Okay, fair enough. But I'm sure I can find people who interact with others frequently and still don't see them. In other words, it's possible that the groupings would not be recreated along the same lines by someone who had never been exposed to Keirseyian theory, but that someone who has might start seeing people fall into those groupings. They might create more than 4, or fewer than 4. I need a way to test that idea, though, and I'll let you know as soon as I come up with something.
I'm determined to either prove you right, or prove you wrong now, as much as you dismiss my ideas on the basis of "not interacting with people enough." If arguments won't convince either of us, I guess we'll need measurable data and repeatable experiments.
If the MBTI is a system of nomenclature, organisation and communication then you actually have to apply it to what you observe for it to exist. The fact that people are different is what exists, what you call the different classifications, how many you have and what their parameters are is more a matter of what system you choose to apply than anything.