Over the past couple of decades, fundamental semantic analysis has been outside of the purview of mainstream natural language processing, but the recent surge of interest in cognitive computing has returned the notion to center stage. However, what exactly is fundamental semantic analysis? What methods can be used to achieve it? What counts as a useful result? How does one measure progress? All of these issues, and many more, define the choices pace for building semantic analysis systems. This paper describes key aspects of this choice space and suggests that published descriptions of systems should explicitly state both the choices made and the positive and negative consequences of those choices. This will counter the current tendency for system descriptions to tacitly reflect choices understood only by insiders without overtly motivating or justifying them.