Making Sense of Sensemaking: Arguing like you are right and listening as if you are wrong.

When something goes wrong, or feels like it is slipping away like a full supermarket trolley on a slight incline, how do you respond? The recent passing of Karl Weick got me thinking about the different approaches to understanding how we make sense of things. I am, to quote Suzy (Eddie) Izzard, quite thinly read in general, but thought it might be an interesting exercise to have a look at a few of the most well-known and influential people in the field and how their approaches differ and what health and safety practitioners can take from each position. Each of those I have chosen offers different concepts, tools and frameworks to explain how we understand the world around us. I’m sure each of them would disagree on many concepts but I think they all would agree with Weick’s ideas of ‘arguing like you are right and listening as if you are wrong.’

Karl Weick: you're already making sense

Weick's central thesis  is that people act and then construct meaning from what they've done. Sensemaking runs backwards, from experience. You're already in it before you know you're doing it.

Think of it this way: people who persist with a plan that's clearly not working aren't irrational; they’re operating on a mental model that's updating slower than the situation is changing. Weick's disaster research (which included Mann Gulch, Tenerife and Bhopal) reveals similar patterns. Not that those involved at the time were necessarily incompetent, just that their ability to create meaning from what they were seeing couldn't keep pace with events. The conclusions he drew are that this is down to the fact that the organisations involved did not have structures in place that could process what was being observed, create meaning and provide feedback to those ‘on the ground’ to help them understand what was happening and allow them to mentally process what was happening.  

The thing to consider here is what do people need that will allow them to notice their current understanding of the context is wrong before the consequences become irreversible. Situational simulations, event reporting, debriefs and pre-task conversations give context to what has and might happen and create spaces where there is a reduced friction opportunity to show unease. 

We all need places to discuss different types of interventions that might be useful in controlling or managing situations that haven’t happened yet. This key is not to build a library of  prescribed, formalised interventions, but because they give ideas and diverse approaches to how to manage similar situations somewhere to propagate.

Gary Klein: Pattern recognition is a powerful thing, and yet poorly calibrated

Klein spent years watching experts working in high pressure contexts.  Fire commanders, ICU nurses and military planners all became areas of research. He found that they don't analyse options and select the best one in a neatly packaged, linear way. They recognise a situation as familiar, run a quick mental simulation of whether the familiar response will work, and act. He called this Recognition-Primed Decision making.

Human mental processes are impressive and have served us well so we need time to remember that tacit expertise deserves respect. When an experienced worker says something feels wrong and can't say why, that's pattern recognition operating. Just because a feeling can’t be articulated doesn’t make it wrong. Dismissing something because it can't be documented is both epistemically and practically costly.

Klein's research shows that familiarity with a situation and making good decisions is not as clear-cut as it appears. Recognition-primed decisions fail not only in genuinely novel situations, they also fail when an expert correctly identifies the pattern type but misreads which features of the situation are helpful in understanding what is happening. Two situations that look identical can have completely different underlying dynamics. Expertise isn't about having a large pattern library to draw from. It's the ability to identify when your pattern recognition is trustworthy and when it isn’t. That awareness is rarely built into the systems or training the health and safety profession develops.

We train for hazard identification, risk assessment and procedure. We rarely train people to ask, “Is this actually the kind of situation I think it is?”  Pre-mortem thinking, working backwards from an imagined failure, is one way of building this skill into a workforce. But the deeper requirement is for organisations to normalise uncertainty, even among their most experienced people, rather than treat expressed uncertainty as a competence problem.

Dave Snowden: know what kind of problem you have and notice when the context moves.

Snowden's Cynefin framework starts with a question many in health and safety skip. “What is the actual nature of this situation?” The argument is that cause and effect have different relationships across different contexts. 

In Clear situations, the relationship is obvious, cause and effect can be seen by everyone. We should apply established procedures that we know from observation and experience, work. 

In Complicated ones, cause and effect are still observable, and we can use best practices to manage risk. The difference is that to know this, we need expert analysis, so we need to find someone with expertise to help us see the link. 

In Complex situations, cause and effect are only legible after the fact. No expert can reliably predict outcomes. We need to test ideas and see what happens in order to develop understanding and guide us.

In Chaotic situations, do things, create structure, get control and then work out what to do next. 

Within health and safety, this framework has real value, both for deciding how to manage risks based on their context, but also as a tool for explaining to positions of influence why risk management needs to be contextual.  Common critical risks, such as working at height, confined spaces and energy isolation, mostly fall under the Complicated domain. They are known hazards which have established controls and procedures that work when followed. Culture, chronic stress, and fatigue sit in the Complex domain, where more rules and better procedures are the wrong response. Probe, observe, adjust. Find where you can go from where you are now. 

Cynefin is routinely misread as a classification tool. Find the right box, apply the right logic, problem solved. That misses the point. Domains aren't stable categories and situations will migrate between them, sometimes rapidly and with little warning. That is why the framework has a cliff edge between what is clear and chaos. You can be going along happily and then fall off a cliff if you are not paying attention. We need to recognise the signals that lead to those transitions, and what effect they will have on our ability to understand the situation. The aim of the framework is not to fit everything into a neat box, nor to achieve the comfort of categorisation. 

Brenda Dervin: the gap between what workers know and what the system hears

Dervin's question is more structural. When safety systems are designed, whose understanding of the situation do they reflect? 

Her research consistently found that institutions treat their own framing as correct and worker knowledge as a deficit. This is something that health and safety often addresses through training, communication, or procedure, rather than as knowledge to be taken seriously.

Reframing worker consultation as a knowledge gathering exercise rather than a compliance exercise is necessary and definitely worthwhile, but insufficient. Many health and safety systems aren't designed to receive what workers actually know.  They are designed to receive worker input in forms the system can process. Hazard reports, incident forms and structured consultation with a predetermined scope. 

The gap between what workers know and what organisations hear isn't primarily a communication problem. It's a power problem. Institutions decide what counts as valid knowledge, and that decision is rarely examined. How ‘work is done’ isn’t an arbitrary decision people make, it is driven by the constraints that are placed on them by the organisation. 

Dervin’s research was interested in whose cognition gets treated as expertise. A system could become genuinely good at highlighting and respecting the tacit knowledge of experienced tradespeople while remaining structurally unable to hear workers who are newer, less confident, work in lower-status roles, or don't fit the profile of credibility that the organisation recognises.  Asking people to describe their situation, their gaps, and what would help, in their own language and without imposed categories (not everything is dumb, dangerous, difficult or different), is designed to reach precisely what most frameworks are built to miss.

Hazard identification that starts with a checklist is already too late. The checklist frames what organisations already know. 

What’s the point? (Is there even one?)

It’s a valid question. Maybe there isn’t. Maybe you can get all you need to know about health and safety management from safety books, safety podcasts, ‘trusted safety voices’ and the LinkedIn safety echo chamber. 

For me, though, true innovation will not come from within the industry. Only by adapting concepts from complexity, psychology, anthropology, etc., and testing their theories in new and novel ways will we progress.

None of the ideas I’ve looked at here resolves into a nice, neat procedure to follow, but hopefully they will spark some curiosity to move beyond entrenched ideas into something more meaningful. 

TLDR (yes, I have put it at the end, so if you didn’t read, you’ll never know!

Weick: You can't choose your sensemaking approach from outside the situation; you’re already involved in it, shaping meaning. The question is whether your organisation can listen to what that process produces.

Klein: The most experienced person in the room has the most reliable pattern recognition and the least reliable sense of when that recognition is failing. Both things are true simultaneously.

Snowden: The goal is staying alert to when the causal logic you're using stops fitting the situation you're actually in.

Dervin: The knowledge your system can't process is probably the knowledge you most need. That's not a communication design problem. It's a question of whose understanding the system was built to receive in the first place.

These are persistent questions to keep asking, and the organisations that ask them seriously will find they're more insightful than any procedure change.