The Powers that Be in a Digital Society
Nation states like to paint themselves as "eternal". There is nothing farther from truth. Most of the modern sovereign states are actually younger than the elders living in them. Decolonization following WWII has a lot to do with this. But even for older states the story is not smooth. Take France for instance, if historically an accepted date of foundation could be 481 AD (Clovis baptism), the current instantiation of France (as a political entity) only dates back 1958 (fifth republic). Just between the French Revolution and today, no less than 13 different constitutions were adopted: a little bit more than 1 by generation. Of course there are exceptions, like the USA that haven't modified their constitution since 1776, but they are the exception, not the rule.
One problem is that it is not immediately clear how to define properly what is a sovereign state. There are many angles to consider. Pragmatically a nation state is a political entity that is able to project power on a given piece of land. Almost every squared inch of the surface of the globe is claimed by such an institution. The interesting case of the Bir Tawil triangle between Egypt and Sudan -both countries *reject* the piece of land- is almost the only exception.
In our current version of sovereign state there are three main ways to express power. Those are the ones formalized by Montesquieu in "L'esprit des Lois": executive, judiciary and legislative branch of power. The structure of the political institutions revolve around how those powers are managed. This is the aim of the most fundamental texts: constitutions.
The recent COVID crisis has revealed that the theory and the practice have diverged a lot. The example of vaccine passports is striking: without the right QR-code (aka proof of vaccination) it became impossible to live a normal life. You couldn't travel freely, go to restaurant or chess tournaments. When you think of it through the lenses of political theory it was an alien form of power at play. Ordinary citizen instantly became enforcers (you could check a QR-code at the entry of the chess competition) and enforced (you had to display your authorization at every street corner). Things were not really examined in parliaments (or retro actively) and could not have been implemented without the active cooperation of multiple private entities: from Internet Service Providers to app stores passing by restaurant owners who were supposed to enforce the regulation for the state. Clearly Montesquieu's user manual on powers needs an update for digital societies.
Where and what are the powers ?
When Elon Musk decided to cut the access to Starlink to the Ukrainian Army - specifically to deter an offensive action - it became clear that how we traditionally understood power was no longer accurate. A sovereign nation, moreover during war time, was subjected to the will of a high tech company. And it was not an indirect thing. It was direct: a guy in an office of Starlink just turned the access down and that was it. This example is illustrative at multiple levels:
Vital functions of the state are subjected to technological conditions that are not under its control.
The environment is more and more digital: either it works perfectly or not at all. There is no real notion of "degraded mode". Things don't "more or less work". They work or they don't.
Borders are not limits that make sense from a technological point of view. They are limits to the sovereignty of states though - by definition.
Linked to 1. is the fact that some things are impossible, mainly for legal reasons, for states, but are very easy for a company. The best example is how three letters agencies and former Twitter, now X.com, have put in place a pretty sophisticated system of censorship. It presumably happened also with the other large social media platform but for X.com we know that it was the case. This partnership is assuring credible deniability on one hand and is creating deflecting excuses on the other hand: no we don't spy on you (three letters agencies) / yes we do spy on you but we were just following orders (big tech)...
So the powers that be are not “pure” as they used to be: orders were executed in the state apparatus following a clear hierarchical structure. Now the powers that be emerge from the collaboration of different entities. It is not true that the three letters agencies directly give an order: they rather suggest or give hints on what to do but the relation is less asymmetric. Big corp provides new services and are granted special rights (or rather no one is going to bother them on some points in exchange of the collaboration). In this sense they are distributed systems of power. Distributed systems work with protocols : those are rules that have to be observed for the system to work correctly. The meddling with the 2020 election is clear: big corp can kill a story (Hunter Biden laptop) because it was suggested to them that it was a hoax. It had a direct impact on the campaign. Were people sincerely believing that the story that it was a hoax was true ? Actually it is not relevant. They acted as if they believed it was true and it was enough. This example shows that the political institutions have to negotiate with big. corp. They can’t just give orders … or they might lose the next elections. It bears some resemblance with negotiations between sovereign states. At the end of the day when you negotiate with Putin or the Iranian regime you can’t give orders. You have to be able to produce credible threats and credible positive incentives. One direct implication is that the notion of “rule of law” is no longer relevant. Because there is no longer rules to be followed. There are links to what I call “the version of the law”:
Things become dynamic and evanescent. This raises serious concerns on to how to limit those powers. It is one thing to put check and balances to powers when they are clearly identified, it is another thing when facing distributed powers that be. For instance: when a distributed system is properly crafted, then eliminating few nodes or links won’t break it nor put any meaningful limits to the system. One of the raison d’être of distributed systems is precisely to avoid single points of failure. It means that controlling the reach of a distributed system is a complex thing. More often than not it has to be done dynamically otherwise the network will adapt and circumvent stationary obstacles. In political terms it means that a fixed constitution is not going to be able to do the job of limiting powers, no matter how smart are the people writing the document.
Meta protocols
A constitution is a meta law. It is used to describe the framework within which the laws can be thought. The idea of law works fine in a written society. I am not so sure it works in a digital society because, as I have just discussed, the powers are no longer in the written words. They are in processes, and even more importantly in the distributed systems build by the interactions of these processes. There is an interesting parallel to be drawn here: there is a jump from programs (executed on a precise machine) to protocols (an agreement between different actors to do things following a common script). No one forces you to follow a protocol, on the other hand you have to follow the user’s manual of the machine if you want your program to work. By nature a program is executed in a controlled environment. A protocol is not like that: you can follow it or not. Generally protocols are respected because they bring benefits. If you incorrectly set up your http or mail server, then you are not going to have interesting results, that is all.
The question is: what would be an equivalent of a constitution in such a context ? I have not yet a clear idea of what an answer would look like. It is easier to see what it can’t be. My vague idea is that the only hope is to have “meta protocols” - for a lack of better word. Of course there are static things to check that come to mind: only allow protocols that are fair, in which there is symmetrical knowledge between parts, etc. The hard part is to imagine dynamic meta protocols that adapt to the different ways protocols might adopt to circumvent static interdiction. The question of how to enforce (or design incentives to make them sufficiently appealing) them is another level of difficulty. In my opinion there will be no realistic version of network state before we have at least some ideas on the matter.