When Humanity Collided with the Problem of Induction
Originally published in the Wet Mountain Tribune on December 24, 2020
In the year 2020, humanity collided with the Problem of Induction. The question is, did we collectively learn anything from the pain caused by our crash? To answer that question, we have to first understand what we ran into, and then decide if the lessons learned can be applied to future problems.
The Problem of Induction can best be understood as the limits of evidence. Longtime readers of this column know that I often refer to the analogy of the Thanksgiving Turkey as described by the probabilist Nassim Nicholas Taleb. The parable goes as follows with a few slight modifications on my part:
For 1000 days a Turkey is fed every day by a friendly butcher. The Turkey has a staff of statisticians who use evidence-based models, charts, and computer models to show the Turkey that butchers love Turkeys. Each day that passes, the statistical confidence of the Turkey grows. Then, a week before Thanksgiving the butcher comes and cuts the Turkey’s head off.
There is a limit to what evidence can show us, but we often forget that fact in the moment. 2020 was a crash course in the limits of evidence. Just like the parable above, statistical experts failed to realize that looking for evidence with COVID-19 was not possible since the disease was new to humans. The timeline of events in 2020 speaks for itself and highlights the struggle leaders found during the spread of the COVID-19 virus.
In January the World Health Organization stated there was no evidence of person-to-person spread of the virus. A few weeks later, the World Health Organization again came out and said there was no evidence that stopping international flights would stop the spread of the virus. By early February the World Health Organization was saying there was no evidence that the virus was airborne. A few weeks later the Centers for Disease Control stated that there was no evidence that masks helped slow the spread of the virus. In late February the Colorado Department of Public Health sent the Tribune a press release stating that there was no risk to Colorado citizens from COVID-19. The Tribune refused to print any such press releases which earned the scorn of the then Custer County Public Health Director, with her sending a nasty email saying “I wish you would leave your opinion out of such things and that you would just print what the experts send you. You know, be a real journalist.” So far, 6,000 Coloradans have died since those original emails back in February stating that there was “no risk from to the people of Colorado.” This author’s grandfather is among those killed by COVID.
The common problem with all the above examples is that just because there is no evidence, does not mean something is not taking place. Probabilists often use the short phrase “absence of evidence is not evidence of absence.” Even now, the same problem continues to dog experts and common citizens alike. A new strain of the COVID-19 virus has been discovered in the United Kingdom and this version is much more contagious than the original COVID-19. The United States has decided not to block international flights from the United Kingdom, citing that “there is no evidence that it is any more deadly.” Planes land at Denver International Airport every single day from Europe, many transitioning through the United Kingdom. Just like the beginning of the outbreak, we are hoping that the new strain is not a problem. The evidence will come far too late for us to do much to stop the spread of a new virus.
At this point, we now have evidence that COVID-19 has a mortality of 1.8 percent with proper treatment. There is hope that the new vaccine will bring the pandemic to a close and we can return to a life where other humans are not kept at arm’s length. But again, this is all based on hope, not on any type of fact or evidence.
Locally, the problem with evidence continues to dog the conversation. People point to the fact that “there are no COVID deaths in Custer County,” implying that since there are no deaths, that the response to COVID was not needed. The question that should be asked is “what would the deaths be like if we had done nothing?” This question is never asked and is a struggle for most to grasp.
2020 is coming to an end, and many are hoping that 2021 will be pandemic-free. But do we know that? No. Are we sure the worst of the virus is over? No. And can we honestly say that doing nothing would have been better? No.
There is a lesson to be learned from this year, and that lesson is learning and exploring the limits of what evidence can show us in times of emergency. We take extreme precautions when a forest fire is raging because we have no real idea where the fire will burn or how fast it can grow. The same precautions we take towards forest fires are easily applied to pandemics. Precaution first, evidence second.
But I do not have a lot of faith this will happen. Our natural instinct to the pandemic seems to be evidence first, precaution second. That is not the lesson we need to walk away with as the year closes.
— Jordan Hedberg