That is indeed strange. And why restrict it to rationals, when it's true for every real (or every complex, whatever)? It's even true for integers, under a suita

That is indeed strange. And why restrict it to rationals, when it's true for every real (or every complex, whatever)? It's even true for integers, under a suitable interpretation of what is meant by dividing by 2 in that case.