2015 Fiscal Year Research-status Report
Investigating a Learner Corpus of Computer-mediated Communication
| Project/Area Number |
26580077
|
|---|
| Research Institution | Gakushuin University |
|---|
Principal Investigator |
MARCHAND Tim 学習院大学, 付置研究所, 准教授 (20645197)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
阿久津 純恵 東洋大学, 公私立大学の部局等, 講師 (20460024)
|
|---|
| Project Period (FY) |
2014-04-01 – 2017-03-31
|
| Keywords | CMC / learner corpus / register / MDA |
|---|
| Outline of Annual Research Achievements |
Research achievements are as follows:
(1) Near completion of the learner corpus. Data collected now amounts to over 500,000 tokens of original learner-generated CMC. The data collection also includes some learner profile and motivational data.
(2) Treatment of learner data for encoding and spelling errors. Tagging tools for pre-treated and post-treated data assessed and validated
(3) Multidimensional analysis of the learner corpus. Both the learner corpus and the reference (native-speaker) corpus show a distinct weighting towards "persuasive speech" (cf Dimension 4 of Biber's Dimensions of register variation). Learner corpus shows a greater tendency towards interactional discourse, in contrast to the reference corpus tendency towards informational discourse (cf Biber's Dimension 1).
(4) Ongoing results shared and discussed with scholars at conferences.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
(1) Data collection has continued at a pace. The size of the learner corpus (approximately half a million words), allows for valid statistical testing on smaller sub-corpora.
(2) The identification of various tagging error types has led to the speeding up of the treatment of the learner data. Over half the learner data has now been cleaned up.
(3) The multidimensional analysis has revealed a striking contrast between the CMC corpora and other registers, and also a clear distinction between learner CMC and native-speaker CMC. This has suggested a new way to approach the issue of proficiency of the learner CMC texts.
(4) Alternative measures of proficiency have yet to be explored in depth.
(4)
|
| Strategy for Future Research Activity |
(1) Complete the pre-treatment of all the learner data. Create a database for the complete learner corpus.
(2) Investigate the extent to which learner CMC develops over the course of an academic year with a longitudinal analysis of register variation.
(3) Compare the results of the longitudinal analysis with an alternative proficiency measure, such as bigrams.
(4) Continue to share the results of, and receive feedback on, the project at conferences.
|
| Causes of Carryover |
Personnel expenditure has not been required until the learner corpus was completed, and an efficient means of cleaning up the data identified.
|
| Expenditure Plan for Carryover Budget |
We anticipate spending money on personal expenditure to help with processing the treatment of the learner data, and to create a database of it.
|
