Protocols at the Limit of Complexity
What protocols are, how they differ from rules and habits, how they shape expertise, and what this reveals about complexity and ethos.
“Civilization advances by extending the number of important operations we can perform without thinking about them.”
- Alfred North Whitehead
I want to continue the conversation from my last post on the complexity budget by extending the idea to an economy or society. I would re-phrase Whitehead’s quote this way
Progress is our ability to take collective responsibility for more complexity, freeing individuals to grapple with new kinds of complexity at different scales.
Last time we talked about products and services as addressing complexity faced by a customer, but now I want to generalise a little, from products and services to technologies and protocols.
Let’s start with protocols
Protocols seem to be all over my feed at the moment, from Model-Context Protocols for AI, to Venkatesh Rao’s Summer of Protocols project. Thanks to the latter, I’ve come to see how they have been a mainstay of movie narratives for most of my life. In fact, it’s not really a stretch to say that I have been raised and live in the era or protocols. My entire adult life has been shaped by protocols: engineering protocols like TCP/IP, public health protocols (slip, slop, slap), protocols of international diplomacy, even protocols of reserve bank independence.(1)
But what are protocols?
Protocols are repeatable, transmissible, mnemonic procedures that reliably shape behaviour under pressure or uncertainty.
Some familiar examples:
Public Health protocols - singing Happy Birthday while washing your hands
Surgical protocols - counting the surgical instruments before closing (2)
Military protocols - Napoleon’s exhortation to troops: if you get disconnected from command, follow the sound of the guns;
Safety protocols - Aviation checklists
Life hacks of the 00’s were classic protocols, including GTD (Getting Things Done) which is a collection of time management protocols.
The point is that protocols are everywhere, and I think they are more interesting than they seem, opening up an important genre of norm-governed behaviour that may be fundamental to understanding what we mean by an ethos, and how that ethos helps us grapple with the complexity of the world around us.
Protocols and Rules
It is quite difficult to articulate the difference between a protocol and a rule.
Cognitively, they are very similar - protocols encourage or discourage specific behaviour in specific conditions - which is to say, like rules, they have normative force, so they can be treated as reasons for behaviour.
Also like rules, failing to follow them has consequences. As John Cosgrove reminded me, you can lose your head for failing to conform to the protocols of a royal court, or face other forms of physical, legal or reputational threat for ignoring them in other contexts.
So recognising that we use the term interchangeably with rules, I want to focus on that class of normative constraints that are deliberately less rigid and more open to interpretation.
The protocols I have in mind are more humble than rules. They are deeply situational, rooted in a moment or a place, and they do not attempt to apply universally.(3) Diverging protocols contrast, but do not contradict each other the way that rules do.
I also recognise that protocols can and do harden into rules, and rules sometimes even soften into protocols, and both interact with each other in everyday behaviour. Nonetheless, I think a distinction between them is helpful, even if they are not independent phenomena.
Let’s look at this distinction from a few different angles.
Psychology
One of the key differences between protocols and rules is the way they relate to the psychology of the individual.
Protocols deliberately preserve the judgement and agency
of the individual applying the protocol.
By contrast, rules almost aspire to be systematised. It would be great for our rules to be habitual, even invisible. We’d rather that no-one ever considered murder as an option, for example.
Rules are supposed to constrain the individual, to reduce their span of control.
Protocols, by contrast, don’t have the same antagonistic relationship with individual freedom.
Protocols and Values
Protocols have a strong connection to the embodiment of values. They help us embody values (safety, fairness, efficiency, connection) by shaping perception - helping us notice situations where those values need to be applied/defended - and enabling judgement - reducing the cost or competence required to express those values.
Rules, by contrast, tend to offer the individual the opportunity to offload judgement - to externalise and commodify responsibilities into a bureaucratic structure - and they release the individual from the need for moral perception - ie to see the world in terms of value or values - replacing this with a purely descriptive perception.
Is X the case? If so, then I do Y. There’s no need in this scenario to assess the value of X in any way. That’s kinda the point of making a rule - taking the judgement out of the hands of the individual.
Workarounds
Often protocols develop as workarounds for rules that are not sufficiently fine-grained for the situation on the ground. Getting around policy constraints in ways that respect the values that inspired those constraints is one example. The “spirit of the rule” is often captured by protocols that are used even when suspending the rule.
Rules can be easily overwhelmed by circumstances - for example, in disaster scenarios or massive trust breakdown. The rule itself becomes too expensive or even dangerous to enforce. At that point, protocols rapidly emerge to coordinate action.
Culture
Protocols are socially legible. They function as a shared script that is recognisable to other community members - indeed that is what makes them both slow to be adopted, and very hard to displace once they have been adopted. Think of greeting rituals: handshakes, bows, cheek kisses. None are mandated by law, but each carries meaning, marking both respect and belonging.
Protocols are a cultural phenomenon, whereas rules are a bureaucratic one.
Rules strip culture down to what can be enforced. Protocols, by contrast, tend to accumulate: songs, gestures, sayings, and rituals that carry values forward. That’s why protocols so often endure long after rules change - they become part of identity itself.
Protocols and Habits
Protocols also need to be distinguished from habits.
Where rules externalise practices or values - taking them away from the individual. Habits are a way of internalising practices or values to such an extent that they are no longer consciously managed.
What habits share with rules is that they no longer engage the individual’s judgement or assessments of value. It’s like the “because” has been dropped from the action. “I don’t do Y because X. I just do Y.”
A third party can still interpret the behaviour as having an intention behind it, but this intention is no longer motivating the way reasons normally are. It’s an imposed interpretation, not a lived intention. Even the individual themselves can do this interpretation of their behaviour, but they may well feel even more alienated from that behaviour as a result. “I guess I do Y because X, but who knows?”
Protocols, by contrast, are offered as motivating reasons for behaviour, but they are a special category of reason. They are reasons why we (or people like me) do this kind of thing.
Expertise
Finally, let’s talk about the relationship between protocols and expertise.
Like habituation, the development of expertise gradually takes the individual’s practice away from protocols, but for a different reason.
From the perspective of an expert, protocols look like rules of thumb that we use to help a non-expert make decisions or avoid blunders.
They are temporary solution to a gap in expertise, or scaffolding to support learning without being overwhelmed.
Chess is full of such protocols:
Control the centre
Don’t move a piece twice until development is complete
Knights before bishops
etc etc
A grandmaster, however, has achieved a level of mastery where they can safely ignore the protocols.(4) It would be a mistake to think of expertise as the internalisation of rules. (5)
Protocols at the Edge of Complexity
And this brings us full circle to complexity. Protocols allow us to engage with a level of complexity that would otherwise be beyond our ability. Mastery is not, however, the hardening of these protocols into universal rules. Nor is it discovering simple practices that can be habituated.
On the contrary, expertise comes when we have developed the ability to confidently grapple with that complexity directly.
An important and perhaps controversial point is that expertise isn’t a necessity. Sometimes protocols are sufficient. We said last time that Terms & Conditions are a way of constraining the amount of complexity we are collectively taking responsibility for. As a society, there is a natural limit to the amount of complexity we can take on, and given that limit, in any specific domain, the protocols we have may be sufficient. After all, expertise is expensive.
Expertise begins where protocols end -
but society can’t afford expertise everywhere
Perhaps the more important point, however, is that expertise is simply not available when we attempt to address a new form or scale of complexity. The best we can do, at that point, is to try to develop protocols that help us assess the situation and avoid blunders.
Ethos
Sometimes you follow all the protocols and bad things happen nonetheless. If you’re held responsible, it’s a moral hazard; if not, it’s bad luck. Where a society or a team draws that line shows the limit of its ability to grapple with complexity.
We glimpse the ethos of a team or profession most clearly in the protocols it develops and adopts, because they show us how that team chooses to grapple with complexity, and the responsibilities it accepts even when habits, systems and expertise are not enough.
Notes:
(1) If this is not obvious, consider the unexpected ease with which the administration is breaching those protocols in the US today, or compare to the earlier part of the 20th century, which was an era of ideologies and grand narratives.
(2) Protocols are closely related to what Dave Snowden calls “enabling constraints”. I mentioned these, and his example of surgical protocols in The Rise of Complexity Work. There I used the idea of a posture to capture the way different protocols are sensitive to both place and time. I also owe the example of Napoleon’s instructions to his troops to Snowden. Protocols may also help to answer Tim Rayner’s question on that piece about how we train people to adopt postures. I have come to think that what I mean by postures are supersets of protocols, combined with other protocols about which posture to adopt where and when.
(3) For example, if the patient is crashing, the protocol for counting the instruments should be abandoned. But no-one expects the protocol to include a comprehensive account of when to abandon the protocol. This is one of the strengths of protocols. Their humility (if that’s the right word) means we don’t end up feeling compelled to dive into a recursive hell of exception classification in order to articulate them.
(4) Grandmasters will often ignore protocols deliberately, precisely in order to destabilise an opponent and generate less predictable game play, because they are confident in their ability to leverage the unexpected opportunities to attack this opens up.
(5) As Hubert Dreyfus taught us.