This study investigates the role of inferencing in cloze test performance and the factors that facilitate that inferencing. Four groups of high school students learning French as a second language completed sets of cloze passages under four treatment conditions. Three conditions provided a potential cue to inferencing, while the fourth was a control condition with no additional aids. The results indicated significant differences in performance attributable to the four classes, the position of the story in the set of passages, and the inferencing treatment condition. A more detailed analysis examined the relationship between experimental condition, type of error committed, and the form class of the original word required by the text. Differences in these patterns were apparent, although non-significant. It was found that certain inferencing cues can have a facilitating effect on students' cloze performance although the pattern of facilitation is different from that found on other criterion tasks. Furthermore, it is possible to encourage students' inferencing behavior through classroom training, thereby improving cloze performance. (Author/AMH)