Some accounts have the wizard Merlin living his life from future to past, in the opposite direction to the rest of us. This meant that to him, what would a final farewell to us would be a joyous first meeting for him, and a first meeting would a sad goodbye. He remembered the future, but the past was a complete mystery to him.
The trolls in Terry Pratchett’s Diskworld had similar ideas. They considered that they travelled through time from past to future, as is normally understood, but they also considered that we were facing backwards in time as we travelled through it. This was conjectured by the trolls, to explain the fact that we can see where we are going when we travel in whatever direction we choose but we cannot see where we are going in time. Similarly we can’t see where we have been when walking from one place to another, but we can see where we have been in time.
Scientists have no difficulties with direction in time. In an equation with a time variable in it, one can trace the changes to other variables from the nominal zero point to see what would occur in the future, merely by incrementing the time variable. A scientist can predict the trajectory of a thrown item (a parabola) merely by substituting later times into the time variable in the equation.
y =ax² + bx + c
The scientist should compare these results to experiment and find, that this more or less works. Lets say though, that the scientist is an astrophysicist investigating an asteroid or other object on a parabolic trajectory around a larger object, like a moon or planet. (A parabolic trajectory is the trajectory which divides objects in hyperbolic trajectories which are not bound by the larger body’s gravity, from those in elliptical trajectories where the object is bound by the larger body’s gravity).
The astrophysicist would be able to track the trajectory of the body backwards in time, simply by substituting negative values for time into his equations. He would be able to see where it had been. However this process of substituting negative time values into the equations only works so far. At some point the body will feel the influence of other objects and the retrograde trajectory will deviate from the values predicted by the parabolic equation.
Of course, if the trajectory is projected forwards far enough similar considerations arise. Eventually some other body will divert the body away from the parabolic trajectory. However in the region in which the parabola applies, the behaviour is symmetrical with respect to time. From a film of the event one could not tell whether or not the film was being run forwards or backwards.
Of course, if a film is run backwards for any length of time, it becomes obvious that something is wrong. Things fall upwards, and broken crockery comes back together again. People walk backwards.
On a closer look, people can intentionally walk backwards and it is possible that a spring or other mechanism could be used to shoot things upwards, but it is a lot harder to imagine a way of reversing the breakage of the crockery. It implies that some process involved in the breaking of crockery is not reversible at a macro level.
It is likely that the process in question is at the molecular level or slightly above. To rejoin two broken surfaces spontaneously would presumably require that the molecules be in the correct positions and that a little burst of energy (equivalent to the little burst of energy that comprises the sound that the crockery makes in breaking and any heat release) be supplied at an instant in time. The weak bonds between the parts of crockery would need to be created, and that is really difficult.
There’s therefore a discrepancy between the scientist’s view of the world, through his equation which time-symmetrical, and the man in the street’s view of the world, which is asymmetrical with respect to time. In fiction this asymmetry is used to good effect, when the protagonist may “wind back time”, to write a wrong or divert history to an alternate course.
When we consider space we usually imagine a three dimensional space. Events happen at locations in this space and three coordinates are enough to locate an event in space. Every possible point in space has its set of unique coordinates. It is common to add an extra dimension for time, making the space four dimension and consequently difficult to imagine successfully.
All events in space and time have a location and time and are represented by a point in the four dimensional space. The path of a particle is a line within this space, and any point on the line represents the position of the particle at a particular time. Positions on the line are either before or after this point, so it constitutes a “now” point for the particle. There is no actual motion over the line, since in the 4-D space all points represent the past and the future of the particle. They are already there.
To move to an earlier time, all we need to do is move to an earlier spot on the line. We would need to either travel back along the line at a speed of so many seconds per seconds where the first “seconds” is seconds measured in the space and the second “seconds” is some other time scale or we could jump out of the space-time completely and return to the requisite point earlier in time. We’d need to do the latter in some other space-time that embeds the original space-time, adding a number of extra dimensions to the mix.
Both options require the addition of extra dimensions, which while possible complicates the situation unacceptably to my mind. The process of adding extra dimensions could be repeated and go on forever, so we end up with infinite dimensions. I believe that it is correct to employ Occam’s razor at this point and declare that it appears unlikely that we could either roll back time or jump to an earlier point in time because of the implication that we would needs infinite dimensions as a result.