In this talk, we will present and discuss a system (hGen) used for the generation of natural language expressions from a base first-order logic representation. This potentially multilingual, scalable system is embedded in the larger inspectGraph system, an application geared at the multimodal (visual and linguistic) presentation of and interaction with otherwise purely graphical information (such as line graphs, pie charts, etc). The architecture of the inspectGraph system is also centered around a FOL (first order logic) semantic description of a graph and hGen is the interface between the user and the graph itself. We believe that the translation and summarization of data graphs into more interactive data structures such as text is a natural topic to relate to software for business intelligence on the one hand, and to an improved access to information in the hands-busy, eyes-busy contexts or by visually-impaired individuals.