The Power Of Three

Three was a natural number for witches. When you had three, you had one to run around getting people to make up when there'd been a row. Without Magrat, Nanny Ogg and Granny Weatherwax got on one another's nerves. With her, all three had been able to get on the nerves of absolutely everyone else in the whole world, which had been a lot more fun. Because, while three was a good number for witches, it had to be the right sort of three. — Terry Pratchett, Maskerade

Pratchett was on to something here, I think. Not necessarily in the very literal sense, because human dynamics are more complicated than that. If they're properly organised, more people can bring more knowledge, perspectives and ideas : diversity really does matter. If they're not properly organized, you get groupthink, dogma, or just a simple massive row or a small war.

Recently I had a brainwave that if science is so darn good at finding out the truth, why not apply some of its hard-won lessons to other arenas ? Like politics, for instance, which seems very badly in need of a good dollop of objectivity and a lot less of tribalism. This post continues my efforts to attack this proposal and root out any weaknesses. Scientists suffer all the same weaknesses of other hairless monkeys, yet, while getting them to agree on anything is like herding cats, this doesn't seem to be much of a problem. In fact, this is even something of an asset, because everyone wants to make the next big breakthrough. In politics, each side just seems to want to break the other.

One lesson from science's truth-finding endeavour is the importance of the triple structure of peer review : author, editor, and reviewer. After finding a skeptical reviewer, the editor themselves doesn't provide extra review for the submitted manuscript, but checks if both author and reviewer are following the rules. In this way they can provide a genuinely independent, impartial check on the whole thing. The editor, in fact, is our Magrat. This works well, and I say this after having endured a year of peer review gone horribly wrong, so if anything I should be biased against the system.

There are plenty of other aspects to the structure and method of scientific inquiry, and plenty of other possible points of weakness to be examined in future, but today I want to look at how this kind of arrangement applies in society and where it's not working. The simplicity and power of this analytical triumvirate is too tantalizing to ignore.


1) The triangle of government

When I was very young, I used to think that the government was In Charge Of Things. They were sort of like elected tyrants in my mind, accountable to no-one and could do whatever they wanted. If a company was misbehaving, they could inflict whatever punishment they liked and that was that. If the dog down the street bit someone, men from The Government would show up and sort it all out.

Very, very slowly, I've come to realise that things are a teensy-weensy bit more complicated than that.

I've written a bit about the importance of understanding the networks that underpin society, and I've mentioned how I want some nice simple diagram to illustrate how the major components of society relate to each other. So it's time to try and start to construct one. At least, very, very crudely. This is all part of my super-secret ambition to work out how to make the world a better place, but don't tell anyone or they'll say it's impossible.

It seems to me that there are two major networks underpinning the governance of society. The first is the body politic itself, which looks a bit like this (you can already find umpteen versions of this elsewhere, of course) :

Okay, that should really be the executive, legislative, and judicial, but I'm a simple man and I like simple terms. In essence, and ideally :

• The government (the executive) carries out the will of Parliament (the legislative). It has some degree of control over Parliament, but is ultimately answerable to it. It has first priority, or even sole ability, when it comes to proposing laws, and carries out things which can't or don't require a vote.
• Parliament (the legislative) makes and repeals laws. Its agenda is largely set by the government, but it has the final say on what the government gets to do via voting. 
• The courts (the judicial) decide how to interpret and apply the laws set down by Parliament. Courts cannot make law but they can set precedents. They decide if both the government and Parliament are acting lawfully. Neither Parliament nor the government have much direct influence over them.

Here I want to be deliberately crude - I only want to understand how the major blocs relate to each other, not how each operates internally*. I've completely ignored how all the nodes access external information - factual data, popular opinion etc (though there are a couple of very nice, slightly more complex diagrams illustrating how political principles are established here). Instead of looking at the details, let me bash out a few ideas here about how these great big nodes are related to each other in practice.

* This runs the risk of missing effects that emerge from the complex underlying structures, but we have to start somewhere. There's value in simplicity.

The connections between all three are fundamentally complex. All sides get to hear each other, so there's always a passive connection between each. But that flow of information's not going to be equal : it's going to be stronger between the government and Parliament, as both sides get access to private information from the other. The courts get some level of private information from each, but I see no reason to think that this flow would be unequal on either side.

Of much greater importance is that the connections are fundamentally different. This is crucial in avoiding a circular firing squad and helping prevent the need for infinite chains of oversight : with everyone performing different types of checks and balances, Plato's problem that it's "absurd for a guardian to need a guardian" is elegantly solved. You do want guardians to have guardians, but they all guard different things. In an academic journal, the editor doesn't act as a second reviewer of the article, but as a reviewer of the reviewer and of the author's responses - they check whether the rules were followed, not the content of the proposal*.

* Though the author doesn't really get to check if the editor is doing their job properly, except in that if they feel unfairly treated they can go somewhere else. Perhaps that's a lesson for research journals.

So it should be with the courts. If the government is the would-be author, then Parliament is the reviewer, and the courts are the editor. They're supposed to be independent, or else there's really no point in them existing at all. Their independence is partly maintained by their separation from Parliament and partly because their authority is fundamentally different. This three-way structure then becomes very powerful : it is robust, difficult to break, but also simple and comprehensible.


In practise...

But this is hardly a universal truth. Bafflingly, in America they have elected judges, which confuses the heck out of me. Worse, their supreme court is appointed by politicians. What the hell the point of that system is, I don't know. I guess if we account for this then we have to redraw the graph :

The court is here little more than a sort of parasite, feeding on whatever scraps the government deigns to give it like a particularly ugly and gruesome deep sea fish.

You might also say, "but hang on, surely the government is only a subset of Parliament, so it's not really independent from it". That's true in many countries : the power to propose a law and the power to vote on the proposals are both allocated to the same people. An easy solution - if this is even a problem - would be to forbid legislators from voting on their own proposed laws, though that's only likely to make a difference in the really knife-edge cases*.

* Alternatively, forbid all government ministers from voting, thus also reducing the number of ministers (a position too often used as a bribe, leading to ministerial-inflation). This would have a much more substantial impact, though there would be a risk of a corrupt government giving incredibly wide-ranging powers to a few individuals instead of lesser powers to many.

This leads me on to Britain, where there's a different problem : the court is indeed independent, but the prevalence of large majorities (lacking in recent years but now restored) for the governing party means that Parliament rarely acts as a properly skeptical reviewer - it can criticise, but it usually can't/won't enforce*. And since both sides rarely actually break the rules, the courts can do little to intervene. The government thus has near total control of Parliament - not quite total, because rebellions among its own MPs do happen, but nearly so. This means that the government is effectively a parasitical predator of Parliament.

*At least in the Commons - things are a bit better in the Lords, and this is a limitation of making these graphs too crude. I read with dismay when people complain about these "unelected peers" without ever checking to see if said peers are actually doing a decent job. We might as well complain about unelected doctors or unelected electricians !

Which is of course inspired by Futurama's mind-controlling brain slugs.

This situation, I'm coming to realise, is as stupid as having the government play a similar role towards the courts. Without having that vital triple-way system of checks and balances, the system is reliant on the goodwill of those running it : it can function, but only if everyone is genuinely committed to acting fairly. What does it matter if the government appoints the courts if it appoints genuinely fair and impartial judges ? What does it matter if the largest party has a huge share of the seats if most of its members are prepared to act against the executive ? Answer : it doesn't, which is an especially insidious obstacle. It's difficult to see the need for a reform when everything's going well, and the structure of the system doesn't guarantee it will fail, it just makes failure much easier.

And that of course is the real problem. If a government does not act fairly, it can appoint toadies to the courts and/or have no meaningful check on its actions from Parliament. In both ways, such a system tends towards my childhood idea of elected despotism. Government control of Parliament means it can make any law it wants; government control of the courts means it can break any law it wants.

Of course any reasonable democracy has provisions that means it's not quite as bad as all that, e.g. judges may have staggered terms of office so aren't all appointed by the same government, altering laws can be a slow process (especially if they are deemed constitutional matters, making any changes subject to judicial review). Ultimately things can only be pushed so far before open revolt breaks out, or more probably before the public will vote the ruling party out of office come the next election.

But therein lies the catch. The point is not that the dysfunctional systems ever reach these tyrannical end states - it's that they can be perverted arbitrarily close to them. In some ways this is worse than abject tyranny. You know where you are with a ruthless tyrant : eventually the people will rise up and may or may not be brutally crushed. It isn't nice, but at least it's simple. Whereas a perverted democracy is a lot more like a toaster that only works if you press the button down in just such a way; it's inconvenient and annoying, but since it basically functions (albeit with a lot of very serious problems*), most people are content to deal with it rather than go and buy a new one.

* Such as burning the toast and/or delivering random electric shocks and occasionally starting small fires.

Much as I love the morality of Doctor Who, when the Doctor says...

The systems aren't the problem. How people use and exploit the system, that's the problem. People like you.

... I think she is only provisionally right. Some bad systems used by good people will produce good results, but not all. And some good systems used by bad people will produce bad results, but again, not all. Systems cannot be so readily divorced by people using them - both the system and its people are inevitably caught in mutual feedback loops that affect them both.


2) The triangle of information

The second key network is on a much larger scale and it looks a bit like this (I've alluded to this before) :

This time the nodes aren't individual institutions - the body politic comprises the entire political system of the first diagram. The whole thing works something like this :

• Laws are enacted, enforced, and interpreted within the body politic. Its laws and rulings are applied to both the media and the general public.
• The media are the primary means of conveying information between the body politic and the public. The media interview politicians and report what they're doing, but they also do the same for members of the general public.
• The public are influenced by but also influence the media and politicians directly. They have limited direct interactions with politicians, except through voting, but have enormous influence over the media via sales.

As I've said previously, anyone hoping to get a truly better system of government had better consider this whole shebang (see link for a nice review of the complexities of this) and not just individual aspects of it. Of course, that doesn't mean we shouldn't stop propoposing corrections to small-scale problems where we find them, because that would be silly.

This system is quite a different beast to the first one. "The public" do not represent a single entity that's capable of making an arbitrary choice - its decisions are emergent from the internal actions of its millions of individual parts. Of course, this is also true of the other nodes, but to a far lesser degree : an editor can decide which stories to run and which to pull; politicians can pick their battles. Whereas whether the public even like something or not is far more complicated. Sometimes people sit back and do nothing except feel faux-outrage. At other times they become, en masse, inexplicably violent.

As evidence of people being weird, I will again cite the case of someone preferring to accidentally murder their best friend than get their grandmother addicted to heroin.

While we can propose endless reforms to the political system or how the media operates, the same cannot really be said of the public. Such an enormous, diverse group is composed of a dizzying array of networks and hierarchies, flows of information, resources, and funneled emotions. Some of those networks are purely consequential, emerging from the choices made by individuals, but others are causal, constraining the actions made by their members. Many are both. Most groups tend to reinforce their own existing norms but rarely with 100% effectiveness.

This is why any kind of proposal saying, "we just need to be better people" is doomed to fail. You can't make the public, or indeed any kind of complex group, behave more nobly just by hoping, or even by winning enough of them over through persuasion. You have to manage the network in which they're embedded, otherwise you're just pissing up the waterfall of information contradicting whatever it is you're trying to achieve.


In practise...

And as before, the above diagram is the ideal case. A lot of people are rightly worried that this will be perverted like so :

It's perfectly obvious why this would be bad : with all of the information flow coming from a government-controlled source, the potential for abuse is huge. Note that there is now no direct link between the public and the courts or Parliament, and the connection between the courts and Parliament is severed. This is another dangerous perversion of the system, allowing the government to claim, "look, we've still got the other essential parts of parliamentary democracy", though in reality it controls them both separately and independently. If the court cannot rule on what Parliament does, which is entirely at the behest of the government, it's rendered impotent - but so much as to provoke any serious kind of public response.

Again, if all parties act fairly, then there's no problem, but as before, if all parties act fairly, then there's little disadvantage to the idealised system. Of course no-one trusts the government not to misuse the power that direct control of the media would bring, so the extra regulations and bureaucracy necessary to separation are more than compensated for by the benefits of keeping the system fair and impartial.

We need not go through the plethora of historical examples of why the media as a minion of government would be bad. I'm more concerned that we've developed something of the opposite problem. The media may already act as a brain-slug controlling parasite of government, or maybe it's even worse than that.

The media doesn't feel so much like a parasite of government so much as a terrifying monster that neither the public nor the politicians feel able to challenge - or worse, are unaware of the need to challenge it. You must understand that though I'm not describing all media outlets here, I am speaking of both right and left leaning press, both of which are equally adept at spinning the same news to support whatever agenda they wish (link has a very nice apolitical example). Or worse. The default setting is not "defend" but "attack" - regardless of the story, there must be something awful about it and someone to blame. 

Even on a good day, all too often impartiality goes haywire.

This may be impartial in the strictest sense of the word, but it's not fair and it's certainly not objective. Even the best interviewers seem to have fallen victim to attacking absolutely everyone before them as though everyone was guilty of something, treating idiots and criminals the same as experts. Is that really being impartial ? Is it in any way sensible to attack experts and lunatics with the same degree of rigour ? How can the public trust anyone in that setup ?

The press seems like a dangerously whimsical beast indeed. Someone can be their darling one day, but should they breathe so much as a word of criticism of the press, then watch as they're immediately cast aside and their once-endearing qualities thrown against them. Too often the media just gets everything wrong, failing to attack what needs attacking, and defending the absolutely indefensible. This is stupid.

And reality is considerably worse than that, because "the media" isn't a single unified entity. Parts of the press will at all times attack the government and any and all who support it as though it was about to bring on the apocalypse, even if it's only trying to alter the taxation rules for strawberry milkshakes. Simultaneously, other parts of the press will defend the government even as it plans a full-scale nuclear attack against Liechtenstein, claiming that everyone hates that stupid tiny country and they're not really people anyway. And both sides will act with stupendous inconsistency : praising and shaming those they like or dislike even if their actions and motivations are identical.

The media is a terrifying monster not just because it's got big nasty teeth, but also because it's got the norovirus. It's continually vomiting from both ends over everything, submerging what's really going on in an acidic fug of atomised bile. And worse, it does this in such a way that people want to keep on buying it.

I could go on, but I won't. What I do want to point out is how incredibly important the media link in the network is. The media is a hugely dominant communications channel between politicians and the public. Even if we had both a sensible public (we don't) and sensible politicians (again, no), then a piss-poor media would still bugger everything up. You can't act sensibly if you don't have sensible information.


This all sounds very bad.

Indeed it is. So is that apparently powerful triple structure really inherently stable ? I would say no : these networks are a consequence of the methods adopted more than they are a cause of the results. A government which wins a strong majority and is sufficiently stupid can easily alter the law to affect the power of the courts however it likes. Unregulated media driven by profit will always grow monstrous.

But this doesn't mean that the triple structure of organising politics, information and so on is unimportant. Far from it. Instead, we should think of this overall structure as the goal, not the method. We should aim to create something like the ideal triangles of checks and balances, but we shouldn't be naive to think that that in itself will be enough for a stable, sensible end result. Rather, we have to consider the internal functioning of each node, otherwise we'll always be at risk : a single weak link in chain of three can quickly go badly wrong.

Ironically, we need look no further then Parliament itself for a good example. The triangle of government, opposition, and Speaker hardly produces constructive debates (though other political debates, with the same people but run in a very different way, are much better). The rules and methods used matter a great deal : the role of the Opposition is to provide an alternative, not act as a skeptical but sincere reviewer; the role of Speaker is to ensure everyone follows the rules, but these rules don't include any fact-checking whatsoever (nor is there even any necessity for politicians to respond to each other's arguments directly - it's perfectly fine to launch a counter-attack or spout meaningless drivel instead). Whereas in that other node of the body politics, the court structure of appellants, judge and jury does seem to basically work.

This hasn't really shed much light on my proposed scientific political system. The methods of checks and balances I've suggested might be enough to keep the thing from collapsing horribly, or they might not; the other unique pressures acting in academia could be crucial, or I may have missed something entirely. Which means this has been a very inconclusive post.

Let me end by summarising some of the major requirements for a better political system. First, it has a number of competing interests it has to balance purely for its own sake :

• Competition and cooperation. Excessive competition leads to hostility, whereas excessive collaboration leads to dogmatic groupthink. But the right amount of each - competing groups who want to outdo but not hurt each other - can be a stunningly powerful combination.
• Diversity and focus. A wide range of opinions and perspectives is essential to tackling a problem from multiple angles - specialists should not all come from the same social backgrounds. Only the smallest number of uncompromising fanatics should ever be included; there are some positions so extreme that no reasonable system can ever handle them.
• Publicity and privacy. The need for private discussions is vital, as no-one can realistically be expected to think sensibly while being throttled by the ubiquitously hostile public/media. Yet at other points we need the system to be as transparent as possible, so that the reasons given can be subject to proper scrutiny.
• Stability and flexibility. Probably the hardest to unify. The system needs to be able to respond to the changing needs of the day without having to completely remake itself into something unrecognisable and potentially unpredictable, being able to wield both the sharper and blunter instruments of society without falling victim to them.

And one more, an especially important one : dedication and detachment. The weirdest extremists of all are the charismatic, analytical psychopaths - the kind of people who can work out in great detail how to solve a problem but have no clue if their solution is a good one, who don't care* who gets hurt in the process, and have a gift for making their stubborness mistaken for being morally principled. Such people are rare indeed, to be sure, but they are disproportionately dangerous. Whereas your regular fanatic is committed to a specific cause or two, these people are dedicated only to their own advancement - or, worst of all, to tearing apart the system because they honestly believe it should be destroyed. These are the kind who just want to watch the world burn. More important than selecting for the extremists willing to die for their cause, the ones of extreme dedication and determination, may be to select the people who aren't even mad keen on working weekends : normal people, hard-working but able and willing to listen to reason and to compromise. The usual tactic of calling such people out when they mess up is usually successful - it's only the uncaring psychopaths who need stronger measures.

* Holding politicians to account by exposing what they've done works well for people who have the decency to care, but it has no effect at all on those who only care about themselves.

There are other things the system must do more for the sake of the society is serves than itself. Without going into what would make for an ideal state, these attributes include :

• Allowing genuine self-determination and meaningful choice. The system must temper bad decisions with expert oversight, but people must be able to make their own mistakes that will even cause them some degree of harm. The trick is to prevent (or reduce) those mistakes from impacting others.
• Combining different methods of decision-making. Democracy, oligarchy, despotism, monarchy - all have their place. Sometimes you need a strongman, sometimes you need maximum diversity.
• Consider both long-term and short-term impacts. Speaks for itself, really.
• Acts in the interests of the whole country. Not just for individual political parties or interest groups. It's going to have to manage competing and conflicting interests, often from groups who don't even realise they're at odds with each other. Somehow it must stand for all of them.
• Willing to learn from past mistakes. U-turns that occur in response to the evidence should be welcomed, while those due to populism serve as warning signs. Trials of different processes should be used, but not the extent of becoming experiments on people.
• Be comprehensible. And even better, be genuinely simple. One of the advantages to Universal Basic Income is that it's so simple it's impossible to cheat - the ideal political system should be like that as far as is possible, not merely accessible to the interested layman.

So that's my rant for today. Some of those aspects I've already tried to include in my suggested system, others I'll look at in the future. Some parts rely more on reforming society itself than they do the political structure. My main goal is to establish if there is, in theory, a way of organising people that will actually work, given people's actual failings and not idealising them. Never mind implementation for now, that's an entirely different question.  Such an organisation would not be Utopian, but only mean that people are basically content, wherein politics was something they would engage with but was fundamentally boring, where there'd be no point trying to seize power because no-one felt the need to trash a working system that already did what they wanted it do. I still feel the current system is far too unstable, but I'm considerably more optimistic that a better one already exists.