I have summarized some of the currently-available evidence in favor of viewing HI as a product of the star formation process rather than a precursor to it, and presented a case that H2 is likely to be present in significant quantities in most galaxies. How much H2 is really present in any given galaxy? I believe this is presently an open question, the answer to which is likely to have important consequences for several areas of astrophysics.
Acknowledgments. I am grateful to my colleagues at STScI for helping to provide a scientifically-stimulating atmosphere, and especially to our theorists Mike Fall and Nino Panagia for discussions on the ideas described in this paper.