Why would a class/function have two overloads, one for lvalue and one for rvalue?

Eg, from this video, it says we have two overloads for vector<T>::push_back

void push_back( const T& value );
void push_back( T&& value );

Why can't we have just one overload by value,

void push_back( T value );

If it was an lvalue, value would be copied and if it was an rvalue, value would be moved. Isn't this the way how it works and guaranteed by the standard?

c++ c++11 move-semantics
Edit
Report