"Phases" are poorly defined. They exist wherever there's a statistical mechanics reason to define them, and they make sense only within the limits of the approximations on which they are based. Glass is "solid" to a mechanical engineer (in most regimes), but it's not a crystal and doesn't behave like one to a solid state physicist. Likewise "plasma" is a great tool to describe a fluorescent light bulb or lightning bolt, but not so much a white dwarf. Even stuff you think you understand, like "liquid and gas are clearly different" turns out to be wrong (http://en.wikipedia.org/wiki/Supercritical_fluid).