I found a minor omission in the latest version of C++ standard. http://eel.is/c++draft/dcl.array, note 6 say:

[ Note: Except where it has been declared for a class (http://eel.is/c++draft/over.sub), the subscript operator [] is interpreted in such a way that E1[E2] is identical to *((E1)+(E2)) (http://eel.is/c++draft/expr.sub). Because of the conversion rules that apply to +, if E1 is an array and E2 an integer, then E1[E2] refers to the E2-th member ofE1. Therefore, despite its asymmetric appearance, subscripting is a commutative operation.  — end note ]

In essence, this says that the following function:

     int foo()
     {
            int a[]={1,2,3,4,5};
            return 2[a];
     }

should compile.

The reference to http://eel.is/c++draft/over.sub makes it clear that what is meant here is overloaded subscript operator within a class, not an operator with a class type as a parameter.

The issue here is that operator [] may also be global, and those would also be an "exception". Perhaps we should rewrite this note along the following lines:

[ Note: Except where it has been declared for a class (http://eel.is/c++draft/over.sub) or in a global or namespace scope, the subscript operator [] is interpreted in such a way that E1[E2] is identical to *((E1)+(E2)) (http://eel.is/c++draft/expr.sub). Because of the conversion rules that apply to +, if E1 is an array and E2 an integer, then E1[E2] refers to the E2-th member ofE1. Therefore, despite its asymmetric appearance, subscripting is a commutative operation.  — end note ]

Another alternative would be:

[ Note: For C arrays, the subscript operator [] is interpreted in such a way that E1[E2] is identical to *((E1)+(E2)) (http://eel.is/c++draft/expr.sub). Because of the conversion rules that apply to +, if E1 is an array and E2 an integer, then E1[E2] refers to the E2-th member ofE1. Therefore, despite its asymmetric appearance, subscripting for C arrays is a commutative operation.  — end note ]