I think the "what's it used for?" question is a good one. I think of RPG Theory as being the way to understand and design games in the same way that Aeronautical Engineering is the way to understand and design aircraft. This was historically complicated by the fact that early RPGs didn't necessarily supply the full game design to the players -- at the very least the GM was expected to design dungeons or adventures, if not figure out how to make a disparate bunch of subsystems interact usefully at all -- so a lot more people needed to be interested in the stuff of game design than just people who called themselves game designers. Also, a lot of early discussions were framed in terms of "How to be a better GM" or "How to be a better player", so there is some expectation that RPG Theory should speak to that. That's kind of at the core of the "System Matters" vs. "System Doesn't Matter" debate -- if each game is its own thing then naturally each game would work differently and would need to be understood on its own terms, but if each game is but a particular flawed attempt to instantiate True Roleplaying then claims about how "roleplaying" work are potentially relevant to everybody.

In my view, and I'm a bit of a broken record about this lately, a lot of the problem stems from the tendency to use theory as prescription. This is not unique to games, but seems to be a greater problem in media that are still comparatively young, vaguely defined, and potentially still insecure in their legitimacy. Instead of saying, "this is what works for me" or "this is what I like," people say, "this is good, that's bad, this matters, that doesn't matter." Sometimes this is an attempt to pin things down, make them less confusing and nebulous, and sometimes it's an attempt to retroactively validate one's own choices. Either way, it does have the tendency to alienate, exclude, and upset people. 

There's a reason why couples counselors and support group facilitators encourage people to use "I feel statements" during points of conflict. "I feel like I'm not being heard" and opposed to, "You are not listening to me." It seems silly and cliché, but it really does serve to frame things in a more palatable, less confrontational, and ultimately more realistic way (the other person may, in fact, be listening, but that has no bearing on the fact that the first person does, clearly, feel like they're not being heard). 

For me, the physics and aerospace engineering analogy ultimately breaks down when poked, because I see tabletop design as an art and not a science (and I also don't tend to draw a strong distinction between art and craft). A big part of the toxicity, for me personally, seems to result from the conflation of the two very disparate modalities of thought. Though I do understand why people with highly analytical minds, many of whom are in the sciences themselves, might be drawn to that approach, at the end of the day I can't effectively talk about what is essentially a free-floating miasma of dreams and wishes held together by golden string as if it were a set of concrete formulae, compounds, schematics, or equations. 

Aerospace Engineering is a field wherein certain immutable laws are impossible to ignore and must be factored in, because they make up the very foundation of physical reality. In engineering, you are trying to make something that "works." In engineering, there are things that are true and things that are false, things that will definitely work, and things that just plain won't. You can't make working jet fuel out of creamed corn and gravel. You can't make a working bridge out of marmalade. Etc. 

In the practice of ttrpg design, as in any art form from musical composition to baking, there are no immutable laws. It's basically Calvinball, and I think that that's ultimately a good thing. There are certainly principles that may make the creation of a certain kind of  "thing" easier or more reliable for a certain kind of person or purpose, just as there are techniques, for example, to effectively paint trompe l'oeil paintings or harmonious combinations of colors, but there are no Laws Of Design. Even those principles are there on a "take it or leave it" basis. There is no universal or reliable formula for a Good Game. And I strongly suspect that any attempt to find one is doomed to failure. 

Things only "work" and "don't work" in relation to a specific individual or a group. Pathfinder simply doesn't "work" for me: to me, its design makes no sense, feels clunky, and has a tendency to direct play in what I perceive to be boring and problematic directions. I personally find Pathfinder to be a poorly designed mess - just a bad game that does almost everything poorly and isn't (to me) worth playing at all (and god knows I've tried). In my eyes, it just plain doesn't work, like a space-shuttle engine cobbled out of used soda cans and fueled by cheese whiz. 

And yet it's one of the best selling games on the market and, more importantly, one of the most played. That means that to the people happily playing it and having a grand old time, Pathfinder just plain works. Some of them might contend that it's a very well designed game that does most things well and is extremely worth playing, whereas some of them may think that it's simply good enough for their purposes: they know it, are used to it, and they don't have the time, energy, or money to learn a new system. 

And they are all correct, and so am I. No accounting for taste. At the end of the day, there is only one condition that I feel may be necessary for a roleplaying game of any kind to "work," and that's mutual buy-in. Honestly, having just typed that I immediately began to wonder if it IS in fact necessary, and, on comparing it to "willing suspension of disbelief" in other media, I'm starting to think that even that may be subverted or ignored. Some designers may want to make games that purposefully alienate the players, that use mechanics, or lack thereof, to intentionally create frustration of confusion. 

And here's the thing, all of that sort of stuff is actually really fun to talk about and theorize about, assuming we don't end up prescribing anything to anyone else, avoid making broad declarative statements about How It's Done, and start focusing on "I wonder if someone might be able to do that," or, "I personally would love to see more games that try this," and so on. "In my opinion, systems don't matter" is still a declarative statement, despite the disclaimer at the beginning. "Systems don't really matter for/to me, in my personal design practice," on the other hand, is less likely to get people up in arms, because you're not telling them that, in your opinion, they are wrong. You are simply saying "I do it this way because it works for me." There's an implied "your mileage may vary" that makes a ton of difference. 

I tend to avoid theory discussions mainly because folks often lean towards the former, and at the end of the day I prefer making games to talking about them, but I would enjoy "shop talk" a great deal more if we all just learned to use "I feel" and "I prefer" more often. Of course, on twitter, it would take more characters to type out "Systems don't really inform my personal approach to creating and engaging with the medium" than it would to say, "Systems don't matter."

So I guess, in conclusion, twitter is, as always, to blame.

PS: I don't playtest any of my games, and I sell them for money. Nothing is true, everything is permitted.