Complexity thinking and the Agile mindset
So, you’re implementing Agile in an organisation. Why can’t we say, “If senior executives back our agile thing it will work”? Or, “If people have the right mindset” it will work?
The Agile community sometimes talks about an “Agile mindset” being very important in order to make sure that Agile works, with the corresponding inference that a lack of this mindset results in a failure of Agile. But what, exactly, is an “Agile mindset”? How is it defined and how do we help people adopt it?
If these sound like the right questions, this blog post is for you! It’s drawn from a grumpy Dave Snowden (meaning a happy Dave Snowden) talking about “Is Mindset Yet Another Agile Buzzword?” at Agile India 2020.
Here’s the outline his argument from the first 20 minutes of the talk. He references cognitive science and complexity theory amongst other things, which reinforces my growing conviction that we in the software development field need to be engaging with ideas outside of what’s traditionally included in our training.
Summary of Dave’s talk
- We fall back to talking about “mindset” when a context-free method (e.g. SAFe) starts failing in a new context.
- The underlying philosophy of “mindset” is the idea that the mind can be programmed, i.e. a computing metaphor of the brain.
- If the mind can be programmed, we can define the mental model we want (the agile mindset) and install it in our people.
The problem is, this is really bad science. The mind isn’t a computer! Here’s why:
- Inattentional blindness (first-fit pattern matching).
- We don’t make decisions in isolation from others (stories are important).
- Post-Cartesian concepts of consciousness (consciousness includes the brain, body, social interactions, common narratives).
Instead:
- People are complex.
- Change isn’t subject to simple cause-and-effect models.
- Sometimes, things just are. Non-western people seem to understand this more easily.
- Each context is unique.
- There’s no predictive framework: you have a dispositional system, not a causal system.
- What matters is: how do you influence that dispositional state?
- This is a co-evolution of method, idea, engagement, and people and everything else.
What you can do:
- Oh look at the time! Dave starts talking about this 19 minutes in. I’ll leave this for a future blog post.
Here ends blog post! The rest is my transcription of Dave’s talk, heavily edited, but hopefully I’ve got the main ideas across.
Edited transcript of Dave Snowden’s talk, “Is Mindset Yet Another Agile Buzzword?”
We fall back to “mindset” when a context-free method (e.g. SAFe) starts failing in a new context.
The idea that “the mind can be set” is developed from the idea that there are mental models. This is a problem because it assumes the human mind is like a computer — it has a series of programs that it can run. You find people talking about the mind using computer language e.g. “We’ve exceeded our bandwidth” and they use a computer metaphor to describe the way the mind works. This is really bad science!
The computer model of the human mind is deeply problematic. It’s almost the language of programming. The language of culture change becomes, “We decide what culture we want and then we’ll instill that culture in people” i.e. “We’ll program them.” It’s a “Brave New World” or 1984 scenario, if you look at the deep underpinnings.
The reality is that human beings make decisions in very different ways from computers. There are 3 key aspects: inattentional blindness, decisions aren’t made in isolation, and Post-Cartesian concepts of consciousness.
Inattentional blindness
This is when you do not see what you do not expect to see. This is a matter of energy efficiency — in all of biological evolution it’s about reducing energy cost, because energy is expensive in terms of lifespan etc. So most of the time we’ll scan about 2 or 3% of the available data (in other places Dave says 5%). That scanning will trigger a series of pattern-based memories, either partial or complete (some of those are body-based memories). These memories are things you’ve lived through, things you’ve heard, things you’ve been taught, including all the things you know but can’t articulate. This includes the 3 types of knowledge: highly explicit, highly tacit, and narrative-based (which sits between the two).
Once the series of memories have been triggered we do conceptual blending: we blend the patterns together and that triggers a course of action. That’s how we make decisions — it’s a first-fit pattern match. Critically, it’s not a best-fit pattern match.
So the key thing in decision support and decision making is what stimulus you give people and what memories it triggers.
“Mindset” is, in the individual, a triggered set of patterns/memories. It’s not fixed — it may form a stable pattern at times — but it’s not a mindset and it’s not a model. It’s an associated set of fused and blended memories.
There are chemicals in the brain that ensure that we never remember anything the same way twice (Walter Friedman’s work). In evolutionary terms, by introducing constant minor mutation, you increase the ability of humans to stimulate or do things in a novel way.
The gorilla radiologists experiment. This invalidates the process by which user requirements are defined, because everyone comes into it with a pre-given mindset/pattern, they don’t see the anomalies.
Decisions aren’t made in isolation
We don’t make decisions in isolation from other people. This is a key finding from both cognitive neuroscience and complexity theory. What matters is how people connect and who they connect with, particularly during times of extreme change and stress. Who you’re interacting with has a more profound influence on things than who you are originally.
Sacred stories: the stories that everyone gets told e.g. the stories people tell you when you join a company. The stories frame the way we see the work. It’s not a mindset of an individual, it’s a mindset of a collective pattern which gets built through experience, repeated in the informal networks of the organization, not the formal networks.
Post-Cartesian concepts of consciousness
The brain is not the sole repository of consciousness. What you see in Western thinking in a mind/brain dualism, which goes back to Descartes. He effectively separated the mind from the body, possibly because he was scared he’d be subjected to a heresy trial. Separating the body from the mind he opened up the possibility that science could be operate separately from religion.
We now know that consciousness is a distributed function. It’s not just the brain — it’s the body, it’s social interactions. If you look at Andy Clarke’s work with scaffolding, these common narratives provide a highly abstract ephemeral framework which actually informs the way we make decisions and is part of our conciousness.
I’m not getting mystical: if the brain gets damaged, you lose consciousness — so the brain is the origin of consciousness but it’s distributed.
When you pull back your hand from the fire, this is an autonomic process. Your brain neurons fire after the action has been taken, which is not to say there’s no free will but that’s because they’re stuck in a Cartesian model: they assume that the brain has to move the body. The reality is that this is called autonomic processing: the brain fires after the event to double check that the result is right this time. If the reaction proves to be wrong, then the novelty receptor part of the brain will come into play and will think about the problem and maybe change the automatic response (this is crude approximation).
The energy cost of using the novelty receptor part of the brain is very high, so we don’t do it unless we really have to.
Summary so far
These are all important to understand, because why and how people make decisions, how they respond to change and novel ideas is inherently very complex, and very messy. It’s not subject to simple, linear cause-and-effect models. You can’t say, “If senior executives back our agile thing it will work.” You can’t say, “If people have the right mindset” it will work…because these aren’t things in a linear causal chain: they’re not deterministic and they can’t be programmed.
This is the basic lesson of complexity theory: if you’ve got a system which has complex dispositional states, but no causality…that’s a scary thing to contemplate. I generally find it easier to explain in parts of India, China, Africa, Asia, because the religions there, particularly Daoism in China, have never lost the concept of “Some things just are, there’s no reason for them to be, it’s just the way things fell out this time.” That’s not necessarily fatalistic, it’s good science: each context is unique. There may be common patterns but there’s no predictive framework within that. You’ve got a dispositional system not a causal system.
What matters is: how do you influence that dispositional state? This is a co-evolution of method, idea, engagement, and people and everything else.
What you can do
(Dave starts talking about this around 19 minutes into the video; I’ll link to this once I’ve written it up).
