Select and Start was how the Atari 2600 did things. At the time, everybody was designing in terms of having one set of controls for when you're in the game, and a set of meta-controls for adjusting stuff outside the game. The 2600 configuration GUI was the dumbest thing in the world. You look at a grid chart of game options in the manual, and you press the Select button 35 times to get to the version that you want.

The Famicom was much more able to draw and interact with a real configuration GUI. But Nintendo's own experience was mostly in making the arcade game "Donkey Kong", where you pick how many players by "pressing" the insert coin button and then Start. Nintendo was selling to a market that mostly knows home games from picking up a 2600 at a bankruptcy sale. So, keeping the separate meta-game buttons and game buttons was natural at the time. Later games developed a better design language for the meta-game UI, so most game studios left the Select/Start interface behind.

(Lol now I see that TubbyCustard said it all, but better)