This one is simple. If, during the early stages of the game, I talk to a character on my own free will and get rewarded with something worthwhile, I’m much more likely to talk to other characters looking for more handouts. A steady stream of dialogue rewards enforces this likelihood throughout the entire game.
If you’ve spent a good deal of time on your writing and truly believe your characters are saying things that your player would enjoy hearing (because you’re funny), this is a good way to motivate a generation of jaded gamers into giving your dialogue a chance. Since the reward is instant, players are already happy. And at that moment you have their attention, so if you’re going to sell your dialogue this is the time.
Games that allow customizable characters sometimes ask you to pick a voice for the character you’re creating. If you’ve chosen a male character, for example, the game may offer you to pick whether you want your character to sound like a scruffy male, crazed male, pleasant male, or alien male. The problem is instantly noticeable. It turns out none of these voices fit the character you had in mind. Odds are you’ll settle for the option that has the least noticeable voice acting or the one that is so absurd you’ll laugh whenever your character talks. No matter your decision, you’ve already lost a degree of connection with your character. He’s now just some dude that came off the character creation assembly line, instead of your in-game avatar.
How did the player get into this position? Let’s rewind time and visit the game designer during the making of this game. The designer decided that players should make noises and grunts when they are hit by attacks or attacking. Then they remembered that they let you create a male or female character when you start the game, so they need two sound files for each noise they want to make – the male and female versions. Then they figured since you can make a big beefy lumberjack guy with a beard or a thin and gangly Ichabod Crane kind of fellow, that they should probably add a couple more sound files per noise; one from a lumberjack and one from a school teacher. Assuming girls can be lumberjacks and school teachers as well, we’re already at four sound files for one noise.