The Philosopher’s Stone has recently posted a 12 part series on the study of society (first one here), which includes an exposition of what it means to be a modern Marxist, the history of social studies and a discussion on what modern social scientists measure about societies. It also makes a fairly damning claim about the study of society as a scientific enterprise: that it is fundamentally impossible to study society impartially, because we perceive society through the eyes of a member of a particular society.
I am going to argue that this is not in fact a hopeless task, although I certainly agree that attempts up to now are capturing only a simple and terribly biased picture. I will first overview the criticisms, then I will give two reasons to be hopeful that we may eventually come up with a truly scientific (i.e. impartial) view of how our society operates. These should be understood as future avenues that social scientists will adopt in addition to current methodology that can give a more complete understanding.
What is the problem with current social science of economics?
I should start by pointing out I am not sure quite what the scope of the perceived problem is for social science. The Philosopher’s Stone criticises “social science” mostly in the context of economics, although notes that “Society is, when all is said and done, nothing more nor less than the totality of the habits, practices, rituals, institutions, beliefs, expectations, and behaviors of a group of people”. As economics is the most rigorous and mathematical of the social sciences, it makes sense to focus on it.
I will write the arguments as I understand them, borrowing quotes only where I fully get the point. So if I am mistaken please correct me; its my hope that my later arguments are robust to small misunderstandings at this stage.
The argument has two main points. “The first point is that society, even at the apparently simplest level, is essentially mystified”. Concepts that we use to perceive society are confounded by our pre-existing understanding of what society means. We define everything in relative terms that are suitable for the society we live in. So for example, a “driver” may be someone who drives for a living, or simply someone who is driving, or who can drive. Not all societies will have people who drive for a living (or anything to drive), and so these categories are somewhat arbitrary. This is true of all categories, including “men” and “women” (which matter for some concepts, such as marriage, but may be counted as either identical or not at all for others, such as employees who may be either male or female, or in e.g. feudal societies females may not even be counted, despite the fact that they also worked the land). Within that argument is the observation that we learn language constructs as children growing up in a society, so concepts within our society are rooted into our way of thinking by language and learning from other members of society. (Although it does seem obvious to me that language shapes our thinking, this is still controversial, as discussed at Replicated Typo,)
The second point is that we can’t index (count) anything about societies: “the concepts we use to get a conceptual handle on a huge, complex, multi-faceted social reality cannot possibly be given a value neutral interpretation without reducing them to tautological triviality”. Any measure that an economist has come up with is fundamentally flawed. A clear example is that we all have experience of inflation rises, but it is impossible to measure because goods and services change price in different directions: oil fluctuates, food rises, electronic goods drop (at the moment). On average, it means very little because we each use these products differently. Even worse, similar things have to be grouped together and these groups are arbitrary: are designer shoes in the same category as practical shoes? Is ice in the same category as a fridge? etc. So indexes don’t even measure the same thing over time, and they do it badly.
Why we can study society scientifically part 1: theory
Firstly, humans are just organisms that interact. At a very basic level, every individual is fundamentally understandable as an agent that does things and interacts with other agents, in ways that are explained by evolution. This is not even a theory – it is a truism (unless you are an anti-evolutionist, but then you should go read some other blog). Therefore if we can understand why people do things, then we can predict what people will do in certain situations. This becomes a theory of society, and it is fundamentally grounded in biology.
The question therefore becomes “what can we learn about ourselves?” I agree entirely that we cannot use the consumer price index to predict very much, because it depends on the individuals in a society in too complex a way. But what we can hope is to be able to predict the behaviour of aggregates of individuals based simply on knowing how they will behave on an individual basis, and in their fundamental interactions. As we move towards a truly mathematical (i.e. precisely defined and falsifiable) theory of how the brain processes information, we may get close enough to the truth in order to predict on average, what people will do. We won’t be able to predict what a specific person will always do (because the brain is too complex to be simulated by anything other than itself) but we can hope to predict them well on average (“when buying a snack, Joe will usually pick chocolate, unless he is feeling bad about his weight in which case he’ll pick fruit”). We should be able to predict what a population of people will do with higher accuracy than we can for an individual.
This approach bypasses the first objection to a large degree: it is no longer necessary in principle to view humanity from the point of view of a society at all. The theory can operate at the level of individuals: if you put the right set of behaviours of individuals in, you will get the right behaviour of the society out. It is still not a “theory of everything” – we still cannot understand a society made up of people we don’t understand. But we can understand the most important type of society: one made up of people who behave as we do. And better than that, we can understand any society where people behave in a way we can conceive.
Of course, we can’t answer all questions in this way. It may be impossible to tell the difference between two theories of how people behave. In statistics, this is known as the identifiability problem. What we can hope scientifically to establish is a range of behaviours that can only be explained by a single sensible theory; an additional range of behaviours that may have multiple explanations; and finally behaviours that are too complex and are unidentifiable. A theory based on detailed modelling will also have computational constraints that will either make it harder to accurately predict what people will do, or possible to only predict small numbers of people.
But even if the theory remains incomplete because of these problems, we can still predict the behaviour itself – we just don’t know which theory predicting it is correct. So we can get an awfully long way towards a theory of society.
Why we can study society scientifically part 2: empirical approaches
The idea of viewing the world though the pinhole of simple demographic indexes is over. In the data-poor world of the 20th century, there we no other options – but with data collection, transfer and processing on a massive scale, we will soon have no choice but to move over to a more empirical worldview, whether we like it or not.
At present, when the government tries to establish if people are better or worse off this year compared to last, they have to rely on the consumer price index, inflation, and other indices of the economy as a whole. But the need for creating arbitrary categories of products to construct these indices is decreasing and will disappear entirely.
Governments will soon (or already do?) have easy access to databases containing the total amount of all products sold within its country: the data can be stored in full. Britons bought x iPhone 4s and y iPhone 3gs this year, they spend z on bananas from Africa, and so on. The specifics of the good is known, and if this data were looked at in the right way we could establish what quality and quantity of everything was being consumed, giving a much better insight into “wealth” (i.e. how financially comfortable we are).
Better (or worse), Amazon, Google and Tesco can all know in great detail what we have individually spent money on. Combine that data and it will be possible to find out for every person whether they are better, or worse off than they were in the past. That is a scary thought and has implications that might make it difficult to implement in practice. However, I’m sure some people will be happy to be paid to act as samples, just like some people have their TV habits recorded so we can rate TV viewing. It will certainly be possible to establish the proportion of people who are better or worse off from year to year; and by how much. That is better than any index, and really does compare like with like. (The one exception is that expectations change. Although everyone may have provably more than they did, if people want more things they can’t have, feel more stressed or overworked, etc, then they will feel less well off).
A very interesting point raised by The Philosopher’s Stone was that unemployment has no real definition outside of a particular “capitalist” definition fit only for recordkeeping. There is an arbitrary choice of who you count as being potentially employable (children? parents? carers?), and who you consider to be employed (part time? full time? mothers? students? etc.). This is before you start to think about how productive people are or could be, and how much work they want to do. The same detailed data approach can be used allowing greater flexibility, but employment is more complex than consumption. More availability of consumer products, healthcare, etc is (all other things equal) a good thing. But different people would be happiest working for a different amount of time, doing different jobs, at different times in their lives. I guess at some point you simply have to ask people how happy they are, and it isn’t obvious that technology is going to solve the difficulties in doing this.
What we can do is measure everything that can be measured and build a psychological theory based on that, again using a sample of people (the same sample as above?). We then know how money is distributed throughout society; how that money translates into the feeling of having what you want, and to some degree how happy this makes people. That is a pretty good basis for measuring a society.
I don’t think that society is fundamentally mystified, meaning that science can’t make sense of it. I agree that current understanding is either obviously inadequate or bestowed a mystical level of belief (the economic crisis epitomised this). We are pretty much at the “leeches and bloodletting” stage of our understanding. But that doesn’t mean that more modern approaches can’t make headway in fundamentally changing the way we understand our society and culture. Society is, in this view, simply an extension of biology and any progress made is real, objective understanding.
None of this changes The Philosopher’s Stone’s fundamental points that a) we can never understand the full range of potential human societies; b) that we can’t reduce society to simple things that we can understand and measure over time. But I think we can make real, objective progress understanding past and present societies, as well as societies that we might hope for that are similar to our own. Additionally, we can measure this well enough to tell each individual whether they really are better or worse off, on an objective scale, and we can change our measures to describe things that we think are important.
Whether these ideas will come to fruition or not is another matter. The same things that I describe above as enabling governments to better serve the public are the same tools that could allow those currently in power to convince us to let them stay in power – by manipulating the numbers to tell us we are happy and prosperous and lucky to have them. This is of course the dark shadow that opponents refer to when they say “capitalism”; I wonder whether any organisation of society can really prevent the powerful remaining powerful? I’m claiming science may eventually have the answer…