Why Clever People Fail to Learn: Chris Argyris on Defensive Routines
Chris Argyris, who lived from 1923 to 2013, spent his career on a question that most management writing avoids. Why do clever, well-meaning people in organisations keep failing to learn from their own mistakes? His answer is uncomfortable, and that is probably why he gets mentioned more often than he gets read.
The core idea: what we say versus what we do
All of us carry two theories about how we act.
Your espoused theory is what you say you believe. “I want honest feedback.” “We decide things on merit.” It is what you would tell someone if they asked.
Your theory in use is the theory that actually drives your behaviour. You cannot see it directly. You can only work it out by watching what someone does.
The gap between these two is usually invisible to the person who has it. Argyris was resolute on this point. It is not simple hypocrisy. People genuinely do not notice. He tested this many times. He would ask managers to write down how they would handle a difficult conversation well, then compare what they wrote with what they claimed to value. Almost everyone showed a mismatch, and almost nobody spotted it in themselves.
Model I: the theory in use that nearly everyone has
When Argyris studied people’s actual behaviour, he found the same small set of underlying values again and again. He called this Model I.
- Get the result you intended
- Win as much as possible, lose as little as possible
- Keep negative feelings hidden
- Appear rational
Model I makes you defensive by design, because the values themselves reward staying in control rather than learning.
Model II is the alternative. It rests on getting valid information, letting people choose freely and with full knowledge, and building real attachment to that choice. Argyris was clear that almost nobody works this way naturally, including the people who teach it. He put himself in that group.
Defensive routines
This is where the idea becomes organisational rather than personal.
An organisational defensive routine is any policy, habit, or action that protects people from embarrassment or threat and, at the same time, stops anyone from finding and fixing the cause of that embarrassment or threat.
The second half is what matters. The routine protects you from discomfort by protecting the source of the discomfort from being examined.
Argyris described the pattern as a bypass, and it has four steps.
- Send a message that contains a contradiction, such as the “constructive” performance review that hides the real problem
- Behave as though there is no contradiction
- Make both of these things impossible to talk about
- Make the fact that they cannot be talked about also impossible to talk about
Step four is what stops the routine. You cannot fix the problem because you cannot name it, and you cannot name the fact that you cannot name it. Everyone knows. Nobody says. Anyone who does say becomes the problem.
What this looks like at work
The meeting after the meeting. The real conversation happens in the corridor between three people. The official meeting produced an agreement that nobody actually holds.
Skilled incompetence. This is his sharpest phrase. People become genuinely skilled at producing the result they do not want. Years of practice at being diplomatically vague make you very good at avoiding the conversation that would actually help.
The comfortable status label. A project sits at “amber” for months while everyone privately knows it is finished. The reporting category exists so that the truth never has to be said out loud.
Blaming the culture. “That is just how things are here.” Argyris was hard on this one. It moves responsibility to something that nobody has to act on, so it is itself a defensive routine.
Fancy footwork. Ignoring the contradictions in your own behaviour, or explaining them away smoothly when someone points them out.
Single loop and double loop learning
His other well-known idea connects directly to all of this.
Single loop learning fixes mistakes without questioning the underlying assumptions. Think of a thermostat noticing the room is cold and turning on the heating. Most improvement at work is single loop: tighter processes, better measures, more training.
Double loop learning questions the assumptions themselves. It asks whether twenty degrees is the right target at all, or whether we should be heating this room in the first place. That means examining the thinking that produced the mistake, which means examining yourself, which sets off all the defences described above.
So defensive routines are exactly what block double loop learning. Organisations become very good at small improvements while the deeper structure that produces the errors stays untouched. That is why change programmes so often deliver real local progress alongside no real systemic change.
The consultants study
Argyris’s most quoted piece of research looked at management consultants. They were young, very bright, and used to succeeding. He expected them to learn well. They were the worst learners in his sample.
His explanation was that they had never failed at anything. With no experience of failure, they had no practice at learning from it. When anyone suggested they had made a mistake, they became defensive and blamed the client, the market, or anything else outside themselves. High achievers, he argued, are often the least able to do double-loop learning, because their sense of who they are depends on being right.
There is a real sting in that for anyone in a senior role.
Where to start reading
- “Teaching Smart People How to Learn”, Harvard Business Review, May 1991. The best entry point. Short, and it covers the consultants study.
- “Skilled Incompetence”, Harvard Business Review, September 1986. The defensive routines argument in readable form.
- Overcoming Organizational Defenses, 1990. The fullest treatment, and the one to read if you want the mechanism properly.
- Theory in Practice, 1974, written with Donald Schön. Where espoused theory and theory in use are set out. Heavier reading.
- Knowledge for Action, 1993. Case based, and honest about the interventions that failed.
