How to tip social systems (for the better, of course?)

In September 2022, the University of Exeter convened an international meeting titled ‘Tipping Points: from climate crisis to positive transformation’.  Part of the conclusion, and the subject of subsequent work (See USS outlines new climate scenarios for improved decision making), is the idea that ‘positive social tipping points’ are probably the fastest and most powerful way of addressing the climate crisis. That is the origin of this thought piece – how might we tip a social system?

Let’s start by creating a model to represent a generic social system. The model will be in the form of a network that has nodes and connections. The nodes will be entities that are capable of making decisions – so individuals and business (but with scope to go more abstract into algorithms, smart contracts and generative AI). The connections between nodes are flows. As we are considering social systems, these flows can include virtues like friendship, help and love, as well as more typical flows like information, money, goods and services.

Complexity science often refers to systems performing computation (they work out the best allocation of resources – the ‘invisible hand’), so let’s use computers as an analogy. A computer has an operating system and the real world equivalent is the ‘rules of the game’. The ‘rules of the game’ is shorthand for a very large set of layered ‘commands’ which govern the behaviour of individuals and corporations. They include international, national and local laws, as well as unwritten values and norms that govern ‘how we do things in this part of the network’[1]. They therefore encode the prevailing ideology (eg capitalism is the best way to organise an economy), amongst other things.

Similarly, the ‘software’ describes how the node processes incoming information and makes decisions. Some of the node’s decision-making algorithm may be standard across most local nodes, but some of the lines of code are likely to be bespoke and dependent on individual context. For example, the majority of nodes at the present time are likely to have a component in their algorithm which encodes the following sentiment: “the products of fossil fuel companies are currently required for internal combustion engines; AND immediately cutting off the supply of those products would be more harmful than beneficial”. We can imagine that the algorithm of a climate activist encodes a conflicting, or opposite, sentiment.

Using this mental model, we now have a line of sight to the answer to our question – how might we tip a social system? We can change the information flowing through the network; we can change the individual decision-making algorithms; or we can change the rules of the game that apply to all nodes.

In essence, our desire to tip a social system implies a number of things:

• The behaviour of the social system is currently suboptimal (against some objective; this is likely to be a value judgement)
• The behaviour of the components of the system might be suboptimal (against the larger, system objective)
• We have identified a mechanism by which we can easily change the behaviour of at least some of the components (a tipping point implies we are looking for small changes that can have a large effect)
• The change in behaviour of those components will propagate through the system, causing other components to change their own behaviour
• The aggregate result of the changed components, will be significant change at the system level.

Our mental model shows how we might attempt to intervene within the system to bring about the change we desire. Most powerful, but most difficult, is to change the ideology (operating system). For example, we could seek to replace “growth is good” with “growth that damages the ecology or the environment is bad”.

We can also lobby for changes to the law. Many countries have already signed into law net-zero emissions commitments, opening the door for further laws to aid its achievement. This would change the societal incentive structure (the rewards and punishments attaching to behaviours). For example, a law that changes the price structure will trigger multiple behaviour changes.

Next, we can try to change the software. Because we are dealing with social systems, this will include a consideration of values and ethics, not just beliefs about how the world works. For example, does a human life in the global south have the same value as a human life in the global north?

I would argue that our current algorithms imply it has a lower value. If that is an uncomfortable, or even abhorrent, thought, then you are free to adjust your own algorithm accordingly – but the change might not produce as much financial return. To push a social system over a tipping point, we are effectively looking for the equivalent of a computer virus – a change in code that spreads through the network, altering the algorithm of each node it ‘infects’. This is what climate activists believe they are trying to do.

Finally, we can seek to change the information flowing though the network (the inputs to the algorithms). In a sense, this is what climate science has been trying to do.

In this thought piece I have only been able to sketch the initial idea. However, it seems to me that the conversation over social tipping points would be greatly enhanced if it included the change mechanism it was seeking to employ, in order to trigger the system change it would like to see.

Tim Hodgson is co-founder and head of research of the Thinking Ahead Institute at WTW.

[1] In the framing offered by Donella Meadows in Leverage Points: Places to Intervene in a System, our rules of the game relate to her three most significant (and hardest) intervention points – the mindset out of which the system arises; the goals of the system; and the rules of the system (such as incentives, punishments, constraints)