Hyland’s Model was the first framework to examine interaction between writer and reader of a text under two key concepts, namely, stance and engagement. According to this model, any writer (un)consciously employs a number of linguistic devices such as hedges, boosters, attitude markers, and self-references to present his or her stance over a topic, and likewise uses some other devices like reader pronoun, imperatives, interrogatives, shared knowledge as well as personal asides to establish reader’s engagement. In this study, a number of book reviews on Persian literary books have been analyzed based on the model proposed by Hyland (2005). The corpus of study includes 100 book reviews; 50 reviews are selected from academic, refereed journals and 50 ones are taken from general journals. When the corpus was prepared, the linguistic devices of stance and engagement were identified, extracted, and counted. The findings showed that for each 11.7 words of the corpus, one word or expression of stance or engagement has been used. The negative attitude markers and imperatives are respectively the most and least frequent devices. It is shown that book review articles, printed in academic journals, tend to be more critical, and, furthermore, the articles published in the 2000s put much more weight on attitude markers. It is also shown that the density of interaction markers is higher in the ending section of book reviews. Taking the gender variable into account, it is revealed that male critics are more explicit in presenting their stance, while female critics are more willing to use engagement devices.