and then announce CJ's new conceptualization of its section on evaluation – the Learning Technology Reviews (formerly known as software reviews).
In language courses, students submit more and more outcomes of their text production tasks to discussion boards, dropboxes, blogs, and wikis. We are testing automating certain components of the feedback learners need to receive, by scoring the proficiency level as displayed in the individual student's text. Proficiency development manifests itself through a nonlinear increase in textual complexity, fluency, and accuracy over time. A detailed multi-faceted proficiency score for written texts, which provides information on the diversity and sophistication of the vocabulary and grammatical constructions, lexical and syntactic accuracy, and appropriate text length, is useful knowledge-of-result feedback for students and enables them to notice learning gaps in future text production tasks, provides pertinent information about the student's L2 development to the instructor as well as supplements their own feedback, and presents relevant data for second-language development research. We use computational algorithms that calculate the probable accuracy of lexical and grammatical constructions. The human language technology employed includes a large learner corpus, spell checker, (non-)annotated corpora, part-of-speech taggers, chunkers, and machine learning packages. Using the information provided by these NLP tools to evaluate orthography, structure of words, and combinations in which words can occur we arrive at a gradient score for the accuracy of the learner text using simple computations.
Based on the principle that effective and sustainable CALL research requires multiple perspectives that emerge from empirical data collection and analysis using a mixed-method approach, the purpose of this symposium is to discuss data and elicitation methods of interaction-based research. The first part of the discussion will be dedicated to theoretical perspectives and conceptual frameworks grounding such research in the context of CALL. The second part will focus on data elicitation methods for learner-task-tool interactions at the computer; more specifically, it will emphasize quantitative and qualitative data collection and analysis by examining the product and the process of the interactions as well as the learner's perception of these interactions. The last part of the symposium will consist of a discussion panel in which all six participants will weight strengths and limitations, successes and challenges, as well as lessons learned in light of their research experiences. By responding and integrating feedback from the audience, the discussion will focus on the ways in which a conceptual framework can guide researchers in producing sustainable CALL research methods, data and tools.
Why would anybody consider the role of number-crunching machines when it comes to the discussion of curricular questions in the context of an academic discipline that studies German language, (often a limited range of exclusively canonical) literature, culture, and cultural history in the twenty-first century? Yet the relevance of such a question is at the heart of this chapter and I will argue throughout that new information technologies have had and continue to have an extraordinary influence on teaching and learning in German Studies. Specifically, my main focus is computer-assisted language learning (CALL), which I will briefly sketch as a field of academic inquiry at the beginning, and its mutual relationship with German Studies. German, as a learned language and as a discipline, has played an important role in CALL discourses. There is not as much evidence, however, for the import of CALL in German Studies. I will outline a number of challenges that arise from this disconnect and suggest ways to address them.
It is with great pleasure that we present issue 29(3) of the CALICO Journal . It contains eight research articles, the first Spotlight article, one book review (Leakey, 2011) and one software review ( E-Tutor , an ICALL web-based courseware for German).
email] email@example.com [Word Count] 2216 [Reference Word Count] 1774 In computer-assisted language learning (CALL), a learner model is a computational data structure that contains information about individual students and thus facilitates individualized instruction through the adaptation of a language learning system to the learner. Based on Wenger (1987), Intelligent Tutoring Systems in general are said to consist of three parts (e.g., Beck, Stern, & Haugsjaa, 1996; Chin, 2001): an expert model which contains the knowledge of the domain, a teacher model which determines the system's instructional content and sequence, and a learner model which stores information about the learners' behavior and draws inferences about their knowledge and beliefs. Despite the need for individualized learning environments, learner modeling has not been a strong focus of CALL because its development and implementation is non-trivial, time-consuming, and labor-intensive (Cerri, Cheli, & McIntyre, 1992; McCalla, 1992) and because the challenging task of representing the domain knowledge of language and language learning is still largely incomplete (Levin & Evans, 1995). Of course, CALL design considers the (individual) learner, but most systems make use of an implicit learner model (McCalla, 1992), which is static, in the sense that the learner model is reflected in the design decisions inherent in the system and derived from a designer's point of view. In tutorial CALL (Hubbard & Bradin-Siskin, 2004) in general, assumptions about students' behavior, knowledge and beliefs are modeled indirectly in the activity design. In Intelligent CALL (Heift & Schulze, 2007), for example, implicit learner models can consist of (grammatical) parser rules which describe aspects of the students' interlanguage, which are not part of the L2 rules, or rules which are part of the learners' L1 system (e.g., Catt & Hirst, 1990), thus the system's view of the learner cannot and does not change in the learner's interaction with the system. An explicit model (McCalla, 1992), on the other hand, is dynamic (Holt, Dubs, Jones, & Greer, 1994) in that it is used to drive instructional decisions and is thus not part of the instructional design structure as such (see e.. Learner models can be used in a number of ways (Elsom-Cook, 1993): Corrective (providing tailored feedback), elaborative (extending the
This chapter provides a historical overview of ICALL over the last decade by focusing on two key areas: resources for the language learning classroom and resources for the researcher. With respect to resources for the language-learning classroom, we discuss and link ICALL developments to contemporary theories in Second Language Acquisition (SLA) by focusing on the importance of interaction and noticing (Gass, 1997; Long, 1996; Schmidt 1990, 1994). The ICALL projects described support a wide range of language learning activities in vocabulary and grammar acquisition, writing and reading comprehension. With regards to resources for researchers, we focus on learner and reference corpora. The chapter concludes with a discussion of new directions in ICALL research.
We discuss the development of courseware in a virtual learning environment for a university- level, German as a foreign language writing course. We view both the language learning of students in this hybrid course and the development as activity systems (Engeström, 1987; Mwanza & Engeström, 2005), describe the individual components of each system in their interrelation and show how the two systems are interconnected. Such an approach to the description of materials development enables us to conceptualize the complexities of the development process in that it goes well beyond traditional software documentation.
This paper addresses several components of successful language-learning methodologies—group work, task-based instruction, and wireless computer technologies—and examines how the interplay of these three was perceived by students in a second-year university foreign-language course. The technology component of our learning design plays a central role in this article. The main part is dedicated to the analysis and interpretation of student data collected in two different groups during two subsequent semesters. After a general discussion of the learning design of the course and task-based language learning, we analyze the interaction between two sets of factors: 1) the students' use of information and communication technologies and their perception thereof, and 2) students' perception of and participation in task-based instruction and group work.
- Jun 2007
This book provides the first comprehensive overview of theoretical issues, historical developments and current trends in ICALL (Intelligent Computer-Assisted Language Learning). It assumes a basic familiarity with Second Language Acquisition (SLA) theory and teaching, CALL and linguistics. It is of interest to upper undergraduate and/or graduate students who study CALL, SLA, language pedagogy, applied linguistics, computational linguistics or artificial intelligence as well as researchers with a background in any of these fields.
The Geroline project comprised the development of and research for two elementary and one intermediate German language courses which are delivered online for university-level distance education. The course materials include a commercially produced textbook package and our online materials which constitute a task-based learning environment by providing guidance and structure to the learning process. We discuss relevant aspects of the pedagogic design of our learning environment, its theoretical foundations and aspects of a study of the efficacy of the learning design. Insights into student perception of different course components and aspects were gained through the analysis of qualitative data from questionnaires and interviews. Our analysis of the numeric test data shows that there is no significant difference in attainment between the group who studied online and largely independently and the groups who were taught on campus. Using a task-based language teaching framework for our online courses helped students who chose to study online to achieve similar learning outcomes as their peers in the on-campus groups.
Attempting to understand and to capture the complex and dynamic nature of language learning processes is a non-trivial task for researchers in Second Language Acquisition (SLA) and Computer-Assisted Language Learning (CALL). After sketching major developments in SLA and student modeling for Intelligent CALL—the intersection of Artficial Intelligence and CALL—this paper proposes a conceptual framework of the dynamic language proficiency of students using the example of the Mocha project. The main project goals are outlined. Dynamic Systems Theory and Construction Grammar are motivated as the theoretical foundation of our thinking about SLA and student modeling. Individualization has been praised as one of the major strengths and advantages of computer-assisted language learning (CALL). Advocates of CALL frequently mention the value of students working at their own pace and receiving feedback immediately and—sometimes—based on an individual context. However, individualization from this perspective often means having the student work individually using a tutorial CALL package or simply lack of instructor control. Individualization in this sense does not imply that individual characteristics of the students are considered and that the computational learning environment is tailored accordingly. Often it does not even mean that the prior learning path—learning events with their contents and the students' achievements—is recorded and has an influence on such decisions as which learning objects are presented or what kind of feedback is provided. I will argue in the course of this paper that a non-mechanical and more humanistic approach to individualization in CALL can only be achieved when the CALL system comprises a student model, a model which is informed through the analysis of learner texts. Let us begin with a cursory look at trends in second language acquisition (SLA) research from which one might seek guidance for development of student models and note the assessment problem that arises in this context. The subsequent discussion of student models will show how different modeling techniques correspond to different approaches in SLA and will highlight their strengths and limitations that pertain to
- Jan 2006
The Canadian Modern Language Review / La revue canadienne des langues vivantes 62.4 (2006) 644-647 Planning and Task Performance in a Second Language is based on a colloquium presented at AILA 2002. In chapter 1, Ellis reviews the research on planning and task-based performance. He distinguishes pre-task planning (rehearsal or strategic planning) and within-task planning (pressured versus unpressured). The pressure in question is exercised through more or less severe time constraints. Ellis situates task planning theoretically in relation to concepts such as attention and noticing (Schmidt 1990, 1994), limitations of the working memory, and focus on form and, more importantly, by reviewing three bases: Tarone's theory of stylistic variation (1983), Levelt's (1989) model of speech production, and Skehan's (1998) cognitive approach to models of task-based language learning. Ellis's discussion of previous research on task planning and of methodological issues prepares the reader for the subsequent sections on different types of planning. In chapter 2 - the only one in the second section on task rehearsal - Bygate and Samuda argue that strategic planning favours content and decreases accuracy, whereas within-task planning is likely to improve accuracy but has a negative effect on fluency. As an alternative, they investigate and discuss the effects of task rehearsal - seen as one form of task planning - on language learning. In their study, EFL learners watched the same short video and retold story twice at an interval of 10 weeks. Bygate and Samuda conclude that significant changes in fram-ing - metacommunicative utterances that are additional to the narrative content - cannot be explained by learning that occurred during the 10 weeks but only by the positive effect of task repetition on language learning. They suggest that such a repetition facilitates the integration of strategic and within-task planning and enables learners to be more creative in their language use. The three chapters of the third section discuss strategic planning. In chapter 3, Ortega reanalyzes her 1995/1999 data and looks at learner strategies (O'Malley & Chamot, 1990; Oxford, 1990) employed in task planning. She concludes that more metacognitive strategies were employed because of the opportunity to plan, but only 59% of learners perceived planning as beneficial. In her summary, Ortega challenges the dichotomy of focus on meaning versus focus on form and argues for a view of form-in-meaning: 'In spite of holding a meaning-oriented interpretation of the task ..., learners paid attention to form during planning without any specific instruction to do so' (p. 106). In chapter 4, Sangarun concludes from an empirical study that it is impossible for learners to separate meaning and form in their planning, in spite of directions to the contrary. 'L2 learners do best when they engage in strategic planning that is focused jointly on meaning and form' (p. 131). This study also suggests that strategic planning can have a positive impact on accuracy, whereas earlier studies suggested that the influence was not significant. In Kawauchi's contribution (chapter 5), the link between strategic planning and task rehearsal is most prominent in that the study uses three kinds of (strategic) planning activities: draft writing, rehearsing, and reading a model. Kawauchi concludes from the empirical study of Japanese EFL learners that strategic planning had beneficial effects on fluency, complexity, and accuracy and that such effects depended on the proficiency level of the learners: 'High EFL learners tended to benefit the most in the case of fluency and complexity while the Low EFL learners appeared to gain most in accuracy' (p. 164). The fourth section focuses on within-task (on-line) planning. Both chapters discuss the importance and the difficulty of gaining insight into on-line planning processes, claiming that it may be the variable of on-line planning that resolves contradictory findings about the influence of strategic (prospective) planning. Ellis and Yuan (chapter 6) investigated the effects of...
This essay provides a snapshot of the change of the German language in the more than fifteen years since the Unification Treaty. It focusses mainly on lexical change and discusses different processes and trends in language change. In addition to certain East German words becoming obsolete in the united Germany, words 'travelling' from East to West and from West to East are identified. A third group is illustrated by providing examples of words which have only entered the German language in the last 15 years. Generally though, the focus is on lexical change change which has some (often indirect) link with unification and post-unification. Overall, it can be shown that the German language proves remarkably robust in integrating and assimilating the words which are new for the entire speech community or at least for a large subgroup (East or West in this case) and that language change has resulted in a wider choice for speakers and thus a better basis for linguistic creativity.
This paper provides examples of student modeling techniques that have been employed in computer-assisted language learning (CALL) over the past decade. We further discuss two of our own systems and show how different types of CALL programs can, nonetheless, share similar conceptual designs of a student model. First, we describe the German Tutor, an Intelligent Language Tutoring System (ILTS) for German as a Second Language which contains a parser and a grammar that analyze student input. The student model is based on student subject matter performance and provides feedback and remedial exercises suited to learner expertise. Second, we provide an overview of Geroline, an online distance education course for ab initio German learners, and its student model. We show how a student model can support computerized adaptive language testing for diagnostic purposes in a Web-based language learning environment which does not rely on parsing technology. Here, student scores from libraries of sentence-based exercises are used as the raw input for the model.
This article discusses selected theoretical aspects of providing error feedback for language learners. The discussion focuses on feedback for grammatical errors, but many of its tenets appear to be of broader relevance. The theoretical considerations concerning the dialog with the learner about linguistic errors are discussed and some conclusions for CALL system design will be drawn. This discussion focuses on errors in written text production and occasionally draws on examples from parser-based CALL, in particular on insights gained during the work on Textana - a prototype of a grammar checker for English-speaking learners of German.
Fall als Zu-Fall? Verbvalenz und Kasusmarkierung Die vorliegende Diskussion von Valenz konzentriert sich auf die morpho-syntaktische Markierung der Verbargumente (Nominal-und Präpositionalphrasen). Dieser Fokus ergab sich im Rahmen eines Forschungsprojekts, in dem eine Parsergrammatik für das Deutsche entwickelt wurde. Dieser Parser ist das 'linguistische Rückgrat' von Textana, dem Prototyp eines Grammatikprüfprogramms für englischsprachige Deutschlerner (Schulze 1998, 1999, Schulze and Hamel 2000). Dieses Programm gestattet Studenten die relativ freie Eingabe von Sätzen, die dann auf grammatische Übereinstimmung mit dem Parserregelwerk und dem (noch kleinen) Wörterbuch hin geprüft werden. Da davon ausgegangen werden kann, dass diese Deutschlerner vergleichsweise wenig Fehler in der semantischen Auswahl der Verbargumente machen, aber Korrekturhilfe bei der morpho-syntaktischen Markierung dieser Phrasen brauchen, ergab sich die Notwendigkeit einer formalen linguistischen Beschreibung der Kasusmarkierung der Verbargumente im deutschen Satz. In diesem Artikel wird zuerst die allgemeine Problematik des Kasus im Deutschen dargelegt. Aufbauend auf der Hypothese, dass die deutsche Sprache, wie andere auch, über morphologische und präpositionale Kasusmarkierungen verfügt, wird ein generelles Schema zur Beschreibung der morpho-syntaktischen Merkmale von Verbargumentstypen vorgeschlagen. Diese werden dann kurz einzeln anhand von Beispielen vorgestellt. Durch die Unterscheidung von strukturellem und lexikalischem Kasus und auf der Grundlage von Obliquitätsregeln ist es dann möglich, sechs Kasusregeln aufzustellen. 1. Kasusmarkierung im Deutschen Ursprünglich hatten indoeuropäische Sprachen keine dedizierten Mittel zur Markierung der Beziehung von Argumenten zum Verb oder von Argumenten untereinander. Einfache Kollokation mit der entsprechenden Intonation wurde als hinreichend angesehen. Komposita wie Hausherr zeugen heute noch von dieser Methode (Paul 1955, Bd. III: 215f.). Man nimmt an, dass morphologische Kasusmarkierungen erst später ins Indoeuropäische kamen. Heute hat das Deutsche vier morphologische Kasus, die einerseits jeweils eine ganze Reihe von semantischen und syntaktischen Beziehungen markieren und andererseits über eine geringe Anzahl von Flexionskombinationen (Determinativ, Adjektiv, Substantiv) verfügen, die eindeutig einem bestimmten Kasus zuzuordnen sind. Diese beiden Eigenschaften der deutschen Oberflächenkasus machen eine Beschreibung der morpho-syntaktischen Seite der Verbvalenz ausschließlich mit Hilfe der vier morphologischen Kasus unpraktisch. Die vier Namen der Kasusformen (Nominativ, Akkusativ, Dativ und Genitiv) sind lateinische Übersetzungen aus dem Griechischen. In deutscher Übersetzung stünde der Nominativ als Namenskasus, der Dativ als Gebenskasus und der Genitiv als 'zu einem Geschlecht gehörend'. Akkusativ ist eine römische Verwechslung des griechischen Wortes αιτιατικη (Instigierung) mit αιτιασξαι (Beschuldigung). Natürlich
Authenticity of language learning tasks, authenticity of learning experiences and the use of authentic language are important characteristics of commu-nicative language teaching and learning. In this article, ways of achieving productive use of authentic language in a computer-assisted language learn-ing environment are discussed. This discussion concentrates on a particular area of CALL — parser-based CALL. The use and adaptation of two exist-ing parsers for two CALL tools that support text production by encouraging learners to concentrate on the linguistic structure — mainly the grammar — of a text they have just produced in a communicative task are outlined to provide concrete evidence for the contribution language processing techniques can make to the field of Computer-Assisted Language Learning. L'authenticité de la tâchelangagì erè a effectuer, de l'expérience d'apprentis-sage et de l'utilisation de la langue sont des caractéristiques importantes dans l'apprentissage et l'enseignement des langues basés sur une approche communicative. Notre discussion porte sur une exploitation que nous croyons productive de cette langue authentique dans des environnements d'ELAO. Elle se concentre plusparticulì erement sur un type d'ELAO, celui base sur le TAL. Nous décrivons l'utilisation et l'adaptation de deux parseurs dans deuxsys emes d'ELAO différents mais dont la fonction commune est l'aidè a la production de texte ecritsò u l'apprenant est amen a se concentrer sur certaines structures syntaxiques qu'il vient de produire. Nous voulons par le fait même démontrer l'apport indéniable des outils de TAL dans lessys emes d'ELAO.
- May 1999
This paper sketches the place and function of grammar in the context of language learning in general and attempts to show the relevance and usefulness of these formal concepts of grammar to Computer-Assisted Language Learning in particular. The approach to grammar described here will be illustrated through a brief discussion of a grammar checker for English learners of German, Textana, which is being developed at UMIST. © 1999, European Association for Computer Assisted Language Learning. All rights reserved.
- Apr 1998
This paper discusses selected theories of Second Language Acquisition and their implications for developing a CALL tool for learners of German — `Textana'. Textana was planned as a generic text production tool, i.e. it was meant to provide help to learners in outlining, editing and post-editing texts in their foreign language, German (Schulze, 1997). It has been decided to concentrate initially on post-editing for grammatical accuracy. This component of the software will allow learners to have their German text parsed for grammatical errors. On the basis of the parse results (and student data that the program has), feedback (canned text) will be given to the learner during an interaction between learner and computer. Thus, the implementation of a substantial `chunk' of German grammar was a prerequisite for all other work on Textana. The coverage of the German grammar and the underlying morphological and syntactical theory will be touched upon briefly at the end of the following section. The main part of this paper is dedicated to a more detailed discussion of the implications of Second Language Acquisition research on the design of Textana.The main objective at this stage of the project is to test linguistic and pedagogic hypotheses. However, this is not to say that Textana is not intended to fulfil a practical purpose in the learning and teaching of German — quite the contrary — but it is far too early to make any claims concerning possible uses of Textana. I can only invite readers to think about possible practical applications themselves.
- Jan 1997
This paper reports on the fast, exploratory, phase of a CALL project for German. The project endeavours to combine experience in the fields of traditional CALL and hypertext applications with the approach of the theory of formulating. The aim is to produce a package that will enable students of German to develop their reading and writing skills, and learn new vocabulary and syntactical structures. At the same time, this program attempts to help them to enhance their knowledge and competence in the production of demanding sorts of text, such as a text commentary. In the first phase an experimental version of this program was produced that comprises a reading comprehension module and a guided tour through the production of a text.
Abstract The collective variables complexity, accuracy, and fluency have been used in second language acquisition (SLA) research to operationalize proficiency of language learners. Given the multifaceted nature of these variables, they are difficult to conceptualizeand measure. In this paper, we focus on the measures for linguistic complexity and propose toutilize a number of robust predictors, which we derive from SLA studies and automated essay scoring, and introduce the collective variable 'balanced complexity', which is shown to be a vector in a four- dimensional vector space determined by mean word length, Carrol's type-token ratio, mean comma unit length, and mean period unit length. Inour case studies—an analysis of electronic text submissions by students of German as a foreign language—we,demonstrate that balanced complexity is likely to have construct validity and can be computedsemi-automatica lly by relying on a number,of (relatively) simple textual surface measures. Keywords
A] Introduction The title of this chapter gives rise to a two-part question: What is Intelligent CALL (ICALL) and what does it have to do with task-based language teaching (TBLT)? I will address this main question and focus on a number of subordinate and related issues such as: What kind of ICALL projects have relied on a task-based design? How was this done? What kinds of task processes have been considered in ICALL systems? What contributions has TBLT made to ICALL, which ones could it make? Which contributions could ICALL make to TBLT? After a brief sketch of TBLT in the context of computer-assisted language learning (CALL), the central goals and approaches of Intelligent CALL – a subfield of CALL which utilizes artificial intelligence techniques – will be described in broad strokes. The main part of the chapter discusses a number of ICALL projects, evaluates trends critically from a TBLT perspective to ascertain what contributions ICALL can make and has made to task-based language teaching and how TBLT can and should inform and/or has informed research and development in ICALL. Doughty and Long (2003, p. 50) argue that: Task-Based Language Teaching … constitutes a coherent, theoretically motivated approach to all six components of the design, implementation, and evaluation of a genuinely task-based teaching program: (a) needs and means analysis, (b) syllabus design, (c) material design, (d) methodology and pedagogy, (e) testing, and (f) evaluation.