I don't know why people keep spreading the misconception that mutation isn't possible in pure languages.

    ref <- newIORef 0
    writeIORef ref 1
    y <- readIORef ref
    print y

> 1

There you go, mutation in Haskell.