Bibliographic Information
- Other Title
-
- コーパスから見える文法
- コーパス カラ ミエル ブンポウ
Search this article
Abstract
An empirical investigation of a grammar, a part of the internal state of a speaker, should be conducted based on externally observable data. With the sizes of corpora increasingly large and the expansion in the amount and kind of data obtainable from them, corpora are now more and more widely used for hypothesis testing. Such corpus data sometimes reveal new facts overlooked so far, leading to the falsification of a widely accepted hypothesis in favor of another which may be counterintuitive. In spite of their usefulness, corpora can be easily abused: With the development and spread of “user-friendly” environments, users tend to pay attention only to the output of software while disregarding the input and the process and not examining whether the data, especially statistical ones, can be interpreted appropriately as evidence for their hypotheses. For the development of corpus-based research, we should examine not only the validity but also the limitations of present methods and potential problems which may be posed by them, as well as often hidden assumptions concerning use of corpora in linguistic research.
Journal
-
- 国際開発研究フォーラム
-
国際開発研究フォーラム 38 23-40, 2009-03-24
名古屋大学大学院国際開発研究科
- Tweet
Details 詳細情報について
-
- CRID
- 1390009224622555008
-
- NII Article ID
- 110007352978
-
- NII Book ID
- AN10436655
-
- ISSN
- 21899126
- 13413732
-
- HANDLE
- 2237/11796
-
- NDL BIB ID
- 10234024
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- IRDB
- NDL
- CiNii Articles
- KAKEN
-
- Abstract License Flag
- Allowed