This work is intended to find replies to practical questions, such as how well native speakers of Spanish are decoded by native speakers of English, which errors interfere with decoding by the listener, and what the implications of the study might be for teaching. Fifteen Chileans were asked to tell stories in English, and several panels of native English-speaking listeners were asked to judge these stories for intelligibility, comprehensibility, grammaticality, and acceptability. It was hypothesized that although correlated, intelligibility (ease of being decoded) and comprehensibility are different concepts. It is also thought that the variables affecting intelligibility and comprehensibility are roughly the same. No correlation was found between intelligibility and comprehensibility, on the one hand, and grammaticality and acceptability, on the other. In spite of a large number of errors, the non-native English speakers proved to be easily decoded by the native English speakers. A synthesis of isolated errors was submitted to additional panels of judges. Pronunciation errors proved the most critical, followed by semantic and syntactic errors. The relatively low criticality of syntactic errors and errors of intonation seems to be of pedagogical interest. Intelligibility and comprehensibility proved not correlated with formal proficiency testing. (Author/AMH)