Well, my exposition of skills wasn't great, but it was okay, but my thesis topics went swimmingly. All three were met with great enthusiasm, and my professor gave me some extraordinarily uplifting feedback! For the first time in months I feel like I genuinely belong here.
Evidently the Unreal 3 editor is a lot more malleable than I initially gave it credit for, and even without a whole ton of scripting it's quite possible to develop some branching dialogue or HUD-based option selection--that being the major feature that would be forcing me to work along the lines of NWN2 or the Elder Scrolls tool set were I to go with... well, practically any of these.
Having viewed sample projects from Level Design Mechanics over the last couple of days I'm actually really impressed with the kinds of things that everybody's been able to do just getting their feet wet with Unreal. I'll be looking into teaching myself the tools as I get the chance--though for right now I've got a lot to do as it is.
Most likely I'll be going with Topic 1: branching narrative, as it has the most doable and most provable project. The idea would be to develop one level following a particular narrative, then do several iterations of it, each one showcasing the particular decision-making paradigms employed by other developers, plus an additional one showcasing my own paradigm for comparison's sake; then I'd have a sample of gamers take the Pepsi challenge to discern which one is the most satisfying. It'd take a ton of effort to build and balance and even more effort to get testers, but it's doable enough that I could have this thing put together before I graduate, if not then not too long after I graduate.