I think Steve's got the basic idea right, at least as far as modern science is willing to state anyway (apart from making a distinction between matter and information... tut tut!).
'Time' is our experience of processes evolving from the way they are 'now' to the way they are in the next instant, and the next, and the next... There is, as far as we know, no lower bound on how finely you can define 'time', although there may be a fundamental limit below which further division is meaningless (kind of the analogue of the Planck length).
We define the second by the number of oscillations of a Caesium atom (usually) and since that physical process depends entirely on your frame of reference, perceptions of time vary from person to person. Steve is also right that distance and time are fundamentally linked - the further away you look, the 'longer ago' you also look. There is, as far as we know, no way for you to influence what you would think of as the past - that is, it doesn't matter how fast you travel, or how massive a body you sit next to etc etc, you can never 'loop back' and change something within your own past light cone.
There might be tricks, of course, if for instance particles are found which travel backwards through time and that we can impart information onto, it could be possible to 'send back' information to a previous observer who was measuring those particles. That's the '12 Monkeys' approach, pretty much, that the person in the past just carves info onto a rock or something and then the person in the future reads the rock and imprints a response onto these particles. This violates causality, of course, so conventional science forbids it, but who knows!