The problem with the concept of RoleModel is that it sets up certain people as worthy of being followed, and distracts attention from the behaviors or patterns themselves that these people exemplify. It is, in short, more directed at identity rather than content, outcome, or effectiveness.
Each of us can have a RoleModel nature to the extent that we model the behaviour that we desire to see in others. It also acts as a check on our own behavior, by giving us a tool to be critical of our own behavior as it might appear in someone else's eyes.
Nobody is perfect and deserves (for good or ill) to be put on a pedestal. Being RoleModels falls to all of us.
Modelling desired behavior crosses the line of the personal and the communal. Unlike the PygmalionEffect, it does not cast expectation on someone else's behavior, and therefore set the stage for a row if (when?) those expectations are not met.
If someone doesn't behave as one wishes or expects them to, is that the fault of that person, or of your own expectations?
Modeling desired behavior is a tool for integrating behavior and expectation--if one doesn't behave as one believes others should, one has only oneself to blame. On the next round, one can then change either one's expectations (theory, if you will) or one's behavior, whichever is appropriate.
Acting as a RoleModel, the founder has to do these things... yet they also seem like power plays, pushing around other members of the community. Instead of seeing it as "Founder X is doing that, I should too!", members think, "It's Founder X's prerogative to do that; I should back off." I see this as a fundamental paradox of the founder role.
This problem is not restricted to founders: On EnglishWikipedia it takes a notable effort to break the feeling that certain "housekeeping" jobs are sysop-only. One must try to explicitly and frequently ask for help.
Asking for someone else to help you do something can be good (sometimes) - Wiki:PairProgramming, BarnRaising, etc. Asking someone else to do something instead of doing it yourself is definately bad. But there is a circle to be squared here.
First, the Founder has to be a leader. He or she has to be articulate and charismatic. A Founder that is avoidant, is overbearing, is malfeasant, is distracted, is incompetent, is narcisstic will not succeed. (Don't worry; we will help you.)
The Founder has to pulse. Every so often, do something dramatic to remind people what they should be doing, and then back off to let others do it. It's tempting to keep going, especially when the Founder has nothing better to do and feels like doing some cleaning, but it's a geometrical relationship: Founder involvement x Community involvement == constant.
Sometimes there is a crisis and the Founder will do some kung fu out of perceived necessity, but that is actually detrimental as it weakens the community. They perceive the community more as the Founder's domain and they lose the opportunity to bond through fighting together against an common emergency.
Of course, it's painful to watch your community tank against a threat, and you have to be ready to step in to end the madness if it looks like your community is drowning. As a corollary, the Founder should not be the first one into the fight as that will invalidate the Founder as an objective authority to appeal to, leaving no one to appeal to except the community--and that fails if the CommunityDoesNotAgree, throwing you into a DeathSpiral?.
You should not lock the FrontPage against vandalism; it helps train people to fix vandalism, even if it is tiresome. Plus it signals clearly (GuidePost) that in fact there is some GodKing who will protect the site when all else fails.
Finally, the Founder should RewardReputation for good behaviour, not just model it or use the PygmalionEffect. For instance, you can give BarnStars, which are for anyone to use on any community, by the way.