I don't really see how signally is bullshit from the examples you've given. Harvard College does make a a decision to cultivate a class composed of the best students, that's sort of the entire point. Acceptance to Harvard College ends up being a signal for those things it selects for, not necessarily those things one actually learns. The gateway is admissions, not what Harvard actually imparts.

Since the admissions standards of Harvard Extension school are lower, it doesn't matter that the students can do the same work. It matters that the students weren't selected in the same way. Selection itself matters to the signal that the education is sending.

As for Harvard (etc) not selecting for Jews and Asians, very explicitly throughout the first 60 years or so (Anyone particularly interested in the topic should read Karabel's The Chosen), that's also, exactly as you point out, a signal sent if you went to Harvard in those years. That doesn't make the policy right or wrong, it just means that if you graduate from Harvard in 1930 you're likely a WASP, and, indeed, that item on you resume is signally to all sort of employers that a good Harvard WASP interviewed you and deemed that your character is fit, that saves a step when they want to hire you at their White-Shoe law firm circa 1935.