In the light of Corona – Part 11: Decision-making in uncertainty

Uncertainty as a term used in coffee tables encompasses all types of haziness regarding tomorrow. In academia the concept is however more nuanced, and infinitely debated. The most impactful categorization of the different types of future haziness was introduced already in 1921 by Knight who distinguished between risk, uncertainty, and what has since been known as Knightian uncertainty. In addition, there are concepts such as equivocality, ambiguity, or isotrophy that capture certain types of the haziness, or combinations of several types of hazinesses, like VUCA (volatility, uncertainty, complexity, ambiguity) that attempt to describe a setting where all is just a blur. 

So what? In a disruption, as decision-making anyway feels like stumbling along in darkness, what does it matter what types of darknesses surround us? In this post I’ll explain both the diverse types of darknesses, and why understanding their differences might matter.

Urns of haziness

Risk. You have an opaque urn that you know contains eight green balls and two red balls. Risk refers to the probability of drawing a red ball from the urn. Risk is the simplest problem, as you know the nature of the components and outcomes, and can calculate the odds with probability mathematics.

Uncertainty. You know that the opaque urn contains both green and red balls, but you don’t know their distribution nor their overall number. The good thing is that you know the contents (the shape of the problem), but the problem is that there is really no way of allocating probabilities to drawing either a green or a red ball. 

Knightian uncertainty. You think there may be an urn, and you stretch out your hand, but it is really not sure whether your fingers touch a ball (of any color), a scorpion, an ice-cream or a bomb. In normal uncertainty we’re at least dealing with known unknowns, whereas in Knightian uncertainty we’re dealing with unknown unknowns. 

Equivocality. You open your hand and find that it holds a red ball. Yet you have no idea how it got there.

Savagean ambiguity. There is someone somewhere who might know the contents of the urn, but you don’t know who and where, nor can you access that knowledge.

Marchian ambiguity. You have no idea whether drawing a green or a red ball would be preferable, and there is no way of finding out prior to picking one up. And even then the desirability is questionable. 

Isotrophy. You are bombarded with all types of information about the contents and desirabilities of the urn content but as there is no way of sorting the relevant information from the irrelevant or misleading information you might as well act without any knowledge. 

Volatility. The distribution of green and red balls in the urn keeps changing at such a rapid pace that even if you could in theory calculate the probabilities, the situation never settles enough for you to do that.

Complexity. You have billions of urns, each containing billions of balls, and each ball you draw either breeds new urns or destroys others so that while in theory you could calculate the probabilities, the volumes of urns, their contents and options make it difficult. 

Dealing with different urnfuls of problems

In our era of digitalization, data and algorithms can be used to mitigate some of the issues. Calculating risks, even when moderated with volatility and complexity, is the bread and butter of computational intelligence. When combined with collaboration and connectivity, they may also assist in tackling Savagean ambiguity: if the answers exist somewhere, they can be searched with mighty resources. 

Both uncertainties and isotrophy are in practice indistinguishable: not being able to tell whether one titbit of information is valid or not is equal to not having it in the first place. The actions to be taken are the same as in equivocality: one just has to do something and see what happens, and hope that as actions lead to consequences, those consequences can then retrospectively be made sense of, which hopefully helps the next decisions. 

Marchian ambiguity is something we tend to ignore because it is the most difficult: there are few objective and collectively shared value baselines along which we could unanimously state that this is good and this is bad. An example of the difficulty: Social distancing saves the lives of people who might have otherwise died of the virus, but at the same time it threatens the lives of individuals suffering from domestic violence – how many should we save on the other risk group to justify risking the individuals in the other risk group? The only way to mitigate these issues is by explicitly engaging in ethical discussions – not to find the ‘right’ answers, but to identify the ones that are acceptable to the society in general. 

The following table consists of examples of diverse types of decision-making barriers and their mitigation possibilities both in general and in relation to the ongoing corona disruption. 

Types of uncertainties

The wake of decisions

As has for quite some time been abundantly clear, any decision-making during this time of disruption happens in darkness. This means that there is simply no way of knowing how the decisions and ensuing actions play out in the long run. However, understanding the diverse shades of the darkness highlights the three possible ways of making decisions and living with the consequences.

  1. There are elements in the corona disruption that can be tackled with the advanced technologies – science. These include diverse algorithmic models tracing the patterns of the spread of virus, and medical technologies harnessed in developing the medical counter-measures.
  2. There are elements that have to be made sense of based on the outcomes of actions that have been taken – policies. This means that the point of a decision is to make something happen so that the something can subsequently be analyzed to create more understanding. With every retrospective analysis the shape of things becomes clearer. 
  3. There are elements that mandate active ethical discussions – principles. They are difficult because ultimately they result in values being positioned on a scale between more and less critical. This evaluation of values in turn paints a picture of us as a society: what (and who) do we genuinely deem important?

The decision-makers of today are in an impossible situation. No matter what is decided and acted upon, the outcomes are still invisible, the cause-effect linkages unknown. This means that even with the best of intentions, such outcomes will happen that will not be liked by all after the dust has settled. However, not making the decisions or not acting are not options either – we just have to keep muddling through. 

I salute all decision-makers. Thank you for shouldering the responsibility, be it on a level of a team, an organization, a nation or the whole globe. 

Milla

© 2020 – University of Turku | Privacy Policy | Website: Sivustamo Oy