On Philosophy 15 (briefly discussed here yesterday), Rob and Scott discussed the dialectical move from object language to meta-language (and then from meta-language to meta-meta-language, and so forth). They call this “weaponized metalanguage.” It’s a nice metaphor, despite its violence, because it captures the idea that the metalanguage of argument gets turned into a tool of argument itself. On a somewhat strained analogy, it’s a bit like using the rules of a game as part of the game (using the referee as a blocker in football, maybe).
Scott and Rob are correct in their observation that a sizable part of political debate nowadays is almost entirely second-order–the subject is not the best policy option but rather what constitutes reasonable talk about what the best policy option is. For some people, the election of Donald Trump is a fundamentally second order affair–“I voted for Trump because I’m tired of hearing people tell me what to think….”)
The trouble with this strategy, however, is that there always seems to be a flanking maneuver available; there’s always one-level up. What constitutes reasonable basis for rules about talk about what the best policy option is.
When that fails–as in the example above (here’s an article on point)–there’s always the tu quoque. My informal guess is that the “leveling up” is done mainly to allege the other person has violated some sort of norm. Naturally, accusers can’t be abusers, so the tu quoque is always an exit strategy.