Syncaine has an extremely well-written post up arguing against accessibility. I don't agree, but it is well worth reading to understand the point of view of the hardcore.
His main argument is that in a game that is too hard, the not good enough players can always work to improve, while in a game that is too easy, the too good players can't do anything to get worse. Let's have a look at that claim!
What is the difference between a casual player and a hardcore player? As one can see from the frequent use of words like "idiots" and "morons", the hardcore believe that intelligence is a factor here. That is almost certainly completely wrong. There is no single agreed upon definition of intelligence, but most scientists agree that intelligence is about figuring things out, being able to adapt to complicated circumstances. A raid encounter in which the boss would use random abilities, and raiders would have to think on the spot about how to counter these abilities would require intelligence. But we all know raids don't work like that. Scripted raid encounters in which your success depends on perfect execution of predetermined moves you can watch on YouTube does not require intelligence. It requires a lot of determination, practice, and motor skills, but little or no intelligent thought.
Thus the real difference between hardcore and casual players is how much effort the two different groups are willing to exert. Gevlon's "slacker" part of the description of casuals, while not nice, is a lot closer to the truth than the "moron" part. For example I did raid BWL up to the last boss during vanilla with a hardcore guild, proving I do have the skill or intelligence to do so if I chose. And then I decided a hardcore raiding schedule was more effort than the fun I got out of this activity was worth, so I continued playing as a far more casual player. The large majority of hardcore players could become casual in a instance if they chose to do so, and the large majority of casual players could become hardcore in an instance if they chose to do so. The number of people who are actually "too stupid to raid" is insignificantly small.
Syncaine is completely right in saying that if somebody applies the typical hardcore effort on a task not requiring that sort of effort, the task becomes trivial. But in games that is called "optimizing the fun out of it". From theorycrafting sites, YouTube strategy videos, boss kill guides, to a plethora of addons, hardcore players use an incredible amount of tools to make raiding easier. That only proves that any game can be beat if you apply enough effort. Pac-man was beat 19 years after release in the first "perfect play" yielding the maximum possible high-score in 6 hours. That tells you a lot about the amount of determination some players have, and very little about whether Pac-Man is "too easy".
Making a game too hard is not about making it "impossible" for the casual players. It is about requiring an effort these casual players are not willing to bring to the table. The G in MMORPG stands for Game, and there are a lot of people who believe that MMORPGs are "just games". If there was a game which would require you to simply press a button 1 million times to get a super reward, some people would do that, while most people would probably say that this isn't enough fun for the reward to be worth the effort.
In real life the possible reward for any given effort is often limited by laws of physics and nature. For example how much effort you have to exert to lose weight and become muscular by working out in a gym is not really optional. That limits the business opportunities of gyms. In virtual life no such limitations exist. A developer can require a grueling 6-hour perfect play for you to get any epic reward, or he can give it to you for pressing a single button, or he can sell it to you. It is true that the 6-hour perfect play epic would be "worth more" to the person having achieved it than the button-press epic everybody got. But that is an optimization problem, with a optimum solution somewhere in the middle. The 6-hour perfect play epic does not motivate a player to play and pay a monthly subscription if that player knows that he will never want to play for 6 hours straight, or do all the training and research required. The single-button press epic doesn't motivate anybody either. So the task for the developer is to design challenges which require enough effort to be considered worthwhile, but not so much as to be considered only accessible by the no-lifers.
MMORPGs are entertainment products. Game companies are in the business of selling these entertainment products. If developers design a game which requires too much effort from the average player for too little gain, the average players will start leaving the game. While Trion never published subscriber numbers, I am pretty certain that this is what happened to them. Rift wasn't nerfed because of some change in philosophy, or to mimic a move Blizzard made 2 years ago, or to do the opposite of what Blizzard did in Cataclyms. Rift was nerfed because it lost too many players who were complaining that the endgame wasn't accessible enough for them. Cataclyms will be nerfed end of this month for exactly the same reason. Accessibility isn't a philosophical concept, it is a business survival strategy. That some people would rather want to see their favorite game fail than to be made accessible speaks for itself.
Of course better solutions than nerfing exist. Single-player games have solved the exact same problem years ago by introducing variable difficulty. Most single-player games wouldn't be profitable if they only existed on their hardest difficulty setting. One day some MMORPG developer will get the risk-reward ratio of a variable difficulty system right and make a killing. Until then going for the largest number of players is more intelligent than catering to the hardcore.