The present paper argues the use of "narratives" as the most appropriate evaluation method in cases of atypical language production. Narrative as a genre has an ecological validity that other genres used in language research and evaluation do not have. Narratives develop naturally from very early, they are independent of education and academic skills, and they are meaningful because of their direct contact to experience. As a result the paper argues that "narrative analysis" is the most appropriate evaluation and research method for atypical language, one instance of which is deaf writing. The paper will present an example of narrative analysis application on a story written by a deaf writer. Via the illustrative example it will be shown that narrative analysis captures deeper levels of the language production by exploring the "content of information", and the "structure of text". (Contains 3 tables, 3 figures and 3 notes.)