Organizations and their Effectiveness (2019)/autocatalysis

From CommunityData

Assumption 1: Complex nodes require either complex inputs, complex transformations, or a conveniently-arranged environment to be reproduced in self-sustaining cycles.

e.g., nodeComplexity ~ f(inputComplexity, cycleComplexity, environmentStructure)
Where f is increasing in each of the three variables and has some degree of substitutability between them

Assumption 2: Simple cycles are more robust (i.e., survivable) than complex cycles

Theorem 1a: Complex nodes are more likely to be produced in low-entropy environments

Theorem 1b: Complex nodes are more likely to be produced in environments rich in complex inputs

Assumption 3: Environmental entropy can only be [is most easily?] lowered by autocatalytic cycles

Theorem 2: Complex nodes are more likely to be produced in cycles that are near* cycles which lower the entropy of the environment (Stigmergy)

Theorem 3: Complex nodes are more likely to be produced in cycles that are near* cycles which produce complex nodes, giving them access to complex inputs (multiple networks)

Assumption 4: A cycle interacting with another cycle is more likely to change than one interacting with a simple environment.

Theorem 4: Therefore, the very conditions that support the survival of complex-node-producing cycles (e.g., stigmergy links or I/O links to other cycles) increase the likelihood of the cycle changing. Also, we should look to these links to identify the source of cycle change.

Jon Comments: A1: What do you mean by complex nodes? That they exist in multiple network structures? And inputs?

A2: True, if simple means the length of the cycle and/or the number of subcycles. And if survival means lasts longer before transitioning to a different, simpler topology.

T1a: Hm, interpreting this with the above definitions is hard.

T1B: Ditto

A3: To me, autocatalysis is interesting when starting with low entropy. Higher entropy implies the self-organization process has already happened. So all cycles should start with low entropy and then we see where it goes from there.

Other As and Ts: prolly better after we talk some more.

In general, I think the 4th (i think) chapter in the book, the multiple network models, is far too speculative and has a tendency to put the horse before the cart in that it assumes the order we wish to explain.


* "near" is imprecise, and Jon had a critique of its meaning]