Thursday, May 1, 2014
Overview of Morning Session
The symposium will begin with Carol Chapelle’s brief introduction to the problem outlining the challenges of developing pedagogies for discipline-specific writing. In the first presentation, Christine Feak (University of Michigan) will describe the state of the art in the teaching of academic writing for graduate students (e.g., through texts such as Academic Writing for Graduate Students: Essential Tasks and Skills, 3rd Edition; University of Michigan Press, 2012) and how this work has been informed by the study of academic language in the MICASE corpus. She will share observations about the pressing needs in the teaching of academic writing resulting from this extensive work. Continuing the discussion of student writing in higher education, Hilary Nesi (Coventry University) will describe how academic writing in higher education has been investigated using corpora such as the British Academic Written English (2004-2007) and findings based on Genres across the disciplines: Student writing in higher education (Cambridge University Press, 2012). To examine writing in one discipline in particular, she focuses on engineering—a theme that continues into the afternoon.
Emerging Issues in Teaching EAP Writing
Christine Feak, University of Michigan
With the globalization of education, research and publication, the importance of writing for academic, scholarly and professional purposes in English continues to grow. Through this writing, students and scholars contribute to, gain membership in, and position themselves in their disciplines and professions. Studies of academic writing over the past decades have transformed our ability to create high quality courses and teaching materials to help students as well as junior and senior scholars to write in ways that are consistent with conventions in their chosen fields. Yet, interestingly, as many English for Academic Purposes (EAP) writing courses become more sharply focused, current trends in academic writing demands, research, employment of Ph.D. holders, grant requirements, and journal publication point to the need for writing course designers to step back and reassess whether and how we are meeting the writing needs of students. Of particular importance in this reassessment is broadening our knowledge of the writing that graduate students must do (beyond the standard course papers, thesis, and dissertation) and for which audiences. A further consideration in re-evaluating our graduate writing courses is whether the traditional division between “non-native speaker” and “native speaker” of English continues to serve a useful purpose at the advanced writing level. Indeed, increasingly faculty across all disciplines complain that graduate students in general struggle with writing, but those same faculty are ill-equipped to provide adequate levels of writing mentoring. These issues will be examined in this session, which focuses on where we are and where we need to be in teaching graduate level writing.
What do British engineering students write, and why?
Hilary Nesi, Coventry University
The corpus of British Academic Written English contains 6.5 million words of proficient university student writing, collected from across more than thirty disciplines, at four levels of study. It was one of the outputs of ‘An Investigation of Genres of Assessed Writing in British Higher Education’, a project which aimed to discover exactly what kinds of writing British students needed to produce during their university studies. Corpus analysis and interviews with teaching staff and students helped us develop a categorisation system for the assignments in the corpus, with thirteen broad ‘genre families’ linked to five social and educational purposes: demonstrating knowledge and understanding, developing powers of independent reasoning, building research skills, preparing for professional practice, and writing for oneself and others (to communicate with non-specialists and develop human awareness). In some disciplines, only a few of these genre families are produced in any quantity, but according to the corpus evidence, all thirteen categories occur regularly in engineering. This places particular demands on engineering students, as although each genre family is structurally and linguistically distinct, they are not necessarily distinguished clearly in assignment rubrics – lecturers often seem to rely on their students having acquired implicit genre knowledge. This paper will provide examples of engineering student writing in the different genres, discuss why these different kinds of assignment are set, and examine the various challenges they present.
Overview of Afternoon Session
The afternoon session will look at linguistic characteristics that differ systematically across writing and examine the process that students go through to become communicators in their respective disciplines. Susan Conrad (Portland State University) will illustrate the use of a statistical approach to detect differences in language used across disciplines based on the quantitative analysis of features that cluster together. The approach has been applied to describing grammatical variation in dialects, historical periods, author styles, and specialized domains (e.g., Variation in English: Multi-Dimensional Studies; Longman, 2001). She is now using this approach for investigating engineering writing. In view of the differences in writing practices across disciplines, Natasha Artemeva (Carleton University) examines communication development within one discipline. She addresses the question of how engineers develop as communicators through their university study and into their professions.
Situating engineering writing within the universe of English discourse
Susan Conrad, Portland State University
In this session, I will describe the use of a quantitative linguistic technique (“Multi-Dimensional Analysis”) to study how writing in civil engineering is different from – and similar to – numerous other kinds of spoken and written texts in English. The analysis is part of a larger project funded by the National Science Foundation, the Civil Engineering Writing Project, in which linguists and civil engineers are collaborating to improve the teaching of writing to undergraduate civil engineering majors.
Using statistical techniques, Multi-Dimensional Analysis considers the linguistic features that cluster together in texts, describing them for the communication functions that they serve (e.g. conveying highly specific information, regulating social interactions, narrating a series of events). The technique allows us to make comparisons among many different types of discourse. Using texts compiled for the larger project, the analysis in this session will situate academic research articles, practitioner writing, and student writing from civil engineering along a wide range of speaking and writing in English. Such analysis helps us understand the challenges that novice civil engineers face in gaining discourse competence. Specifically, I will use the analysis to address the following questions:
• In what ways are professional registers of civil engineering writing similar to and different from other registers of English such as academic writing generally, conversation, and fiction writing?
• In what ways do the registers written by civil engineering practitioners (e.g. design reports, site visit observations) differ from research articles written by academic faculty?
• When students write papers for assignments that are designed to mimic practitioner registers, in what ways does the student writing differ from the practitioner writing?
In the session I will also discuss the strengths and weaknesses of this approach, especially when it is applied to the development of instructional materials.
From school to work: The formation of an engineering communicator
Natasha Artemeva, Carleton University
Over the past three decades, professional communication scholars have explored various ways in which students and novices develop as competent communicators within their professions, including the complex processes of the academy-to-workplace transition, and professional and organizational acculturation (e.g., Artemeva, 2005, 2008, 2010, 2011; in press; Dias, Freedman, Medway, & Paré, 1999; Dias & Paré; 2000; Lave & Wenger, 1991; Wenger, 1998, 2005). The research has often indicated radical differences in the types of communication used in the academy and in the workplace, thus questioning the efficacy of the traditional professional communication classroom and the possibility of transfer of communication strategies not only between academic and workplace contexts, but even from one academic course to another (cf. Brent, 2012; Tuomi-Gröhn &Engeström, 2003). Other studies have reported on some successful pedagogical interventions in the training of future professionals, such as engineers, who need to develop specific communication competences (e.g., Artemeva, Logie, & St. Martin, 1998). By drawing on the findings of a longitudinal study of engineering students’ trajectories in becoming professionals (Artemeva, 2005, 2008, 2010, 2011), this presentation focuses on the formation of engineering communicators as well as on possible approaches to bridging academic and workplace contexts to assist engineering students in this process.
Friday, May 2, 2014
Overview of Morning Session
On the second day, we will examine the state of the art of computational analysis of students’ written language and look forward to how this research can be extended to the analysis of students’ writing in STEM disciplines. In the first session on Friday, Scott Crossley (Georgia State University) will describe his research on development and use of natural language processing tools in assessing writing quality and text difficulty, with a focus on lexical development and implications for analysis of discipline-specific language. Jill Burstein (Educational Testing Service) will describe the first steps in efforts to apply research in natural language processing as applied in evaluation tools such as E-rater® to discipline-specific language.
Natural language processing tools and automatic essays scoring systems
Scott Crossley, Georgia State University
Automatic essay scoring (AES) systems have quickly developed from theory to functioning systems in the past 20 or so years. This presentation will provide an overview of studies conducted to provide summative and formative writing feedback to users in the Writing-Pal intelligent tutoring system. The presentation will focus on automatically assessing textual features using computational tools such as Coh-Metrix, Linguistic Inquiry and Word Count (LIWC), the Writing Assessment Tool (WAT), the Tool for the Automatic Analysis of Lexical Sophistication (TAALES), and the Tool for the Automatic Analysis of Cohesion (TAACO) and how such assessments can inform AES systems and automatic writing evaluation (AWE) systems. Future developments for both AES and AWE systems will be discussed along with the potential for computational tools to analyze discipline-specific language.
Automated evaluation of writing: Expanding the body of knowledge for writing in the disciplines
Jill Burstein, Educational Testing Service
Automated writing evaluation (AWE) has made rapid advances in the past 20 years, and has developed into a discipline in its own right, now having a history, a modern presence, and a well-motivated research agenda which is strongly guided by efforts associated with the Common Core State Initiative to support college- and career-readiness. Professional effective communication is a critical competency for college- and career-readiness. At this moment in time, AWE capabilities focus on evaluation of the traditional academic essay which is a core genre in K-12 and college; however, as students move into college and the workplace, other writing genres are required. In this talk, AWE current capabilities will be illustrated in the context of e-rater®, Educational Testing Service’s automated writing evaluation system. The relationship between the writing construct and system development will be discussed. In addition, results from two recent survey studies administered to several hundred university faculty will be presented, highlighting a greater need for writing evaluation capabilities that support the evaluation of core competencies (for writing) that are critical for discipline-specific writing genres, such as research proposals, which are more common requirements in college and workplace settings.
Overview of Afternoon Session
In the afternoon session, Nick Pendar (Skytree Inc.) will describe how machine learning techniques are used in projects where content and sentiments are inferred from text. Elena Cotos (Iowa State University) will describe the genre-based approach used in the Research Writing Tutor software, designed to incorporate research and resources from linguistics into pedagogy aimed at discipline-specific writing of research articles. General discussion will follow to advance ideas for creating a foundation for the generating of innovative practices in the teaching and assessment of STEM writing.
An Overview of the State of the Art of Text Data Mining
Nick Pendar, Skytree Inc.
Text data mining is the practice of extracting structured information from unstructured textual data. Given the breadth and complexities of human language, the variety of information that one may desire to extract or infer from text, and the increasing amount of textual data available today, text data mining touches upon a multitude of disciplines. This presentation provides a cursory survey of the field and its major challenges along with the main ideas and techniques stemming from linguistics and statistics, to computer science, bioinformatics, and AI. Many text mining techniques have direct applications in pedagogical settings. Others may not seem obviously linked to the types of inferences made in the analysis of students’ writing. Any usable and useful computer-assisted pedagogical application that attempts to analyze text must be able to seamlessly incorporate a hierarchy of several complex techniques. This presentation aims to encourage and facilitate an interdisciplinary discussion of the state of the art, and perhaps new possibilities for combining existing practices in a way that addresses current objectives.
Technology driven marshalling of genre theory and research for the teaching of discipline-specific writing
Elena Cotos, Iowa State University
The relation of corpus and genre analysis to disciplinary writing has received considerable attention in English for Academic Purposes (EAP) research and pedagogical applications. Concerned with the communicative needs and practices of language learners in academic contexts, EAP practitioners draw from a rich research base and an assortment of theoretically grounded techniques. Those working with advanced L2 graduate students, in particular, have widely adopted genre as a tool in teaching discipline-specific writing (Johns, 2003), examining and deconstructing examples of genres compiled into specialized corpora. Graduate writing courses are often organized into units that reflect the IMRD (Introduction, Methods, Results, Discussion) structure of research-reporting genres and tend to rely on analytical models rooted in genre analysis, one of the most influential being Swales’ (1981, 1990) framework of moves or ‘bounded communicative act[s] […] designed to achieve one main communicative objective’ (Swales & Feak, 2000, p. 35). Nevertheless, EAP practitioners are still waiting for a cross-disciplinary guiding framework and for instructional tools to adequately address the needs of their traditionally heterogeneous academic writing classes, where students from a variety of majors need to learn the discursive practices of their particular discipline.
Targeting this practical challenge, this paper demonstrates how the construct of communicative moves can be operationalized computationally to facilitate genre-based discipline-specific writing instruction. It presents an innovative model for applying comprehensive move/step frameworks, devised for the research articles genre and validated across thirty disciplines (Cotos, Huffman, Link, & Paben, forthcoming), to the development of an intelligent AWE application – the Research Writing Tutor (RWT). RWT pushes the envelope due to its unique capability to analyze students’ individual research article sections, generate discipline-specific feedback based on the rhetorical conventions of this genre, and provide different forms of corpus-based scaffolding. Overall, this paper makes the relationships between genre theory, genre analysis, and genre instruction explicit, at the same time demonstrating that move analysis is a powerful and promising theoretical, analytic, and teaching construct for discipline-specific writing instruction. By illustrating a technology driven marshalling of genre theory and research, I take Swales’ vision to a new dimension of conceptualizing EAP pedagogy.