I understand now. I think seeing the code really helps to that end. I thought you were arguing that there was no delay at all.
Any which way they do it (which apparently isn't a global delay), there definitely needs to be a longer minimum length of silence between the lines of monologue. Since they haven't yet implemented the apparently random NPC chatter, they could still make them not talk so much.
Honestly, even if they did have more lines of monologue (which certainly would be less annoying), the frequency of it would still be annoying. I also feel like the frequency lowers the realism, as I don't think people are shouting quite that much in a real market (and I never hear people checking their shopping lists so audibly). Even the guys selling water out of coolers on city sidewalks aren't shouting that much, and they get way more foot traffic going by.