In the current research project the functions of self-initiated self-repairs in an advanced Japanese language classroom were investigated. From the reviewed literature it was found that repairing is possible through monitoring, which includes error detection, and involves attention and memory. Therefore, data was collected on the abilities of the participants in the current research project to repair, monitor and their attention levels and memory. There were seven methods used to collect data; participant observation, classroom interaction tape recordings, a questionnaire, stimulated recall interviews, attention test, memory and attention test and proficiency level assessments. From the participant observation, classroom interaction tape recordings and stimulated recall interviews data was collected on the repairs that the participants made and the way in which they monitored was explored. The questionnaire revealed relevant background information, such as, number of years the participants had studied Japanese, which supplemented other information collected. The attention test and memory and attention tests were used to obtain information on the participants’ perceptions of their levels of attention and their actual levels of attention and memory respectively. The final data collected was on the participants’ own perceptions of their speaking proficiency levels in Japanese and an independent judge’s evaluation of their levels. The results of the data collected on the way in which the participants repaired and monitored showed that overall the participants repaired and monitored in ways that had been discussed in previous research on repairing and monitoring. However, for the first known time, phenomena related to repairing and monitoring, which the researcher terms as communication strategies, are used frequently by the participants and also, that these communication strategies used and repairs made by the participants were not needed to be made. In other words, the participants in the current research project were often repairing errors that had not been made. As for the data collected on the participants’ attention levels and memory, no significant differences were found between the participants and neither did any differences reflect on the ways in which the participants repaired. From the proficiency level assessments, further evidence was found that supported previous research done on repairing and monitoring as well. Also, it was found that the participants under-estimated their levels of proficiency in comparison to the independent judge’s evaluations. Therefore, the participants both over repair and under estimate their Japanese speaking abilities. This is seen as detrimental to the participants’ performances in Japanese. Recommendations are made to use this data provided by Japanese language learners when repairing to guide instruction and to encourage learners to gain fluency by repairing less often than is thought necessary.