Error Analysis: Activity 3
Case marker accuracy on different tasks
Watch Anna B’s Retell and pay attention to her use of case markers. Does she make the same errors with case markers in the Retell and the Interview tasks, in terms of types and tokens (recall that a count of error types omits repetitions of the same error, while a count of error tokens includes such repetitions)? Watch Sophia’s Retell and pay attention to her use of case markers. What kinds of case markers does she use?
Do you think just counting number of errors is a good measure of proficiency? Why or why not?
Anna B Retell
Transcript (PDF)
Sophia RetellTranscript (PDF)
Please type your answers to the questions in the box below.
When you have finished typing your answer, click to compare your response with the Learner Language staff response.
Anna B produces quite a number of error tokens -- that is, a count that includes repetitions of the same error -- in case marking in the Interview, but her case marking is highly accurate in the Retell task. She mostly uses nominative and accusative case markers in her utterances. The differences in accuracy between the two tasks might be due to the fact that she practiced using case markers while doing the three intervening tasks (Interview, Narrative, Question).
In contrast, Sophia rarely uses accusative case markers in her utterances. She uses an accusative case maker erroneously in the Retell task as follows:
12 S: 아팥(.) 아파트를 못(.) 가요.
The accusative case marker ‘-를’ needs to be replaced by a locative case marker ‘-에’ here. This error is the only use of an accusative case marker in any of Sophia’s six communicative tasks. In terms of fluency, Sophia sounds more spontaneous and error-free than Anna B. However, she omits most case makers across tasks. Although the omission of nominative or accusative case markers is common in spoken Korean, Sophia’s excessive omission of accusative case markers could cause us to question her proficiency. Sophia’s errors in omitting case markers suggest she is less proficient, while her fluent, native-like speech patterns suggest she is more proficient. Counting her errors just lets us see part of her proficiency.
