Cultural Context and Multimodal Knowledge Representation: Seeing the Forest for the Trees