"This is not a democracy, it's a meritocracy."
The statement comes from the Ubuntu Governance page, but you can find similar statements in the Fedora Release Notes, the Why Debian For Developers page, and just about anywhere else were free and open source software (FOSS) projects discuss their core values.
Clearly, meritocracy is one of the myths that the FOSS community tells itself. By "myth," I do not mean that the claim is a lie, but that meritocracy is part of the story that community members tell themselves to establish and maintain a common identity.
In other words, the idea that FOSS is a meritocracy is equivalent to the idea that America is the land of opportunity, or that scientists are objective. For members of the FOSS community, the idea that hard work is rewarded with recognition and the opportunity for more responsibility is central to their belief system and their sense of who they are.
For people to continue to believe such myths, they must have enough truth in them that no one calls them into question, and exceptions to the myth can be explained, or even denied.
Yet if a community's myths are not a lie, they are not necessarily the whole truth, either. Often, they are over-simplifications of a much more complex situation.
That, I believe, is the situation with FOSS and meritocracy today. If you have the right background, and contribute in the right area in the right way, FOSS actually can act as a meritocracy, and frequently does.
However, the idea that the community runs on merit alone does not completely describe the vastly more complex situation in which merit is only part of the reason for recognition and responsibility. Instead, many other considerations come into play, often unrecognized.
One of the problems with any claim of meritocracy is that it can become a circular argument justifying how power is already distributed. How do you know that the people in power deserve their position? Because they are in a meritocracy, and they are in power. If their actions had not merited their positions, they would not hold those positions.
Yet you do not have to search too far to see that merit is not the only reason for holding power in the FOSS community. In particular, project founders tend to hold influence regardless of the value of their recent contributions -- or whether they continue to contribute.
For instance, when Ian Murdock founded Progeny Linux Systems (a company I used to work for) in 2000, he had not been active in the Debian Project for several years. Yet when the company started to become active in Debian affairs, his status remained undiminished within the project. Despite the increasingly evidence that he would not actively involved in the project personally, he was even offered the chance to skip the usual process for becoming a Debian Maintainer, although he turned it down.
More recently, Mark Shuttleworth became benevolent dictator-for-life of Ubuntu and Canonical -- not because of his contributions to free software, but because he had the energy and money to create the position for himself. Few people in Ubuntu or Canonical begrudge him his position, but the fact remains that it was not obtained through merit (as the community defines it) so much as the exercise of existing power.
Nor are leaders the only ones who gain influence through means other than pure merit. In projects in which some contributors are paid and some are volunteers, paid contributors almost always have more influence than any volunteers. In some cases, like OpenOffice.org, the paid contributors can almost shut out the volunteers entirely.
In others, like the Fedora project, power is more equably distributed, but the paid contributors often step into positions of responsibility. For instance, seven out of ten of the current Fedora board members are Red Hat employees. In much the same way, three of the five members of the openSUSE board are employees of Novell, the project's main sponsor, and another is a consultant specializing in Novell products. The situation is much the same in most projects.
Of course, you could argue that paid staff has more time to devote to responsibilities. Yet, while true, that is irrelevant. The point is that paid staff tend to hold responsible positions in projects more often than volunteers. What matters is that the assumption that a level playing field exists in which effort alone determines status within a project does not hold up to even the most casual scrutiny.