Many of our basic political conflicts center around the question of what rights we truly possess. In the United States, we have the Bill of Rights which attempts to lay out an initial list, but it's obvious after brief thought it's woefully incomplete and doesn't address many rights we would quickly agree should be protected. It's also clear some rights we grant are perhaps too strong, and can be easily captured and abused, such as in the case of intellectual property rights. Even our simplistic assumptions about land ownership could be much too strong, since those rights could in theory allow small groups to hold unreasonable amounts of control for unreasonable amounts of time.
There have been attempts to define a more complete set of rights, such as those by Franklin Delano Roosevelt or the United Nations. However these lists are still frustratingly arbitrary, and we don't have a rigorous framework to evaluate them and understand their trade offs.
I propose in this chapter just such a framework. I've attempted to make it as general and rule-based as possible. The position of any right in this framework has direct consequences on what kinds of public action can be taken regarding it, and how it is balanced with other rights.
In the future I intend to formally specify and verify this framework in a proof assistant, so at that stage it will necessarily become fully detailed and logically precise. However for now I hope you'll forgive me for only laying out some loose descriptions and conjectures for what I intend to prove and how I intend to prove it.
Rights are all about understanding how different beings interact with one another, deciding what kinds of interactions should be disallowed, and how we should disallow them. We need a model for what a Being is, how it can act individually, and how multiple Beings interact.
We might be tempted to make a full model of the entire physical universe and the full complexity of human biology and psychology, but luckily we don't need to. It's very often useful to describe complex components in a logical model with "abstractions", a sort of behavioral shell, and not concern ourselves with why that shell exhibits the behaviors it does. We can ignore all internal details, and simply treat the shell as a "black box" that we can act upon and see outputs from. It's safe to do this when the inputs and outputs to a component are the only thing we care about, such that our analysis wouldn't change even if we completely understood every internal detail.
My basic model of a Being has these properties:
With this very simple definition, we can build up some more useful abstractions:
All these abstractions only concern a single Being, and make no assumptions about the nature of the world they exist in. Since the Qualia and Welfare of Beings is entirely internal to them and can't be experienced directly by others, Actions are the only way that internal state can manifest itself.
In order to usefully model interactions between Beings, we need to assume they exist in a shared, durable, consistent universe. We can model multiple Beings existing in such a universe by describing the whole system as a "loop", where each cycle of the loop goes through these steps:
Since the actions of all Beings are added together and effect the same universal state, this implies Beings can coordinate with one another by incrementally building up shared information theoretical codes to communicate arbitrarily complex ideas to one another.
Welfare is the only part of our model that structurally assumes any kind of preference or "goodness", and is just another term for the concept of utility from Utilitarianism, an ethical philosophy stating that we should seek to create the greatest amount of subjective happiness or freedom from pain for the greatest amount of people. Since Utilitarianism is simple and relies on very few assumptions about the nature of consciousness and reality, it's a perfect fit for developing a theory of minimum necessary rights. However it often becomes very controversial when people discuss how to put it into practice.
Utilitarian philosophers often discuss different ethical thought experiments, such as the famous trolley car problem. These thought experiments often try to determine what kinds of decisions are ethically justified when some group of people must be harmed for the benefit of another group of people. Simplistic representations of utilitarian thought then sometimes try to perform "calculations" about the aggregate utility of many different people to decide how to make decisions on their behalf. However our model points to a potential problem with these kinds of utility calculations.
We've made very few assumptions about the nature of Welfare for a very good reason. Every Being might have a completely unique experience of Welfare, and might prefer completely different experiences! This makes it unrepresentable to simply "add up" the Welfare values of different Beings. Further, since Qualia fuzzily contains all the past experiences and memories and internal ideas of a Being, even if their Welfare type was the same as ours, we can't completely predict or understand their current internal Qualia. This means it's impossible for us to predict what Welfare some other being will experience from any Action we take, we can only guess.
But even if it's impossible for us to predict what a Being's Welfare for any arbitrary Action might be, we can always communicate to them some finite list of possible Actions, and give them a way to signal to us what Welfare they predict they will experience from those Actions relative to one another. There's still imperfection here, and each step of communication and prediction introduces possible error or inaccuracy. But at least we have a principled approach, and can work to gradually remove uncertainty from each of these steps. Voting is simply a way to apply this principle when making decisions in a group.
With this realization in hand we can state a useful principle:
With any group of Beings seeking to maximize their own personal Welfare, it's obvious they will routinely interfere with one another. Let's define interference as a situation where a Being makes a Prediction about the results of some Action, takes that Action, but as a result of Actions taken by other Beings the final perceived state of the universe will differ from their Prediction such that they don't achieve the Welfare they were intending. The opposite will also routinely be true! Sometimes the Actions of other Beings will actually help some Being experience more Welfare than they expected, which we'll define as cooperation.
Both interference and cooperation effect the Welfare experienced by all Beings in a group, so it will often be rational for them to choose some Coordination Plan, a function intended to minimize interference and maximize cooperation between them to generally help each best maximize their Welfare. A Coordination Plan takes as input the Actions of the coordinating Beings, and outputs for each Being either some new Plan specifying what they should do next, or no Plan if nothing is asked of them. Of course any Being may choose to ignore a specifying Plan, but let's discuss that later.
Let's consider some examples of Coordination Plans to get a better grasp of this idea:
These are three extreme examples that illustrate some potential problems we need to avoid. They bring an important question into sharp relief: how do we decide if a Coordination Plan is optimal?
Here's my conjecture:
Let's call this the Original Position property. The Original Position property seems equivalent to this restatement:
For fun, let's call this the Trading Places property.
This is an interesting property, since it implies such a Coordination Plan will both be the most rational choice for most Beings most of the time, and is the best choice to make on behalf of those who can't signal their preferences to us. This includes future not-yet-existent Beings, or any Being who is sufficiently distant from us in some way. The latter implication strongly suggests such a Coordination Plan is the most ethical one we could choose.
But how do we design such a Plan?
It seems intuitive that in order to maintain the Original Position property, a Coordination Plan must inherently consider the preferences of different Beings equally. The Plan might make decisions that give more control to different Beings, but only after all aggregate preferences have been considered. The root preferences must be considered equally.
Let's classify different outcomes of Actions based on their Welfare results:
It should be obvious an optimal Plan will prefer both Strict Improvements and Mixed Changes over Strict Declines, and should always encourage Strict Improvements and discourage Strict Declines. But when trying to decide between outcomes where neither of those rules applies, we need more complex rules:
Obviously voting is one way we can gather relative preference strengths for outcomes, but there's another: direct communication of consent! If a Being gives their voluntary consent to some Action, that necessarily implies they consider that Action to be among their best options for Welfare improvement. This is one way we can discern between Strict Improvements and Mixed Changes without having to conduct a vote.
But how can we know if the Actions of a consenting group will interfere with other Beings?
The only way to know with certainty that some Action won't interfere with other Beings is to somehow solicit their preferences, either through consent or a vote. However it's obviously inefficient to do this for every single potential Action, since the effort spent conducting a vote itself creates changes in Welfare. If the Welfare cost of conducting a vote routinely outweighs the possible Welfare increase gained from the coordination of the vote, that vote does more harm than good. I'll call this idea Coordination Efficiency, or how well a particular vote is able to improve outcomes compared to its cost.
So we need some way of improving Coordination Efficiency. To do this it's useful pre-classify Actions into categories depending on their predicted outcome:
Of course these classifications are still floating in the air: how do we actually draw the line between them? And even then, it will still be inefficient to conduct a vote for all the potentially interfering Actions.
These dual classifications suggest a very efficient concept we already routinely use: mutually excludable ownerships. An ownership is a classification of an abstract space of Actions, and a designation of an Owner of that space.
Ownerships delegate voting power to an Owner. The Owner is given sole deciding power within their ownership, representing some kind of prediction by everyone else that such a delegation will produce Strict Improvements. The group is essentially deciding that any vote conducted about any Action in that space would almost always simply defer to the wishes of the Owner. It also represents a judgment that any Mixed Change Action within that space could be substituted for a Strict Improvement, so the group simply disallows Mixed Changes within the ownership and instead requires any potential actor to find a different but similar Owner willing to give consent.
A group choosing to designate some list of ownerships is in a very real way still giving every member of the group the power to signal their preferences for every possible Action by all other members. They are simply doing so indirectly and in advance.
Ownerships can even be useful to decide when votes must be conducted. Obviously votes must be conducted to define the nature of ownerships, but a dividing line between public and private ownerships can do the rest. The Owner of a public ownership is the entire group, so a vote must be conducted; whereas the Owner of a private ownership is an individual or group of individuals, so consent must be gained.
Ownerships are very efficient, but they still must be assigned in a principled way. All people must have the same capabilities for ownership, ownerships must not collide with one another, and they should structurally imply the necessary predictions of interference.
What sources does Welfare come from? It's useful to understand these sources, since they structurally imply production of Welfare, and so likely define efficient ownership boundaries.
Nicely, these sources align with John Locke's concepts of Life, Liberty, and Property, although not intentionally or perfectly.
These sources suggest two different essential ownerships, which I'll show soon are efficient:
Let's also define a concept of Exclusion, where the possession of some ownership necessarily decreases the potential ownerships of all other Beings. This is a useful concept because it points toward how ownerships relate to Strict Improvements and Mixed Change substitutions. If we can grant some ownership to everyone equally without producing exclusion we should do so, since such ownerships will maximize improvements.
Ownership of Body is immediate and intrinsic. Every person is granted ownership of their Body when they are born, and don't have to do anything to qualify for it. A Body is equivalent to having Qualia and Force of Will. If someone didn't own their Body, they could still experience negative Welfare and would have no inherent capability for non-interfering Action. This is obviously an inefficient situation where potential Strict Improvements aren't structurally allowed and Mixed Change substitutions not required. Ownership of Body is also not exclusionary, since we can give every person ownership of themselves equally without taking an inherent source of Welfare from anyone.
Ownership of Universe however is necessarily exclusionary! This means any ownership of Universe must be mediated in some way to push toward Strict Improvements and Mixed Change substitutions. Since the essential Ownership of Body implies some small degree of Ownership of Universe, it's not difficult to arrive at some kind of principle of universal common ownership, requiring that all exclusionary ownership of the Universe must in some way compensate all other Beings for that ownership. In future chapters I'll discuss how the combination of a novel Common Resource Tax and normal private markets achieves that goal.
Finally we can address this important question: why will any individual Being follow a Coordination Plan when it isn't in their personal interest to do so? Simply put, the Coordination Plan is equivalent to a body of laws! To create a Coordination Plan protecting rights in the form of ownerships, we make the Plan accept inputs in the form of accusations of rights infringements, and the Plan decides on some response to those infringements. These responses can be anything, from restorative mediation, to deterrent punishments, to required restitutions. In creating these responses, we want to know what restrictions we should place on ourselves. How can we in a principled way decide on responses that won't themselves destroy more Welfare than they protect?
It's useful to pull apart our two essential ownerships, and be more specific about how they're defined. The first thing to notice is how the aspects of the ownerships relate to possible interference.
It's also interesting to note that the rough categories of Life, Liberty, and Property seem to transitively depend on one another. Liberty and Property cannot give Welfare without Life, and Property is rendered moot without Liberty. Combining the interference aspects with this dependence order, I propose the following matrix as defining the different categories of ownership.
|Necessary Interference||Possible Interference||No Interference|
|Life||Resources for Life (universe)||Action for Life (body)||Mind and Body (body)|
|Liberty||Resources for Liberty (universe)||Public Action (body)||Private Action (body)|
|Property||Finite Common Resources (universe)||Dangerous Items (universe)||Personal Items (universe)|
The ownerships of Private Action, Personal Items, and Dangerous Items are in more protected categories than is strictly necessary, since their possible interference was less severe than the category they strictly belong to. Whether to use these exact categories or others would be up to the group defining their system of rights.
Columns designating interference possibility determine what ownerships can be regulated or taxed:
Rows determine what ownerships can be suspended as punishment for infringement:
Finally we need to somehow define how severe an infringement is. It should be obvious that we should respond to any infringement proportionally to the infringement's severity, so a brutal assault shouldn't receive the same response as a light scratch on the finger.
I'll have more to say about this in the future, but for now I'll simply suggest a few spectra we can combine to determine severity.
We desperately need a more precise framework for reasoning about rights. I hope this theory could be a foundation for further work.