- #101
JesseM
Science Advisor
- 8,519
- 16
No, that's not how I was thinking of it. Your "seed" comment suggests you're thinking in terms of picking a random set of initial conditions, then evolving them forwards using dynamical laws. I'm thinking of generating entire histories at random, with no laws constraining them whatsoever. I used the analogy of a chess board earlier--instead of picking a starting configuration for the board and evolving it forward using only legal moves, imagine simply picking a random number of moves that the game will last, then making a slot for each position on the board at each move, and then assigning each slot a piece (or designating it empty) completely at random, like "A4 on move 12 will contain a black rook" or " E11 on move 3 will be empty". The vast majority of histories generated by this method won't look like a legal chess game at all, the number of pieces and their positions will be changing randomly from one move to another, but the computer can then go through and throw out every history that does not obey the "laws of chess" from beginning to end. Similarly, I am imagining something like a computer which generates a spacetime manifold whose curvature varies in a completely random way, with the worldlines of objects also drawn at random, and then all of these are thrown out except for the spacetimes where the metric relates to the density of matter/energy according to the rules of GR, and where the worldlines also obey the correct laws of physics (being geodesics in the absence of non-gravitational forces, for example). Of course, the complete description of a "history" in quantum gravity may be something other than worldlines on curved spacetime, but whatever the basic description you should be able to come up with an analogous notion of "random histories".moving finger said:
Yes. These histories are not logically inconsistent, but they contain points where the laws of physics are not obeyed. This would probably be because the laws of physics are not obeyed throughout the entire history, just zigzagging worldlines and changing spacetime curvature that follow no laws at all, but there would also be occasional histories where the laws of physics were mostly obeyed but that contained specific points where they weren't, like if I went back and killed my mother but then at some moment her dead body suddenly disappeared and she was suddenly walking around with no memory of having been visited by a time traveler.moving finger said:
Yes, but this is just an analogy, the point is just that the histories we see the computer return as output will obey the laws of physics throughout, and will thus be consistent. I'm not suggesting that there is some entity sorting through random histories and then picking one where the correct laws of physics are obeyed throughout, waving his magic wand over it, and making it a "real universe"; it's just a way of thinking about what it means for the "laws of physics" to exist as timeless constraints on entire histories (like general relativity, or the principle of least action in classical physics) as opposed to dynamical rules that start from some initial conditions and evolve them forward.moving finger said:
