How to Use Corpora in Contrastive Linguistics
Linguists use corpora to compare features of more than one language. Corpora consist of large chunks of language, such as speeches or conversations, rather than sentences or individual words. The benefit of this kind of contrast between languages is the natural, contextual nature of the corpora. The data sets carry information about why or how words or phrases would be used, in context, instead of being empty vocabulary. However, this can be a problem when the context of each data set is too dissimilar.
Other People Are Reading
Things You'll Need
- Native speakers of each language
- Recorder
- Computer
- Printer
- Pens of different colors
Instructions
-
-
1
Decide exactly what aspect of language you are trying to analyze. This may be passive sentences or past tense verbs or personal pronouns.
-
2
Establish a prompt for speakers of each language. Ensure that the answers or responses will produce the language aspects you need to analyze and will be equal to one another. In other words, the prompts should result in the same answer from each speaker in his or her respective answer.
-
-
3
Present the prompts to each speaker and record the response.
-
4
Transcribe each response either with a computer word processor or a paper and pencil. Print the responses, if you're using a computer.
-
5
Highlight the particular features of each language with colored pens. This will make obvious the differences or similarities in each language.
-
6
Analyze the data, looking for features that are different or are shared between the languages. Record your findings.
-
7
Repeat the data collection and analyze, to further verify your findings. If the same patterns emerge, even with different prompts and different answers, you can be better certain your data is accurate.
-
1
Tips & Warnings
You can also use a corpus that has already been produced, such as the National British Corpus. However, the context and speaker of the compared language should be similar enough to the corpus to ensure there are no additional factors you'd need to consider in your analysis.
Although you may need to ask questions in order to garner the type of language you need, do everything possible to ensure the native speakers are producing language as naturally as possible.
