Adriana’s talk offered less in the way of practical solutions than I might have liked and less perhaps than the write-up promised, but this did not diminish a fascinating talk by someone clearly absorbed in her topic.
The following write up is based, in order of importance, on her notes which were quite thorough and provide the bulk of words, my own thoughts and recollection, and from watching the video in the process of editing it. I could have simply posted Adriana’s notes, but wanted to provide a flowing readable narrative.
Adriana is fascinated by Networks – especially a particular kind, called “distributed networks”, which are special. Distributed networks have no centre that controls them. This puts them at odds with a common and mistaken mental model of networks as a kind of bicycle wheel with each node reporting to a centre. This is a mental model that gets in the way.
Rather than offering practical help, Adriana set out to point her audience in an interesting direction. Perhaps a little like Newton did when he described gravity. First, she noted that the Internet – a successful distributed network – offers a hint about the truth of organisations and social structures. This hint is much more obvious in distributed networks.
Who is Adriana?
Adriana was (is?) as well paid consultant specialising in disrupting companies. It sounds odd that people would value being disrupted, but her role consisted of going in, finding the talent and forming a new structure to get something done outside of heavyweight processes and conventional structures. Adriana discovered that underneath all the “flat departments” and “theory Y” management corporations and businesses are ultimately hierarchical organisations, and very good ones. Processes and rules she said do not empower nodes (employees) but rather act as an efficient implementation of authority.
She set out to discover if there are alternatives to hierarchies.
What is a hierarchy?
There are many ways of defining a hierarchy. Examples include the traditional family, monarchy, democracy, church, army, corporation etc, which are complex but the fundamentals still apply authority.
The mathematical fundamental – derived from graph theory – is a network with no peer connections. Whatever the node is (person, computer, device) each node is connected only to a node above and below. A node above another, by definition, is superior and possesses authority to command the lower inferior nodes.
In a corporate environment, this may not directly match your experience. Things may indeed get done without direct recourse to a manager. Does this observation falsify Adriana’s theory? No. Adriana pointed out, quite reasonably, that processes (I imaged examples like “Agile”, or “Petty Cash”) replace the need to constantly seek permission each time from superior nodes. This delegation or proxy for authority also allows a hierarchical organisation to function day-to-day in a reasonably efficient way.
Relative to the individual
Another way of defining a hierarchy, Adriana observed, is that in a hierarchy the individual is always subjected to the system. It does this by offering a faustian bargain. We think “collectives have benefits, surely?!” but in fact hierarchies don’t empower individuals, they confer power on individuals within the system but not autonomy. Neither a king or CEO are autonomous, but are dependent on the system that grants them power. The system’s grant of power (providing legitimacy) and the utility of directing other nodes (providing labour) are together essential to the leader.
Adriana went as far as to venture that, because of the above, any organisation is a hierarchy, and she says this is something she would like to justify through further research. She then drew on direct experience to point out that revolutions merely lead to other types of hierarchy, with different individuals and different processes.
Are hierarchies “hardwired”?
It sometimes seems as if hierarchies are pervasive and natural, that the evolution of a an order – however spontaneous or flat or whatever – was ultimately a hierarchy. Adriana used the talk itself and the meetup as an example of a hierarchy that had spontaneously errupted. This leads to me to mention my only objection to her talk which is that all the petty hierarchies (the non-state hierarchies) are accountable to an ultimate over-arching non-hierarchical structure – the market – and so the spontaneous order of one person facilitating a regular meetup does not feel like a hierarchical enterprise and certainly not like a problem. Ultimately, it’s me and my customers.
The point of mentioning this was, it seems, to demonstrate that hierarchies may indeed be hard wired into us. Adriana was giving the status quo its due credit, suggesting that the spontaneous way they can emerge and their prevalence is evidence that they are unavoidable.
Adriana proposed that it is possible to tweak hierarchies and that not all hierarchies are equal.We can observe that some are more or less desirable “in some respects”. Democracy is better than slavery or monarchy, for example.
She also shared an example from her quantified self meetups, where she tries to avoid placing people in a theatre or auditoriam style seating layout (speaker at the front, with rows of seats facing the front). She says that you never get a natural discussion happening in that layout, but that you might if you arrange seating in a circle.
Another situation in which hierarchies fail to blossom is on the internet. It’s ability to route around obstacles gives it connectivity, it’s lack of centre gives it resilience. I don’t think Adriana mentioned it specifically but I often think of this as an open network of “ends”, where all ends have equal connectivity, anyway, I think most of us know what the Internet is and know that it is really very much not a borign stifling hierarchy.
The existence of these examples shows, Adriana argued, that there may be an alternative set of structural norms that can work non-hierarchically. If we can work out what’s common to these examples then you can present an alternative and this is what Adriana did. She calls it a heterarchy:
n. – a network where each node is or can be connected with another, without central coordination.
The Five Laws of Heterarchy
These are the factors that describe what was special about the Internet and the other examples of social phenomena that work without hierarchies.
Adriana’s notes are very thorough in her laws, though they are what they are which is a statement of current research which is very much “in progress”. I will quote them directly:
1. Collapse of Function and Balanced Asymmetry
Collapse of functions at the individual node level, which says that for a heterarchy to exist and persist, each node has to be able to perform certain functions. It means there are no hardwired distinctions of role and functionality among the nodes.
In terms of three technologies and one online phenomenon that I found share the same heterarchical characteristic. TCP/IP itself, BitTorrent, public key cryptography and the blogosphere.
TCP/IP turns any server into an originator, relaying party or recipient of a message. The relationship between nodes is not symmetrical but each server can at any time perform any of the roles.
BITTORRENT is another technology native to the network. It was born out of scarcity of bandwidth but it uses the distributed peer-to-peer nature of the internet to share files among nodes. In short, once a user starts downloading a file, he/she is automatically uploading too.
PUBLIC KEY cryptography was conceived for communicating in a distributed network. The traditional symmetrical keys approach couldn’t solve the problem of how to communicate with someone you trust in an untrustworthy environment.
For the purposes of this example, public key cryptography also exhibits ‘asymmetrical balance’ where the recipient of the message is part of encryption. So Alice uses Bob’s public key to encrypt her message for Bob, thus making him part of the process. Again, this was necessary in order to solve a problem that arose from new type of environment – i.e. distributed network, a heterarchy.
2. Freedom/Ability to bypass nodes
This law is not dissimilar to what John Gilmore said about Internet censorship: “The Net interprets censorship as damage and routes around it. It is the freedom and ability to bypass an obstacle.
A very important law, it ‘dissolves’ any attempts at hierarchy as it makes it impossible to lock-in the nodes and the flow of information or resources into a particular situation or formation without their cooperation.
3. Decentralised and Distributable Resources
It is the difference between having to go to the post office to make a phone call and the ability to make a call at home, difference between making a call from home on a landline and being able to make it anywhere using a mobile phone. Another example, having to go to specialist shop to have things printed or copied and being able to print it at home from your own printer.
Both previous examples were taken from communication or information distribution but when we start looking at 3D printers, things get really interesting. This law is important as having to use a centralised resource, apart from restricting nodes’ autonomy, opens up opportunities for control of that resource and imposition of a hierarchy.
… of resources or at least sufficiency of the most important resource that enables us to create and maintain the network. For internet its connectivity and bandwidth.
In other words, don’t try to build a heterarchy around a scarce resource as once that is controlled, heterarchy disintegrates. An example of an abundant resource would be information online, as digital format makes information [easy to duplicate] and therefore abundant.
5. Low marginal cost of communications
This is important because in a distributed peer-to-peer network it takes a lot more information exchanges to negotiate transactions. Think of trying to arrange a lunch meeting with a friend – agreeing on a day, then the venue, then timing, going back and forth because you are negotiating as peers, not top down… This of course, requires some information infrastructure with zero or low-cost communication.[After] 20+ years of the internet – very few network native technologies based on its distributed nature. TCP/IP itself, BitTorrent, public key cryptography. There are more, BitCoin, TOR etc but only in niches.
And we are imposing hierarchies even there. Think of content filtering, large platforms like Facebook etc that mediate our access to other nodes.
Let’s put them to work
With explicit understanding of heterarchy, there is potential to build systems and technology truly ‘native’ to the internet as a distributed network. Which would be more robust, powerful and also disruptive. Just think of how much (positive) impact TCP/IP, public key cryptography and BitTorrent have already had.
There is a potential to find better models for operating in networks and in virtual environments. And finally, there is potential to come up with alternatives to our hierarchical organisational defaults. That would be a great news for all those trapped in stifling and disempowering organisations.
Adriana dealth with a number of common objections to her theory about the superiority of heterarchy and to her implication that heterarchy is a better political order and organisational order. Again here notes here were thorough, so I will quote directly:
ORDER = HIERARCHY, LACK OF HIERARCHY = CHAOS
Hierarchy is the most common organisational model, people assume any order is by definition a hierarchy. Even if they recognise flaws of hierarchy they are equally worried about the lack of hierarchy leading to chaos.
OTHER TYPES OF ORDER, for example, ranking or POWER LAW.
Power law distributions tend to arise in social systems where many people express their preferences among many options. They became a popular concept during the rise of blogging when some blogs achieved ‘disproportionately’ greater popularity than others. Another example would be ranking twitter users by the number of followers or retweets or replies – there are people with more of each but that wouldn’t make such order a hierarchy (remember definition of hierarchy)
The important difference here is that while the popularity of blogs, for example, might follow a steep power law distribution, that popularity only offers an incentive to deal with that blog. Popularity does not confer power over less well-trafficked blogs, or prevent their autonomy.
INDIVIDUAL AUTONOMY = SELFISHNESS/ATOMISM
“if you give individuals independence and autonomy, wouldn’t it all disintegrate into selfishness or even atomism?” On the contrary, hierarchies are far better configured for selfish people, they allow them to gain control over others, manipulate them for their purposes. And as for atomism, aren’t we supposed to be social animals?
SOFT EXPLANATIONS FOR INEVITABILITY OF HIERARCHY
Adriana had heared it claimed that “people need security and certainties that hierarchies provide” but argued that other people do not need such security and argued that it is equally valid to optimise for those people instead.
Adriana also dealt with the objection that societies and institutions have evolved as hierarchies for social, political reasons. This is arbitrary and there is no reason we should not continue to evolve them in new ways.
The last objection implicitly argues that we are hardwired and for a good reason. The claim was “human evolution favoured hierarchies because they optimised survival and efficient distribution of resources”. But the economy we enjoy today has moved beyond a state of scarcity and now there is an abundance of most day to day resources for most Western people. The need for this optimisation is gone.
The very fact that there are structural reasons for hierarchy and stories to be told about how to encourage heterarchical behaviours and examples of vibrant heterarchical systems is encouraging. Explicit knowledge of some of the underlying causes of that success can lead to efforts to build organisations and systems that exploit and encourage heterarchy.
…this means hierarchies may not be inevitable and we don’t have to resign ourselves to hardwiring them in our minds