I think you've answered your own question (unless I've misunderstood you).
I don't think it's a drafting error. I think it's a deliberate feature of the OGL. It's how it works in a "viral", "genie can't be put back in the bottle" fashion: every licensee in the SRD ecosystem is authorised by licensors (who, unless they are WotC issusing the SRD, are also licensees) to sub-license their OGC to future parties, provided those parties in turn agree to be bound by the terms of the OGL (which includes granting this authority in respect of their OGC).
I think I follow you. But the definition of Use relates to Derivative Material of Open Game Content, not Open Game Content itself. This is where I mostly lose my footing, in the definitions of "Derivative Material" and "Open Game Content". If the OGL's leaning in part on the U.S. Copyright Office's definitions of "derivative works", then the intent might be to allow licensees to only sublicense any Derivative Material (derivative work) that the licensee contributed to the OGC, excluding the OGC material contributed by WotC, i.e. the material in the SRD itself.
The hypothetical that sparked this part of the thread contemplated that an OGL 1.0a licensee could sublicense OGC found in the SRD in perpetuity, provided that the licensee had published SRD-based OGC prior to the "non-authorization" (to use your term) of the OGL 1.0a . That's the scenario I've been asking about, anyway.
Also, I'm not as confident in the integrity of the OGL's drafting. In Section 4, there's a misplaced comma between "Use" and "the Open Game Content." In Section 12, there's a reference to "Open Game Material", a seemingly defined term that isn't actually defined. The definition of "Distribute" includes the word "distribute," which is arguably circular. In Section 9, You are permitted to "distribute" any Open Game Content (not using the defined term), while in Section 10, You are required to include a copy of the OGL in every copy of Open Game Content You
Distribute (now using the defined term).
"Sub-licensee" is not, in itself, a particularly technical term. It refers to someone who (i) is licensed in respect of X, and (ii) was so licensed by a party who was also licensed in respect of X and whose licence included an authority to sub-license. It's similar in structure to the concept of a sub-lease.
I would expect the main issues in sub-licensing, in US law, to be ostensible and actual authority of the sub-licensor. In the present context, I think that would be determined mostly if not exclusively by the terms of the OGL. There may be some relevant general principles too.
Although my credibility has been questioned on this point, I am (and was) aware of that.
The effect of (i), as I'm envisaging it, would be to mean that all existing publishers who want to reuse their stuff would have to reissue it under the new licence. Where publisher A's existing stuff is heavily intertwined with publisher B's OGC, that may be tricky. I don't know if WotC would care about the trickiness, and hence do something different from my (i), or not care. The most different thing they could do is try to set up a one-way door, so that everyone can bring everyone's existing OGC under the new licensing regime, but stuff licensed only under the new licence can't go back the other way. To my mind, that might be a bit tricky given the wording of section 9 of the current OGL, but I imagine WotC will be able to retain better drafters than me!
When Publishers A and B are both licensees who have published OGC under OGL 1.0a, this makes sense to me. But what if Publisher B is WotC and the OGC in question is the SRD? This is where the nuances of Derivative Material and Open Game Content lose me, again, along with the particulars of how the Copyright Office defines the rights of original copyright holders versus those held by the authors of derivative works. The community has long held that OGL 1.0 and 1.0a licensees have the right to continue using the SRD (any version published under any prior version of the OGL) and - in the hypothetical I've been referencing - the ability to sublicense SRD content to which WotC holds the copyright. Maybe that's right, but has it ever been tested in court? WotC now seems like it's in the mood to do just that.
Second - and this is moving further from my fields of expertise, but is still something on which I have a modest intuition - I would expect that there is a plausible argument, in US contract and licensing law, that if WotC were to exercise a power of at-will revocation so as to deprive (say) Paizo of the benefits of any licence, then WotC would simultaneously lose its rights to an in-perpetuity royatly-free licence to Paizo's work. Or to put it another way, I doubt that the only tenable construction of the licence terms would be that one party is able to revoke at will while retaining the full benefit, in perpetuity, of the contractual promise made by the other party. (I am not saying that this is knock-down in any sense, only that I would expect there to be a plausible argument here. Of course everything will turn on the details of the drafting of the OGL v 1.1.)
This also matches my intuitions and presents a potential pothole on the road to 6e. The playtest documents released so far for 6e have similarities to Pathfinder's designs and rules. WotC's hired a bunch of Paizo designers to work on 6e, so that's not entirely surprising. Under this theory, Pathfinder could arguably sue WotC for infringement if 6e's final form ends up incorporating elements of Paizo's design in a way that's not sufficiently transformative. I don't think there's a high likelihood of this. But the idea illustrates how the lawyers and executives making decisions about the OGL might not understand what's going on in the actual design of 6e itself.
Maybe they do. After all, they could just say that they're going back to 3e design principles and expressions. You'd still have to do the textual analysis to see if a copyright infringement occurred, but the argument is probably strong enough to take into court.
