▼ This thesis examines the effects of the language used in an idea generation task on the quantity and potential quality of ideas developed during a timed idea generation task; quality here being how developed the ideas are. This was done by having two Japanese ESL students, of differing proficiency levels, engage in three rounds of idea generation tasks using their native language (Japanese), their second language (English), or the language of their choice depending on the round. When the participants finished a round of idea generation, they then wrote a short essay, in English. To code the data, the Japanese was glossed into English. Then all idea generation tasks (English, Japanese, Choice) were coded using a modified version of episodic units (see Brice, 2005). This thesis corroborates findings of Wang and Wen (2003); which seems to suggest that the language used in idea generation may correlate to a writer’s level of English proficiency. Also, it was discovered that the participants of this study appeared to develop their ideas more thoroughly in English, as opposed to Japanese. This may be because of differences in the languages “communication mode(s)”(Scarborough, 1998). Advisors/Committee Members: Reichelt, Melinda.