This study is going to consider the factors vital for comprehension. The Graded Salience Hypothesis is the framework for this study. Persian indirect requests will be tested and the relation between different contexts, familiarity level and reading times will be studied. At first, figurative and literal contexts were prepared. A software for measuring reading time in self-paced reading experiments was designed. In the first pretest, participants defined the familiarity of expressions. The second pretest aimed to confirm the sameness of context bias. In the first part of the main test, participants read each indirect request in a figurative context and reading times were recorded. In the second part, participants read each indirect request in a literal context and reading times were recorded. After comparing the reading times, it was concluded that Graded Salience Hypothesis predictions were not confirmed and sometimes, context was a more important factor than salience. Therefore, instead of a parallel process, a semi-serial process was witnessed. Therefore, among Persian Indirect requests, salient meaning in familiar and less familiar figurative expressions was figurative meaning. In unfamiliar indirect requests, the salient meaning was figurative and literal meaning. Therefore, literal meaning was not salient meaning and this finding challenges the modular-based views.