Susan McClary is one of the most prominent musicologists alive today, at the forefront of the turn of the discipline towards the postmodern. She has published on a variety of musicological topics, but one of her books strikes me as particularly D&D-relevant. Her book, Conventional Wisdom (from which I stole my title), chooses to study the conventions of musical practice, instead of the typical mode of inquiry which focuses upon the strange and inexplicable. Her argument for doing so is that the foundation of expectations from which scholarly inquiry proceeds requires discussion and analysis - if we are to judge something as "different", we need to have a very good and clear idea of what constitutes "normal".
This brings me to Joseph Manola's post from earlier today. Manola argues that RPG material with a price tag needs to be more creative material than a DM might produce on their own. He uses the term "conceptual density", by which he means how many "good" (i.e. useful and nonobvious) ideas exist per page. I can't imagine someone sanely opposing Manola's point, and that's not what I want to do today.
Instead, I want to talk about his criteria for "good" ideas - their utility and nonobviousness. Both depend quite strongly upon ideas of normality - i.e. convention. "Good" ideas challenge the expectations held by DM, who judges the information, and then player, to whom it is presented in play.
But if every event is unconventional, is the power of these ideas blunted? Or is it cyclical, that after a while a dwarf who enjoys mining is itself strange and exotic? Can we meaningfully differentiate between "good" ideas and novel ones?
Just as DMs induce the play they are willing to facilitate, so too do worlds (as presented by the DM) construct a horizon of expectations for the player. In a "vanilla" setting, one presumes that these expectations are pre-built. In a snowflake setting like Manola's or mine, part of the game experience lies in constructing that horizon of expectations, against which ideas are judged "odd" or "normal".
We find an interesting liminal space where the snowflake setting butts against the "vanilla" one - most commonly when a snowflake world uses "vanilla" concepts (like the word troll or dwarf) or a "vanilla" setting has thematically separate areas (the Frozen North functions very differently from Waterdeep). In these borderlands, expectations are all over the map, attempting to map the evidence presented into one or the other horizons of expectations.
What's important, ultimately, is that one of a DM's conscious responsibilities lies in constructing their players' horizons of expectations, whatever they are. Without a horizon of expectations, the odd seems banal and the commonplace, mysterious.
I'll finish this up with a short story about a game I ran a while ago. As part of Alexis' world-generation exercise, detailed in How to Run, I had prepared three different worlds and run my players through a one-shot in each. One of them was a version of my Prodigy setting, fast-forwarded several hundred years into a cyberpunk dystopia. I made many mistakes in that running, and the biggest of these was failing to construct a horizon of expectations. To get the game started with a bang, the session opened with my players at a rally run by one of the minor political parties. After a particularly anti-government speech, the government security drones (who had been monitoring the protest) attempted to kill everyone present. I had wanted the deaths of so many civilians after little perceived provocation to shock my players, to galvanize them against the government forces, but they just looked at me and asked, "is this normal?"
One of my strategies for first sessions is to jump into some kind of action as quickly as possible - to put my players into a situation where they have to start making choices, quickly (beginning the session usually immediately follows a great deal of exposition on my part to get everyone through the character creation process, so I want them to talk instead of me for a while). It usually works quite well, getting my players used to the idea of being in the driver's seat in the world. But trying to have this intrusive event shock my players, right off the bat, is an ineffective strategy - they have no conventions against which to measure the events transpiring.
In short, boring, conventional ideas do serve a purpose. They are crucially important in creating a world and running a game. Once everyone knows what's expected, you can start throwing the curveballs.