This paper summarizes non-native speakers' reading-time results for two types of constructions. In the first experiment, native Chinese speakers were faster to read relative clauses in Japanese when extraction was from the subject position compared to when extraction was from the object position. In the second experiment, native Japanese speakers slowed down when reading sentences in English containing a subject-verb agreement violation. In both experiments, the results for learners are similar to what has been reported for native speakers, suggesting that non-native readers process sentences in ways that are similar to what has been reported for native readers. The paper also discusses methodological issues such as advantages of reading-time techniques over questionnaires, issues specific to non-native readers such as proficiency, as well as issues such as spillover effects, which are likely to be particularly problematic when collecting data from non-native speakers.