In the context of my writing concerning division of power, I want to make a distinction between personal power and collective power.
That is not the same as the distinction between absolute power and limited power. Absolute power can be collective, for example if a state is under the control of a committee, and limited power can be personal, if an individual has control over a particular department or aspect of policy.
There is a continuum of collective power, depending on the amount of personal influence. At one extreme there is a situation where a group of two or three people who know each other can make decisions by discussion; at the other is the ordinary voter, whose opinion is aggregated with those of millions of strangers.
Towards the latter extreme, collective power is no power at all. A collective does not reach decisions the same way an individual does. An individual can change his mind, but that has small chance of altering the action of the collective. To change the action of a collective, some more significant force than an individual impulse normally has to act on it. That’s why, when we attempt to predict the action of a collective, we do not talk about states of mind, we talk about outside forces: media, economics, events.
In many cases, we can predict the action of the collective with virtual certainty. The current US presidential election is finely balanced, but we can be sure Gary Johnson will not win.
This feature of collective power has implications for the consideration of divided power, because in the right circumstances a collective power can be completely neutralised. An absolute ruler is not omnipotent, in that he depends on the cooperation of many others, most importantly his underlings and armed forces. But as a rule they do not have personal power; they have collective power. Any one of them can be replaced. An individual can turn against the sovereign, but if he would just be dismissed (or killed) and replaced, that is not a realistic power. If too many of them do not act as the sovereign orders, he would be helpless, but that requires a collective decision, and one which with a bit of work can be made effectively impossible.
There are exceptions to this. If the sovereign is utterly dependent on a single particular individual, that individual has personal power. There have been historical cases of sovereigns in that position, and it is observed that that constitutes a serious qualitative change in the nature of the government.
Where a person can covertly act against the sovereign’s power, that is a personal power. Competent institutional design is largely a matter of making sure that rogue individuals cannot exercise power undetected by anyone. As long as there are any others who can detect this abuse, then the power once again becomes collective power, held by the individual and those placed to stop him. Again, where collectives do act in this way, it is a sign of a breakdown of government institutions. As an example, see this article describing the upper ranks of the army working together to deceive the president. If the president had absolute power and a moderate amount of sense, this sort of conspiracy would be suicidally dangerous. Once power is formally divided, then the capability to prevent this kind of ad-hoc assumption of power is massively eroded.
That is the fundamental reason why division of power is bad: whatever division of power is formally made, these gaps for further informal division will tend to be opened up by it, because limited power denies the power to enforce necessary limits on others. If anyone has power to punish those who take powers they are not formally entitled to, then that person effectively is absolute. If nobody has that power to punish, then any ambitious crooks can run wild.
If there is no single person other than the sovereign who has personal power, then I would call the sovereign absolute. His power is not infinite: he has to maintain control over the collectives which necessarily have power, but that is a lesser constraint than having to cope with personal power held in other hands. It is more akin to the other constraints on his power imposed by such things as the laws of physics and the existence of foreigners and wild animals.
Note that the nature of feudalism is that feudal aristocrats are not replaceable, and do have personal power—limited, but not collective. Feudalism is thus not a system of absolute power even under my refined definition.
The great significance of collective power is that it is subject to coordination problems. Or, since from the point of view of the sovereign, the problems of coordinating a collective can be an advantage, I will call them coordination obstacles. That is why it is not voters who have power, it is those who mediate the coordination of the voters: parties and media. A change in the way that voters can be coordinated is a thoroughly material change in what I have called the Structure of the state. The US does not have the Structure that it had 25 years ago, because (among other reasons) social media is part of the current Structure. That is an actual revolution, and why the fights over use of social media for political coordination are so significant. Note that since the Constitution doesn’t say anything about social media, the constitution in itself obviously does not define the Structure.
It also means that for a formally absolute ruler, obstructing collectives from coordinating is an important tool. In the period of formally absolute monarchy, any attempt by people of importance to coordinate in confidence was suspect: prima facie treason. The most basic right claimed by parliaments was the right to meet: simply allowing aristocrats and city leaders to meet together and discuss their interests was giving them a power that they wouldn’t otherwise have.
This is the problem with the formalism that Urielo advocates: formally establishing any power that anyone in a given Structure happens to have. Power that is held collectively and is not legitimate is often neutralised by coordination obstacles. If you make that power legitimate, that goes some way to dissolving the coordination obstacles, and thereby increases the effective collective power.
Modern political thought does not generally respect the idea that coordination by those with informal power is not legitimate (though we retain the historical unfavourable associations of the word “conspiracy”) but it went without saying for most of history. Organisations that have existed in England for hundreds of years, such as guilds and the older schools and colleges, generally have royal charters: the charter is their permission to exist.
There are a couple of interesting exceptions to the modern toleration of conspiracy: one is anti-trust law, and another is insider trading law. Those both deal with economic activities.
They do show, however, that legal obstacles to coordination are not obsoleted by technological effects. Indeed, modern communication doesn’t mean that coordination obstacles are easily overcome, especially if the obstacles are considered legitimate. No matter what messaging options are available, if you need to identify yourself for the communication to be useful, and you cannot trust the other party not to expose your attempt to conspire, then attempting to conspire is dangerous.
Here is another example: in investment banks, it is generally not permitted for employees to coordinate on pay. It is a disciplinary offence to tell anyone how much you are paid. This is taken seriously, and is, in my experience, effective. That is an example of an obstacle to coordination imposed as part of a power structure.
Legal obstacles to treasonous coordination were removed for ideological reasons, because division of power and competition for power were considered legitimate. Effectively, “freedom of association” was one more way to undermine the ancien régime and unleash the mob. As with the other historical destabilising demands of progressives, things are starting to change now that the progressives have taken permanent control of the central power structures.
You no longer need a Royal Charter for your golf club or trade association, but that doesn’t mean you are free to coordinate: if you don’t have sufficient female or minority members, you may need to account for yourself in the modern Star Chamber. The Mannerbund is the same kind of threat to today’s status quo as a trade union was to that of 1799.
The useful point is that it is not proved that you can run a stable society with complete freedom of association. That makes it more acceptable for me to recommend my form of absolutism, where people other than the sovereign inevitably have the capability to act against his policy by acting collectively, but such collective action is both illegitimate and made difficult by deliberate obstacles put in their way.
Update: just come across this 2004 piece from Nick Szabo, where he talks about dividing power to produce “the strategy of required conspiracy, since abusing the power requires two or more of the separated entities to collude”. However, as I see it doing that is only half the job: the other half is actually preventing the separated entities from colluding.