@BenLim Yes, this is the definition of a parabolic subgroup equivalent to the fact that it contains a Borel subgroup. The point here was that Q is not necessarily parabolic. The intersection P\cap Q is also not necessary parabolic. So you need some extra arguments to see this.
