Well, observation. If a person routinely acquires status even though it DOESN'T further any of their goals, they presumably value status as a terminal value. If someone only acquires status WHEN it's instrumentally valuable, then it's presumably not a terminal goal.

This seems slightly circular to me. For both of the observations to work, you'd already have to know some of the person's terminal values. I believe this doesn't get any easier even if you're observing yourself.

For the other points, I appreciate your elaboration, but I was more like hoping to invoke the consideration that social status is a very fuzzy concept, and probably too high level to be optimized by evolution anyway. Once you invent the concept, then I guess you can shortcut some social computation... and get some cognitive biases as a side effect.