Pica, T.  (1994).  Questions from the language classroom:  Research perspectives.  TESOL Quarterly, 28.  pp.49-79.

7. To what extent does error correction assist the L2 learner?
 

• Errors are often interpreted as learners experimenting with L2 rules and patterns - eventually they figure out how they work and how to use the language appropriately.

 
• However, some researchers (e.g. Schachter, 1984) argue that lack of correction implies to the learner that the utterance was accurate.

 
• Not all teachers use "correction" in the same way - it is a diverse phenomenon.

 
• Brock, et al (1986) found that while correction had no effect in non-classroom conversations, it was very effective in games in a classroom setting.

 
• Schmidt and Frota (1986) found that for learners who are attending, being able to hear a corrected version helped them understand what they were doing wrong.

 
• There are two key features that make correction effective:
• It must bring students' attention to their own errors.
• It must do so in meaningful, communicative contexts.

 
• Research on correction places considerable doubt on the claim that comprehensible input is all that is needed for language acquisition.

 
References
Brock, C., Crookes, G., Day, R.R., & Long, M.H.  (1986).  The differential effects of corrective feedback in native speaker-nonnative speaker conversation.  In R.R. Day (Ed.)  Talking to learn (pp. 229-236).  Rowley, MA:  Newbury House.

Schachter J.  (1984).  A universal input condition.  In W. Rutherford (Ed.), Universals and second language acquisition (pp. 167-198).  Amsterdam, Netherlands:  John Benjamins.

Schmidt, R.W., & Frota, S.N.  (1986).  Interaction, acculturation, and the acquisition of communicative competence:  A case study of an adult.  In N. Wolfson & E. Judd (Eds.), Sociolinguistics and language acquisition (pp. 137-174).  Rowley, MA:  Newbury House.