The present article focuses on the role of mental cognitive processes in creating novel imaginative characters in stories. To do this we apply the basic concepts of Conceptual Blending Theory (CBT) i. e. Input Spaces and Blending Networks. CBT accounts can provide a methodological tool for cognitive analysis of literal texts. Fauconnier and Turner (2002) provide an overview of how blending affects the course of a human life, and more specifically, how young children are engaged in building complex blends in very early stages of their life. In the present article we determine to show the cognitive bases for creation of "emergent characters" in stories of elementary students' textbooks (Farsi Khandari). Our main question is what are novel/emergent characters in stories introduced in each grade and what are the available cognitive tools for the writers of each grade to make theirs characters creatively different from other writers. The result can provide a methodological cognitive framework for cognitive analysis of the process of novel character creation in children literature.
 1. Introduction
Cognitive semantics is the study of the relations among experience, conceptual system, and the encoded semantic structure throughout the language; thus utilizes the language as a methodological means of studying cognitive phenomena. Cognitive semantics conveys a methodological means to answer the question that 'how do we use our mind to generate meaning in stories. Conceptual blending theory is regarded as a useful way for semanticization, in the case of the emergence of cognition in literary texts as well as their creative aspects.
Here, we investigate the blending of main characters in the elementary school textbooks (Farsi Khandari); we try to find the blending network representation for each character as well as finding the 'creativity' in 'new-appearing characters' of each school level to get to a cognitive analysis of characterization process in elementary school textbooks and, as a result, to answer the following questions: Based on the blending principles, which kind of characters have been created in comparison with the previous and proceeding levels; and, what different choices have the authors taken to have novelty in creating new-appearing characters.
To choose samples, we used purposeful sampling, a kind of a sampling to achieve representativeness or comparability. This kind of sampling has six ways that survey maximum variation sampling, the maximum variation in qualities, and the characteristics of the studied sample as well. The importance of the present study is both theoretical and applied aspects: in the theoretical aspect, it shows the role of mental space and the blending in semanticization procedures; in applied aspect, on the other hand, it reveals the importance of the blending roles.
This study is based on Fauconnier's (1997) blending theory, followed by Grady (1999), Coulson and Oakley (2000), and Alonso (2003).  The main question of this theory is that the emergence of meaning contains a blending structure that creates a meaning much farther than its components. The blending network has inputs which are connected by mappings.  The mechanisms of these mappings are as follows: selective projection, composition, completion, elaboration, backward projection, human scale, and vital relations. Depending on what type of blending space we face, we have simplex, mirror, single-scope, double-space, and multiple blending networks.
Results show that textbook writers have chosen different input spaces to create new characters. Input spaces turn more abstract, and the blending networks become more complicated as the level increases. The results of this paper give us a framework of cognitive analysis of the creation process in children's stories.