A person's self-identity is defined by who they think they are, how other people say who they are, and how they can be generally understood to be human. It would be nice if the previous two were linearly correlated, because that would make it easier for people to work together in the same direction. If they are not linear, the other illustrations can be a feedback loop, depending on who the other person is, or a hall of mirrors, which is about self-perception. The third can be the most useful, if one can be addressed as a human being, then things would be common or not uncommon.
Linear regressions are the most common type of correlation, because there is no up and down, or the suggestion is that the up and down can be ignored. This is the crucial step in building a network when any two of the items are seen as either coexisting (synchronicity) or linearly correlated.