13 September 2016


No matter how big you grow, you are still vulnerable to a single accident. This includes a single self-inflicted accident.

For robustness, growing is helpful but not sufficient. You need to reproduce.

However, reproduction is not merely making copies. That is barely different from growth.  Again, redundant structures and information help, but they're not sufficient.

To survive longer periods and greater risks, you need to duplicate and separate.

The bigger you are, the further you have to separate.

Your "size" is not your mass, it is the space you occupy. If you are frequently highly mobile, that is like being large, and means you have to separate further.

There are two ways to separate: either you use a different mechanism of movement for separation than for all other purposes, (like a plant seed blowing on the wind), or you make a sustained determined effort to escape, to run far away from all of your kind, with a high speed and consistency of direction that you do not use for any purpose other than separation.

11 September 2016


At last I have set the necessary prerequisites to discuss Urielo / @cyborg_nomade's discussion of constitutions.

It is possible I could have been more concise about the prerequisites: what it really amounts to is:

  • Division of power is dangerous and to be avoided
  • It's better to have less division than more
  • Sometimes that isn't possible

Within the context set by those propositions, the difficult parts of "neocameralism and constitutions", as well as Land's "A Republic, If You Can Keep It", start to appear at least relevant. So too the considerations of control and property in Land's "Quibbles with Moldbug".

Let's say that in some given situation, it is impossible to effectively unify power. The next best thing is to nearly unify power. Some small number of people have some small amounts of power, but the main power-holder can set rules about how they are allowed to use that power, and threaten to crush them like a bug if they break them. That's workable too, provided the mechanisms of supervision and bug-crushing are adequate.

However, that's not always the case. Sometimes, power is too divided, and crushing like a bug isn't on the table. That's when the hard bit starts.

What you need to do is find a pattern of division of power that is stable, and compatible with effective government. The second implies the first: if the pattern of division of power is unstable, then those in power will be incentivised to protect and expand their power, rather than to govern effectively.

Part of setting up this stable pattern might be to write a lot of rules on a long sheet of paper. I can't see, though, how you could ever start with the paper and get to the actual division of power.

"Actual division of power" is such a mouthful. The word I wanted to use for this is "constitution", but I suppose I will have to give in and call it something else. (I had this idea that the original sense of "constitution" meant  what I mean, and the idea of a constitution as a higher set of laws was derived from that. But it seems my idea was completely wrong). Let's just call it the "Structure".

So how should one design a Structure? You have to start from where you are. If at t=0 one power is effectively unchallenged, then they should just keep it that way. You don't need a Structure.

Urielo really hits the nail on the head here:

eventually, a constitution always arise out of a multiplayer game, because conflict eventually ends with an agreement - @cyborg_nomade

A non-autocratic Structure is the the result of a peace settlement between potential or actual rivals, and a Constitution represents the terms of that peace settlement.

The aims of the settlement should be that it will last, that those who came into the settlement with power are willing to accept it, and will be incentivised to maintain it into the future and to preserve those things that incentivise the others to maintain it into the future.

The simplest peace settlements consist of a line on a map. What happens on one side is the responsibility of one party, and on the other is the responsiblity of another. The two (or more) sides invest appropriately in either defensive or retaliatory weaponry, to provide incentive to each other to keep to the agreement.

This is not normally what we think of as a Structure within a society, though it is an option. https://en.wikipedia.org/wiki/Partition_(politics) . If the powers of the participants cannot be easily separated by a line on a map, a more detailed agreement is necessary.

Another of Urielo's tweets:
pretty much all working societies recognized some sort of power division. the estates of the realm being the European version - @cyborg_nomade

I've written before about the vital elements of feudalism as I see them: It resembles somewhat the "line on the map" kind of settlement: each feudal vassal had practical authority over a defined region, subject to certain duties he owed to his Lord. The Lord would spend his time travelling between his vassals, resolving disputes between them, collecting his share of the loot, and checking that they weren't betraying him.

This worked practically, most of the time. As I wrote before, the crucial fact that necessitated a settlement between the King and his vassals was that he wasn't physically able to administer the whole kingdom, because of limitations of communication and transport. Whoever he sent to run them, would in fact have considerable autonomy (whether the constitution gave it to them or not), and so the Structure had to accommodate that fact.

I say it worked most of the time, but it didn't work all the time, or even nearly all the time. Conflict between King and nobles was pretty common.

If we're talking estates of the realm, of course, then there's more than nobles. The Medieval English Structure basically treated the church as a sort of noble. Bishops and Abbots had similar rights to Barons, but fewer duties. (That meant it would be a problem if their power increased relative to nobles.) The other group to be recognised with power within the Structure were the small landholders. At a guess, I'd put their claim to power as follows:

Fighting enemies was the responsibility of the King, and in the King's interest. His vassals were required to supply men and/or funds to him to do this. The actual fighting would be done by Knights and men known to and under the direct control of Knights. It was therefore in the King's interest that the Knights be incentivised to fight effectively, and would see honour and/or profit in doing so. However, to the Lords the Knights were just farmers and taxpayers; it was not in the Lord's interest to have his Knights flourishing and strong. Therefore, the King had an interest in defending the status of Knights against their Lords.

That's kind of a just-so story; I'm open to disagreement on specifics. In any case, this Medieval English Structure obviously depends on an agricultural economy, and military technology that relies on a relatively small number of expensively-equipped, skilled soldiers. It's not coming back.

The commoners and serfs basically have no power recognised by the Structure. That's probably an oversimplification, at least after the Black Death when their economic power became more significant (and serfdom faded out). But in any case, the point of the Structure is not some abstract fairness, it's stability and efficiency.

The Structure was quite flexible and changed significantly over time. Burghers were accepted into it once trade became economically significant enough for their power to need to be preserved. But even there the simple fix was geographic: towns were made Boroughs, lines were drawn around them on the map, and the Burghers were allowed to run the towns, with a limited and transparent set of rights and duties with regard to goings-on outside the borough.

The King, Nobles and Knights form a triangle: that's popularly considered to be stable, for the reason that if any one of the three starts to get too strong (or weak), the other two can see it and attempt to correct it with superiority. With two or more than three large power centres, it's too easy for a theoretically weaker coalition to unexpectedly show itself strong enough to reconfigure the Structure. That's a guideline of Structure Design that one might expect to be durable. One wonders whether Structures that are designed to have many powers (Neocameralism, bitcoin) might coalesce into three. Just a thought.

Now we come to Parliament. I don't see the medieval English parliament as "part of government" in the sense that the modern UK Parliament is. It wasn't responsible for law, or for any routine act of government. Its role seems to me to have been the constitutional watchdog, checking on behalf of the Lords and Knights (and later Burghers too) that the King was sticking to the constitution. Running the country was the job of the nobles, within their lines-on-the-map, and of the King, regarding defence. The power of parliament didn't come from any constitution; it came from the fact that it could reach an agreement, and then go to the country and say "The King is infringing on his subjects' rights". (Or, conversely, it could say "Lord Splodgeberry has defied the King and the King is justified in going and kicking his arse"). It makes sense as a transparency mechanism rather than as a power in its own right.

Transparency, even more than Triangles, seems like a durable guideline for Structure Design. You want people with power to be working for good government, not for enhancing their own power, and you need to be able to see that that's what they're doing.

Having said that, I don't think there are many general principles for Structure Design. I've spent this piece looking in detail on one historical Structure, to say why it was they way it was and why it worked. I think that's what you have to do: Structure Design is a boundary value problem. You have to start from where you are.

But then again, Structure Design is a thing. Where two or more powers come together, reaching an agreement is more than just recognising their existing position. It may mean one or both giving up some power that they really hold to cement a durable deal. The establishment of rights of Knights I described above follows that pattern: the King needed it to happen so it was added to the Structure by negotiation. (That may be a stylised version of what really happened, but it could have gone that way).

So I think you can say a bit more than this:
the estates of the realm don't arise from nowhere. they were supposed 2 formalize the *actual* structure of power that underlied sovereignty - @cyborg_nomade

What you can't do is just dream up some "constitution" and assume that anyone will follow it. The half-life of a Structure designed that way is generally measured in weeks. Even a constitution that worked somewhere else will fail immediately if the power on the ground doesn't match the Structure that the constitution is designed to support.

Decolonisation of Africa produced a number of experiments to demonstrate that process.

Once the holders of actual power have been identified, "constitutional design" can take place to create an arrangement by which they are incentivised to participate in an efficient government. However, "constitutional design" in a vacuum is worthless. Democracies with deviations from "one-man-one-vote" have been moderately successful in the past, but I do not think this example is rooted in any realistic assessment of power.

Similarly, various people from time to time (including even myself, long ago) have suggested random jury-type selection of decision-makers. This has attractive efficiency features, but nobody with vested power would have a clear interest in keeping it running fairly, and the scope for corrupting it would be enormous.

The way to think of creating a stable government Structure where there is intractable division of power is midway between diplomats negotiating a peace and lawyers negotiating a contract. Neither of those are trivial or negligible occupations. (At the completely rigorous level, Structure Design is a matter of game theory, but I doubt real-world situations are tractable to mathematics).

Constitutions need to resemble contracts in that they have to cover detailed interactions unambiguously, but they need to resemble peace treaties in that they need to provide for their own enforcement.

The whole Godel amending process is a bit of a red herring. In the words of Taylor Swift, nothing lasts for ever. Circumstances change, and new Structures have to accommodate them. A new Structure can be built out of an old one--such as representatives of Boroughs being included in the House of Commons alongside Knights--if the parties with power agree they are necessary. Making a constitution change is not the hard bit; making the Structure stay the same from one year to the next is the hard bit.

Sometimes a Structure has to go. Gnon has the last word.

07 September 2016

Unchecked Power

In my previous post, I explained why Neocameralism is not a division of power in Montesquieu's sense, but rather a special case by which the benefits of power can be divided without dissolving responsibility.

However, while dividing power is not desirable, there is no Ring of Fnargl, and power is never perfectly concentrated. A real sovereign still has to deal with forces beyond his control, most obviously those beyond his borders; the loyalty of his subjects is always a real issue. Sufficient incompetence can destroy anything.

The reason that division of power is undesirable is that it erodes responsibility. Government is responsible if whoever has the power benefits from exercising it well and is harmed by exercising it badly. If the single absolute sovereign owns all the extractable product of his realm forever into the future, then it is in his interest to make it a successful, functional, realm. His interests may not be perfectly aligned with those of his subjects, but they are not all that far away. It is better to live under a secure sovereign who rules in his own interest than under a chaotic parliament which attempts to rule in yours. This is an analogous argument to the superiority of for-profit services to government-provided services in other spheres.

If power over the corp is divided, each individual with power now has two sets of incentives: to maximise the value of the corp and its product, as for an absolute ruler, but also to maximise their power over and benefit from the corp. Division of power is harmful to the extent that the second set of incentives exist and contradict the first.

The two largest classes of undesirable incentives are to extract value from the corp for oneself, and to increase one's power over the corp at the expense of one's rivals. The first is more obvious, and the second, in historical experience, more extensive and more damaging. Conversion can be restricted if the number of participants in power is reasonably limited, as it tends to be obvious. However, if power is distributed flexibly, then it is easy to provide rationalisations for a change in policy that is actually directed at increasing the power of one participant.

The fundamental problem is that power, whether formal or informal, is fungible. As I wrote in 2011:
A realistic chance of power is power in itself. It can be traded, borrowed against, threatened with. A "politician" is one who holds "Virtual Power", and tries to increase it, just as a fund manager tries to increase the assets he holds.

If making power formal doesn't help, then what is "formalism"? Formalism is Neocameralism. Formalism's solution to persons with practical but informal influence over the government is not to formally define and legitimise their influence, it is to buy them out. It is to put a value on their influence, and to have them give up that influence in exchange for dividend-bearing securities.

As described in my previous post, the point of that is to take away their incentive to steer management in one particular direction or another, and to give them instead an incentive to have the management maximise shareholder value.

Clearly, then that is not a perfect solution to all problems of politics. It only works to the extent that a participant's power, whether formal or tacit, is seen as legitimate. If a participant's power is informal but legitimate (which is a common situation in the Modern Structure), it should indeed be made formal, but only as a preliminary to removing it.

It follows that formalism does not solve the problem of necessary division of power: the fact that however legitimate power is defined, there are those outside it who have influence over those inside it. It doesn't solve, in general, the principal–agent problem. (The CDCC is designed to partially solve one particular instance of the principal–agent problem, of the armed forces openly defying rightful instructions; by providing a specific solution it implies that there is no general solution).

What formalism does is to leave the fundamental problem unsolved, and then insist that it is the fundamental unsolved problem, and that as a matter of day-to-day competence it must be limited at all costs. Take a moment to see how far that is from the conventional wisdom, which celebrates and actively encourages all division and distribution of power.

If any slope is slippery, it is the division of power. Division proceeds from division. Complete power is inviolable, small allowances of outside influence can be monitored, limited and reclaimed, but once substantial centres of power become strong enough to defend themselves, the remaining power will be shredded in the inevitable conflict.

The problems of people trying to influence a near-absolute ruler are not a different kind of problem to those we are used to. They are the normal problems; the exact same problems that utterly cripple any kind of competent government of modern states, only much smaller and more manageable.

There is no magic formula which will make good government out of an unviable realm. The possibility of concentrating power sufficiently for stability is the sine qua non of independent government.  What is the ideal form of government for Mauritania? What is the ideal form of government for Marsh Farm? In both cases, it is for them to be ruled by outside forces that are strong enough to be secure.

Compromising the integrity of the structure of centralised power is to be avoided. Take for example, the hypothetical case I raised when I discussed the issue before, in Aretae's day: the Pineapple Computer Co who want the King to appoint a judge under their control, to get them out of a PR problem.

By the logic above, the worst thing the King could do would be to agree to Pineapple's request. That is giving away power, and there is a danger of not ever getting it back. Telling them to go fuck themselves would be better. Offering to match Queen Tamsin's duty-free zone would be better.

A formalist answer, if instead of a King there was a Neocameralist CEO, would be to hold merger talks: if the sovcorp buys out Pineapple in a stock-for-stock transaction, then the interests of the sovcorp and the factory are henceforward aligned. I'm not convinced it's a good idea for a sovcorp to own too many nationalised industries, but if the factory is genuinely essential to the wellbeing of the state, that is a reasonable solution.

(If the King is really a King, but the Pineapple company is privately owned, the same end could perhaps be achieved by having the owner of Pineapple marry the King's daughter).

03 September 2016

Checked Power

The latest from cyborg_nomade at antinomiaimediata is a wide-ranging poking at the cracks of the neoreactionary/Moldbuggian concepts of Sovereignty and Responsible Government.

As I said on twitter, cyborg_nomade is, from my point of view, picking up from where Aretae left off all those years ago, not in that he is the same: as their respective aliases suggest, Aretae rooted his arguments in Classical philosophy, while cyborg_nomade is more Continental. But cyborg_nomade, like Aretae before, is challenging details of neoreactionary theory from the left, and that's a more productive critique for defenders to concentrate on than the intra-far-right discussion that takes most of our time.

So, "neocameralism and constitutions" is quite a wide discussion, and I'm first going to pick off some low-hanging fruit concerning the role of stockholders in neocameralism.

I'm not going to talk about "conservation of sovereignty"--to me that is an unclear concept, so I'm going to try to be more concrete. I'm going to talk about the "corp", meaning both joint-stock corporations as we know them today, and sovcorps as envisaged by neocameralism.

Moldbug repeatedly denounced "separation of powers" as a principle. no sovereign can be subject to law . On the other hand, cyborg_nomade points out, is it not true that modelling a neocameralist government on a joint-stock company implies a separation of powers:

The controllers have one job: deciding whether or not Steve is managing responsibly. If not, they need to fire Steve and hire a new Steve.

That quote is from Open Letter VI, and cyborg_nomade quotes more, but it is actually necessary to read the whole thing.

In particular, the paragraph immediately following cyborg_nomade's selection:

What happens if the controllers disagree on what "responsible" government means? We are back to politics. Factions and interest groups form. Each has a different idea of how Steve should run California. A coalition of a majority can organize and threaten him: do this, do that, or it's out with Steve and in with Marc. Logrolling allows the coalition to micromanage: more funding for the threatened Mojave alligator mouse! And so on. That classic failure mode, parliamentary government, reappears.

The introduction of stockholders is not a matter of checks for checks' sake. Nowhere in OL-VI is there a suggestion that dividing power is a good thing in principle. The purpose of stockholders is a very narrow one: to fix the location of responsibility.

The corp exists for the benefit of the stockholders; if it is run well, they benefit, if it is run badly, they lose out, therefore, they should have the power. All of it. Choosing to exercise that power via at-will appointment of a Chief Executive is an implementation detail, but a well-tested one, and, other than for sovcorps, an almost universally accepted one.

Why multiple stockholders rather than one? Because with a single owner, the purpose of the corp becomes unclear: it is whatever that single owner chooses. However, if the corp has a large number of diverse stockholders, their idiosyncratic interests cancel out or become negligible, against their single shared interest in ROI.

Note that this is not a guaranteed state of affairs. A corp with a joint-stock structure can, as described by the quote above, decay into politics. For existing non-sovereign corporations, this is very unusual, but that is because many measures are taken to actively prevent it. In Anglosphere corporate law, it is not considered sufficient that stockholders can replace management by a majority vote of stock. It is in principle illegal for management to work for a goal other than return on stock, even if it has the support of holders of a majority of stock. There are also restrictions on how concentrated stock ownership can be, at least for corps for which stock is publicly traded.

So it turns out that the purpose of a joint-stock structure is not to distribute power across a larger number of humans, but to concentrate power on a single non-human "virtual" decision-maker, the shareholder-value maximiser. To the extent that a joint-stock structure does not do that, it is always considered defective, and frequently illegal.

(The parallel to bitcoin, converting individual miner decisions of transaction validity to a single non-human abstract "blockchain" decider, is obvious).

Compared to the essential feature of responsibility, the preference expressed by Moldbug for joint-stock versus monarchical sovcorp structure is marginal:
A family business is a great idea if your business is a corner store or an auto-body shop. If you have a continent to run, you want professionals.

The next question to answer is: why? Why is it good to have a corp run in the interests of this non-human abstract, "maximisation of shareholder value"?

The answer is that this is a clearly definable, constant goal that is usually consistent with the long-term continued existence of the corp. As Moldbug explains, if you want some other goal, then first maximise shareholder value, then spend the proceeds on whatever goal you want; that is a matter of consumption, not effective management.

As an aside, cyborg_nomade suggests that "customers" constitute another check on the power of management of a corp. I don't think that is a useful way of looking at things: we are talking about the management of a corporation, or a nation-state, and any such thing, unless it is the whole universe, exists alongside other things beyond its management, and has to interact with them. Good management means good management in connection with customers, suppliers, neighbours, and competitors, and no change to the organisational structure of the thing being managed makes any difference to that fact.

This whole defence of neocameralism leaves some obvious gaps. First, enforcing shareholder voting rights on a sovereign joint-stock company absolutely requires the cryptographic-weapon-lock scheme. Moldbug in OL-VI is explicit about that:
The neocameralist state never existed before the 21st century. It never could have existed. The technology wasn't there.
It is because I am sceptical of the practicality of that scheme that I tend to advocate for what I call "degenerat formalism", which is right back to that old family business. Nevertheless, my position is that assuming a working cryptographic decision and command chain, neocameralism is good.

Second, the CDCC provides for shareholder voting rights, but not for the extra minority-shareholder rights that are provided by modern corporate law. If those are actually necessary (and they may well be), then some other mechanism has to enforce them. Note that those rights in part predate the actual corporate law that now enforces them: they were provided in the rules of the company, because it was understood people wouldn't want to buy into corporations that did not have them. Moldbug's solution to these problems is Patchwork: Not only are sovcorps structured according to the neocameralist design, but they exist in a competitive marketplace, and the forces of competition apply the remaining necessary constraints on management.

As I said, this is only picking on one part of the argument in "neocameralism and constitutions", the part that is easiest to deal with because I think it is a clear-cut error. The more interesting part, about constitutions as spontaneous order, or products of selection, remains to be answered.