If code shadows an intrinsic but gives it a different interface then fparser currently raises a SyntaxError as the interface does not match that of the Fortran intrinsic and it does not countenance any other options.
It would probably be better if the match on the intrinsic were simply allowed to fail at this stage.