This is a way to remove duplicate info in the table - but in this situation a space must be added to the end of each entry AND they must be set in a decreasing length order - PLUS there is no entry for 'me'.
This reduces the program by 5 lines (all of which are data) but adds about 5 lines of programming. The first way is superior in my opinion, because it's simpler and the version below would need a special case to identify when the grammar calls for 'I' to be 'me'. Also, having more lines of data rather than code tends to be superior, since data never needs to be 'debugged'.
DECLARE FUNCTION changeperson$ (tochange$)
IF tochange$ <> "" THEN
i$ = changeperson$(tochange$)
FUNCTION changeperson$ (tochange$)
s$(1) = "myself yourself "
s$(2) = "mine yours "
s$(3) = "my your "
s$(4) = "i you "