"Rhetoric" is the virtual use of sentences to create secondary meanings. In children's books, the image is one of the most important grammatical tools of sentences. Conversely, sentences also change the meaning of images. According to Roland Barthes, written text affects visual text, and vice versa. The purpose of this study is to investigate how text and image interact from a narrative and semantic perspective. The present article seeks to answer three questions about children's literary-visual texts, especially in illustrated stories: To what subdivisions are the five semantic relations of the image with the text (listed by Scott and Nikolajeva)  divided? How does typography anchor their content? Morphologically, what interactions can there be between two texts based on image code? The rhetorical function of the image in picture books can be classified from two perspectives: one from the semantic perspective and the other from the narrative perspective. In this paper, first, from the semantic point of view, the types of interactions between the image and the written text, inspired by the Scott-Nikolajeva division, are introduced and also divided into smaller subsections; then, the rhetorical function of typography and the rhetorical function of pictography (an image formed like a line-drawing of letters) in children's narrative texts are examined, and finally, in the final section, the actions of the pictorial subtexts on each other are classified.  The required data for this paper was gathered through library research method. The purposive sampling was used to select the samples. The research method is descriptive analysis, and its field of study is children's stories. Since the art of illustration is an integral element of children's texts and that the most objective language for the preoperative child is the image, the need for interdisciplinary views of the image and visual literacy is felt. To date, the author has not observed the morphological classification of image functions, which is introduced in this article, in a similar study (both internal and external). Also, for the first time in Iran, this study examines the relationship between image interaction and image, as well as the relationship between typography and text.
 Golden (1990) offers five categories for text-image interaction: 1. Text and image are symmetrical; 2. The clarity of the text depends on the image; 3. The image is the expander of the text; 4. There is a semantic burden on the text; 5. The main burden is on the image Nikolajeva and Scott did not consider the Golden classification to be comprehensive and proposed a fivefold classification: symmetrical, complementary, enhancing, counterpointing, and contradictory. This study first tries to open the mentioned five semantic relations by giving examples and shows that these five semantic interactions can be considered on a continuum that starts from the complete harmony of the image with the text (symmetry relationship) and tends to intensify inconsistencies and create gaps between text and image; the other end of the continuum ends in complete inconsistency (contradictory relationship). This study shows that "contradictory interaction" is based on the contradiction between image narrative and written narrative and has a high frequency in humorous and ironic texts. Morphologically, it can be divided into two categories: scattered contradiction and anecdote contradiction.
"Scattered Contradiction" can also be used in two forms, "regular" and "irregular". In the scattered pattern, there is a contradiction between the image and the text throughout the text. This scattering can have a regular pattern or no specific pattern. For example, if in all episodes of the text, the contrast of the text with the image is repeated, the contradiction will be "regular scattered".
Unlike "scattered contradiction", in the "anecdote contradiction" pattern, a contradiction occurs in the narrative ending between the written text and the visual text.
Typography can be used in two contexts: in the context of text, in the context of image. Playing with typefaces that are printed is called typography, and playing with typefaces that are handwritten is called hand-lettering. Whether the wording is printed or handwritten, it creates a different semantic interaction in the narrative. If the text code is placed in an image context (not just an anchor on the image), it is called "image writing" or "word image". For example, in the "Vāv" (Zanjanbar, 2022), a picture of a cow's head is drawn with the word "cow".
An Image-based text is an image centered on an image (whether or not text is anchored to that image).
One of the results of the study is the introduction of image narrative functions (Syntagmatic interaction, Nesting, Substitution, Parallelism, Paratextuality).
Syntagmatic interaction is of two types: time-oriented and space-oriented. In Time-Centered Companionship, images along time form a narrative.
 Images in collage, logo-collage, episodic segmentation based on place-orientation.
In "Substitution Interaction", an image text replaces the old image and can narrate different functions such as "animation function" and "metamorphosis function".
In the "substitution interaction" pattern, part of the image changes and part remains the same. In preschool age group picture books, with the help of creative book-making, graphic, and drawing tricks, it is possible to magically suddenly turn one image into another (play with images). In fact, the new image is replaced by removing part of the old image. Substitution interaction can narrate different functions such as "animation function" and "metamorphosis function".
The animation technique can be seen in the two-volume science-fiction series "Ricky Ricotta's Mighty Robot Adventures" by Pilkey (2020). Mosquito wants to destroy the planet by capturing mutant mosquitoes, but Ricky fights him with the help of his robot. The book depicts the exciting scenes of their confrontation with animation. Before animation, the reader is instructed on how to place the hand on the page of the book and how to hold the sheet containing the image and shake it. The two pictures (which are slightly different) are on two consecutive pages of the book (i.e. on two pages with consecutive odd numbers). By holding and shaking the first sheet, the child creates an interaction between the two images and achieves a synthesis called animation. In this animation, the robot appears to have killed the monster and is shaking its arms triumphantly like a gladiator over its corpse.
"Metamorphic substitution" can be manifested in the form of "single frame" or "multi frame". The pattern of "metamorphic substitution at the multi-frame level is one of Ali Khodaei's favorite tricks. In "The Same and the Same" by Khodaei (2016), the every other pages of the book are cut in the middle and parallel to the width of the book. That means half the page can be flipped. When half a page is flipped; the images change their nature, for example, what is seen as a mane in the image of the roaring lion, turns into hedgehog blades with a half leaf. Transformations that occur at the single-frame level; they are usually based on the "image ambiguity" array.
"Nested interaction" consists of two image-based texts, each inside the other, and the semantic gaps are filled by these nested images (along with their text anchors).
In "parallel interaction" the same interaction occurs for two visual texts that narrate two narratives in parallel.
Contextual elements such as "material texture, how paper is cut on each page, or die-cutting on the cover" can play the role of catalyst and mediator in spatial-oriented accompaniments to create a narrative function.