Given that little research has been conducted to date in the classroom about the exploitation of aspects of "visual grammar" for the teaching of literacy, the purpose of this study is to provide research data to support the adoption of a common image/text relations metalanguage in educational practice as an effective tool for critically negotiating, and mainly, for composing intermodal meanings. Forty-six sixth-grade students who attend a state primary school in the city of Ptolemaida, northern Greece participated in the study. The materials used in the study consisted of: (a) informational, print-based multimodal texts and, (b) compositions produced by students, at first individually and then in groups. The overall research is designed on a pre-test phase, an instructional intervention phase, and a post-test phase. The qualitative comparative analysis of student compositions manifests that the metalanguage of "visual design" constitutes a truly promising, pedagogically utilizable tool for the description, interpretation and comprehension of the interactions among the various semiotic modes co-existing in multimodal ensembles. This entails the development of multimodal and visual literacy skills by the primary education students. These findings highlight the need for adoption and incorporation of such a metalanguage for the design of curricula facilitating the teaching of literacy, in order to reframe the monomodal nature of communication.