Research and discovery seem to be driven by a certain tension between our ideas and our observations. On the one hand, our observations provide facts about reality, and it is very pleasant when these facts agree with our understanding. On the other hand, there will often be cases where the observed facts cannot be readily explained. The latter will either not fit our ideas or be incompatible with our basic understanding and, while this can be an unpleasant experience, it forces us to rethink our ideas in different and perhaps more productive ways. The tension between ideas and facts, i.e., between "something going on" and "something being wrong", will be illustrated through the research findings that led to a new approach for estimating parameters of the human vocal-tract via a simple and seemingly absurd, linear transformation of parameters of speech spectra.