Why is nominal interest defined the way it is?

The nominal interest rate is defined the way it is because, along with the compounding interval, it is a succinct way of describing how interest is computed.

If, for example, the nominal interest rate is $6$ percent, and it is compounded monthly, then we can simply divide the nominal rate by the number of months to obtain $0.5$ percent, and now we know that each month, the principal goes up by a factor of $1.005$.

The actual effective interest rate is about $6.1678$ percent, since $1.005^{12} \approx 1.0061678$, but it would be a rather ungainly way of expressing the same thing. What's more, it's likely to be only approximately correct, unless you want to carry this out to $28$ places or whatever.

To be sure, of course, we could have started with the effective interest rate, and then worked out what the nominal interest rate must be. But this requires us to compute a twelfth root, and people back in the day of hand calculators (and before that, hand computation) were understandably loath to do that. And just imagine what would happen if you were to go to daily compounding. (In many ways, compounding continuously is easier, though it requires taking a logarithm.) It was simply easier to deal with the nominal interest rate.

Also, from a marketing perspective, it was easier to tell people that their effective rate was higher than their nominal rate (sounds like they're getting a compounding bonus) than that the rate they actually got each month was less than the effective rate divided by $12$ (sounds like compounding costs them money).

It's simply a convention. It doesn't allow for precise comparison of nominal rates for different compounding frequencies, but the nominal rate is roughly the same order of magnitude as the actual annual rate. This provides some justification, but beyond that it's arbitrary.