A query to the Kauffman Foundation expressing my confusion about the recently-announced academia/government Open Collaboration Principles produced the following response from project participants:

There is absolutely no intent to disallow the GPL or edict the GPL. In fact, we always assumed the GPL would be the license that would be used where we are working on certain projects (e.g contributions to the Linux kernal). The bottom line is that the open collaboration principles are OSI license neutral and are in no way intended to exclude the GPL.

One more comment about a provision that seems to be causing a little confusion. There is a provision that explains when a member of the public's right may be terminated. The word "may" was chosen to leave open that someone could choose a license with defensive termination. The "may" is not a "must". If collaborators choose to distribute under the GPL, for example, then they "must" follow the GPL and that license can not be terminated except where permitted under the GPL.

As noted in my original post, there is a contradiction between the statement's focus on the importance of potential commercialization and the provisions of the the GPL, but it would appear that an essential pre-condition of producing an agreement was to ignore this.

Of course, any work on the Linux kernel must indeed be put under the GPL, but there are many other possible interactions with Linux or other GPLed programs where the situation is considerably more ambiguous.