There's some interesting insight into the working of ICANN, although not all that surprising. Liz Williams was on one of the panels evaluating new TLD proposals; her committee recommended rejecting all but two (.CAT and .POST.). Her statement says:
20. ICANN learned of our conclusions in advance of our final report, and was clearly disappointed that we intended to recommend that so few applications should proceed to the next step in the process: registry agreement negotiations. At that time, certain ICANN staff, including Kurt Pritz, an ICANN Vice President, and John Jeffrey, ICANN's General Counsel, were very clear that they wanted as many applications as possible to advance to the next stage of the process, and encouraged us to be more positive in our assessments in order to achieve that goal. I believe that they viewed the new sTLD application process as a means to demonstrate ICANN's prestige and importance as an international organisation; therefore the process had to be seen as successful--- meaning lots of new sTLDS--- so that it would appear that ICANN was in control of a valuable resource, and so that future rounds would attract more applicants. At the time, it was perhaps more important to them that the process result in as many new sTLDs as possible that it was that the criteria be applied correctly to the applications.
I think this gets to the heart of the problems with the new TLDs. They don't have community backing, nor wide acceptance in practice. Their creation is not guided by the interests of Internet users, but by rent-seeking on the part of both registries and ICANN.