‘Explanations’ in complex evolving systems

Cause or effect?

The recent debate about the role of ‘moralizing gods’ in the development of complex societies, sparked by a 2016 paper by Norenzayan et al, has focused on the use/misuse of data (from the SESHAT database). The consensus seems to be that if we fix some data analysis issues we can get back to business as usual, arguing about the role of religion as a causal element in the rise of complex societies.

A more important question (in my view) is what role ‘explanations’ have in complex evolving systems, whether living organisms, ecosystems or societies. Indeed, what it means to ‘explain’ something in the context of emergent properties like complex societies, sedentism, agriculture and so forth.

We now know enough about complex evolving systems to know that properties and structures appear (‘emerge’) in them that are not predicted by lower-level theories. For example, there is nothing in living organisms that is incompatible with our understanding of (say) chemistry, but in no sense is life predicted by chemistry. In societies, some structures have historically tended to emerge together. Complex sedentary societies tend to co-occur with agriculture, organized religions, legal systems, trade specialization and so on. However, hypotheses that one or other of these variables ‘caused’ the others usually founder in a mire of examples and counter-examples, with differing opinions from different authors based on their individual focus and experiences. The responses to Norenzayan et al’s paper are a good example (if a rather exhausting read).

A second important point is that such systems (whether organisms or societies) are constantly changing in response to changes in their environment, which they also influence in various feedback loops. These systems ‘evolve’ in the broadest sense of the word.

Most people love a good explanation, and preferably the simpler the better. Journal editors love papers that present simple explanations: A caused B. But based on what we know of complex evolving systems, what can we realistically expect?

Some explanations are clearly possible and valid. Here are some examples:

•    The dinosaurs were killed by a meteorite
•    The Archduke Franz Ferdinand died because he was shot by Gavrilo Princip
•    There are limits on the size of land animals imposed by gravitation and the properties of muscle and bone

I chose these examples because they have/had wide consequences for evolving systems and are often considered to be important. Most people would agree with them. Why do they work as explanations? There are some common factors: they either concern proximate causes (A was shot by B, the dinosaurs were killed by a meteorite) or else they concern external constraints (gravity, physics). The dinosaurs were an evolving system, but the meteorite was not, and there was no time for dinosaurs to respond to being killed by one. Similarly, land animals have evolved a variety of body shapes and sizes and methods of locomotion, but physics imposes some limitations that cannot be out-evolved, and the physics itself is unaffected by evolutionary forces.

The general form of these explanations is:
[external event or constraint, non-evolving] x [evolving system] -> outcome

Properly formed and researched, these can be valid statements and thus constitute true explanations. They are not necessarily the whole story (why did Gavrilo Princip shoot the Archbduke?) but they work within their own scope.

However, in discussions concerning religions and moralizing gods and suchlike the types of ‘explanation’ we are asked to consider are of a different type:
[emergent property A] x [emergent property B] -> outcome

In this case both property A and property B are part of the same system, and both are evolving in tandem, and are interacting with a great many other parts of the system that are also evolving. This brings me to my main point: such relationships are, at best, statements that certain things tend to be associated with each other. They are not explanations, even though they may appear to be so, or may be associated with some intuitive notions of 'causality'.

I am aware that this is an unpopular view, since it spoils a lot of entertaining games, such as the writing of interesting books that pin the history of humankind on one particular emergent phenomenon (wars, religion, trade, brain size, opposable thumbs, swimming ability and so on). These are ‘Just So’ stories, gently satirized by Rudyard Kipling in the children’s book of the same name. They may be written with great erudition, and may contain many curious insights and anecdotes, but the premise that one emergent property can be explained on the basis of another is flawed.

This does not mean we should not look for associations and correlations … quite the contrary ... rather that we should consider constraints in the first instance, and then seek correlations. The first may yield partial explanations, the second will give us a rich account, but few satisfying explanations.

The study of the emergent properties of complex systems, which means the study of the increasing differentiation and modularity and new types of connectivity that gradually emerge in such systems, alongside their evolutionary trajectories, offer a more promising route to understanding evolving societies. This is a steeper and rockier road, and one that is unlikely to be marked by neat explanations such as ‘moralizing gods support complex societies’ (or vice versa), but it is more likely to yield insights in the long run.