Stigmergy and Collective IQ
It is so often the case that powerful discoveries crop up ahead of their time, only to be buried again to wait for the minimum necessary connectome of related ideas to form around them. This repeating pattern, I believe, justifies the present state of stigmergic systems research. Before we can concern ourselves with what that means, we must first share a definition of what a stigmergic system is.
Evidence for Stigmergy in Natural Systems
The primary motivating example for stigmergy must, unfortunately, come from a seemingly unremarkable creature. Perhaps researchers would have seen more interest and funding if they had a more alluring mascot than the humble termite. As researchers we have to subsume our natural preferences for elegance and beauty at times in order to uncover new phenomena. In the study of termites, the evidence points to a valuable secret underlying their phenomenal success. How else could a nest-building wood-eating insect carve such success from so many environments around the globe that it outweighs the human biomass by three to one?
Beyond termites, stigmergic patterns are central to coordination among other organisms, including ants and bees, and perhaps even simpler organisms including fungus and slime molds also employ some primitive coordination principles that will one day fit into the domain of study. Many of the most successful grouping organisms rely on stigmergy to process and decide on information it can acquire from its environment with the goal of more efficiently allocating its own resources — including time, agent effort, attention and energy. In the canonical example termites site and architect their cavernous homes via a completely autonomous process that begins when a termite believes they have found a good site for a new mound. This critter collects a mudball and laces it with its unique pheromones and places it as a sort of vote at the site. These mudballs are placed randomly in the termite colony’s environment at first, but the presence of pheromones increases the allure of that location — compelling other colonists to consider placing more mudballs atop the first, and if consensus forms in this manner then these mud balls will eventually form the mounds, arches, tunnels and chambers in an unfolding collaborative architectural project requiring no leader or prearranged plans.
Ant colony optimization using pheromone trails to organize workers is another noteworthy example illustrating much the same point — with a few instinctual rules complex patterns that consistently beat random can be embedded in social behaviors. Ants are another extremely successful social organism, and despite its tiny individual brain has shown such amazing coordination that it can conduct organized warfare, engage in insect husbandry and generally forage its environment with an efficiency that makes our most successful management and industrial systems theories look quaint and ineffective. A central theme across stigmergic systems is ditching a leader or central planner results in effortless coordination that dramatically outperforms anything a single agent within the stigmergic system could conceive of. This is the draw for studying these simple creatures — to uncover a science for human governance and collaboration in the absence of domination or needlessly over-complicated rules.
The world of today looks very different than even the world of 20 years ago, the Internet has supercharged everything regarding the speed and throughput of human communication in an unprecedented paradigm shift of possibilities. So where is the complementary explosion in the way humans can integrate knowledge, make sense of our world and our place in it? I believe it is emerging now, and with conscious effort coordination systems that outperform anything the ancient Greeks, Marx or Hayek could have ever imagined will be within the realm of possibility.
Principles of Natural Stigmergy
The origin of the term “stigmergy” was afforded to termite behavior by a French Biologist. Pierre-Paul Grasse used the Greek *stigma* meaning “mark” or “sign” to capture the “Stimulation of workers by the performance they have achieved”. Stigmergy exists as a pattern language within the broader domain of self-organizing systems, where rigorous and complete first-principles are not available. This means we must rely on description and themes, which in our case yield a rich but incomplete fabric of possibilities.
The original use of the termite to motivate this principle does not do full justice to how beautiful it can be in its own right. Since Grasse named it, the phenomena has been identified in many places beyond worker insects. Perhaps the most beautiful example, and the one that inspired the naming of this venture is the fantastic flocking behavior of a starling swarm. Their mesmerizing murmurations have been reproduced by Craig Reynold’s Boid simulation using only three core steering behaviors: Separation, Alignment and Cohesion. The advantage of this high degree of coordinated flocking cannot be understated, compared to their non-swarming peers starlings are more efficient in transit, have higher survival rates and can even maneuver to confuse a swooping Hawk into diving through the entire flock without catching a single starling.
Stigmergic Collaboration
“The detour into insect and bird societies is nice, but what does this have to do with Blockchain?”, you might ask. To understand how ant strategies apply to wikis and how termite nesting relates to decentralized finance we have to make concrete definitions. For that, we can use the work of Dr. Mark Elliot, whose contributions to the art center around a formalization:
Collaboration is dependent upon communication, and communication is a network phenomenon.
Collaboration is inherently composed of two primary components, without either of which collaboration cannot take place: social negotiation and creative output.
Collaboration in small groups (roughly 2–25) relies upon social negotiation to evolve and guide its process and creative output.
Collaboration in large groups (roughly 25-n) is enabled by stigmergy.
The purpose of this formalization is to enable us rigorously speak about the phenomena of stigmery in communication networks. It is an important distinction to make that every network that fulfills these criteria has some expression of the phenomena of stigmery, the degree to which it is intentional and its effect on the network are variables which I believe can be tuned to some degree once the nature of the network is roughly constrained.
A Stigmergic Wiki vs Co-Authoring
To properly understand the distinction between collaboration and stigmergy, consider a wiki co-created by many parties. The research suggests that without any help from technology, 2–8 persons is the ideal collaboration size and the upper limit is around 25. This prompted Elliot to break the formalism at 25, with any group larger than that enabled by some form of stigmery. Wikipedia’s scale makes it impossible to rely on social negotiation to progress edits. Wikipedia, being a distributed but still centralized service relies on a CRDT to reconcile parallel edits , this technological solution functions to enable co-authoring, and may technically increase the stigmergy of Wikipedia, but if we take stigmergy directly as a design concern we come up with a different solution.
Following along with Elliot, if we want to redesign Wikipedia we must address locations where social negotiation is relied upon, with the intention of replacing that with a method of coordination that does not require participants to converse in order to collaborate. This could, for example, remove the co-authoring need for a human reviewer and approval period in every case. The need for discussion in co-authoring needs to be balanced against the cost of discussion and how overall disruptive it is for asynchronous collaboration. This can be seen in their own editing policy . The Wikipedia of today still in some ways expects editors to work independently and compete to have their entire edit accepted or rejected, suggesting that a better pattern of peer to peer knowledge generation might be attainable with technology improvements alone. In conclusion:
Networks that allow such user groups to self-organise, known as ’group forming networks’, have been identified in research as being one of the more powerful drivers of network value which may have contributed significantly to the growth of giants such as Ebay, Wikipedia.org, the Open Source software movement and even the Internet itself.
Starling Foundries’ Stigmergic Charter
Starling Foundries’ tagline is:
Realizing the power of hyperconnectivity to reshape collaboration.
This won’t be accomplished all at once, but there is a distinctly under-served branch of cyber-social systems called collective intelligence to draw under-valued insights from, and a new industry of DLT solutions popping up with some similar ideas. If we are ever going to intentionally design and adapt our established patterns of coordination to this new hyper-connected world where a tweet can be read a billion times in the first minute of its existence, we have to have a history of prior experiments in order to draw any meaningful conclusions.
It is for this reason that Starling Foundries was founded as an educational and experimentation venture devoted to the continuation of this vital work and the exploration of its applicability to Distributed Ledger Technologies, and Decentralized Autonomous Organizations. Successful outcomes may include a robust methodology for studying and cultivating stigmergic DAOs. Stay tuned!
