For a language, x, to be Fortran complete, it must be possible to transform any code written in language x into code that's as ugly and unreadable as the average Fortran program.
It has been proven that:
given some language x and programmer y, there exists at least one y such that the Fortran completeness function F(x,y) evaluates to true.
Therefore, F(x,y) is always true for sufficiently bad values of y.