As text understanding pipelines mature through the advent of richer and deeper semantic representations, one of the largest unaddressed questions remains how to cope with the shifting dynamics of language, micro-languages, and more specifically, content that is not in English. This talk aims to explore the limitations of current approaches from translation to Latent Semantic Indexing, among others, and to propose interesting avenues for further exploration. This talk is based on research in collaboration with several EU FP7 funded projects, DFKI, and collaborations with CMU, Stanford, and Columbia University.