Complexity Science and Social Systems

The New England Complex Systems Institute recently had a a webinar titled, “2024 Election and Complexity Science: Insights for Collective Action and Governance.“

As I do with most YouTube videos, I asked Claude to summarize it for me and then asked questions about it. I often keep notes on interesting conversations, but this seems relevant enough to post here. I’ll put the prompts in bold to stand out. The first one is to summarize the transcript.

---

Let me provide a comprehensive summary of this lecture about complexity science and its implications for elections and societal systems.

The speaker presents several key arguments:

1. Complexity Science and Systems

• Complexity science studies how components work together and their interdependencies
• Uses concepts like multiscale understanding and the “complexity profile” (measuring information needed to describe systems at different scales)
• Introduces Ashby’s Law: an entity must have greater complexity than what it’s trying to control to be successful

2. Organizational Evolution

• Traditional hierarchical organizations are limited by the complexity of individual leaders
• Society’s complexity has exceeded individual human capability since the 1980s
• Network-based organizational structures can achieve higher complexity than hierarchical ones
• While a smooth transition to networked systems might be expected, we’re seeing resistance

3. Democracy and Current Challenges

• Representative democracy, with periodic voting for central control, is described as an oversimplified decision-making system
• Current political polarization is creating a pattern where:
• Moderate candidates don’t necessarily win anymore
• Abstaining voters can lead to more extreme outcomes
• The population is increasingly split into two distinct groups

4. Adaptation Strategies
The speaker identifies two competing responses to increasing complexity:

• Adaptation through network-based systems that can handle higher complexity
• Reduction of environmental complexity through:
• Monopolization
• Media control
• Conflict/warfare
• These methods maintain power but fail to address complex challenges like pandemics, climate change, and poverty

5. Proposed Solutions

• Suggests moving toward cooperative rather than competitive models
• Advocates for creating network systems for decision-making
• Warns that current social conflicts may be a “divide and rule” strategy diverting attention from underlying issues
• Emphasizes that ethics naturally improve in systems with strong collective action and consequences

The speaker concludes that we cannot rely on existing systems to provide societal safety and structure, and must actively work to create new network-based systems for public good.

---

Prompt: Elaborate on what network systems for decision-making would look like. Do they mean voting or something else?

Based on the transcript, the speaker actually makes a clear distinction between traditional voting and network systems for decision-making. They explicitly state that voting is a “very oversimplified dynamic” that “fails to integrate and leverage the diverse strengths and capabilities of individuals.”

Instead of voting, network systems for decision-making would likely have these characteristics:

1. Distributed Control

• Rather than having a single person or small group making decisions, authority would be spread across many interconnected nodes
• This allows the system to have greater complexity than any individual member
• Similar to how modern corporations have moved away from top-down management to more distributed decision-making processes

2. Functional Differentiation

• The speaker uses biological development as an analogy, where cells differentiate into different tissues and organs that serve different functions
• In a social context, this would mean different groups could take on different roles based on their strengths
• Instead of forcing consensus or having one group dominate others, groups would operate independently while cooperating

3. Cooperative Rather Than Competitive

• Groups would focus on “you do you and we’ll do us” rather than trying to control each other
• Different groups could help each other while maintaining autonomy
• This would avoid the “divide and conquer” dynamics that the speaker suggests powerful interests use to maintain control

4. Collective Action

• Decisions would emerge from the interaction of many participants rather than from single votes or individual leaders
• Success would depend on collective rather than individual action
• This would naturally encourage more ethical behavior since benefits would come from cooperation rather than exploitation

However, the speaker doesn’t provide specific examples or detailed mechanics of how such systems would work in practice. They acknowledge that complexity science is “still early in its development” and that creating these network systems is something we “have to step up and figure out.”

It’s worth noting that while the speaker criticizes current democratic voting systems, they don’t entirely dismiss electoral reform – they mention that ranked choice voting and mandatory voting could help with current problems. However, they suggest these would be incremental improvements rather than the fundamental transformation they believe is necessary.

The closest existing analogies might be:

• Open source software development, where many contributors work independently but cooperatively
• Scientific research communities, where different labs and researchers contribute to collective knowledge
• Some modern corporate structures that use agile, distributed decision-making processes

But the speaker seems to be suggesting we need to develop new forms of collective decision-making that go beyond these existing models.

---

Prompt: Would liquid democracy fit into this paradigm?

Analyzing this through the lens of the complexity science framework presented in the transcript, liquid democracy would partially align with what the speaker envisions, but may not fully capture their concept of network systems.

Ways Liquid Democracy Aligns:

• It moves beyond simple periodic voting to a more dynamic system
• It allows for specialization (people can delegate votes to experts in different domains)
• It creates a network structure of delegations rather than pure hierarchical control
• It can theoretically handle more complexity than traditional representative democracy since expertise can be distributed across the network

Ways It May Fall Short:

1. Still Fundamentally Vote-Based

• The speaker explicitly criticizes voting as “a very oversimplified dynamic” for combining people’s input
• Liquid democracy, while more sophisticated, still reduces complex decisions to votes
• It doesn’t fully “integrate and leverage the diverse strengths and capabilities of individuals”

2. Competition vs. Cooperation

• Liquid democracy still operates on a winner-takes-all principle for individual decisions
• The speaker advocates for systems where different groups can operate independently while cooperating (“you do you and we’ll do us”)
• They suggest moving away from having any groups impose decisions on others

3. Complexity Handling

• While liquid democracy can handle more complexity than traditional voting, it may still be limited by the need to reduce issues to discrete voting choices
• The speaker seems to envision systems that can handle higher levels of complexity through genuine collaboration and emergence, similar to how biological systems develop

That said, liquid democracy could potentially be seen as a stepping stone – an intermediate form between current representative democracy and the more fully networked decision-making systems the speaker envisions. It introduces some network principles while remaining within a familiar voting paradigm, which might make it easier to implement as part of a transition.

---

Prompt: How would an open sourced software paradigm translate into a political process?

Let me analyze how open source software development principles could translate to political processes, based on the complexity science framework from the transcript.

Key Translation Points:

1. Distributed Decision Making

• In open source: Issues and features are proposed by anyone, discussed openly, and implemented by those with the capability
• Political translation: Policy proposals could emerge from anywhere in society, with transparent discussion and implementation by those with relevant expertise
• Instead of voting on pre-packaged platforms, people could directly contribute to or modify specific policies

2. Fork and Merge Dynamics

• In open source: Projects can be “forked” to try different approaches, with successful innovations merged back
• Political translation: Different regions/communities could experiment with policy variations
• Successful policy experiments could be adopted by other communities, similar to how states sometimes copy successful programs from other states but with more systematic sharing

3. Modular Architecture

• In open source: Complex systems are broken into manageable, interconnected modules
• Political translation: Break governance into more granular domains where different groups can focus on their areas of expertise
• Policies could be more modular and combinable rather than all-or-nothing packages

4. Merit-Based Contribution

• In open source: Influence is earned through quality contributions rather than position
• Political translation: Policy influence would come from demonstrated expertise and successful implementations
• Leadership would be more fluid and domain-specific

Challenges to Address:

1. Speed vs. Deliberation

• Software can be quickly rolled back if bugs are found
• Political changes often have longer-term, harder-to-reverse impacts
• Would need robust testing/simulation processes for policy changes

2. Participation Barriers

• Open source requires technical knowledge
• Political participation would need to be more accessible while maintaining quality
• Would need systems to translate between technical policy details and public understanding

3. Coordination Problems

• Software has clear functionality tests
• Political success is often more subjective and hard to measure
• Would need new ways to evaluate policy effectiveness

4. Resource Allocation

• Open source software can be copied infinitely
• Political decisions often involve finite resources
• Would need mechanisms for resolving resource conflicts

5. Accountability

• Software bugs have clear technical impacts
• Political decisions affect human lives
• Would need robust accountability mechanisms while maintaining distributed control

This aligns with the speaker’s framework because it:

• Creates higher complexity than individual decision-makers could handle
• Enables functional differentiation
• Allows for cooperation without central control
• Creates collective rather than individual incentives
• Maintains transparency and accountability

The transition would likely require:

1. New tools for policy development and testing
2. Education in collaborative decision-making
3. Cultural shift toward evidence-based experimentation
4. Infrastructure for sharing and evaluating policy outcomes
5. Mechanisms for ensuring broad participation and representation

This could potentially address some of the speaker’s concerns about current systems while maintaining democratic principles in a more sophisticated form.

---

Prompt: Even open sourced projects have control mechanisms that either involve project owners or voting. How would they propose decision-making not involve voting when any changes would affect an entire underlying population?

This is a crucial question that gets at a fundamental tension in the speaker’s argument. While they criticize voting as too simplistic and advocate for network-based decision making, they don’t fully address how binding decisions would ultimately be made in such a system.

Looking at the transcript, there appears to be a gap between:

1. The theoretical argument that network systems can handle more complexity than hierarchical ones
2. The practical reality that collective decisions affecting everyone still need some mechanism of final approval/implementation

The speaker suggests that different groups could operate independently (“you do you and we’ll do us”), but this doesn’t address:

• Shared infrastructure decisions
• Resource allocation
• Policy changes affecting everyone
• Resolving conflicts between groups
• Setting common standards

Even in their example of biological systems differentiating into organs, there are still central coordinating mechanisms (nervous system, endocrine system) that help regulate the whole.

The closest they come to addressing this is suggesting that groups could “help each other” while maintaining autonomy, but this sidesteps how binding decisions would actually be made. Even in open source:

• Core maintainers ultimately decide what gets merged
• Foundations often provide governance
• Many projects use voting mechanisms for major decisions
• “Benevolent dictators for life” often have final say

This suggests that while network systems might improve on current decision-making in many ways (better information sharing, more distributed expertise, greater ability to handle complexity), some form of voting or hierarchical decision-making may still be necessary for final implementation of changes affecting everyone.

The speaker might argue that new forms of consensus-building or emergence could replace voting, but they don’t specify what these would be. This seems to be a key area where the theory needs more development to address practical governance needs.

---

Prompt: If you had to propose a solution to the gap, what might you suggest? Pick your top three solutions ranked by feasibility.

Let me propose three potential solutions to bridge the gap between network-based decision-making and the need for final implementation authority, ranked by feasibility:

1. Hybrid System with Domain-Specific Authority
   Most Feasible: Works within existing frameworks while adding network complexity

Key Features:

• Divide governance into specific domains (education, infrastructure, healthcare, etc.)
• Each domain has:
  • Open network participation for policy development and testing
  • Domain experts who evaluate technical feasibility
  • Affected stakeholder groups who must approve changes
  • Clear metrics for success/failure
• Final decisions require consensus from three groups:
  • Technical feasibility approval from domain experts
  • Stakeholder approval from affected groups
  • Traditional democratic approval for major changes
  • Smaller changes can be approved at domain level
  • Failed policies automatically sunset

Advantages:

• Builds on existing structures
• Clear accountability
• Balances expertise with democratic input
• Can be implemented gradually
• Allows for experimentation in less critical areas first

2. Algorithmic Consensus System
Moderately Feasible: Requires new technology but based on existing concepts

Key Features:

• Continuous rather than discrete voting
• Weighted input based on:
  • Proven expertise in specific domains
  • Track record of successful contributions
  • Direct stake in outcomes
• Machine learning to:
  • Identify emerging consensus
  • Flag potential negative impacts
  • Simulate policy outcomes
• Automatic policy adjustments based on real-time feedback
• Override mechanisms for emergencies

Advantages:

• More dynamic than traditional voting
• Can handle greater complexity
• Provides rapid feedback
• More objective than pure human decision-making
• Can integrate multiple data sources

3. Emergent Decision Markets
Least Feasible Now: Requires significant technological and social evolution

Key Features:

• Policy changes proposed as “contracts” with clear success metrics
• Anyone can contribute to policy development
• Implementation requires:
  • Proof of positive expected value
  • Insurance against negative outcomes
  • Stake from implementers
• Successful policies generate returns for contributors
• Failed policies trigger insurance payments to affected parties
• Natural selection of effective policies

Advantages:

• Self-regulating system
• Strong incentives for good outcomes
• Natural experimentation
• Clear accountability
• Rewards expertise and successful innovation

All three systems attempt to:

• Increase decision-making complexity beyond individual capacity
• Maintain accountability
• Allow for network-based contribution
• Provide clear implementation authority
• Balance expertise with stakeholder interests

The first solution seems most feasible because it:

1. Works within existing frameworks
2. Has clear lines of authority
3. Can be implemented incrementally
4. Maintains democratic legitimacy
5. Allows for learning and adjustment

This could serve as a transition step toward more sophisticated systems as technology and social structures evolve.