This paper attempts to establish a relation between the role played by metapragmatic awareness in oral comprehension and the skill to produce written narrative texts. It is also part of a greater project that looks at metalinguistic and/or cognitive abilities that might associate with the development of oral comprehension of non-literal language in school ages and the possible differences in the development of reading and writing in boys and girls with distinct comprehension levels of figurative language. It is believed that metapragmatic awareness starts at six years of age and that it involves verbalized conscience of the difference between language and context plus a reflective control over this relation in a given communicative situation (Gombert, 1992). Verschueren (2002) argues that the indicators of metapragmatic awareness work as (a) mechanisms that seek linguistic forms relative to the context; (b) reflective interpretation signals of the activities language users do; and (c) contextualization cues. The participants were 82 third-graders who took a metapragmatic awareness test as well as a writing test, which was graded according to a focalized holistic scale. Findings show that the better performance in composition, the more metapragmatic responses, and it is concluded that metapragmatic awareness is not systematic in all communicative situations but depends on the way the context relates to the language used.