Conclusion

This dissertation has undertaken to probe into EFL writing pedagogy with advanced Hungarian university students. After reviewing relevant theory and empirical work in the related fields of writing pedagogy and corpus linguistics, it has presented a possible ethnography of advanced writing at university, and given details of the lexical and discourse features of the JPU Corpus. It has aimed to synthesize pedagogy and linguistics by bridging the gap between process and product approaches--an area that remained virtually unexplored before this endeavor. The course of investigation has been framed by current understanding of writing processes (Zamel, 1992; Zinsser, 1998) and by the increasing prevalence of developing and exploiting representative and specialized computer corpora (Sinclair, 1997; Kennedy, 1998).

A number of factors have remained beyond the scope of the analysis. For example, there has been no space to position the theory and practice of the institutional assessment and evaluation of writing skills, which represents one of the outcomes of the writing process. This field is well worth further investigation. Also missing from the evaluation of the writing process have been the wide range of classroom and study guide materials developed during the past years and the assessment of the practice in out-off-class meetings with students. However, it seems that the original three-fold aim of the study, that is, to collect evidence of advanced students' language use, to apply this data to research, and to apply the results in writing pedagogy, has been met.

In terms of processes, we have seen the development of the writing course syllabus that gradually moved in the direction of focusing on the process of developing writing skills, according equal importance to the products of each step during that process. A number of pedagogical innovations and new task and text types have been introduced and evaluated. Their analysis has been special in that it has had to be predominantly self-reflective: as no concurrent outside observation took place, the study describing and evaluating EFL writing pedagogy at JPU has employed mainly qualitative data.

In terms of products, this dissertation has been the first in Hungary to present the case for the need to collect data on written language performance by advanced students of EFL. It has described and explained the design and development of the JPU Corpus, and provided a sample of the lexical and discourse analysis made possible by the scripts. We have seen the results of writing pedagogy in the work of pre-service and in-service students of EFL. The study of the ten hypotheses has provided evidence of the uniqueness of learner scripts. One area where the investigation may have opened up a new perspective of corpus application has been its limited focus on error. Instead, it has attempted to capture some of the characteristic elements of student writing in a non-prescriptive manner, much in the tradition of how corpora of L1 texts are dealt with. The pedagogical exploitation of the scripts has been shown to include the design and application of several types of tasks that are to guide students' acquisition of lexis and discourse patterns.

Further empirical work is also to be conducted. There are scores of areas that can be studied for a fuller understanding of learner writing, and, eventually, how new generations of EFL students acquire and reflect on written discourse, whether personal or academic. Of these, I will undertake to recommend what appear to be most needed and relevant, divided into three groups according to discipline: those that are primarily concerned with writing pedagogy, those that focus on the analysis of a corpus, and those that would aim to synthesize the two disciplines.

Implications for writing pedagogy research

Of the most relevance would be cross-sectional studies, based on classroom observations, structured interviews, the comparative analysis of school curricula and think-aloud protocols, exploring writing pedagogy in Hungarian secondary-school programs in the native language and in FLs. This appears to be among the most pressing needs. What is currently known about this practice indicates that much is to be done for students to find writing a meaningful and empowering activity (cf. Nikolov, 1999). The qualitative and quantitative study of L1 writing processes and products by Hungarian secondary-school and university students could begin to explore contrastive rhetorical issues.
To establish what factors contribute to success in writing in a FL, case studies are necessary. For researchers and teachers to validate theory, such investigations can supply the data on which to plan syllabus development. These studies would need to be conducted on intermediate and advanced levels of proficiency.
Innovative designs, with students involved to enhance the validity of the research agenda, could result in an ethnography of writing, similar to the chronicle-based report of cross-cultural and pedagogical issues in writing instruction by Pally & Diallo (1995). A nationwide project of action research in writing pedagogy could facilitate the development of such projects and the publication of the results. The project sets the context of systematic work in studying academic writing, including such activities as sharing and evaluating syllabuses, evaluating the curricular status of writing courses, establishing writing centers at participating institutions, developing new team-taught courses, and developing new teaching materials.
The expertise of the writing tutor and the role of training and continuous development are two closely related concerns well worth further research. Case studies are necessary to establish the pedagogical variables that contribute to student achievement. This focus could help reveal how various learner styles match and mismatch with teaching preferences. The results may also have implications in developing pre-service and in-service teacher education.
An exploration of current issues in testing writing skills in EFL at Hungarian secondary schools and universities would contribute to collecting data on the validity and reliability of various types of assessment instruments. More research is to be initiated on the writing component of the secondary school-leaving exam, on marker training, and on validation procedures. Teams and individual teachers could conduct action research and validate progress and proficiency tests. Experiments measuring the effects of item and task type, length, and time variables would require concerted effort and sustained funding.

Implications for corpus studies

One of the areas of such further research may seek to study the correlation between performance in oral and written tasks. We still have little empirical evidence of Hungarian EFL students' performance in these areas. Especially promising would be the assessment of communicative skills in authentic academic oral and written tasks across several higher-education institutions. The development of a parallel oral and written learner corpus could provide the sample for piloting such an investigation.
Also of potential interest would be the development and piloting of further sets of study guides based on parallel L1 and L2 corpora. This could be achieved by extending the framework of John's (1997b) remedial data-driven tuition by supplementing a course with these types of materials and tasks, assisting individual students and groups with lexis, collocational use, grammar, discourse, and even punctuation.
Clearly, the international extension of the use of the JPU Corpus would play a key role in the future. This would mean the contribution of a 200.000-word subset of the JPUC in the ICLE and the LonLC. Comparative and contrastive studies of lexis and discourse could be developed.

Implications for a synthesis of pedagogy and corpus studies

To analyze, using both qualitative and quantitative approaches, the processes observed in writing conferences and to inspect the text of teachers' written feedback we will need cross-institutional cross-cultural studies of response to student scripts. One could investigate the factors underlying feedback by native and non-native English speaker teachers. This would necessitate piloting the procedures for developing a learner corpus annotated by three types of comments: the writer's commentary on purpose and content; peers' comments; and the teacher's response. By incorporating these comments in the corpus, further research could investigate specific features of L2 writing processes, which, in turn, could be applicable in writing pedagogy. As suggested in the last chapter, the development of a corpus with PCA added to it could reveal patterns of discourse as writers and readers negotiate meaning (Horváth, in preparation-a).
Finally, the study of the writing of the tutor would provide a new perspective on writing pedagogy. This research domain may include the investigation of how teachers' preferences in their own writing transfer to syllabus design and classroom procedures. This factor represents an entirely new vista in the ethnographic study of writing. Data used for the analysis could comprise authentic scripts by teachers and a sample of syllabuses.

It remains to be seen how many of these suggestions for further inquiries into the three areas will meet with support. Clearly, awareness of, and interest in, the need to raise standards in Hungarian writing pedagogy and the potential outcome of improved levels of student performance are among the motives that determine the sustainability of any educational proposal. In submitting my work on the processes and products of advanced writing in EFL, I hope to have laid the necessary basis and shown some direction for these and other studies that aim to achieve those ultimate goals.

Furthermore, the JPU collection of scripts can become the basis of establishing a Hungarian EFL learner corpus, so that college and university students' scripts may be collected and studied by a national team. One outcome of this study may be that writing teachers with a similar concern may cooperate in syllabus development and corpus analysis. We need such a collection to capture the essence of what goes into the writing process and how its products can be appreciated.