The Local Rationality Principle

Local Rationality

“People do reasonable things given their goals, knowledge, understanding of the situation and focus of attention at a particular moment.”

We do things that make sense to us at the time. And a fundamental premise of psychological safety is the belief that you will not be punished for making a mistake if your intentions were good. A key approach to making this real is through applying the Local Rationality Principle. Local Rationality is not a claim that a decision was objectively correct. It is a claim about perspective. When an action looks irrational or puzzling, the Local Rationality principle treats that as a signal that we are judging it from the wrong vantage point — usually an outsider’s view, informed by hindsight, broader system knowledge, and time to reflect.

What is Local Rationality?

We’ve mentioned Local Rationality in previous newsletters but I thought it would be worth diving a bit deeper into the concept. Local rationality is a concept that emerges from the “new view” of workplace safety and organisational performance. The concept stems from an idea that was first introduced by Karl Popper in a 1963 Harvard Lecture, and published in his book “Myth of Framework“, and further built on by Herbert Simon in 1957 with the similar concept of “bounded rationality” where people make decisions that are good enough – we satisfice instead of optimise.

Earlier ideas such as bounded rationality still implicitly compare human decision-making to an imagined standard of full or optimal rationality — something humans fail to reach because of cognitive limits. Local Rationality goes further. It drops the idea of full rationality altogether. There is no hidden, correct decision waiting to be discovered; only decisions that make sense locally, given the goals, knowledge, attention, and constraints present at the time.

Central to safety practices such as human and organisational performance, safety differently, safety-II, resilience engineering, and human factors, Local Rationality helps us to understand the thought processes and motivations behind people’s decisions and actions, especially when incidents occur. It’s based on the premise that people do their best given the constraints they’re operating within. We shouldn’t “blame” someone if they weren’t adequately trained, if they misunderstood what was happening in the moment, or if the environment was such that they couldn’t focus on the task at hand. 

Someone else’s decision often doesn’t make sense to us, because we aren’t making decisions in their context, and Local Rationality helps us examine that context. Pressures from other people, including managers, peers or customers influence how we make decisions or act in a certain way, as do multiple other factors, including awareness of our past successes and failures, the availability of resources and time pressures. Only by examining all of these can we ‘make sense’ of the decisions and actions taken by the person doing the work.

Decisions are also shaped by shared organisational contexts. Teams, professions, and organisations create implicit “decision zones” in which certain actions feel normal, others risky, and some unthinkable. What makes sense locally is therefore not just an individual judgement, but a product of shared assumptions, norms, incentives, and histories within the system.

The underlying idea of local rationality is to understand why a decision made by an individual seemed sensible (rational) to them at the time. Traditional safety measures tend to focus on identifying who failed, whereas approaches incorporating local rationality are more inclined to question what failed, encouraging a more empathetic, learning-oriented perspective. In ourHealthcare Practice Playbook, we include a Local Rationality Investigations Template that helps professionals to think about useful questions to ask and perspectives to consider in investigations. The Local Rationality Principle has also been adapted into the Agile Prime Directive:

“Regardless of what we discover, we understand and truly believe that everyone did the best job they could, given what they knew at the time, their skills and abilities, the resources available, and the situation at hand.”

People are the solution, not the problem.

One important principle of this new view (encompassing Safety II,Resilience Engineering and others) is the belief that people are the solution, not the problem. Traditional safety approaches focus on making sure people comply with rules, whilst “new view” safety considers violations as expected and sometimes necessary. The concept of Local Rationality underscores this, emphasising the importance of understanding the decisions made by employees at the “sharp end” – where the actual work takes place, and where risks and hazards are real. This is vital post-incident or failure, as the aim is to learn why the event occurred so that we can maximise learning from it.

Local Rationality is inseparable from power. What people notice, question, or stay silent about is shaped by authority, hierarchy, professional status, and dependency. In steep power gradients, silence, workarounds, and rule-bending are often locally rational responses — even when they later appear as “failures to speak up”.

Motivation

Local Rationality explains sense-making, not morality. Understanding why an action made sense to someone at the time does not absolve intent, remove accountability, or replace ethical and legal judgement.

Local Rationality doesn’t mean that nobody can ever face consequences for their actions – but it does strongly signal that their intent matters. An interesting, and sometimes overlooked, aspect of Local Rationality is the consideration of a person’s goals. If someone’s intention or motivation is to cause harm, intentionally sabotage something, or recklessly endanger people or a project, we should take appropriate action. We should be careful of assuming reckless or ill-intent where there is not, and conversely, careful of always assuming positive intent. For example, we may assume that healthcare professionals all share a goal of saving, or improving the quality of, lives. In fact, this may not be the case 100% of the time, as we’ve seen with cases such as Harold Shipman. He was acting according to the Local Rationality Principle – however, until the cases came to light, the world didn’t know that he possessed a malevolent intent. 

The lesson here is not that Local Rationality excuses harm, but that intent is often inferred — sometimes wrongly — and must be examined carefully rather than assumed.

Hindsight Bias

As Gareth Lock points out in this Human Diver blog post, Local Rationality recognises that after the event, we have much greater insight and knowledge than people did during the work. Local Rationality therefore helps us to acknowledge our hindsight bias. A good example of this is the investigation into US Airways Flight 1549, where after a bird strike, pilots Chesley “Sully” Sullenberger and Jeffrey Skiles decided to land the plane into the Hudson River off Midtown Manhattan. Initially, investigators attempted to show via simulations that the flight could have returned safely to LaGuardia or diverted to Teterboro. The NTSB report however stated that these simulations “did not reflect or account for real-world considerations, such as the time delay required to recognize the bird strike and decide on a course of action.” [NTSB] In more realistic simulations, in which a 35-second delay was inserted to allow for those factors, the plane crashed. In hindsight, it seemed as if the pilots could have made it to the airport in time, but in the real world, that was impossible due to the constraints they were operating within.

Importantly, hindsight bias is not just a cognitive quirk — it is also a methodological problem. Investigations often reconstruct events using information that was unavailable at the time, then judge past decisions against that enriched picture. Local Rationality requires us to deliberately subtract what we now know, and reconstruct the situation as it appeared then — even when that feels uncomfortable.

Local Rationality and Psychological Safety

Local Rationality is closely linked to psychological safety. Applying the Local Rationality principle in retrospectives or investigations demonstrates to people that they are welcome, even expected, to speak up with questions, concerns, and mistakes. It implicitly states that “human error” is not a useful category of cause, and instead we are looking for systemic causes of failure, attempting to take blame off the table.

From a Local Rationality perspective, “human error” explains nothing — it merely restates that an action made sense locally, but had unwanted consequences in a larger system.

Systemic Improvements

As well as enhancing understanding of people’s decision-making processes, studying Local Rationality offers valuable insights into the workplace systems and the compromises we have to make when working: do we work with extra care and attention, or focus on hitting a tight deadline? Do we ask for help with a heavy lifting task, or get it done by ourselves? What trade-offs are we making when carrying out work?

Focusing on human error can often lead to futile attempts to eliminate it. Instead, by understanding Local Rationality, we can identify and adjust the factors (such as incentives, training, tooling, environment and culture) in the workplace system that influence people’s decisions. This allows for system improvements that enhance the organisation’s performance despite the occurrence of human errors, and in doing so, enhance psychological safety in the workplace – facilitating a virtuous cycle of continuous improvement.

How do we put the Local Rationality Principle into practice?

Practising Local Rationality is not just an analytical technique — it is an ethical stance. It commits us to treating people as sense-making agents rather than problems to be fixed, and to designing systems that support good work rather than punishing inevitable trade-offs.

1. Make the Local Rationality Principle explicit. State it, read it out loud, print it off, include it in resources. Make sure people understand and accept it, and if there are questions or concerns about it, address them.
2. Listen to people. Hear their stories and what their experiences were at the time. Ask for their perspectives and experience in their unique context. Empathise.
3. Understand motivations and goals. Discuss incentives, compromises, trade-offs, and people’s plans and expectations. Are there any motivations and incentives that distract from doing the “right” thing?
4. Examine the environment in order to understand their focus of attention. Concentrate on knowledge “at the time” rather than after the event. Look at the level of noise, distraction, time pressure and other environmental factors. Check the ergonomics of tools, equipment and systems: are they intuitive or difficult to use, especially under pressure?
5. Investigate how things appeared to people at the time. Seek multiple different perspectives and perceptions of events as they occurred. Consider the implications of these different views on how events occurred.
6. Identify systemic improvements to systems, practices, environments, culture and language that surface through the above discussions.

The Local Rationality Principle encourages an understanding of why a decision, action, or error that seems irrational in hindsight, was actually the most rational choice to the person at the time, given their understanding, skills, and constraints. This approach promotes a culture of empathy, continuous learning, and psychological safety, shifting the focus from blame to systemic improvement. Ultimately, understanding and applying the Local Rationality Principle is not just about preventing mistakes, but also about fostering a more understanding, empathetic and psychologically safe work environment.

Links:

Carsten Busch on Heinrich and Local Rationality

Psychological safety at work

This is a great piece on neurodiversity in the workplace, making the very good point that if you’re not creating an environment that is welcoming for neurodiverse people, you’re not only missing out on great talent and opportunity, but you’re likely fostering an environment where people are afraid to disclose their neurodiversity. “All employees thrive in cultures with a strong sense of belonging and psychological safety, but these factors are especially important to neurodiverse employees. Stephanie explains, “If a neurodivergent employee feels like they may be negatively treated due to the culture of the organization, they’re less likely to disclose their condition in the first place, or to seek the support they need to be successful.””

This is a fab paper from 2022 in relation to aviation and aerospace, specifically psychological safety between crew members in the cockpit: The Persistence of Safety Silence: How Flight Deck Microcultures Influence the Efficacy of Crew Resource Management, by Kimberly Perkins et al. The authors show that despite Crew Resource Management (CRM) training, safety silence persists (“safety silence”, in a sense, is the opposite of psychological safety). The study found that 75.8% of pilots feel psychologically unsafe every year because of a negative tone set by the Captain, and 57.3% of pilots feel silenced after raising a safety concern. The paper also provides examples of First Officers feeling undervalued, micromanaged, and uncomfortable because of the behaviour of Captains, and examines Power Distance – which negatively impacts crew dynamics and potentially leads to safety silence and an decrease in psychological safety.

If you’re in tech, this is an excellent rundown of a number of tech failures, their causes, and resolutions, from Gergely Orosz at The Pragmatic Engineer: Interesting Learnings from Outages. A DNS mystery at Adevinta; A failover causing an outage at GitHub; The challenge whether to do a rollback at Reddit; and the difference between public and internal postmortems. George highlights some great takeaways for tech teams too.

And remaining in tech, here’s a great talk from Michiel Rook: “Congratulations, you Failed!” from DevDays Europe last year. Michiel dives into a lot of the stuff we geek out on in this newsletter, from human error and complexity, to Cost of Delay and Test-Driven Development.

Finally, here’s a fab Tedx talk from Joaquín Viñas who attended the Barcelona workshop last week, about Happiness at Work and the competitive advantage.