The neurophilosophical project, as envisioned by Churchland, involves interrheoretic reduction, moving from (or eliminating) theories formulated in terms of common sense and folk psychology, to theories that have stood the test of scientific experiment. In her view, folk psychology, as well as introspective phenomenology, will be eliminated in favor of neuroscience. Neurophenomenology holds that phenomenology (as a practice) is not only possible, but is in fact a useful tool for science; and that phenomenology is ineliminable if the project is to pursue a neurobiology of consciousness. Clarification of these issues rests on an understanding of how phenomenology can be an alternative source of testable theory, and can play a direct role in scientific experiment. Rather than talking in the abstract about the role of theory formation in science, I consider two specific issues to show the difference between a neurophilosophical approach and a neurophenomenlogical approach, namely, the issues of self and intersubjectivity. Neurophilosophy (which starts with theory that is continuous with common sense) and neurophenomenology (which generates theory in methodically controlled practices) lead to very different philosophical views on these issues.