Having mulled over everyone's opinions for a few days after starting this, I think I've come to a better understanding of the fundamental problem. I see where both Raven and Kraus & Hussar are coming from, and at this point (if I had a different gaming group) would be highly inclined to test out some of the advice of pro-social skill posters. In the event that removing the social skills proves to be a mistake, I'll add them back and have taken some notes on how to make them work better.
That said, the fundamental problem seems to be two-fold. One is the discrepancy in ease of simulation between the top 3 attributes and the bottom 3. Irl, I'm about 5'6" and weigh a 110 lbs. I've been that height and weight for over a decade now and it suits me just fine. However, roleplaying a character that is larger, stronger, healthier, or more agile than -I- am is almost automatic. Let's pretend that combat isn't part of the game to avoid that arguement prematurely. If my character is strong and physically resilient, it translates very easily into good athletic skills, high carrying capacity, strong resistance to physical harm and disease. He generally volunteers to carry heavy things, probably takes the lead in physically dangerous situations, and is probably on the town's fire-fighting brigade. He plays out as much more physically intrepid than I would be simply because I, and therefore he by extension, know he's capable of handling physical situations. If I want to get more deeply into the character's role, those physical qualities could certainly have an effect on his psyche. Perhaps he's careful about his size, aware of the fact that it makes him very capable of hurting other people around him if he's not careful and doesn't maintain control of his temper, or perhaps he's a bully.
The bottom 3 attributes aren't as easy to simulate. Intelligence is aptly named and from a mechanical stand-point it simulates rather well. Knowledge checks or checks involving mechanical or systematic operations (e.g. Disable Device and Search) suit the ability just fine. A wizard's spell-casting is properly dictated by Intelligence because it represents his systematic approach to magic, that this phrase with this gesture logically leads to this magical result. However, it's difficult to properly roleplay a character who is dumber (at least substantially dumber) than the player is; it's difficult to keep the players native intelligence from intruding on the character. On the other hand, it's nigh impossible to roleplay a character that's substantially smarter than the player. How does the mind of someone with a 16+ Int score work exactly? What about creatures (or characters with stat-boost items) with an Int score in the mid-high 20s? How can your average player (probably smarter than your average joe off the street) play out that level of clarity in reasoning and logic? Wisdom, as defined by the RAW, is easier to simulate because it's mostly corresponds to an all-around awareness of the world about you and doesn't necessarily have anything to do with how "wise" the character is. A cleric's spell-casting is best dictated by Wisdom as it is an awareness of the will of the divine power he serves and from which his own magical ability stems.
Charisma is the stickler, though. It's just as well-defined as the other two, and in a purely mechanistic game it functions well. A lie by a PC will result in a Bluff check opposed by the other party's Sense Motive check; the result of those rolls dictates the outcome. Sorcerer magic, being an intrinsic part of the sorcerer himself, is appropriately tied to Charisma by showing his ability to project himself onto the world around him. Perhaps unfortunately, as this thread shows, few people play a purely mechanistic game because D&D is, at some level, about sitting around a table and talking. That talking can get in the way. While I appreciate where the notion to get a description of what the PC is trying to do and then providing a modifier to the check based on plausibility (pretending to be the king vs pretending to be a visiting dignitary to borrow someone else's example), you don't do that with other skills or checks because you don't have to. What do you do about a player who's neither silver-tongued nor particularly crafty? Do you give his character an Intelligence check to come up with a good lie, and then use that as the basis for your bonus on the Bluff? A Jump check functions under the assumption that the character gives it his all, and a Search check does as well. Does the Bluff skill merely describe a character's ability to make a lie sound believable, or does it also indicate his ability to select and tell a convincing lie? It seems like the designers set up the functions of Charisma to work like the functions of any other ability (and I don't have a problem with that as it only makes sense), assuming that most groups would simply use them as written, just as they do all other skill checks, but the pure roleplay parts of the game get in the way of the rules (or vice versa depending on your PoV), and it just so happens that the social skills are the ones that turn up in those scenes most frequently.
As an aside, one of my players works as a personal protection officer (read: liscensed bodyguard), which means he gets paid to assess threats. He hates the way Intimidate keys off Charisma because perceptible threat in another person is an issue of "totality of circumstance." It's an issue of how big the person is, how fast and agile he appears, and how healthy. How intelligent or shrewd does he look, and does he seem observant. His arguement is that Charisma should be a derived stat, though he admittedly has never come up with a satisfactory way to do this.
The other problem is simply an issue of design assumption. The general view of how to play D&D, and rpgs in general, has come a long way since Chainmail, but the codified rules lag behind. The suggestions about how to award xp for good roleplaying (and it is just a suggestion, not a core-mechanic) are rather ambiguous. It's also called a bonus, leading to the assumption that the bulk of xp will still come from handling encounters. Encounters themselves are rather loosely defined, only becoming specific when dealing with encounters where combat-like rules (actual combat and traps) and resource management are involved. Theoretically, an encounter handled with skillful Diplomacy should result in less xp and probably doesn't result in any treasure (unless skillful thievery takes place as well) because there is less risk, and at the end there's no dead monster to loot. The more focused your group is on the core objective of the game (to gain xp, levels, and treasure), the less incentive there is to employ non-combat solutions to encounters. As such, it seems the abilities that lend themselves to non-combat solutions got the short-shrift design-wise, and most of those abilities are social skills. Everyone here admits that Diplomacy by the RAW is poorly written. The DCs are too low, and what you can do with it is too powerful, especially considering there's no opposed roll. How many spells can you think of in the core books that affect another character/monster and don't either require an attack roll from the caster or allow the target a saving throw? This is a pretty universal facet of design in 3.x, but where is that design elegance in Diplomacy? Most spells are written in a manner so explicit as to border on over-kill so that we will know exactly how they work, and by and large they are carefully considered to make sure they are appropriately powerful for their level. Going from 3.0 to 3.5, I would guess that more spells changed schools than received a serious overhaul because they were over or underpowered. On the other hand, the social skills were not much changed at all. Innuendo got shunted into Bluff, the NPC reaction table was placed in the PHB with Diplomacy so players could make more informed decisions about the skill, and that's about it. We didn't get anything about how to better set up Gather Information tables, like someone very generously displayed above, no revision to Diplomacy's many faults, and not so much as a sidebar about how to handle the gap between social skill check results and player meta-gaming, which I got for the asking here and in abundance. Why not? Because for the core-game, played by core-mechanics, with the core-objective in mind, the social skills are little more than fluff. They exist for rogues and clerics with too many skill points, to give bards (a 5th wheel character class if ever there was one) a unique niche, and for those situations where the PCs hit an encounter they can't win through combat. It is precisely because so many of us don't play the game in quite that manner that these skills become problematic.
