Good example of different rod weights from the same company, I have a Scott G series 3 weight & 4 weight. The 3 is 8'8", the 4 is 6' 10". It's nearly half the diameter of the 3 weight also, and bends/loads waaaaaay more even though it's much shorter. It throws a 3 weight line with a slow action. It'd throw a 2 weight in what I would call "normal" fashion. It will also throw a 4 weight in a little slower fashion. Why they chose to call it a 4 instead of a 3, I can only assume is based on math (static deflection) & not observation.
I have no idea, of course, how they (or any other company) does it, but my hunch, based on conversations, experience, etc. is that "observation" is probably actually closer to the truth than "based on math". In other words, since there is no standard metric for determining how to "weight" a rod, I really suspect the companies do it in a fairly subjective way.
Although I certainly manage with the current system of line and rod weights, and action descriptions, I personally wish there was in fact a consistent and repeatable system to describe rod behavior. I'm not holding out for it, but to me, the ideal system would be one such that if a company said something like "This rod is our X,Y,Z", where some information replacing each of those letters would tell me how it would behave with lines of a certain grain value, then I would know that if another company said they had an "X,Y,Z" rod, it would behave in exactly (or almost exactly) the same way. And if there was something about the manufacturing process for each of these companies that meant each's rods were not identical in behavior...no problem: just give them different labels (e.g. One company would have X,Y,Z, and the other would have W,Y,Z).
ALthough my understanding is that a lot of people really don't like the Common Cents system, especially when applied to the big rods designed for large files, long casts, etc., the one thing I do like about that system is that they have continuous metrics for different attributes of a rod.