The research reported in the study was undertaken to measure English language teachers' ability to interpret second language learners' intended meanings in idiosyncratic utterances in written English. In doing so, it also aimed at verifying Corder's (1981) speculation that language teachers can correctly guess the intended meanings of learners in a large number of erroneous sentences if they are familiar with learners and their first language. Given the size of the data, the study suggests a cautious interpretation of the results, which show that teachers' interpretations of learners' intended meanings are, in many cases, off the mark. The study implies that the traditional practice of checking learners' written work in their absence and providing correct versions of incorrect forms or expressions contained in the work may have little learning value for learners. The study therefore suggests following more effective means of correcting learners' written work and providing feedback.