This paper presents a provisory, social semiotic-based definition of the emerging multimodality research field, as the recognition of the fact that diverse meaning-making modes like speech, gesture, writing and image, constantly reformed in social interaction and by culture, are joint in texts and communicative events, that may or may not imply language. It further considers some of its core notions, such as mode, resource and medium, and proposes three orientations in the field: systemic functional-multimodal discourse analysis, social semiotic-multimodal discourse analysis, and multimodal interactional analysis (sharing interest with pragmatics).