Defining what is/can be included in a description of what constitutes the universe is a frequent topic of interest. Most of us generally agree it consists of 'all that is possible', but have difficulty achieving a consensus on what is vs is not 'possible'. While I tend to steer clear of purely philosophical approaches, I am attracted to efforts to marry logic with mathematical structure. So [after much adieu], I thought some similarly afflicted souls might find this interesting: http://www.anthropic-principle.com/preprints/spacetime.pdf. Nick Bostrom has written some very engaging and well grounded philosophy of science stuff over the past decade. He has achieved some degree of repectability in more traditional circles having coauthored a couple of papers on Arxiv with some familiar names: http://arxiv.org/abs/astro-ph/0512204 Title: How unlikely is a doomsday catastrophe? Authors: Max Tegmark (MIT), Nick Bostrom (Oxford) http://arxiv.org/abs/gr-qc/9906042 Title: Cosmological Constant and the Final Anthropic Hypothesis Authors: Milan M. Cirkovic, Nick Bostrom I'm also fascinated by Lucien Hardy's work on causaloids, so take that into consideration.