If you were to travel to the past, there would be the possibility that you might meet yourself, your former self, in the past; it might also happen that were you to travel to the future you might meet your future self.  Further, you might find that the self you meet is not you--that is, that this meeting never happened for the older version of you when he was the younger version.  Such other selves are called temporal duplicates or doppelgangers, and are another form of temporal anomaly which occurs in time travel.

Under fixed time theory, the resolution of such an anomaly is relatively simple:  if the meeting occurs, it will have occurred for all versions of the traveler.  If you did not meet your older self when you were younger, you will not meet your younger self when you are older.  It is not something to be avoided; it is something that cannot be caused to happen.

Parallel and divergent dimension theories also treat the matter simply:  you exist in each such universe, and if your other self has not also left for yet another universe you might meet him and change his life significantly, but since he is not really you, this does not matter except in that you are demonstrably an alien visitor to this universe.

It is under replacement theory that temporal duplicates become an issue.  Under this theory, there must be an original history of the world in which the time traveler did not meet himself; then the time traveler moves to the past and can meet his younger self, even though he did not when younger meet his older self.  This will change his history.  His younger self must then (to prevent an infinity loop) make the same trip when he is older to meet in turn his younger self, but because having met himself has changed him and he may remember the meeting, this event might be different from what he remembers.  Again the younger self must make the same trip in turn.  As long as each time traveler eventually makes the same trip as the original, there is a high probability that the sawtooth snap will resolve to an N-jump, and time will continue.  In the final version of time, the time traveler will have played both roles in the meeting, but the "two" meetings will have been identical.

The problem is less complicated with temporally duplicated objects, but that again the "younger" of the two objects must in its turn make the same trip made by the "older" object.  This means that temporally duplicated objects (or people) can exist only from the moment of the arrival of the older to the moment of the departure of the younger.

The problem is similar with travel to the future, but with an extra wrinkle.  If a traveler goes to the future, in the original history he will find that he vanished from the universe at his point of departure and was never seen again; if he then returns to the past, he restores himself to the timeline, such that it is possible for his younger self coming forward to encounter him as his older self.  This again alters the younger self, who must now return and become the older self for the younger self to meet, again with the hope that time may eventually stabilize such that "both" meetings are identical for the one traveler who experiences the meeting twice.  Ultimately it is travel to the past that creates the duplicate, as the traveler backtracks on his own history; it is experienced differently if it begins with a leap to the future.

Next time we will address what is known as Niven's Law, one specifically about time travel.