The expression of modal meanings presents a difficulty even for advanced speakers of a foreign language and it has been shown that beginning and intermediate learners produce few modal auxiliaries. To investigate the phenomenon, a functional/semantic framework was used to determine what forms these students use to convey a particular meaning. The framework was based on Palmer's "Modality and the English Verb." In that book, two basic kinds of modality are distinguished--deontic and epistemic. Deontic deals with expressions of obligation; and epistemic, with expressions of probability. Interlanguage data for this analysis came from two groups of students receiving special instruction in conversational English. In one group the three speakers were native Japanese speakers; in the other group, there were two Arabic speakers, a Chinese speaker, and a Spanish speaker. Tape recordings made of free discussion and of presentations were analyzed by the teachers and by the researchers. A tally of forms occurring at least three times were subjected to further analysis. It was found that for each modal function, each subject seemed to have a preferred expression. Also, the ESL learners expressed only a limited number of functions. In conclusion, the advantage of the functional/semantic framework is stressed. Relevant tables are appended. (AMH)