Realizing the Original Position
by Ali Sina S
Technological advancements that have been made in the past two centuries have altered our society in such a radical way that life today would be mostly unfathomable to someone born in the start of the 19th century. In this time frame, the standards of living of the average person increased considerably. However, this is not to say that the consequences of recent technological advancements have been fair. As I will provide further support for later in the paper, while the increased productivity and mobility along with the fast transmission of information that technology has afforded society has certainly been a big factor in the massive economic growth and enrichment that correlates with that period, its effects have not been felt equitably. It has led to further disparities in wealth, and further widened the chasm between the lives of the wealthy and those at the bottom of the economic ladder to the point where the poorest segment of the world population effectively live decades behind the global wealthy. Black Mirror, a science fiction tv series, frequently addresses the ethical issues surrounding technology, often through dystopian depictions of society in the near future. A society very similar to ours, save for the technologies in question. The negative consequences of the new technologies that have entered our lives in the past two centuries are not due to the nature of the advancements themselves. In other words, there is nothing inherently bad about the specific technologies that were developed. The main issue lies with how those in power were able to implement them in a way that disproportionately promoted their interests over the rest of societies’. After the initial rush of pessimism sits in, however, it is possible to see ways in which we may reconcile the unethical implications of technological advancement in the past, present, and future, with basic moral principles of equality and fairness. In the Black Mirror episodes USS Callister, San Junipero, and Hang the DJ, there are virtual reality simulations where individuals have access to all five senses and are essentially indistinguishable from reality. These technologies can allow us to make models/simulations to empirically study counterfactual situations. I think that an effective use of such capabilities would be to model the original position, as its outcomes can outline how we can form a basic structure to remedy the economic inequality, along with various other concerns such as privacy, brought upon by new technologies.
Before I move onto questions specific to the original position, I would like to support my previous claims about the state of the economic order. Thomas Pogge proposes that we evaluate an economic order by the extent of absolute poverty, the extent of inequality, and the trend of the previous two factors over time. I think that this is a good way of assessing the morality of an economic order as it not only looks at the amount of absolute poverty and inequality, which viewed together also show how avoidable the poverty is, but also how they have evolved over time. This allows us to see whether the current state of the economic order has been effective in trying to remedy poverty and inequality, instead of only focusing on the present circumstances. The latter approach may be misleading as the existence of poverty or inequality does not necessarily mean that an economic order is morally flawed as if the rate of poverty and inequality is significantly better than it was, say, 10 years ago, it would mean that the economic order has been effective in dealing with these issues. After describing the present circumstances of the global economic order in his book World Poverty and Human Rights, Pogge gives concerning figures showing the trend of how the income gap between the world’s richest 20% and the world’s poorest 20% has gradually widened from 3 to 1 in 1820, to 11 to 1 in 1913, to 30 to 1 in 1960, to 60 to 1 in 1990, and to 74 to 1 in 1997.
Keeping in mind the scope of this paper, I will limit my own arguments to the national level. This also means that I will mainly be looking at A Theory of Justice when referencing Rawls. However, this limited scope does not mean that Pogge’s observations are any less relevant. Figure 1 shows how the share of national income has changed between the top 1% and the bottom 50% of the US population:
Figure 1
We can see that the global trend Pogge has described can also be observed in the US from 1980 to 2015. I am using the US as an example due to the availability of data, and because they have been pioneers in the development of most of these technologies. Moving onto how technological advancements may relate to this trend, we can consult figure 2:
Figure 2
Figure 2 measures the number of transistors that fit into a microprocessor from 1979 to 2015, roughly the same time frame as figure 1. There is not a direct way of measuring the rate of technological development, but the number of transistors that fit into a microprocessor is a good representation as it reflects how computing power has increased over time, which is often the main limiting factor in technological research. An important observation to be made between figures 1 and 2 is that the curve representing the incomes of the top 1% of the US population has roughly the same shape as the curve in figure 2 (figure 2’s y axis is not to scale, but the trend is nonetheless the same), and the curve in figure 1 representing the incomes of the bottom 50% is roughly the inverse of figure 2. This shows that over the same period, the incomes of the top 1% have increased roughly parallel to the rate of technological advancement, while the incomes of the bottom 50% have steadily decreased. There is a plethora of factors that affect income growth in an economic order but given these figures, we can reasonably assume that the technologies developed during this period is a major factor. The causation I am supposing can be attributed to factors such as the automation of jobs and the subsequent displacement of workers seen in industries such as automobile manufacturing, though the strong correlation and the mere existence of some degree of causation is adequate for the rest of my argumentation.
Seeing as how there seems to be a strong correlation, and at least some degree of causation, between technological growth and economic inequality, over the past 40 years or so, I see no reason to assume that this trend would be reversed going forwards. The solution I proposed earlier for reconciliating this technological growth with ethics was the realization of the original position through a model. The original position is a thought experiment proposed by John Rawls as a fair and equal method for citizens, represented by agents under the veil of ignorance, to agree on principles of justice on which to base the basic structure of their society, which refers to the social and political institutions. I would now like to address a potential objection to my proposed solution, which is that thought experiments are precisely what the name suggests, and that they are not meant to be performed. I do not think that all thought experiments need to be realized. Some fulfill their purpose in their hypothetical form, and we would not gain much if we were to perform them in real life or in a simulation. An example of such a thought experiment would be the Chinese room argument, devised by John Searle. Briefly put, the thought experiment puts a non-Chinese speaking person into a room where they must respond to notes in Chinese slipped under the door using a computer program for manipulating the Chinese characters to create appropriate responses to the notes. Searle argues that those slipping the notes under the door would be led to believe that there is a Chinese-speaking person in the room. The implications of this, according to Searle, is that this is the only kind of ‘understanding’ a computer may have of human language, and ultimately that the notion that human brains act like computers is wrong. If we were to actually perform this experiment, we would not gain anything out of it as the whole thought experiment is a mere description of a procedure with a constant outcome. The only room for variance is how one interprets the outcome, such as how there is a considerable amount of discord on whether the outcome of the Chinese Room Argument have the implication Searle proposes. A distinction that seems to separate the original position from the Chinese room argument and other similar thought experiments which I would classify as not in need of realization is that the original position has the two following characteristics:
It relies on normative principles as rules
The procedure includes variables (agents)
In the original position, the agents are governed by a set of normative principles, such as the assumption that the agents are reasonable and willing to cooperate. Using different agents would result in different outcomes due to their varying comprehensive doctrines, which refers to the set of values and beliefs that an individual holds. In other words, thought experiments I would classify as in need of realization can be viewed as mathematical functions, where different inputs generate different outputs. Thought experiments in the former category instead describe a scenario with no variables, and with a constant outcome. This does not mean that there is no use in realizing those kinds of thought experiments, as even if they have a single result, we cannot empirically know what it is until we perform the experiment, and lots of thought experiments have contested results. However, the product of their realization would not be a model which can be tested in multiple conditions and produce different outcomes like the original position, but merely a proof.
I would now like to sketch an image of what such a simulation would look like. The original position has a well laid out procedure that would make it relatively easy to model. Each agent needs to be free and equal. Rawls suggests that we can attain this by symmetrical positioning, where agents cannot threaten each other or hold out for a deal that better suits their interests. Another key feature is the aforementioned veil of ignorance. The agents will only know relevant facts that would not lead them to be biased in their judgements. This means that the agents in the simulation will be withheld information regarding their citizen’s socioeconomic status, race, age, etc. but will know that different agents have varying comprehensive doctrines. For the purposes of this demonstration, I will take the example of a closed liberal society. In the first stage of the original position, the parties are asked to choose a conception of justice. At this stage, the agents that represent the citizens of this liberal society would agree on the conception that has the principles most in line with the comprehensive doctrine their citizens hold, with the equalizing effect of not knowing what position their citizen actually has in society due to the veil of ignorance. We will go with Rawls’ assumption that they will accept the principles of justice as fairness as their conception of justice. Whether or not they would make such a choice is not important for our discussion as the purpose of the model is to validate such assumptions. The next stage of the original position, which is to check that social and political institutions created by those principles would be stable over time, is where a computer model would thrive. Rawls’ idea of stability is an overlapping consensus, which is essentially the idea that different citizens will accept a conception of justice for different reasons depending on their individual comprehensive doctrine. In this stage, the veil of ignorance is gradually lifted, and citizens agree on increasingly more micro and practical aspects of their societies’ basic structure. Essentially, agents start by agreeing on a principle of just savings, which determines how much each generation saves for future generations. Next, they agree on a constitution based on the principles of justice they have adapted. They then agree on legislation and apply it to cases to make sure it is coherent with the previous steps. With a simulation, these steps can be run in a brute force manner, where each agreed upon part of the constitution, piece of legislation, application of the law etc. is tested against the set of core principles and the products of the previous stages. For example, an aspect of the constitution would be tested against the principles of justice that were agreed upon, while a specific application of the law would first be tested against the legislation, which had already been tested against the constitution, which had already been tested against the principles of justice.
In Rawlsian terms, this would allow for a perfect reflective equilibrium on a societal level. Reflective equilibrium is achieved when an individual’s beliefs and values are coherent with each other from the broadest fundamental moral principle to the most specific conviction regarding a moral situation. In perfect reflective equilibrium, though believed to be possible only in theory, an individual’s most basic moral principles, such as “every individual has the right to free expression”, are not contradicted by any of their more narrow convictions, such as “individuals have the right to practice any faith or no faith at all.” The reason why perfect reflective equilibrium is unattainable is because individuals have way too many convictions. It would be impossible to list every conviction one has, and check if it is compatible with their broader beliefs. If one were to achieve perfect reflective equilibrium, they would be able to make judgements consistent with their comprehensive doctrines in any given situation. The original position model would allow us to achieve that at a societal level, as it can check each specific aspect of the basic structure to the less specific principles they were formed by. This would mean that the social and political institutions of that society would make accurate and fair judgements, based on principles agreed upon by everyone, in every situation.
The simplified application of the conditions of the original position to a potential simulation I demonstrated are most likely idyllic. There would inevitably be lots of challenges and dilemmas in the programming of such a model as expressing normative principles in code is not an easy task. However, if we ever reach a point where people can be fully represented in a virtual environment with all their convictions and values in-tact as suggested in Black Mirror, coding normative principles into these simulations likely will have already been done. There is a lot of controversy surrounding the outcomes Rawls predicted would be drawn from the original position. It is hard to say whether even liberal societies would accept justice as fairness in its entirety if at all without trying it. But I think that it is safe to say that the basic structure of society citizens would choose in a fair and equal playing field would be significantly different than ours. It is questionable if being able to empirically show that our society is in no way near the optimal point would have any substantive impact, though I believe that it would be an empirical demonstration of just how immoral our basic structure is, and forcefully remove the wall allowing states and, to an extent, corporations to paint a portrait of an ideal society.
Sources:
John Rawls, A Theory of Justice
Thomas Pogge, World Poverty and Human Rights
https://plato.stanford.edu/entries/rawls/#ReaCit (sections 2.4, 3.2, 3.5, 4.6)
