Why are different intervals used when restricting the domains of the sine and cosine functions in the process of defining their inverse​ functions?