(This answer may come across as a bit of a rant, and to anyone who interprets it that way, I'll apologize in advance for that's not my intent. My intent is to make a coherent argument supported by current evidence that SE, Inc. ought to revise their software and/or rules so as to allow for the option of some committed members to uncommit from certain beta communities under certain circumstances.)
I think @ThomasOwens has an excellent point with this question. The Network Engineering proposal and beta community is a perfect example going on presently of why there ought to be a way out of a commitment after a site goes into beta and before the commitment is fulfilled.
The NE proposal itself and at least one solicitation for recruits (see "Update 2" which reads in part, "Now, we just need people to commit to being active (writing questions and answers) for the beta site to launch, essentially signing a petition. We can really use people who are already active on other Stack Exchange sites, so please spread the word to friends and colleagues in sister fields.") gave no indication whatsoever that SOHO networking would be way off-topic in a community proposal titled "network engineering."
Indeed, had the site proposal description been written explicitly with such a limitation, I think it likely would have stayed closed as a proposal rather than getting re-opened and moving on to beta as it did. Consider that when I proposed CopyrightX (with the explicit proposal description, "initially: 500 students of Professor William Fisher's CopyrightX course at edX.org, first taught in spring 2013. Ultimately a medium where everyone in the world can ask questions and get answers about existing and ideal copyright laws in all nations."), I did so with the vision of attempting to recruit an initial cohort of experts comprised of 500 of my fellow students in a Harvard Law School MOOC (with many more outside that cohort enthusiastically welcome to join too). But <24 hours after proposing it, Robert Cartaino closed it with the comment, "...we cannot create a closed-group site. Instead proposals should identify an area of expertise to a broader public interest." (my emphasis.) One Area51 user elaborated on Robert's behalf that, "we don't want the wording in the proposal to sound like it may be intentionally excluding people," and that my initial proposal description "sounds like you're excluding people."
So even creating the impression of excluding people is enough to get a proposal closed. Yet on Day 1 of its private beta, Network Engineering essentially became, in Robert's words to me, the verboten "closed-group site" in that it was essentially off-limits to non-L2/L3 professional Network Engineers, if not through any explicit software enforcing its exclusivity, then certainly through peer pressure. For example, one NE user wrote about a wifi troubleshooting question, "Most of the people that are going to be populating this SE are network engineers - either L2 or L3 by trade. While there will most likely be some people who have more experience with wifi from a signals level, troubleshooting at the OSI layer that you're having problems with is much less common. However, even at a very quick level, I don't think anything with the word 'linksys' (which is strongly correlated to consumer-level gear) will get a lot of traction on this SE. Perhaps a wifi signals question dealing with large build-outs will get traction..."
Despite explicit private blog posts to the contrary (see "Update 4" which reads in part, "Everyone is welcome to join and contribute."), the closed-group status of the community is implicitly clear in several respects, with some users who are not quite full-blown L2/L3 professional Network Engineers feeling intimidated within the community about asking questions that are perhaps "too amateurish," and also in the rapidly falling question-per-day ratio which started above 22 and which twelve days later is about half of that despite many members openly admitting to reposting their own questions and answers from other forums (perhaps infringing those other sites' copyrights on their user-submitted content) in an apparent attempt to artificially inflate their Q/D ratio on SE.
Now I'll admit to feeling some resentment about this community that seems to be implicitly a closed-group site hosted by SE, Inc., but that's not my main point in this answer. My main point is that I knew on Day 1 of the NE private beta that SOHO networking questions were overwhelmingly off-topic there, and I had committed to the proposal with the firm impression that such questions would be perfectly on-topic, and with lots of clear indicators to support that impression. So regardless of whether or not NE would have stayed closed as a duplicate, I certainly would never have committed to the proposal had there been an explicit indication or even a hint anywhere (in the Example Questions or proposal description or Discussion Questions) that SOHO networking would be strictly off-limits in a community explicitly described as being, "For network engineers".
Having held one impression about the community based on Example Questions et. al., I committed to it enthusiastically, and on Day 1 of the private beta, I spent more than an hour composing a question that's been subtle and quite interesting to me for several years. Within minutes of posting that question it was very clear that my impression about the community was completely incorrect. Nonetheless I participated as described by SE staff in various questions scattered throughout meta.SO with 10 questions and answers on main (in spite of a rather hostile reception to same on NE) and several questions, answers, and comments on meta. I had one main question closed and another almost closed before I voluntarily deleted it (apparently a desirable thing to do on SE as I earned a "Peer Pressure" badge for having done so? does that count as participation?).
So even on Day 1 of the NE private beta, it was clearly a very bad fit for me (as @ThomasOwens also mentioned about the Ubuntu community and himself). In spite of this, I fulfilled my commitment to NE. However, 5 days into NE's public beta and my commitment token still hasn't cleared so as to allow me to be able to commit to another proposal. And I just read that I may have to wait as much as six months to get it back?! If that's the case then I'd like to uncommit from NE but that is currently not an option, and I think it should be. Fortunately my own community proposal is not near to being commitment-ready, for I fully intend to commit to it, but I never imagined I'd be prevented from committing to my own proposal because of these limited-commitment rules.
Like @ThomasOwens described for the Ubuntu community, I think that in cases where a minority is unable to shape the direction of the community (as was the case for me in NE and for him with Ubuntu), and where the majority repeatedly shuts out attempted contributions from such individuals (as was the case for me with NE), I think there ought to be a threshold level of attempted participation where the shut-out minority members eventually gain the option of uncommitting from the proposal/community without suffering any reputation penalty. This presents a small disincentive to would-be gangs from intimidating minority opinions out of the community while still keeping those same minority members' commitment tokens hostage. If the majority is strong enough then the minority truly doesn't belong there and shouldn't have their commitment token held hostage there. But the majority at least needs to reckon with the loss of a significant portion of their members if they repeatedly shut out minority opinions. I think that such a rule would be a significant improvement upon existing SE community rules, for it encourages a tolerant attitude in nascent communities which creates a welcoming atmosphere for potential new community members which would ultimately increase the probability of every new beta going live.