>>> A range error occurs if the magnitude of (positive or negative) finite x is too large. Why not A range error occurs if the magnitude of finite x is too large. What does (positive or negative) add, especially for an even function like cosh? The finite seems appropriate to exclude infinite x.