In our societal moral discourse we often claim to be able to justify our values by recourse to some empirical facts. We say that we care about group x because group x is oppressed; we fight against y because y is harmful. Here we are trying to justify our values (that we care for x and fight against y) in terms of empirical facts (that x is oppressed and y is harmful).
But as I have argued on here on multiple occasions recently and as is well known in philosophy, there is no strictly empirical way of of finding out exactly who is oppressed and what is harmful. In fact, that quest doesn’t even make sense, as terms like “oppressed” and “harmful” are highly normative terms (meaning that they have values built into their meaning, unlike terms like “chair” or “moon” etc) and normative facts cannot be discovered empirically.
This result is called in the philosophical literature “the value-ladenness of theory”. It means that theories are not strictly descriptive and empirical, but have values built into them. There is also a related result called “the theory-ladenness of observation”, which means that all gathering of evidence already presupposes some theory and so there is no way of gathering data in a purely detached manner and letting it guide us to a conclusion. Rather, in the choices that we make for the data-gathering and the ways in which we choose to interpret the data, we are bound to presuppose some theory.
In several of my recent posts I have given examples of how terms standardly used in sociological discourse are heavily value-laden. But it is worth repeating, as it is a result that is widely ignored even by educated people and which has far reaching consequences for our political and moral discourse.
The first thing to notice about society is its sheer complexity. Society is the most complex entity in the universe by far. Nothing in the universe – even complex star clusters, chaotic weather and turbulent flow – come even close to the complexity of society. In fact, we have good mathematical models to deal with some very complex systems, but we have no such models to deal with society as a whole. We cannot predict when pandemics will happen, when war will break out, or even who the next ruling party in government will be.
How do we make sense of such complex systems? How do we go about explaining what’s happening in society? We do this by reifying (bringing a concept into existence) highly abstract metaphysical entities and by artificially isolating and cutting off causal paths. Let’s look at these two in turn.
Strictly speaking, there are no men, no women, no races, no classes. There are only people, over 7 billion of them. Each of these billions of people are different from each other in an infinite number of ways. This is all we can say about humanity if we want to be precise and accurate and not presuppose any metaphysical, non-empirical entities. But this view isn’t very helpful, as we will need infinite pieces of information in order to describe people’s behaviour on this level. That’s why we reify sociological concepts.
We pretend that there is such a thing as “a woman” and we ask what it’s like to be a woman, what are some things we can say about women, etc. Strictly speaking, there is nothing what it’s like to be a woman; there is only what it’s like being person x at time t and person y at time t. Feminist philosophers have long questioned the meaningfulness of talking about “women” as a well defined concept. Let’s break it down.
Asking what it’s like being a woman seems like too broad. It seems obvious that your experiences will be impacted by other factors besides for womanhood. So let’s try being more specific: what is it like being a black woman in the UK? But this still seems too broad. Surely other factors besides for your gender, race and country will impact on how you experience the world. Let’s try again: what is it like being a black, working class woman in the UK? Still too broad, as different black, working class, UK women will experience the world differently depending on whether their (dis)ability, health, income, age, sexuality, etc.
Ok, so how about we try factoring all these in? What is it like being a black, cis, abled, straight, working class, …, woman living in the UK? Are we done now? Let’s ask this question differently: Do all black, cis, abled, straight, working class, …, women living in the UK experience the world identically? Well, obviously not. If we want to be more precise with our categories we are going to have include more identity factors. Once we have finished including all factors for which we have names in our language (such as gender, sexuality, class etc), we can now include factors for which we have no names. Here are examples of some factors that will have a bearing on one’s experience: having 3 children; having 3 children aged, 2, 5 and 13; having 3 children – a boy aged 2, a boy aged 5 and a girl aged 13; having 3 children – a neurotypical boy aged 2, a boy with autism aged 5 and a girl with back-problems aged 13; having 3 children – a neurotypical boy aged 2 who likes to play with friends, an autistic boy aged 5 who doesn’t like cereal and a girl with back problems aged 13 who is bullied in school.
As you can see, we can be more and more specific in describing a person’s experience and in trying to isolate what it feels like to be them. Eventually, we get down on the level of individuals and we are simply describing an individual’s life. So what has happened to the category of “woman”, or of “black woman”, or of “black, working class woman”? We have seen that there is no such thing as a woman or a woman’s experience and there is no such thing as a black woman or a black woman’s experience. But, conversely, having 7 billion sociologies, one for each individual member of society is completely uninformative.
What we therefore do is we reify metaphysical concepts (essentially, make them up) to help us simplify society. We imagine that there is this entity called “womanhood”, or “being a woman” and we try to find common experiences for all those that fit in with certain criteria who we are happy to call women. We do the same thing with race, sexuality, etc. The crucial point, however, is that these “carvings up of society” into these categories are not unique. Who to count as “black” for example is not unique: how dark does your skin colour need to be in order to count as black? How much black ancestry do you need to have in order to be black? Etc.
A second way in which these “carvings” are not unique is in making sense of the data. A topical example is police violence. We can show data that there is disproportionate police brutality against blacks. But the same data interpreted differently will say that it isn’t disproportionate violence against blacks, but against men. Yet a third interpretation will say that it’s about income; a fourth interpretation will say that it is black, men of a certain income; a fifth interpretation will find a psychological component to it as well; finally, a sixth interpretation will just enumerate all the victims of police brutality, name this group of people as “people likely to experience police brutality”, or “brutalisable” in short, and claim that the people whom police brutality really affects are “brutalisable” people.
If we want to ask which hypothesis is most supported by the data, we will find that the last one is. In fact there is 100% correlation between “brutalisable” people and people who are victims of police brutality. So it isn’t really about race, gender, income etc, but about “brutalisability”. Of course this is utterly uninformative. But here we are again at our result that there is always a trade-off between informativity and precision. What level of precision to sacrifice for what level of informativity is not a trivial matter and there is no unique way of doing it right.
So far we have looked at the reification of abstract entities (womanhood, race, class, etc) in order to make sense of the complexity of society. We have seen that there is no one unique way to do it and that there are an infinite number of possible entities that we can reify (such as “brutalisables”), depending on how it helps us understand society. Now I want to look at the second way in which we analyse society and that is through isolating and cutting off causal chains.
Suppose I ask for the reason of a specific sociological phenomenon. To give you a full picture of the reason, I can look at some causal chains, by asking what caused this phenomenon (very often asking for a reason, is asking for a causal chain). Strictly speaking, the cause of any phenomenon is the Big Bang. If not for that, then this phenomenon would not happen. But this is probably not informative and not what you asked for. You may ask for a more proximate cause, or for a cause that would explain why x happened instead of y, rather than a cause that would explain why there is something rather than nothing. I may then move on to the formation of the solar system as a more proximate cause. Still not informative. The evolution of life? Slightly more informative. The evolution of culture? A bit more informative. The history of Western society? Even more informative.
We see here the same trade off between giving the full picture and being informative. The fullest explanation of any phenomenon would be giving the full causal chain starting off from the Big Bang. But that is very uninformative. The more informative we want to be, the less of the causal chain we should be including. But deciding what bits of the causal chain to include and what to omit is not a trivial matter. Let’s make this more concrete through an example:
If I ask why single mothers are often poor, I can give any one of the causal components in the following list: because they have no one to feed them whilst they look after their children; because they got divorced; because they are not supported by the state; because we live in a brutal, capitalist society; because we live in a patriarchy; because we have an institution of marriage in the first place; because they chose to have kids; because they chose not to have abortions; because they haven’t given up their kids for adoption; because their exes are dicks who don’t pay their share; because thy don’t work hard enough; because they don’t kill themselves; because they don’t kill their children.
All of these listed are part of the causal chain in the sense that it is true that if one of these hadn’t happened then the plight of single mothers would have been different. It is true that it single mother all committed suicide then they wouldn’t be poor. Should we then include “not being willing to commit suicide” amongst the causes of poverty for single mothers?
But this brings us back to the idea that there is no unique way to conceptualise societal causes. Ultimately, what causes to include as the reasons for a phenomenon is a choice and there is no unique, or empirical, way of doing it.
I think that the way in which we settle these questions is by introducing a paradigm (what I am proposing is itself a paradigm). How do we decide what entities to reify and what parts of the causal chain to include in our analyses as informative? I think that we build up a paradigm, or a way of looking at things and we interpret the data through the lens of this paradigm. This paradigm is both theory-laden and value-laden in the following sense. For the theory-laden part, we use past results and conceptualisations to help us understand the data. The reason why we see police brutality as a race issue and not as a “brutalisables” issue, is because we already have a “theory” or conceptualisation about race and so we use the lens of race to interpret the data. We have no analogous theory of brutalisables.
For the value-ladenness, the reason why we don’t include the lack of suicide as a cause of poverty is because we don’t believe that poor people SHOULD commit suicide. Or, in other words, we don’t believe that our society SHOULD solve poverty by encouraging suicde. Whether we include capitalism or lack of government support as amongst the causes of poverty will depend on whether we believe that we SHOULD bring down capitalism and whether we believe that the government SHOULD support the poor. Thus our theories are value-laden.
Coming back to ideas of harm and oppression, I claimed that there is no uniquely empirical way of determining those without value-ladenness. If we want to decide who is oppressed, we need to have presuppositional values about how our society SHOULD look like, how rights SHOULD be distributed, what inequalities SHOULD be avoided and so on. We don’t count it as oppression of white people when a disproportionate number of mega-rich football players are black. We don’t count is as oppression that plus-sized people are underrepresented amongst Olympic athletes. Do we count it as oppression that rich people live more comfortably? That will depend on our values. If we have a communist bent, then we might count that as oppression. If we are more on the libertarian side of things, then we are unlikely to. What we count at oppression depends on our values and on how we envision the ideal society to be. There is no empirical way to determine this.
Likewise with harm. We don’t count it as harmful to throw sex offenders into jail, even though they will definitely suffer harm there. We don’t count it as harmful to drive a car, even though there is a statistical probability that you will kill someone. What we count as harmful is, again, dependent on what we envision the ideal society to be like and, ultimately, on our values.
Coming back to where I started, empirical data cannot be used to justify values. You cannot justify your values by appeal to what is harmful or oppressive, as to interpret these concepts you are already presupposing values. Ultimately, the data and our values are in a symbiotic relationship where they both need each other and both reinforce each other. Your values tell you how to interpret the data and the data – having been interpreted by your values – further shapes your values. To think that the data empirically vindicates your perspective and your paradigm is to misunderstand this relationship.
Ultimately, we think in paradigms. We build up a metaphysics of society based on our values and based on what conceptualisations and reified entities we find most conducive towards our understanding and our political aims. We are never talking about objective facts (unless we are referring to the raw, uninterpreted, data-points), but always about constructions and conceptualisations. Internalising this can help us understand better where disagreements in the moral, political and social realms come from and why they are here to stay. It can also help us be more compassionate and understanding towards those who see things differently to us. Rather than seeing them as “bad” or “wrong”, we can understand that they are operating within a different paradigm and are conceptualising things differently. There is no way of showing that they are wrong and we are right.
(Btw, none of this is novel. These results have been well known at least since Quine’s work in the 50s.)