abstract the present study investigated the effects of task types and involvement load hypothesis on incidental learning of 10 target words (tws) in junior high schools (jhss) in givi, ardabil. the tasks deployed in this study were two input-based tasks (reading plus dictionary use with an involvement index of 3, and reading plus gap-fill task with an involvement index of 2), and one output-based task (sentence writing with an involvement load of 3). accordingly, the study is two faceted. part a tests the claim made by the ilh in order to see if this hypothesis plays a role in the learning and retention of ten tws by students in jhss; and part b surveys the different effects, if any, of the two types of tasks with the same involvement index. eighty seven participants initially underwent two pretests, one proficiency test to ensure the homogeneity of the participants and one vocabulary pretest to ensure that the tws were unfamiliar to the participants. only 60 participants were chosen and randomly assigned to three experimental groups, two input-based groups and one output-based group. two unexpected passive recall vocabulary posttests were administered to examine the learning and retention of the 10 tws. the results of paired sample t-test, in part a, partially supported the ilh for elementary learners because while in the immediate posttest, task one (reading comprehension plus dictionary use) yielded higher vocabulary learning as compared to task two (reading comprehension plus fill-in), in the delayed posttest the reverse held true. and the results of part b revealed that the participants in task three (sentence writing) outperformed those who had taken input-oriented task (reading plus dictionary use). the results of the study suggest that greater importance should be attached to the effect of kind of tasks in interpreting the role of the ilh. keywords: input-based tasks, output-based tasks, incidental vocabulary learning, involvement load hypothesis, input hypothesis, output hypothesis.