I always think of indexes as index of a book.
Why would one ever want to have the index of a book twice ?
Let's say, someone took out the index from a book, it would be hard to find your pages in the book.
But , with a second index you still could find your pages faster.
For databases this could be the same.
For example, a large table, for which the index has to be rebuilt, it would be better to build a second (same) index first and then drop the original index.