Sorry, actually sound files are stored onto the microcontroller and then, once the user says something, like "access file one", the microcontroller will play that audio. Instead of just linearly playing the audio, I wanted to give the user the ability to rewind and fast forward. The only way I could think about how to do this was create a variable which could be decremented (for rewinding) and incremented (for fast forward) a user specified number of times. This would be much more preferable than having to wait until the entire program is done before repeating. That would be bad.
So, once the program has read a section of text, the user is prompted to say "continue", "stand by", "go backwards" and "go forwards". Then once the user says "go backwards" or "go forwards" he will again be prompted for the number of sections of audio he wants to go back or forward to. So I thought I would create a variable (DeltaIndex) which is initially set to "0" and then accesses an array of functions that reads the text, the zeroth one being the first section of audio. Then, once the user says "continue", the variable is incremented and the subroutine that tests for the words mentioned before, goes back to the main program, where the array is again accessed by the program with an offset of "DeltaIndex+1"(assuming the user said "continue"). Then, that same variable could be decremented and incremented for going backwards or forwards.
Before, I did not keep track of where the user was in the text - it just played sound files.