1 month, 24 days ago

Cohesion Spectrum

Link: https://www.strategicstructures.com/?p=2775

This essay is part of the Autonomy and Cohesion series.

We can’t deal with most things in life on our own. We don’t have a large enough repertoire of responses to all the stimuli of the environment. We get together and form tribes, communities, companies, networks, states, or, in other words, wholes. And the whole is greater than the sum of its parts, as the saying goes.

This aphorism is often attributed to Aristotle, while, more likely, it came in this form much later, possibly through Gestalt psychology. Whatever the source, it seems popular and further amplified by those trying to explain concepts such as system and emergence.

While widespread and evocative, the saying doesn’t hold up under scrutiny. It yields at greater and parts. When it comes to parts, for social systems, it’s not whole and parts but rather a whole of other wholes. System of systems. And systems formation does not work with a whole-part scheme.1

Regarding greater, the “whole is neither more nor is it less than the sum of its parts: it is different,” wrote Heinz von Foerster in 1976.2 Later, in an interview, he offered another clarification by simply modifying the popular saying and giving an example:

I would add the following correction to this principle. You require an additional measurement function for this: The measure of the sum of the parts is greater than the sum of the measures of the parts. One is the measure of the sum; the other is the sum of the measures. Take, for example, the measurement function “to square,” which makes this immediately apparent. I have two parts, one is a, the other b. Now I have the measure of the sum of the parts. What does that look like? a + b as the sum of the parts squared, gives us a2+2ab+b2. Now I need the sum of the measures of the parts, and with this I have the measure of a(=a2) and the measure of b(=b2): a2+b2. Now I claim that the measure of the sums of the parts is greater than the sum of the measures of the parts and state that: a2+b2+2ab is greater than a2+b2.

With “the measure of the sum of the parts is greater than the sum of the measures of the parts,” the aphorism makes more sense, and when the whole in question is a social system, variety is a good measure. The whole has (potentially) more variety than the sum of the varieties of the agents making up the whole.3

Agents can get more protection, respond more successfully to external stimuli, and achieve more ambitious goals. That’s, of course, the law of requisite variety.

Forming a whole brings one of the biggest problems every society and generation has faced since there have been people on the planet: how to balance the needs of individuals and groups. But let’s keep this moral box closed for now and focus on a related question: how does the system maintain identity, function, and viability?

The whole, be it a tribe, organization, network, or state, stays a whole as long as it has cohesion. Cohesion is brought about by natural forces and artificial tools and technologies, as elaborated in another essay from this series.

The way cohesion is maintained varies. For some social systems, it is through command and control. On the other extreme are systems where coordination mechanisms are technical standards and protocols, and autonomy is only reduced by participatory interoperability constraints. Between these extremes, there are different zones corresponding to idealized versions of actual social systems, forming a cohesion spectrum.

The CABIN model

The CABIN model proposed here distinguishes five zones in the cohesion spectrum.

Coercive (Authoritarian)

  • Examples: Traditional command-and-control military or corporate structures with rigid hierarchies.
  • Dominant cohesion mechanisms: Top-down directives, strict rules, and punishments for non-compliance.
  • Autonomy level: Extremely Low

Bureaucratic (Directive)

  • Example: Bureaucratic organizations with defined procedures and regulations.
  • Dominant cohesion mechanisms: Formal processes, guidelines, oversight and performance reviews.
  • Autonomy level: Low

Normative (Guided)

  • Examples: Housing associations, cooperatives, and tightly-knit communities with strong social norms.
  • Dominant cohesion mechanisms: Collective agreements, social pressure, expectations of conformity, and reputation systems.
  • Autonomy Level: Moderate

Adaptive (Responsive)

  • Example: Marketplaces
  • Dominant cohesion mechanisms: Feedback systems, ratings, and market forces adjusting the behavior of the participants.
  • Autonomy Level: Moderately High

Interoperable (Coordinated):

  • Examples: The web, decentralized social networks, and open-source projects.
  • Dominant cohesion mechanisms: Technical standards and protocols.
  • Autonomy Level: High

The acronym is an anagram of CABIN, hence the name.

Each zone in the spectrum is characterized by dominant cohesion mechanisms. These mechanisms are only dominant but not sufficient for the overall system cohesion. For example, there are feedback loops and protocols in all zones. Furthermore, no system is in a vacuum. Markets are in the Adaptive zone but can’t work without currencies maintained by institutions in the Bureaucratic zone.

The movements in the CABIN spectrum have their own concepts and narratives. When the starting point is in the Coercion zone, the movement toward a better balance between cohesion and autonomy is led by talks about delegating, subsidiarity, and increased autonomy. When autonomous agents need to work together, the movement towards increased cohesion is led by talk about interoperability.

At first glance, the CABIN may look like a spectrum from highly centralized to highly decentralized systems. That’s not always the case. A system in the Normative zone of the spectrum may be more decentralized than a federated social network and yet keep a lower level of autonomy. For example, in most housing associations, individual choices are limited by maintaining uniformity and adhering to certain aesthetic conventions. There are other cohesion mechanisms experienced as community pressure that limit autonomy. The pressure could be created by explicit conformance expectations, as it is in reputation-based systems, or implicit expectations for adopting specific language and references in order to participate, as it is in some meme culture communities.

The spectrum also invites judgments. The movement from command-and-control to standards-and-protocols-based cohesion can be seen as a movement from bad to good. However, it depends on the system type, the situation, and the maturity of the cohesion tools. A few words on each.

Hierarchical command-and-control structures are hardly a good way to manage a contemporary company, yet they still seem appropriate for an army or a ship. Even the biggest agile zealots would not board a ship managed by an agile crew. That also doesn’t mean the current practice is the best it could be. Staying with the ship example, pirates elected their captains and quartermasters, the latter also serving as arbiters during disputes between the crew and captain, and all that was agreed in the articles signed by every crew member. These were very progressive social experiments back in the 17th and 18th centuries.

It could also be the same type of system but in a different situation. The pandemic showed that the majority of people were willing to forgo a lot of their autonomy to get more protection.

Last, there is the maturity of the tools. Since the common market cohesion mechanisms are now being replaced by protocols in the case of crypto-currencies, new protocols could replace some of the cohesion mechanisms in systems currently belonging to other zones of the CABIN spectrum.

First published on Link&Think.

1    When it comes to social systems, system formation is a matter of system/environment differentiation, which also applies to subsystems. It “does not involve the decomposition of a “whole” into “parts,” in either the conceptual sense (divisio) or the sense of actual division (partition).” See Luhmann, N. (2013). Theory of Society, Volume 2 (R. Barrett, Trans.). Stanford University Press. https://www.sup.org/books/title/?id=16878
2     von Foerster, H. (2003). Objects: Tokens for (Eigen-)Behaviors. In Understanding Understanding: Essays on Cybernetics and Cognition (pp. 261–271). Springer. https://doi.org/10.1007/0-387-21722-3_11 (Original work published 1976)
3     This follows the mainstream notion that people are part of the social system. We can entertain another possibility, defended by Luhmann, that people, while structurally coupled with the social systems, are part of their environment, while the social systems emerge from closed networks of communications. The LORV still applies, but this time, it is due to the variety of communications compared with the variety of the system and the irritations it has to respond to.