A mindset for not-knowing

13/11/2023 ☼ not-knowing ☼ risk ☼ uncertainty ☼ strategy

This is the reading for the 11th episode in my monthly discussion series on not-knowing.

tl;dr: A mindset is a set of assumptions which shape what we perceive about the situation, how we interpret what it means, and how we decide to act on it. A mindset appropriate to not-knowing begins from having different words to describe different types of not-knowing. This makes both risk and non-risk types of not-knowing in a situation equally noticeable, and favors explicit reasoning about which type/s of not-knowing a partial knowledge situation should be interpreted as being. Only with such a mindset is it possible to identify 1) broad approaches that work for different types of not-knowing, and 2) the type-specific tools that can become part of a toolkit for not-knowing.

The biggest insight so far from over a year of investigating not-knowing is that words matter because words shape how we think. Using the same words (“risk” and “uncertainty”) to refer to different types of partial knowledge makes it hard (or impossible) to clearly distinguish between these types of partial knowledge situations.

(This is why I use “not-knowing” to refer to all types of not-knowing, and qualify the type of not-knowing. Risk is a type of not-knowing, as is not-knowing about possible actions and outcomes, causation, and valuation.)

Actively distinguishing between different types of not-knowing makes it possible to see that they emerge from different sources and have different effects on how we choose to act (strategy).

This insight — that words shape how we think about the unknown — is the foundation of a toolkit for relating productively with not-knowing. The way into this is via the concept of a mindset.

In this context, a “mindset” is a set of assumptions which shape what we perceive about the situation, how we interpret what it means, and how we decide to act on it.¹

How mindset shapes perception, interpretation, and decisionmaking about a situation

A risk mindset is dysfunctional in an uncertain world

Our world today is clearly filled with unknowns. Some of these are risks (where actions, outcomes, and probabilities are well-understood and quantifiable) but most of these unknowns are not risky. They are various forms of true uncertainty.

A risk mindset assumes that all unknowns are risky. A risk mindset

Makes risk-like aspects of a situation more noticeable (while making uncertain-but-not-risky aspects less noticeable).
Favors interpreting partial knowledge situations as simply risky situations.

How risk mindset shapes perception, interpretation, and decisionmaking about a situation

There is a common protestation in reaction to the assertion that the risk mindset is a problem, or even that it exists at all: “people just use the word ‘risk’ as a shorthand. No one seriously believes that all these unknowns are only risky.“

This is wishful thinking. The risk mindset doesn’t merely exist, it is extraordinarily widespread. The evidence is in how people make decisions when they confront unknowns: they use decisionmaking strategies that are only appropriate for true risk situations.

These strategies are invariably forms of cost-benefit analysis or expected value analysis. Both families of analytic strategy require clear knowledge of all possible actions, outcomes, action-outcome probabilities, and relative valuations of outcomes. (Such precise and accurate knowledge about what is unknown is available only in true risk situations.)

A cost-benefit analysis or an expected value analysis can only be interpreted and can only guide decisions sensibly if it is applied to a situation where it is possible to obtain the type and nature of knowledge needed for the analysis.

Let’s consider a few examples:

The risk mindset makes risk-like aspects of a situation more noticeable (while making uncertain-but-not-risky aspects less noticeable). E.g., during the early phase of a novel viral pandemic, the risk mindset might focus attention on verified information about disease transmission dynamics from former pandemics while limiting attention paid to evidence of the virus’s unusual virulence and transmissibility from cities where the first outbreaks had happened. Could this have happened during the first couple of months of COVID-19, when the World Health Organisation recommended not limiting travel or trade in any way at all after doing “a careful risk assessment”?
The risk mindset favors interpreting partial knowledge situations as simply risky situations. E.g., during an industry-wide rush to create complex financial derivatives (especially complex mortgage-backed securities), the risk mindset might lead bankers and investors to interpret such complex derivatives as being calculably risky and therefore viable assets for investment — instead of interpreting them as being embedded in a matrix of financial instruments so complex and interdependent that the behavior of the whole becomes unpredictable. Could this have happened during the run-up to the 2008 Global Financial Crisis?

Even if people genuinely believe that they have a nuanced view of the unknown, they act as if all unknowns are truly risky. This is not because risk-appropriate decisionmaking is generally appropriate for all types of unknowns. On the contrary, risk-appropriate decisionmaking creates enormous problems when applied to non-risk unknowns.

An appropriate mindset for not-knowing

A risk mindset assumes that everything which is partially known is partially known in the specific sense of being risky. The assumption that everything is comfortingly quantifiable and calculable risk contaminates perception, interpretation, and then decisionmaking. It is dysfunctional, even delusional, in a world where unknowns are clearly not all risky. The root of the problem with the risk mindset is its failure to distinguish between different types of partial knowledge.

An appropriate mindset for not-knowing is one which explicitly distinguishes between different types of not-knowing. Such a mindset must have and use different words to refer unambiguously to different types of not-knowing. The simple act of distinguishing — which begins from using different words to refer to different types of not-knowing — automatically leads to paying closer attention to not-knowing.

Such a mindset would

Make the different types of not-knowing (i.e., both risk and non-risk types of not-knowing) in a situation more equally noticeable.
Favor reasoning explicitly about which type/s of not-knowing a partial knowledge situation should be interpreted as being.

How an appropriate mindset for not-knowing shapes perception, interpretation, and decisionmaking about a situation

Mindsets point to broad approaches and specific tools

Words matter because they shape the mindsets we hold. The mindsets we hold determine the processes/actions we end up developing for relating to the world.

Mindsets affect what tools are developed.

The risk mindset conflates under the word “risk” many different types of not-knowing. This makes it privilege a broad approach of treating possible actions, outcomes, and causation as being fundamentally knowable. This leads inexorably to the intensive development of a toolkit ultimately based on the logic of expected value/cost/benefit that relies on this assumption of knowability and is fragile when this assumption doesn’t hold.

Having a mindset which recognises different types of not-knowing is the the starting point for identifying broad approaches that work for different types of not-knowing, and then for identifying the type-specific tools that can become part of a toolkit for not-knowing.

This framing of mindsets for not-knowing is an evolution of my earlier work on the uncertainty mindset. I adapted the original idea of a mindset as an important basis for perception and action from Carol Dweck’s (and her various co-authors’) distinction between people with a fixed mindset (who believe that abilities are static and can’t be changed) and people with a growth mindset (who believe abilities can be changed through effort and learning). Dweck and coauthors focus on mindsets in relation to intelligence, but mindsets apply much more broadly — including in relation to perceiving and acting on partial knowledge.↩︎