In 1960 Popper published his paper Knowledge without Authority where he conjectured that the central question of political theory should not be Who should rule? but something entirely different.
What he put forward has become known as Popper's Criterion - a criterion I believe IBDD satisfies with flying colours.
Popper's Criterion
In response to the who should rule? question, Popper writes:
This political question is wrongly put and the answers which it elicits are paradoxical. It should be replaced by a completely different question such as 'How can we organise our political institutions so that bad or incompetent rulers cannot do too much damage?'. I believe that only by changing our question in this way can we hope to proceed towards a reasonable theory of political institutions.
The question about the sources of our knowledge can be replaced in a similar way. It has been asked in the spirit of: 'What are the best sources of our knowledge — the most reliable, those which will not lead us into error, and those to which we can and must turn, in case of doubt, as the last court of appeal?' I propose to assume, instead, that no such ideal sources exist — no more than ideal rulers — and that all 'sources' are liable to lead us into error at times. And I propose to replace, therefore, the question of the sources of our knowledge by the entirely different question 'How can we hope to detect and eliminate error?'
Karl Popper, Knowledge Without Authority (1960)
There are two important political questions here:
How can we organise our political institutions so that bad or incompetent rulers cannot do too much damage?
How can we hope to detect and eliminate error?
Question 1 has morphed to become known as Popper's Criterion:
Popper's Criterion — Good political institutions are those that make it as easy as possible to detect whether a ruler or policy is a mistake, and to remove rulers or policies without violence when they are.
David Deutsch, The Beginning of Infinity (2011)
There are two important ways we must consider IBDD in the context of Popper's Criterion. The first regards the nature of authority and how IBDD is fundamentally anti-authority, and the second is a direct application where we investigate exactly how well IBDD is able to 'detect whether a ... policy is a mistake, and to remove ... policies without violence when they are'.
I would also like to extend Popper's Criterion to include the system of decision making itself. If we are unable to change the underlying system without violence — seeing as the system itself is a policy — then such a system must fail Popper's Criterion.
Authority and IBDD
A note on the use of 'good' and 'bad' when describing policy and explanations: I am using these words in the same manner as Deutsch uses them in The Beginning of Infinity. He does a great job of explaining them in detail, so I will summarise: good explanations are those which are hard to vary and account for the phenomena they purport to. Not all good explanations are the most correct, but the most correct explanations must be good. Since policy is a way to take an explanation relating to some problem, and provide instructions on how to solve it, policy inherits this property. Policy which is based on a bad explanation is bad policy, and vice versa for good policy.
Given that 'who should rule?' is 'who or what is the authoritative source of good policy?' in another form, the concept of authority plays a strong role in current political systems.
At a fundamental level Issue Based Direct Democracy (IBDD) attempts to acknowledge that there is no authority on good policy (or knowledge). In competing democratic systems, the answer is often in the form:
The party or parties elected through a given electoral system. (Representative Democracy (RD))
The people, or the will of the people. This could be put more exactly: the source of good policy is the measurement of public opinion where each voter is treated equally. (Pure Direct Democracy (DD) and Liquid Democracy (LD))
The groups able to purchase the most votes given a quadratic pricing relationship (Quadratic Voting)
While IBDD is a form of direct democracy, we do not claim that any one source of policy is particularly better than another. Rather, we reason through the process a little differently.
All systems of decision making must rely on all people or some subset of people (if it involves explanations)
Thus it is not useful to say that simply because a system involves all people at all times, it must fail Popper's Criterion
However, we can institute a system which gives different weights to different sources of knowledge at different times
The set of such systems must include those which are very bad at biasing sources of knowledge, and systems which are very good at biasing sources of knowledge
Thus the question we'd really like to answer is how can we make the best guess of which source of knowledge is best for a given problem, without invoking authority?
Furthermore, we can see that different sources are appropriate for different problems, even if those problems coincide, and therefore an optimal system must treat and weigh relevant sources of knowledge independently.
While majoritarian systems (RD, DD, LD) do incorporate many sources into a ruling faction — either through a single party or a coalition of entities — they are tightly bound, and without political disturbance, or specific arrangements of representatives in a legislature, they are inextricable from one another. Often this collection of knowledge is referred to as a policy platform.
As explained in the Flux Philosophy White paper Redefining Democracy, the reason for this centralisation is due to the fundamental nature of majoritarian democracy.
This centralisation is not exclusive to representative democracy. Rather, it is a result of one particular rule many people feel strongly about: one person, one vote. Traditional democratic intuition holds that one vote should be distributed to each voter and no redistribution should occur: that the set of voters should be static. However, this ignores that the distribution of interest in different issues is not even. When a voter is
not seriously interested in an issue they are able to use their vote as leverage against those voters who are interested. This leads to a curious result: to be sure you can pass a bill in a diverse group of voters you must form a bloc large enough to be the deciding factor in whether the bill passes or not. If you are not in that bloc, someone else will be, and they will use your bill as leverage on other bills that you aren’t interested in, forcing you to participate anyway. Thus the optimum strategy in a static democracy is to become the largest bloc, as it gives you the greatest chance of being responsible for whether any given bill passes or doesn't.
Kaye & Spataro, Redefining Democracy (2017)
Thus, in order to treat solutions to social problems independently we cannot have a system ignorant of distribution of knowledge in society. That is to say it must weight different sources of knowledge differently.
The selectorate theory of political power pioneered by Bueno de Mesquita et al. teaches us that the current weights on sources of knowledge is determined by political leaders and their need to satisfy key supporters (the selectorate). However, biasing sources of knowledge needed to keep a leader in power does not strongly bias good explanation. This is accounts for why these sources are largely inextricable from one another in such systems, and why there are fundamental limits on the speed of progress of RD, LD, and DD.
The problem we are concerned with now becomes how to weigh sources of knowledge without authority? and it is to this question that IBDD has a direct answer.
IBDD delivers its answer in the form of a specially crafted persistent and inclusive market. This allows the distribution of political expression on any one issue (or rather, on a particular policy conjectured to solve that issue) to be decided by the aggregate input of all voters considering that conjecture in the context of all present and future issues.
This allows voters to individually weigh the value of their conjectures and criticisms by their perceived benefit at a given time in a highly dynamic way.
It is possible, at this point, to argue that this is just authority in another form: that really we are answering the question of what is the source of good policy? in the form the dominant weighted source of knowledge as determined by dynamic participation of all voters though a market.
However, this answer neglects other important aspects of IBDD, and particularly neglects the effect of such a system over time.
If we factor time in to the above answer, we reach a more compelling answer: the source of good policy is the result of repeated conjecture (by proposing policy) and criticism (by voting against it) by those voters who participate through creative self selection whereby their criticisms are weighted through their dynamic participation in an all-inclusive market. By replacing the source of good policy is with one way to ensure good policy is continually produced and bad policy is continually challenged is through we can answer this question in a manner compatible with fallibilism, and indeed one which directly contains many core components of it.
Note: the bulk of this essay concerns the 'policy' part of Popper's Criterion, and not the 'leaders' part. This is because IBDD does not have leaders in the sense that RD does. There is no Prime Minister or President (at least not in the current formation), however, this does not mean that leaders will not arise. While I don't deal with this in the main body, I've included an addendum that discusses this concern.
IBDD as a system for correcting mistakes
A system for correcting mistakes must necessarily act over some time period, and thus a key question becomes how quickly are mistakes able to be fixed?. It is trivial to see that a slower system (RD, for example) is less useful than a faster system (which I conjecture IBDD is).
Note: this is not to say that a system which instantly enacts legislation is superior, as this itself may be a way to introduce mistakes. I am in favour of a mandatory delay period for most legislation passed in IBDD, which would help protect methods of correcting mistakes without impacting the speed at which conjecture and criticism is offered.
Popper's question How can we hope to detect and eliminate error? has been explained in general quite well by both Popper and Deutsch. Their conclusion is a cycle of conjecture and criticism (or conjecture and refutations as put by Popper).
Seeing as we've included this near verbatim in our answer to what is the source of good policy?, I will not stress that point again.
Instead, let us think about the rate at which conjecture and criticism are applied in various democratic systems.
Representative Democracy is a highly exclusive system, where the ability to provide real conjecture and criticism (by proposing new policy or blocking the implementation of new policy) is a right given to very few actors. Without being the governing party, or having a majority through an arrangement of smaller parties, it is near impossible to do much beyond making a point. Additionally, in situations like a bicameral system where the same party has a majority in both houses, the ability to propose conjecture and criticism is reserved for those in the ruling party. It is the case that a minority party may be given a chance to propose criticism when given permission by some subset of the ruling party in the case of minority defection on some issue by members of the ruling party, in the form of a temporary alliance.
There is, of course, a matter of mass conjecture and criticism when the terms of some members of either house expire, whereby the aggregate of all voters may determine some other party should be the ruling party (the second party acts as conjecture by its very existence, and voters provide the criticism by voting for it over the ruling party).
The fact that conjecture and criticism is instantiated in RD to some degree does imply the possibility of progress, but because the set of conjectures and possible criticisms are so small we should not expect this process to be efficient.
Thus to increase the rate, volume, and quality of conjecture and criticism we must remove this exclusivity and simultaneously introduce the possibility of specialisation — allowing some conjectures and criticisms to be weighted more highly than others.
Furthermore, because we cannot predict the source of the best conjectures (as new knowledge is unpredictable) we cannot exclude any voter from this process.
These two conjectures in concert imply that an efficient system must be 'direct' (as in sourced from individual voters) in some way.
This is reassuring in that it creates a sense of 'permissionlessness' around the ability to propose new conjectures and apply criticism. A property we observe in many instances, such as the progress of science and the success of some businesses over others (though not all environments for business remain equal, and so there may be some external bias applied in the economic case).
Using Markets to spur the creation of new options and good policy
Introducing a market for votes is not sufficient to provide good quality conjecture and criticism, though. It is easy to imagine some market which includes a rule like Person A is allocated 100x the resources of all other participants. It is clear that in such a case Person A has been selected as a better source of knowledge than other participants. This is a problem because it is not possible to consistently predict which sources of knowledge will be most correct (though I'm sure we can make good guesses), and to assert otherwise is prophesy or authoritarianism.
However, it is possible to embody this uncertainty in a market by weighting all sources equally, at least initially. Through a well crafted market we will see that this produces interesting incentives aligned with the production of good policy, and crucially, the creation of new options.
We also need to lay out some ground rules for this market. Particularly that suggesting some new policy requires opportunity cost. At a minimum suggesting policy should be restricted over time, such that a bad actor cannot flood the system with bad conjecture. I believe it is, however, better to unify the opportunity cost of suggesting policy with the opportunity cost of voting on policy such that they use the same resources. In this latter case there is a slight bias against the proposer of a policy, increasing the likelihood of bad policy not passing, and incentivising the creation of better policy from the get-go.
An easy way to facilitate this is through a one-to-many market, where a central liquidity token (LT) mediates value between all issues and the ability to conjecture new policy.
Furthermore, the introduction of this LT allows us to manipulate the supply, and when we apply some constant inflation (say, 30% per annum) and distribute the newly created LTs evenly among the voting population. This provides an equalising force, constantly pushing political power back towards an even distribution, and also incentivises voters to act sooner rather than later (introducing opportunity cost over time as well as conjecture of policy).
By auctioning off the right to propose new policy (in the same way votes for various issues are auctioned, except there is no recipient for LTs spent in this way) we can control the rate of conjecture and ensure we maintain some sensible political environment (whatever that means). Since the number of spots for new policy conjecture is limited, we can adjust this up or down depending on how we perceive the system to be functioning. The exact method of this adjustment is an unsolved problem at this time, though I do not believe it is that relevant to the current discussion, provided edge cases are excluded.
In order to understand this setup, let us consider some cases that may arise in a voting market.
Case 1: Bad Policy Conjectured
If a proposed policy is based on a bad explanation then its predictions will not hold. In extreme cases it is easy to see that a majority of voters will be harmed and thus they will not relinquish their vote on that issue. In this case it is probably not worth the cost of even proposing it.
However, the case also exists where most people are not harmed, and the group in favour of said bad policy (A) is strictly larger than the group to be harmed by it (B). In this case A and B will both bid on the auction for votes on this policy, and thus both expend some LTs in an effort to pass or block the proposal respectively.
In the case such a policy passes, it is in group B's interest to make a counter-conjecture, at a minimum undoing the policy put forward by group A. In order to defend against this group B will once again need to expend LTs (like group A) to defend their policy.
Group B can continue this tactic indefinitely, which eventually ends in a 'war of attrition' where both group's ability to affect change is severely limited.
If we presume that group B has some good policy (if their policy is bad also we will look at that in case 5 and 6), we can see a pattern emerge:
As long as group A attempts to maintain their bad policy they will need to expend LTs, limiting their ability to affect change in other areas. Unless their policy is so beneficial that this is worth it (which is dubious as we've already presumed group B, who are directly harmed, is small) the path to be as politically productive as possible involves not continuing to put forward this bad policy.
Case 2: Bad (but Good) Policy
There is also the case where a good explanation is used to craft policy (perhaps by a leader aligning or acquiring key supporters, and thus benefiting some group disproportionately), but it is presented under the guise of a bad policy.
This is similar to case 1, whereby the maintenance of such a policy becomes a burden to those who benefit from it. In the case of a political faction this can stress the internal relationships as it requires a contribution of opportunity cost from all involved.
If some subset of this group is able to create good policy that provides an equivalent benefit to them (without the harmful externalities of the original policy) then they are incentivised to break from the main group and put forward their policy instead. Their optimum strategy, then, is to create new knowledge (in the form of their good policy) and become independent.
In this way IBDD acts as a decentralising force on the political landscape, and encourages specialisation.
Case 3: Good Policy simpliciter
In the case that good policy is conjectured that has very little in the way of negative externalities (granted, problems are inevitable so really this is a case where those negative externalities are seen as less important than the externalities of other policies) there is little incentive for any group to oppose such a policy, as doing so removes some of their ability to change other policies that would result in a greater improvement to the group.
However, this case is probably far less likely to occur than case 4, where some group already has some bias in their favour, and a good policy would remove or reduce that bias.
Note: for the purposes of this analysis I include removing a bad policy as a good policy roughly equivalent to don't implement this bad policy because this good explanation shows it is harmful.
Case 4: Good Policy that harms some group
A good example of this might be increasing or reducing taxation on some strata of society. It is easy to conceive of a taxation system that unfairly biases some group, and a good policy may be to reduce or increase some tax bracket.
In this case we see a similar pattern to case 2 (Bad (but Good) Policy). The group that unfairly benefits (A) is incentivised to defend the past policy that biased them, and the group that is harmed (B) is incentivised to put forward some good solution.
A similar equilibrium exists, whereby the continued defence of bad policy by group A restricts their ability to affect change. In the issue of taxation this does include a monetary component (though you can argue all policy includes some economic component) and so it might be a little 'stickier' than other issues.
If policy just goes back and forward in a manner such as adjusting tax brackets it's easy to see this continuing forever in some kind of back-and-forth rally, or possibly converging (in the manner of a compromise) to some middle ground. But, this does not preclude the possibility that there is some better solution out there, and incentivises its creation in the same way as case 2.
Note: I personally suspect a far superior method of taxation exists than income/sales tax where tax is leveraged implicitly through deliberate inflation, and prices are expressed in terms of percentage points of M1. This is an example of a possible conjecture that would break a taxation rally as described above (provided it really is a good policy).
However, if the conjectured policy really is good the equilibrium lies in favour of the good policy. That is group A has less incentive to criticise the policy, and group B has every incentive to conjecture it. The case for both groups is that passing such a policy results in a net increase in their ability to affect other issues.
Case 5: Two opposing sides, both with Bad Policy, unwilling to cooperate
In the case where two opposing sides are competing for control of some policy area (environmentalists vs coal companies might be an easy example) they might both put forward bad policy.
The result of such a conflict (as explored above) is an expensive stalemate, due to the perpetual need to acquire votes to blockade the opponents, where neither group really gets what they want, and their ability to act on other issues is diminished.
Thus we would expect to see both groups either with very few, or quite a lot of LTs. One possibility is an active standoff, where each group continually puts forward their policy and we oscillate between them, and the other is a passive standoff where both groups stockpile LTs, ready for the 'final confrontation'.
In both possibilities it is clearly not an optimum strategy for the groups.
Case 6: Two opposing sides, both with Bad Policy, but willing to cooperate
If we take case 5, but add in their willingness to cooperate, then they are likely going to either seek a compromise or create some new option.
In the case of compromise they might manage to pass a new policy, but neither side will really be happy about it, and so there's always an incentive to resume the standoff in ambition of a more beneficial policy.
However, if the two groups have had a passive standoff and found some new option which is a good policy, they are now unified as one force, and both have substantial stockpiles of LTs they can use to further their agendas in new ways.
Thus the optimum strategy for groups in this position is to focus on creating new options, and removing the bad policy which came before it.
Case Conclusions
In each of the above cases we observe that the optimum strategy involves creating new options, cooperating with other groups, and producing good policy.
In answer to Popper's two questions:
How can we organise our political institutions so that bad or incompetent rulers cannot do too much damage?
How can we hope to detect and eliminate error?
We can see now that IBDD answers both quite well. We are able to avoid damage by introducing constant criticism and aligning incentives in the direction of preventing damage (and producing good policy, which should lead to progress and prosperity), and the detection and elimination of errors is enabled by direct participation, and incentivised via market dynamics.
Popper's Criterion and IBDD as a system - what happens when IBDD becomes the mistake
There is, of course, a final matter we must deal with. Because problems are inevitable, it is necessary that even if IBDD is far superior to our current systems of governance, it will one day become a prescient problem, and voters will once again need to support some new system.
I do not think the method determining and implementing that new system needs to be discussed here, as IBDD applies to a wide and diverse set of possible voters, and thus there will be many instances of it, it is entirely reasonable to presume that once some new system is tested somewhere and appears more desirable it will be rapidly implemented elsewhere. What such a system will look like is, of course, unpredictable, though I suspect it may better enshrine Fallibilism than IBDD does.
The particular thing we must discuss here is the difficulty of replacing various systems from within without violence.
Let us modify Deutsch's definition of Poppers Criterion such that it applies to systems:
Good political institutions are those that make it as easy as possible to detect deficiencies within themselves, and to remove those deficiencies, improve themselves, or replace themselves without violence when they exist.
I do not see any reason that this should be an invalid statement, or subject to some criticism that Popper's Criterion is not.
When viewed in this light, representative democracy does not fair well. If it did, democracies around the world would converge to a common implementation.
Part of the problem with RD is that such a change requires a referendum or overwhelming consensus in most cases, and additionally that the consent of the representatives who may be harmed by such an improvement is required. In the best case a referendum must be won, and so our best threshold for RD is the smallest set of voters larger than 50% of the voting population.
We can therefore predict that a system which fulfils Popper's Criterion better than RD should enable modification to the system with less than 50% of the voter base. Similar to some of the cases above, I predict a modification to IBDD constructed via a bad explanation should not fare well within IBDD.
However, in the case that such a modification would really improve IBDD, then the threshold for that improvement should be far lower than 50%, as it is good policy, and therefore those who agree with it are not incentivised to participate except in the case it may fail to pass.
In this way, error correction is so strongly embedded in IBDD that we should expect IBDD to exclusively improve itself more efficiently than the swap from whatever came before IBDD.
I know of no better way to satisfy Popper's Criterion than this.
As a final note, it is possible that IBDD does not act as predicted, and that returning to RD will result in better policy. However, even in that case, we see that at worst IBDD has the same threshold to change as the best case for RD. Thus, if IBDD does turn out to be a mistake, it will be no more difficult to change back, as it was to make the change in the first place. However, I doubt anything as extreme as this will come to pass. It's unlikely that IBDD would reach dominance without being criticised fatally, if such a criticism exists. Given it is likely to first appear in small communities (such as municipal governments) or hold the balance of power in some legislature, any such fatal flaws will likely come to the fore far earlier than a complete transition into IBDD could occur.
Conclusion
We've now seen that not only does IBDD satisfy Popper's Criterion (by providing mechanisms to remove bad policy and bad systems without violence), but also embodies many of the core philosophical conclusions Popper himself came to when thinking about the problem of progress, authority, and knowledge.
In this way, it is possible to view IBDD as an embodiment of Popper's Criterion: a system which takes the removal of bad policy and systems so seriously as to optimise for this effect.
Through the dynamic redistribution of political power we diversify groups, spurring specialisation and breaking down factions. This diversification then also increases the ability of any group to criticise established policy and affect its removal.
Furthermore, by carefully constructing an egalitarian market for political power, we actively incentivise the creation of new options by increasing the value of new conjecture.
Finally, due to the above properties we predict IBDD should improve itself far more effectively than any system it replaces, and will be replaced by a system that excels even further in this manner.
It occurs to me that the transition from RD to IBDD should mirror the transition from anti-rational memes to rational memes. It was not that anti-rational memes prevented progress absolutely, but that they were far less powerful at producing progress than rational memes (including by restricting the manner and form of such progress), and the groups which adopted rational memes progressed faster and became far more resilient. Although rational memes still require constant defence (as I imagine IBDD will to), they are sufficiently embedded in western philosophy that an immense change would be needed to extinguish them - unless that change were an improvement, of course.
Is it possible that IBDD embeds the values and explanations of our current rational memes more deeply than other forms of democracy? It is certainly the case that the rational meme of Fallibilism has compelled this author to design, improve, and implement IBDD.
It is my hope that IBDD is able to prove itself, is able to persist once instantiated, and that it is able to help spread rational memes. For if IBDD works, people will attempt to understand it, and if they do they will undoubtably come up against the rational meme of Fallibilism, and at such time they might themselves find that it compels them into action.
Addendum: Leaders in IBDD
It’s important to note ... that IBDD is not a system which denies the importance of leadership amongst human societies, but rather, embraces it. IBDD allows leaders to be tested in a way where the best continuously rise to the top, while poor leaders are quickly removed . Interestingly, we start to select [19] leaders based on different criteria. Presently representatives are often chosen for characteristics such as charisma, tenacity, and their promises. Since the challenges faced by leaders under representative democracy are not present in IBDD, they can be selected by a far more significant criterion: their ability to create good policy.
Footnote 19: The ability to remove bad leaders without violence is not specific to IBDD, but is certainly enhanced. Karl Popper has argued that this ability is what has allowed democracies to progress and remain stable in spite of the authority held by Government.
Kaye & Spataro, Redefining Democracy (2017)
I do not envision IBDD having leaders in the same way our old political systems required leaders. Rather, leaders will naturally arise through their ability to affect change, and as we have seen in a system like IBDD this requires producing good policy.
It may seems slightly circular, then, to apply Popper's Criterion to leaders in this case, as they are by definition not bad. Those aspiring leaders that are bad are heavily restricted via market mechanisms from ever making much of an impact.
However, it's easy to imagine (and wise to presume that) a leader may arise through the production of good policy, and then attempt to abuse their position, in effect transitioning into a bad leader, and, by Popper's Criterion, we must have an effective way to remove them without violence.
I predict this will unfold in a similar way to the cases of bad policy we considered earlier:
A leader who was previously good, but has now become bad, will have a track record of producing good policy. However, the new policy she produces will not adhere to this track record.
The only way in IBDD to obtain far greater political power than the average voter is to convince voters to delegate to you. Because voters are free to change delegates at any time (or revoke it entirely), leaders will need to continue providing some benefit to the voter — if they do not, there is every incentive for the voter to select a new delegate.
In this case, the act of becoming a bad leader actively diminishes that leader's ability to affect change, totally satisfying Popper's question 'How can we organise our political institutions so that bad or incompetent rulers cannot do too much damage?'
However, there is the case where a group may follow a leader without requiring progress from them. Examples of this might be extremist groups, interested in only their ideology, and prevented from accepting new options due to anti-rational memes.
In such a case, it is unfortunate but a reality of IBDD that they are a near-permanent inefficiency (I say near-permanent as they may have some capacity to produce good policy, and over time their anti-rational memes should weaken). However, if this opposition really does embody an anti-rational meme (which it probably will if IBDD acts as predicted) then the problem of convincing them to abandon a particular anti-rational meme is the same problem we've faced since the enlightenment.
There is, of course, the case where such a group has overwhelming power, and uses this to enforce their anti-rational memes. However, such a problem, just as before, is not unique to IBDD. Its potential is a constant in all societies, and all democracies. Critically, though, it would suffer from the same problem of all static societies — suppressing the ability to create the right knowledge to solve problems. As it stands, I consider this scenario no less likely in RD as IBDD, and crucially, the meme of IBDD works harder to propagate rational memes than RD does (though that may just be my bias or perspective).
Addendum: A Hypothesis from the Selectorate Theory
While investigating how selectorate theory, IBDD, and Fallibilism interact I noticed an extension of a pattern in selectorate theory.
As states transition from dictators to democracies, a key factor is the increased industrialisation of that state. Such progress necessitates new key supporters (perhaps in new industries), and as the number of key supporters grows dictatorships become less stable.
In the end, the only choice is to transition into democracy, though that doesn't guarantee the democracy will be stable at that point.
However, we are also witnessing a growing discontent with how democracy is done today. (Redefining Democracy covers this in greater detail). While transitioning back to dictatorship cannot support more key supporters (keys) and so it's entirely reasonable to conjecture that these keys are looking for a solution better able to accomodate them.
Due to IBDD's ability to incentivise the creation of new knowledge, it seems likely that IBDD possesses a far greater potential for satisfaction of keys, and we may, for the first time in human history, witness a transition from centralised leadership, where the leader wrangles the keys, into a decentralised leadership where keys are supported by the very system itself.
From this line of reasoning I created a hypothesis. I believe that if this hypothesis is true (for the most part) IBDD should inevitably succeed over RD, since it is predicted to support more keys.
That hypothesis is:
The system of governance that eventually dominates is the one able to accomodate the most key supporters.
Whether this is true or not remains to be seen, but it certainly feels like it has merit.
Addendum: Quadratic Voting and Popper's Criterion
I mention Quadratic Voting (QV) briefly in the above essay, and have mentioned it in several other places, but I have not yet actually formed an argument as to why QV does not satisfy Popper's criterion as well as IBDD.
The setup for QV is this:
Votes for every issue are available to all voters
Voters start with 0 votes
Voters may exchange money (which goes into a pool) for votes
The price per vote for each voter depends on how many they purchase, this schedule increases quadratically (e.g. $1 for 1 vote, $4 for 2, $9 for 3, etc).
Votes are cast as normal
The pool of money is then either split equally between all voters, or distributed to the entire voter base
I have a few problems with QV. There are also some points I think make it superior in particular ways to RD, DD, and LD.
One advantage of QV is that it breaks down majoritarianism and allows for specialisation.
However, this specialisation is far more available to wealthy voters than to the vast majority of voters.
Furthermore, in the same way that splitting Bill Gates' $88 billion fortune between all people on Earth would only result in $12.57 per person, socialising the result of a quadratic vote does not substantially redistribute wealth from the point of view of other voters. It does not substantially increase their ability to participate.
A related problem is that by splitting a kitty between multiple voters you obtain far cheaper votes, even if this is only between 2 people. $2y*2y for 2y votes is twice $y*y + $y*y for 2y votes. Thus there are potential economic attacks.
QV is good at ensuring that (when dealing with equally wealthy advocates) the proposal with wider support wins, and good at preventing minority rule to some degree, but this comes at the cost of decreased specialisation, and thus a diminished capacity to effectively use the knowledge spread throughout society.
IBDD has some similar properties, but also has (in my opinion) some vastly superior properties. For example, in IBDD the cost (in liquidity tokens) to acquire 1 vote in any given issue is the same regardless of how many you'd like to acquire (presuming it's not so much that supply and demand comes into play, but each vote is still just as valuable as each other vote regardless of the acquiring party).
Additionally, because IBDD uses a closed economy cut off from real world wealth, it is much more difficult for the ultra-wealthy to manipulate the outcome in their favour.
Now, QV does satisfy Popper's criterion in some cases, it's definitely possible to remove bad policy without violence.
However, it may also be possible for QV to violate it. Consider the case where an ultra-wealthy individual employs many thousands of people to take part on their behalf. These workers buy votes at a vastly reduced rate (compared with our wealthy person purchasing it all themselves).
The effect of this is a mostly economical attack to prevent the removal of bad policy. In addition to less money being included to begin with, the socialisation of proceeds also is spread far more in favour of our rich attacker, and so their risk and overheads are lower.
While this does still result in some redistribution of wealth in the direction of the dissenters, we've set up an economic equilibrium between the benefit of the policy and the cost of maintaining it. It is therefore theoretically possible such a system would prevent the removal of that policy indefinitely, or at least long enough for it to matter.
In the case of IBDD, however, gathering such a following requires opportunity cost on the follower's behalf, and additionally reduces the ability for the attacker to interact with other legislative processes. Thus I consider IBDD to more effectively solve this problem than QV.
Addendum: IBDD and Budgets
This is an unsolved problem.
One possible solution is simply to average the opinion of all voters. As far as 'wisdom of the crowd' goes, this doesn't seem too bad a place to start.
It does not seem like a budget is the sort of thing one person or source would be able to produce well consistently, and competing budgets aren't guaranteed to account for everything as well as we may like.
Is it appropriate for budget requirements might be attached to policy and automatically granted when the policy passes? Could we design some taxation system to facilitate this?
I am sure that if IBDD begins to succeed this problem will receive attention, and all problems are soluble.
Addendum: IBDD and War
This is also an unsolved problem. War is never a simple thing, and seems to lend itself more to centrally controlled systems than decentralised ones.
The naive approach seems to suggest a referendum, and certainly many people would like it if their governments were required to hold a referendum before going to war, but it's not clear to me that this would necessarily produce better or worse results.
In the limit, IBDD is predicted to unify political systems into a global system of governance (note: not a global government). Perhaps it is the case that this will never need to be solved? What if we later encounter some outside threat we did not predict, and then don't have the capacity to deal with it?
I believe this problem to be soluble.
Additional Reading
The Beginning of Infinity, David Deutsch, 2011
The Dictator's Handbook, Bruce Bueno De Mesquita, 2012
Rules for Rulers, an excellent summary of the selectorate theory by CGP Grey, approximately 20 minutes long, Youtube
In 1960 Popper published his paper Knowledge without Authority where he conjectured that the central question of political theory should not be but something entirely different.
What he put forward has become known as Popper's Criterion - a criterion I believe IBDD satisfies with flying colours.
Popper's Criterion
In response to the question, Popper writes:
There are two important political questions here:
Question 1 has morphed to become known as Popper's Criterion:
There are two important ways we must consider IBDD in the context of Popper's Criterion. The first regards the nature of authority and how IBDD is fundamentally anti-authority, and the second is a direct application where we investigate exactly how well IBDD is able to 'detect whether a ... policy is a mistake, and to remove ... policies without violence when they are'.
I would also like to extend Popper's Criterion to include the system of decision making itself. If we are unable to change the underlying system without violence — seeing as the system itself is a policy — then such a system must fail Popper's Criterion.
Authority and IBDD
A note on the use of 'good' and 'bad' when describing policy and explanations: I am using these words in the same manner as Deutsch uses them in The Beginning of Infinity. He does a great job of explaining them in detail, so I will summarise: good explanations are those which are hard to vary and account for the phenomena they purport to. Not all good explanations are the most correct, but the most correct explanations must be good. Since policy is a way to take an explanation relating to some problem, and provide instructions on how to solve it, policy inherits this property. Policy which is based on a bad explanation is bad policy, and vice versa for good policy.
Given that 'who should rule?' is 'who or what is the authoritative source of good policy?' in another form, the concept of authority plays a strong role in current political systems.
At a fundamental level Issue Based Direct Democracy (IBDD) attempts to acknowledge that there is no authority on good policy (or knowledge). In competing democratic systems, the answer is often in the form:
While IBDD is a form of direct democracy, we do not claim that any one source of policy is particularly better than another. Rather, we reason through the process a little differently.
While majoritarian systems (RD, DD, LD) do incorporate many sources into a ruling faction — either through a single party or a coalition of entities — they are tightly bound, and without political disturbance, or specific arrangements of representatives in a legislature, they are inextricable from one another. Often this collection of knowledge is referred to as a .
As explained in the Flux Philosophy White paper Redefining Democracy, the reason for this centralisation is due to the fundamental nature of majoritarian democracy.
Thus, in order to treat solutions to social problems independently we cannot have a system ignorant of distribution of knowledge in society. That is to say it must weight different sources of knowledge differently.
The selectorate theory of political power pioneered by Bueno de Mesquita et al. teaches us that the current weights on sources of knowledge is determined by political leaders and their need to satisfy key supporters (the selectorate). However, biasing sources of knowledge needed to keep a leader in power does not strongly bias good explanation. This is accounts for why these sources are largely inextricable from one another in such systems, and why there are fundamental limits on the speed of progress of RD, LD, and DD.
The problem we are concerned with now becomes and it is to this question that IBDD has a direct answer.
IBDD delivers its answer in the form of a specially crafted persistent and inclusive market. This allows the distribution of political expression on any one issue (or rather, on a particular policy conjectured to solve that issue) to be decided by the aggregate input of all voters considering that conjecture in the context of all present and future issues.
This allows voters to individually weigh the value of their conjectures and criticisms by their perceived benefit at a given time in a highly dynamic way.
It is possible, at this point, to argue that this is just authority in another form: that really we are answering the question of in the form .
However, this answer neglects other important aspects of IBDD, and particularly neglects the effect of such a system over time.
If we factor time in to the above answer, we reach a more compelling answer: . By replacing with we can answer this question in a manner compatible with fallibilism, and indeed one which directly contains many core components of it.
Note: the bulk of this essay concerns the 'policy' part of Popper's Criterion, and not the 'leaders' part. This is because IBDD does not have leaders in the sense that RD does. There is no Prime Minister or President (at least not in the current formation), however, this does not mean that leaders will not arise. While I don't deal with this in the main body, I've included an addendum that discusses this concern.
IBDD as a system for correcting mistakes
A system for correcting mistakes must necessarily act over some time period, and thus a key question becomes . It is trivial to see that a slower system (RD, for example) is less useful than a faster system (which I conjecture IBDD is).
Note: this is not to say that a system which instantly enacts legislation is superior, as this itself may be a way to introduce mistakes. I am in favour of a mandatory delay period for most legislation passed in IBDD, which would help protect methods of correcting mistakes without impacting the speed at which conjecture and criticism is offered.
Popper's question has been explained in general quite well by both Popper and Deutsch. Their conclusion is a cycle of conjecture and criticism (or as put by Popper).
Seeing as we've included this near verbatim in our answer to , I will not stress that point again.
Instead, let us think about the rate at which conjecture and criticism are applied in various democratic systems.
Representative Democracy is a highly exclusive system, where the ability to provide real conjecture and criticism (by proposing new policy or blocking the implementation of new policy) is a right given to very few actors. Without being the governing party, or having a majority through an arrangement of smaller parties, it is near impossible to do much beyond making a point. Additionally, in situations like a bicameral system where the same party has a majority in both houses, the ability to propose conjecture and criticism is reserved for those in the ruling party. It is the case that a minority party may be given a chance to propose criticism when given permission by some subset of the ruling party in the case of minority defection on some issue by members of the ruling party, in the form of a temporary alliance.
There is, of course, a matter of mass conjecture and criticism when the terms of some members of either house expire, whereby the aggregate of all voters may determine some other party should be the ruling party (the second party acts as conjecture by its very existence, and voters provide the criticism by voting for it over the ruling party).
The fact that conjecture and criticism is instantiated in RD to some degree does imply the possibility of progress, but because the set of conjectures and possible criticisms are so small we should not expect this process to be efficient.
Thus to increase the rate, volume, and quality of conjecture and criticism we must remove this exclusivity and simultaneously introduce the possibility of specialisation — allowing some conjectures and criticisms to be weighted more highly than others.
Furthermore, because we cannot predict the source of the best conjectures (as new knowledge is unpredictable) we cannot exclude any voter from this process.
These two conjectures in concert imply that an efficient system must be 'direct' (as in sourced from individual voters) in some way.
This is reassuring in that it creates a sense of 'permissionlessness' around the ability to propose new conjectures and apply criticism. A property we observe in many instances, such as the progress of science and the success of some businesses over others (though not all environments for business remain equal, and so there may be some external bias applied in the economic case).
Using Markets to spur the creation of new options and good policy
Introducing a market for votes is not sufficient to provide good quality conjecture and criticism, though. It is easy to imagine some market which includes a rule like . It is clear that in such a case Person A has been selected as a better source of knowledge than other participants. This is a problem because it is not possible to consistently predict which sources of knowledge will be most correct (though I'm sure we can make good guesses), and to assert otherwise is prophesy or authoritarianism.
However, it is possible to embody this uncertainty in a market by weighting all sources equally, at least initially. Through a well crafted market we will see that this produces interesting incentives aligned with the production of good policy, and crucially, the creation of new options.
We also need to lay out some ground rules for this market. Particularly that suggesting some new policy requires opportunity cost. At a minimum suggesting policy should be restricted over time, such that a bad actor cannot flood the system with bad conjecture. I believe it is, however, better to unify the opportunity cost of suggesting policy with the opportunity cost of voting on policy such that they use the same resources. In this latter case there is a slight bias against the proposer of a policy, increasing the likelihood of bad policy not passing, and incentivising the creation of better policy from the get-go.
An easy way to facilitate this is through a one-to-many market, where a central liquidity token (LT) mediates value between all issues and the ability to conjecture new policy.
Furthermore, the introduction of this LT allows us to manipulate the supply, and when we apply some constant inflation (say, 30% per annum) and distribute the newly created LTs evenly among the voting population. This provides an equalising force, constantly pushing political power back towards an even distribution, and also incentivises voters to act sooner rather than later (introducing opportunity cost over time as well as conjecture of policy).
By auctioning off the right to propose new policy (in the same way votes for various issues are auctioned, except there is no recipient for LTs spent in this way) we can control the rate of conjecture and ensure we maintain some sensible political environment (whatever that means). Since the number of spots for new policy conjecture is limited, we can adjust this up or down depending on how we perceive the system to be functioning. The exact method of this adjustment is an unsolved problem at this time, though I do not believe it is that relevant to the current discussion, provided edge cases are excluded.
In order to understand this setup, let us consider some cases that may arise in a voting market.
Case 1: Bad Policy Conjectured
If a proposed policy is based on a bad explanation then its predictions will not hold. In extreme cases it is easy to see that a majority of voters will be harmed and thus they will not relinquish their vote on that issue. In this case it is probably not worth the cost of even proposing it.
However, the case also exists where most people are not harmed, and the group in favour of said bad policy (A) is strictly larger than the group to be harmed by it (B). In this case A and B will both bid on the auction for votes on this policy, and thus both expend some LTs in an effort to pass or block the proposal respectively.
In the case such a policy passes, it is in group B's interest to make a counter-conjecture, at a minimum undoing the policy put forward by group A. In order to defend against this group B will once again need to expend LTs (like group A) to defend their policy.
Group B can continue this tactic indefinitely, which eventually ends in a 'war of attrition' where both group's ability to affect change is severely limited.
If we presume that group B has some good policy (if their policy is bad also we will look at that in case 5 and 6), we can see a pattern emerge:
As long as group A attempts to maintain their bad policy they will need to expend LTs, limiting their ability to affect change in other areas. Unless their policy is so beneficial that this is worth it (which is dubious as we've already presumed group B, who are directly harmed, is small) the path to be as politically productive as possible involves not continuing to put forward this bad policy.
Case 2: Bad (but Good) Policy
There is also the case where a good explanation is used to craft policy (perhaps by a leader aligning or acquiring key supporters, and thus benefiting some group disproportionately), but it is presented under the guise of a bad policy.
This is similar to case 1, whereby the maintenance of such a policy becomes a burden to those who benefit from it. In the case of a political faction this can stress the internal relationships as it requires a contribution of opportunity cost from all involved.
If some subset of this group is able to create good policy that provides an equivalent benefit to them (without the harmful externalities of the original policy) then they are incentivised to break from the main group and put forward their policy instead. Their optimum strategy, then, is to create new knowledge (in the form of their good policy) and become independent.
In this way IBDD acts as a decentralising force on the political landscape, and encourages specialisation.
Case 3: Good Policy simpliciter
In the case that good policy is conjectured that has very little in the way of negative externalities (granted, problems are inevitable so really this is a case where those negative externalities are seen as less important than the externalities of other policies) there is little incentive for any group to oppose such a policy, as doing so removes some of their ability to change other policies that would result in a greater improvement to the group.
However, this case is probably far less likely to occur than case 4, where some group already has some bias in their favour, and a good policy would remove or reduce that bias.
Note: for the purposes of this analysis I include as a good policy roughly equivalent to .
Case 4: Good Policy that harms some group
A good example of this might be increasing or reducing taxation on some strata of society. It is easy to conceive of a taxation system that unfairly biases some group, and a good policy may be to reduce or increase some tax bracket.
In this case we see a similar pattern to case 2 (Bad (but Good) Policy). The group that unfairly benefits (A) is incentivised to defend the past policy that biased them, and the group that is harmed (B) is incentivised to put forward some good solution.
A similar equilibrium exists, whereby the continued defence of bad policy by group A restricts their ability to affect change. In the issue of taxation this does include a monetary component (though you can argue all policy includes some economic component) and so it might be a little 'stickier' than other issues.
If policy just goes back and forward in a manner such as adjusting tax brackets it's easy to see this continuing forever in some kind of back-and-forth rally, or possibly converging (in the manner of a compromise) to some middle ground. But, this does not preclude the possibility that there is some better solution out there, and incentivises its creation in the same way as case 2.
Note: I personally suspect a far superior method of taxation exists than income/sales tax where tax is leveraged implicitly through deliberate inflation, and prices are expressed in terms of percentage points of M1. This is an example of a possible conjecture that would break a taxation rally as described above (provided it really is a good policy).
However, if the conjectured policy really is good the equilibrium lies in favour of the good policy. That is group A has less incentive to criticise the policy, and group B has every incentive to conjecture it. The case for both groups is that passing such a policy results in a net increase in their ability to affect other issues.
Case 5: Two opposing sides, both with Bad Policy, unwilling to cooperate
In the case where two opposing sides are competing for control of some policy area (environmentalists vs coal companies might be an easy example) they might both put forward bad policy.
The result of such a conflict (as explored above) is an expensive stalemate, due to the perpetual need to acquire votes to blockade the opponents, where neither group really gets what they want, and their ability to act on other issues is diminished.
Thus we would expect to see both groups either with very few, or quite a lot of LTs. One possibility is an active standoff, where each group continually puts forward their policy and we oscillate between them, and the other is a passive standoff where both groups stockpile LTs, ready for the 'final confrontation'.
In both possibilities it is clearly not an optimum strategy for the groups.
Case 6: Two opposing sides, both with Bad Policy, but willing to cooperate
If we take case 5, but add in their willingness to cooperate, then they are likely going to either seek a compromise or create some new option.
In the case of compromise they might manage to pass a new policy, but neither side will really be happy about it, and so there's always an incentive to resume the standoff in ambition of a more beneficial policy.
However, if the two groups have had a passive standoff and found some new option which is a good policy, they are now unified as one force, and both have substantial stockpiles of LTs they can use to further their agendas in new ways.
Thus the optimum strategy for groups in this position is to focus on creating new options, and removing the bad policy which came before it.
Case Conclusions
In each of the above cases we observe that the optimum strategy involves creating new options, cooperating with other groups, and producing good policy.
In answer to Popper's two questions:
We can see now that IBDD answers both quite well. We are able to avoid damage by introducing constant criticism and aligning incentives in the direction of preventing damage (and producing good policy, which should lead to progress and prosperity), and the detection and elimination of errors is enabled by direct participation, and incentivised via market dynamics.
Popper's Criterion and IBDD as a system - what happens when IBDD becomes the mistake
There is, of course, a final matter we must deal with. Because problems are inevitable, it is necessary that even if IBDD is far superior to our current systems of governance, it will one day become a prescient problem, and voters will once again need to support some new system.
I do not think the method determining and implementing that new system needs to be discussed here, as IBDD applies to a wide and diverse set of possible voters, and thus there will be many instances of it, it is entirely reasonable to presume that once some new system is tested somewhere and appears more desirable it will be rapidly implemented elsewhere. What such a system will look like is, of course, unpredictable, though I suspect it may better enshrine Fallibilism than IBDD does.
The particular thing we must discuss here is the difficulty of replacing various systems from within without violence.
Let us modify Deutsch's definition of Poppers Criterion such that it applies to systems:
I do not see any reason that this should be an invalid statement, or subject to some criticism that Popper's Criterion is not.
When viewed in this light, representative democracy does not fair well. If it did, democracies around the world would converge to a common implementation.
Part of the problem with RD is that such a change requires a referendum or overwhelming consensus in most cases, and additionally that the consent of the representatives who may be harmed by such an improvement is required. In the best case a referendum must be won, and so our best threshold for RD is the smallest set of voters larger than 50% of the voting population.
We can therefore predict that a system which fulfils Popper's Criterion better than RD should enable modification to the system with less than 50% of the voter base. Similar to some of the cases above, I predict a modification to IBDD constructed via a bad explanation should not fare well within IBDD.
However, in the case that such a modification would really improve IBDD, then the threshold for that improvement should be far lower than 50%, as it is good policy, and therefore those who agree with it are not incentivised to participate except in the case it may fail to pass.
In this way, error correction is so strongly embedded in IBDD that we should expect IBDD to exclusively improve itself more efficiently than the swap from whatever came before IBDD.
I know of no better way to satisfy Popper's Criterion than this.
As a final note, it is possible that IBDD does not act as predicted, and that returning to RD will result in better policy. However, even in that case, we see that at worst IBDD has the same threshold to change as the best case for RD. Thus, if IBDD does turn out to be a mistake, it will be no more difficult to change back, as it was to make the change in the first place. However, I doubt anything as extreme as this will come to pass. It's unlikely that IBDD would reach dominance without being criticised fatally, if such a criticism exists. Given it is likely to first appear in small communities (such as municipal governments) or hold the balance of power in some legislature, any such fatal flaws will likely come to the fore far earlier than a complete transition into IBDD could occur.
Conclusion
We've now seen that not only does IBDD satisfy Popper's Criterion (by providing mechanisms to remove bad policy and bad systems without violence), but also embodies many of the core philosophical conclusions Popper himself came to when thinking about the problem of progress, authority, and knowledge.
In this way, it is possible to view IBDD as an embodiment of Popper's Criterion: a system which takes the removal of bad policy and systems so seriously as to optimise for this effect.
Through the dynamic redistribution of political power we diversify groups, spurring specialisation and breaking down factions. This diversification then also increases the ability of any group to criticise established policy and affect its removal.
Furthermore, by carefully constructing an egalitarian market for political power, we actively incentivise the creation of new options by increasing the value of new conjecture.
Finally, due to the above properties we predict IBDD should improve itself far more effectively than any system it replaces, and will be replaced by a system that excels even further in this manner.
It occurs to me that the transition from RD to IBDD should mirror the transition from anti-rational memes to rational memes. It was not that anti-rational memes prevented progress absolutely, but that they were far less powerful at producing progress than rational memes (including by restricting the manner and form of such progress), and the groups which adopted rational memes progressed faster and became far more resilient. Although rational memes still require constant defence (as I imagine IBDD will to), they are sufficiently embedded in western philosophy that an immense change would be needed to extinguish them - unless that change were an improvement, of course.
Is it possible that IBDD embeds the values and explanations of our current rational memes more deeply than other forms of democracy? It is certainly the case that the rational meme of Fallibilism has compelled this author to design, improve, and implement IBDD.
It is my hope that IBDD is able to prove itself, is able to persist once instantiated, and that it is able to help spread rational memes. For if IBDD works, people will attempt to understand it, and if they do they will undoubtably come up against the rational meme of Fallibilism, and at such time they might themselves find that it compels them into action.
Addendum: Leaders in IBDD
I do not envision IBDD having leaders in the same way our old political systems required leaders. Rather, leaders will naturally arise through their ability to affect change, and as we have seen in a system like IBDD this requires producing good policy.
It may seems slightly circular, then, to apply Popper's Criterion to leaders in this case, as they are by definition not bad. Those aspiring leaders that are bad are heavily restricted via market mechanisms from ever making much of an impact.
However, it's easy to imagine (and wise to presume that) a leader may arise through the production of good policy, and then attempt to abuse their position, in effect transitioning into a bad leader, and, by Popper's Criterion, we must have an effective way to remove them without violence.
I predict this will unfold in a similar way to the cases of bad policy we considered earlier:
A leader who was previously good, but has now become bad, will have a track record of producing good policy. However, the new policy she produces will not adhere to this track record.
The only way in IBDD to obtain far greater political power than the average voter is to convince voters to delegate to you. Because voters are free to change delegates at any time (or revoke it entirely), leaders will need to continue providing some benefit to the voter — if they do not, there is every incentive for the voter to select a new delegate.
In this case, the act of becoming a bad leader actively diminishes that leader's ability to affect change, totally satisfying Popper's question 'How can we organise our political institutions so that bad or incompetent rulers cannot do too much damage?'
However, there is the case where a group may follow a leader without requiring progress from them. Examples of this might be extremist groups, interested in only their ideology, and prevented from accepting new options due to anti-rational memes.
In such a case, it is unfortunate but a reality of IBDD that they are a near-permanent inefficiency (I say near-permanent as they may have some capacity to produce good policy, and over time their anti-rational memes should weaken). However, if this opposition really does embody an anti-rational meme (which it probably will if IBDD acts as predicted) then the problem of convincing them to abandon a particular anti-rational meme is the same problem we've faced since the enlightenment.
There is, of course, the case where such a group has overwhelming power, and uses this to enforce their anti-rational memes. However, such a problem, just as before, is not unique to IBDD. Its potential is a constant in all societies, and all democracies. Critically, though, it would suffer from the same problem of all static societies — suppressing the ability to create the right knowledge to solve problems. As it stands, I consider this scenario no less likely in RD as IBDD, and crucially, the meme of IBDD works harder to propagate rational memes than RD does (though that may just be my bias or perspective).
Addendum: A Hypothesis from the Selectorate Theory
While investigating how selectorate theory, IBDD, and Fallibilism interact I noticed an extension of a pattern in selectorate theory.
As states transition from dictators to democracies, a key factor is the increased industrialisation of that state. Such progress necessitates new key supporters (perhaps in new industries), and as the number of key supporters grows dictatorships become less stable.
In the end, the only choice is to transition into democracy, though that doesn't guarantee the democracy will be stable at that point.
However, we are also witnessing a growing discontent with how democracy is done today. (Redefining Democracy covers this in greater detail). While transitioning back to dictatorship cannot support more key supporters (keys) and so it's entirely reasonable to conjecture that these keys are looking for a solution better able to accomodate them.
Due to IBDD's ability to incentivise the creation of new knowledge, it seems likely that IBDD possesses a far greater potential for satisfaction of keys, and we may, for the first time in human history, witness a transition from centralised leadership, where the leader wrangles the keys, into a decentralised leadership where keys are supported by the very system itself.
From this line of reasoning I created a hypothesis. I believe that if this hypothesis is true (for the most part) IBDD should inevitably succeed over RD, since it is predicted to support more keys.
That hypothesis is:
Whether this is true or not remains to be seen, but it certainly feels like it has merit.
Addendum: Quadratic Voting and Popper's Criterion
I mention Quadratic Voting (QV) briefly in the above essay, and have mentioned it in several other places, but I have not yet actually formed an argument as to why QV does not satisfy Popper's criterion as well as IBDD.
The setup for QV is this:
I have a few problems with QV. There are also some points I think make it superior in particular ways to RD, DD, and LD.
One advantage of QV is that it breaks down majoritarianism and allows for specialisation.
However, this specialisation is far more available to wealthy voters than to the vast majority of voters.
Furthermore, in the same way that splitting Bill Gates' $88 billion fortune between all people on Earth would only result in $12.57 per person, socialising the result of a quadratic vote does not substantially redistribute wealth from the point of view of other voters. It does not substantially increase their ability to participate.
A related problem is that by splitting a kitty between multiple voters you obtain far cheaper votes, even if this is only between 2 people. $
2y*2yfor 2y votes is twice $y*y + $y*yfor 2y votes. Thus there are potential economic attacks.QV is good at ensuring that (when dealing with equally wealthy advocates) the proposal with wider support wins, and good at preventing minority rule to some degree, but this comes at the cost of decreased specialisation, and thus a diminished capacity to effectively use the knowledge spread throughout society.
IBDD has some similar properties, but also has (in my opinion) some vastly superior properties. For example, in IBDD the cost (in liquidity tokens) to acquire 1 vote in any given issue is the same regardless of how many you'd like to acquire (presuming it's not so much that supply and demand comes into play, but each vote is still just as valuable as each other vote regardless of the acquiring party).
Additionally, because IBDD uses a closed economy cut off from real world wealth, it is much more difficult for the ultra-wealthy to manipulate the outcome in their favour.
Now, QV does satisfy Popper's criterion in some cases, it's definitely possible to remove bad policy without violence.
However, it may also be possible for QV to violate it. Consider the case where an ultra-wealthy individual employs many thousands of people to take part on their behalf. These workers buy votes at a vastly reduced rate (compared with our wealthy person purchasing it all themselves).
The effect of this is a mostly economical attack to prevent the removal of bad policy. In addition to less money being included to begin with, the socialisation of proceeds also is spread far more in favour of our rich attacker, and so their risk and overheads are lower.
While this does still result in some redistribution of wealth in the direction of the dissenters, we've set up an economic equilibrium between the benefit of the policy and the cost of maintaining it. It is therefore theoretically possible such a system would prevent the removal of that policy indefinitely, or at least long enough for it to matter.
In the case of IBDD, however, gathering such a following requires opportunity cost on the follower's behalf, and additionally reduces the ability for the attacker to interact with other legislative processes. Thus I consider IBDD to more effectively solve this problem than QV.
Addendum: IBDD and Budgets
This is an unsolved problem.
One possible solution is simply to average the opinion of all voters. As far as 'wisdom of the crowd' goes, this doesn't seem too bad a place to start.
It does not seem like a budget is the sort of thing one person or source would be able to produce well consistently, and competing budgets aren't guaranteed to account for everything as well as we may like.
Is it appropriate for budget requirements might be attached to policy and automatically granted when the policy passes? Could we design some taxation system to facilitate this?
I am sure that if IBDD begins to succeed this problem will receive attention, and all problems are soluble.
Addendum: IBDD and War
This is also an unsolved problem. War is never a simple thing, and seems to lend itself more to centrally controlled systems than decentralised ones.
The naive approach seems to suggest a referendum, and certainly many people would like it if their governments were required to hold a referendum before going to war, but it's not clear to me that this would necessarily produce better or worse results.
In the limit, IBDD is predicted to unify political systems into a global system of governance (note: not a global government). Perhaps it is the case that this will never need to be solved? What if we later encounter some outside threat we did not predict, and then don't have the capacity to deal with it?
I believe this problem to be soluble.
Additional Reading