What minimum criteria would you suggest for an entity to be considered a person, subject to all laws and possessed of all rights that this entails? How should these criteria be measured? (One example that seems particularly important, if considered, is sentience.)
This is an interesting and difficult question. Personhood is not binary, as we see it, but rather a function over a multidimensional space, where some regions may have a high personhood value and others may have a low value. We might visualise it as a ‘personhood test’, where your answers to many different questions are aggregated into a single score. This is purely epistemological, since determining this function or even what many of the dimensions are are beyond us at this time.
Albeit, this is contentious. If anything, it is an oversimplification. What we consider ‘personhood’ may be several incombinable aspects, incombinable because different parts of what we afford to persons may matter differently to these different aspects. Nevertheless, as far as we are restricting ourselves to a single label of ‘person’, this cannot matter, since we cannot divide this label.
Legal personhood, however, is binary, at least for now, and so we need a cutoff, a point above which we consider an entity to be a person and below which we do not; this point is not a claim of truth, it is merely a practicality. Or we could divide up our scale into different classes of person, where we may have a class for full personhood, one for some limited rights, and so on. Nonetheless, we don’t believe that this is necessary.
We posit that on the scale of personhood, most entities are clustered into two groups: clearly person and clearly not person. Intuitively this much is clear, simply because there are so few entities that might fall in a zone of uncertainty. Tulpas, as a class, fall clearly into the person camp, as do hosts - we will explain this reasoning below. Meanwhile, kitchen appliances and such fall clearly into the nonperson camp. As far as ‘servitors’ and such are concerned, judgement is discretionary. We have seen servitors in others that act exactly like tulpas, but call themselves servitors. And we have seen servitors that are merely automata, a habit given a name. There is little that unifies the category with regards to personhood. But there are few servitors to begin with, and most will be clearly in one camp or another, if not all the same camp.
In a certain sense, this might allow us to dodge the question entirely. It does not matter where in the grey area we place the boundary, because the entities there are few and sparse compared to the tulpa population. And moreover, whatever is excluded is low in personhood value regardless. It matters far more to assert that the clearly people are clearly people.
But, we will not dodge. 'Sentience’, as you suggest, is, perhaps, less important than you think. By comparison, few believe that cattle should be afforded full personhood, just treated with compassion. And there may be reasons to afford personhood to nonsentient beings.
A from-principles motivation of personhood might be required to address this fully, which is beyond us now. To skip to its conclusion, we say that we ought to take the utilitarian view, and worry about the cost of a particular assignment of sentience (or a particular cutoff), the cost being an ethical one, or in any case whatever is valued by the asker of the question. This is not in general reducible to a question of the cost of assigning personhood to individual entities independently, because there may be interplay between different assignments affecting the costs. However, we shall suppose that these are unimportant in scale, and consider independent questions of whether to afford personhood to a particular class of entities. The most important considerations here are:
The wellbeing afforded to the entities in question. This is, of course, a matter of their sentience, partially; a nonsentient entity could make little use of extra welfare. But there are other considerations. How well can the entity actualize their personhood? By this we mean, why should we afford them the proactive benefits of personhood, such as owning property, if they are merely passive. We may allow cattle to own their fields all they wish, but they will not derive any benefit from it. All we wish for cattle is that they are treated with compassion, which may be achieved without recourse to personhood.
The wellbeing afforded to other entities. It may be that we wish to implement a policy that all tulpas are sentient, as a precaution; we may not wish to expose sentient tulpas to the suspicions that they are not sentient. On the other hand, it might be that attempting to assign personhood to too many things may detract from our causes, such that even if we believed sea anemones were as much people as the day is bright, we would not wish to campaign for that in case we should discredit ourselves.
The economic cost of the policy. We are less concerned here with the costs of implementing the policy, and more the costs of keeping it. Implementing only happens once, the cost is finite, and in the long term will approach a negligible fraction of the ongoing total costs, so long as the cost is still small enough compared to the global economy that there will not be enormous lasting impact, or the policy be unimplementable. However, we may not wish to afford personhood to kitchen appliances because it would greatly slow down cooking.
All of these must be set against each other. Obviously, a policy may still be implemented at great cost if the benefits are greater.
Because pure personhood is nonbinary, we must make practical considerations when deciding where to place a boundary. In a utilitarian calculus where we care about wellbeing, the major factors here are
The sentience of the entities - we may call this the conversion rate from welfare to wellbeing.
The ability of the entities to gain from their personhood - we may call this the conversion rate from personhood to welfare. This is quite a broad group, but in practice we mean the entities’ ability to self-actualise and act as agents of their own interests.
The welfare gains/costs to other entities of implementing their personhood.
The global economic costs of implementing their personhood.