So yes, magic is very important in DnD. No, it's not everything to the game, just as it's not every page in the rule books. But it's a very big part of the game, just as it's a very big part of the rule books. Likewise combat. And social activity... isn't as much.
I think that there is a huge hole in this thesis and it has to do with the difference between modelling the physical skills of a player character versus modeling the mental skills of a player character.
If you have a physical skill like "jump" or "run", then a the players physical skill is completely unrelated to that. No matter what the physical capabilities are of a player, in a table top RPG they don't extend into the imagined universe (they could if we were running a combat LARP of some sort). The players ability to run can't influence how the character runs, jumps, or swings a sword.
The same is not true of mental abilities, including social abilities. No matter what the mental capabilities are of the player character, the player's mental abilities always extend into the game universe. A player that lacks good judgment will have a very hard time playing a character that has good judgment. A player that isn't a humerus wit, will have a very hard time playing a character that is one. A player with poor memory and reasoning skills will have a very hard time playing a character with great intelligence. We can sort of help the character along, but it is absolutely important to realize that not only can we not fully bridge this gap through some sort of simulation of mental and social abilities, but we would not want to. If in fact we could fully simulate the mental abilities of the player character, then the game would cease to be a game and become a simulation. The player character would make it's own choices based on dice rolls, and the player would cease to be a player and become an observer of the character only.
So it is absolutely important to understand that there is a difference between the goals we have for the game with respect to physical abilities of the character and mental abilities of the character. For example, for the game Pendragon there is no 'Intelligence' score for a a character at all. It should not be inferred from this that problem solving or making intelligent choices is deprecated as core goal of game play.
Combat abilities and athletic abilities are not easy to model in an imagined game world, but they are easier and less information dense than mental abilities and social abilities. Most RPG systems have some degree of what I call 'cinematic experience' as a core goal of game play. By that I mean that they want the experience of play to tend to produce a transcript in the player's mind which resembles a movie. That is to say, when the players have a fight with a monster, they want the player to imagine the sequence of exciting events that transpired as the players fought the monster. When trying to achieve a cinematic experience in combat, it aids the imagination of the players if various concrete events happen during the fight - someone throws a spear, someone is bitten, somone is hurled to the ground, a healer heals someone, a caster hurls some magic, or whatever. Different systems go into different levels of granularity to help achieve this concretely imaginable result that is the exciting transcript of combat. The idea is to as closely as possible realize the imagined fight without going into so much bookkeeping it distracts from the excitement of the event by becoming the most salient play activity.
What about social abilities? What thing can we do at the table that most resembles social interaction and is most cinematic in the same way combat rules achieve cinematic results? The answer turns out to be social interaction. The thing that is most like a scene in a movie where characters talk, for whatever reason that they talk, is simply acting out the scene by talking. It turns out that quite the opposite of what happens with combat, the more rules simulation that you add to a social conflict, the less it actually produces a transcript of dialogue and the less it resembles the thing you are trying to simulate. Social interaction is in fact the most cinematic way to simulate social interaction.
So what this means is that there turns out to be almost no direct relationship between how many rules you have for adjudicating social interaction and how much the game intends for social RP to be a central pillar of play. And in fact, in my experience the game designers that have gone all in with the intention of making social interaction the central pillar of play by sort of intuitively taking the sort of rules we have for combat systems and creating social combat systems complete with maneuvers, reputation points, and simulated social combat actually end up creating an incoherent system where social interaction is minimized in importance. After all, if social interaction is simulated as combat, then the transcript of play resembles combat by other means and not actual role play. All that is really need for a game that involves a lot of social interaction is thinking about the game being about social interaction and making that rewarding. As far as rules go, I've never seen anything actually be more effective than just a simple test to see if a proposition challenging an NPC's beliefs (or sometimes a PC's beliefs) worked and maybe to what degree. Anything more than that actually detracts from social focused play.
