Kahneman

Actually, Dan Gilbert's argument was that people are extremely good at reporting their emotional present. In fact, he argued that people couldn't remember or forecast their emotional states because they couldn't suspend their current emotional state the same way that they could suspend their current sensory inputs. They can vividly remember or even forecast a visual or tactile experience because their brain allows them to suspend their current visual and tactile inputs. The brain stubbornly refuses to allow you to suspend your current emotional sensations.

In fact, Gilbert's prescription for predicting your future experiences is to find someone to whom those experiences are currently happening. He argues that the differences between people and their preferences are generally much smaller than the errors introduced by trying to predict your future emotional outcomes.

We deal in a combinatorial world on one hand and a linear world on the other.

Our ability to remember events is more closely related to our ability to track some number of herd events, locally, over space and time. We remember a finite set, throwing away the least informative event, in our combinatorial world. The least informative event is either the farthest in time or weakest in strength.

The number N events we track result in the number N in which we assume binomial holds. It relates to the universal error ratio, or herding constant.
It defines the "rank" at which a herd member has to move between quantum states over time in order to approximate the binomial system in a combinatorial world.

When you move a herd of cattle to the rocky mountains and wait a few hundred years they look like mountain goats. The reason is the Solow growth residual, cows seek better ways to move from grassy knoll to grassy knoll so they can better approximate a gaussian distribution. They become more efficient at switching between quantum states under quantum restrictions.