In general relativity, time and space are a set of variables that can be used to parametrize the geometry of spacetime and the kinds of geodesics that are possible. But they are not the only kinds of variables that form a set of four coordinates that 'span' the dimensionality of spacetime. In probing the mathematics of black holes, physicists have discovered other sets of coordinates that are even better. For example, the event horizon appears in the mathematics as a 'coordinate singularity' if you use the coordinate set (x,y,z,t) or (t,r,theta,phi) but if you use the 'Kruskal-Szekeres' coordinates, it vanishes completely. There is only one true singularity in a non-rotating Schwartzschild black hole solution and that is the one at r=0, at the event horizon, the curvature of space is non-infinite so that means that coordinate singularities are not real singularities and can be mathematically transformed away. Now, if you study what happens to the Kruskal-Szekeres coordinate system as you pass inside the black hole event horizon, nothing unusual happens, but in the conventional newtonian (x,y,z,t) system, if you look at the formula for the so-called 'metric' you see that the space and time parts reverse themselves. This means that just inside the horizon, space becomes time-like and time becomes space-like. What we call time does change to something with the mathematical properties we have normally associated with space.

This sounds pretty bizarre, but considering that we are using a non-proper coordinate system in the first place, I would not put much stock in how to physically interpret it. It is possible that time changes somehow inside a black hole, but that is an experiment we will never be able to test because we can never receive information from inside a black hole. Because (x,y,z,t) is a 'bad' coordinate system in the first place for describing black hole geometry, I do not feel compelled to ask it to make too many predictions.

Return to Ask the Astronomer