I'm Paul Stevens. For more than 35 years, my work has been in aviation communication — particularly the teaching and testing of English for pilots and air traffic controllers working in an international environment. That meant spending a lot of time thinking about communication. Why do two people hear the same words and understand something different? Why does an experienced person notice something a novice misses? How do we know when enough understanding has been achieved? What happens when misunderstanding isn't noticed? And how can training make any of this better? Over time, I found myself becoming increasingly interested in what happens before communication. Before we can communicate something, we have to notice something. Before we can decide what to do, something has to become available to judge. And because our time, attention and other resources are limited, we can't investigate everything. Those questions eventually took me well beyond aviation.

Wait a second...

NL didn't begin with a plan to build a framework. It began with things that caught my attention. A question would snag. I'd follow it. That would expose another question. Something I'd previously accepted would stop looking quite so obvious. I'd pull an idea apart, try to understand what it was doing, put it back together and see whether the new representation was more useful. Sometimes it was. Sometimes it wasn't. And sometimes an idea I thought I'd finished with came back months later because something new had made a difference visible. Eventually, the pieces began to connect. I started calling the result the The Normal-Model Framework. This eventually became The Noticing Loop.

I haven't discovered THE way

That's important. NL is a representation. It's one way I've found of looking at how we notice, allocate limited resources, judge, choose, communicate and learn. A way. Not the way. At this moment, I believe there is something useful here. I have reasonable grounds for that belief. But believing something doesn't make it immune from challenge. Quite the opposite. Build the model. But remain open to questioning the model. That applies to NL too.

Can we get a little better?

This is ultimately what interests me. Not perfection. Not eliminating uncertainty, disagreement, misunderstanding or mistakes. And certainly not telling other people what BETTER should mean. Something much smaller. Can we notice one useful difference that would previously have passed us by? Question one assumption? Make one judgment a little more deliberately? Communicate sufficiently better with one receiver? Learn something from getting it wrong? And can some of that be trained? I believe it can. And small doesn't necessarily mean insignificant. One small improvement may make very little difference on its own. But we notice, judge, choose, communicate and learn again and again, throughout our lives. Across thousands of decisions, millions of people — perhaps billions — a very small difference can become a very large difference at scale. NL doesn't need to make us dramatically better at everything. A little better, a little more often, may be enough. That's enough reason for me to keep exploring.

This is an open door

I'm putting NL into the world before it is finished. I'm increasingly doubtful that “finished” is even the right objective. I want others to encounter it. Use it. Question it. Find differences. Tell me where it doesn't work. Show me something I haven't noticed. Perhaps some of those differences will change NL. I hope they do. Because if the framework asks us to remain open to changing our models when something better becomes available, it ought to be capable of doing the same itself. A framework for better. You decide what better means.