Thread by @JonathanShedler

1/ Getting fed up with scientific "experts" giving info about coronavirus risks that is, scientifically, unknowable. The most "scientific" information about spread of virus & risk is: we do not know. Some reasons:
a) we cannot know mortality rates because this would require

2/ a representative sample of people with the virus. We have not tested representative samples & do not know the characteristics of the samples we have tested. Skewed toward people who are already symptomatic? Self-selection? Availability of tests? When we calculate mortality

3/ rate, there is no credible denominator. We are therefore working on guesses.
b) We cannot know what systemic failures may result (e.g., food supply, infrastructure, financial systems) because there is nothing to extrapolate from. Society more interconnected than ever before.

4/ It is a a 'complex system.' No one can anticipate 2nd-, 3rd-, & Nth-order effects of variables in a complex system. Maybe nothing happens. Maybe we have catastrophic infrastructure collapses. No one knows.

This is the ultimate human illusion... that somewhere there is

5/ an expert who can knows things that are inherently unknowable.

We humans seem unable to tolerate the cold fact of uncertainty. We instead construct narratives that appear to make sense of the unknowable. We adjust the narratives retrospectively, to account for every new,

6/ previously-unforeseeable fact that emerges. Our cognitive processing system is designed to see patterns & causes, including patterns & causes where none exist. Experimental psychology has shown time and again.

7/ The relevant question is: What decisions would you make if you truly believed no one knows?
#AcademicTwitter #AcademicChatter #coronavirus
with gratitude to @nntaleb