DownwardCausation

ThoughtStorms Wiki

Context: PhilosophyOfScience / Epistemology

We assumed sciences or our models of the world are organized in a hierarchy of abstractions.

But what does it mean to say that the higher level can explain what goes on at a lower level?

Do these higher level entities or abstracta even exist? Or are they merely convenient fictions?

Could there be downward causation whereby the behaviours of the abstract cause the behaviours at the underlying level? This seems weird, right? If the higher level is just aggregates of their lower level components, do we think that the higher level component can force the lower level entities to behave contrary to the laws governing them? If so, that seems very unscientific. But if not, then what does the higher level add? The lower-level and its laws already constrains, predicts and explains all the behaviour of the aggregates. So should we just eliminate high level?

https://www.quantamagazine.org/the-new-math-of-how-large-scale-order-emerges-20240610/

Software in the natural world: A computational approach to hierarchical emergence : https://arxiv.org/abs/2402.09090

Another thought is that maybe the higher level expands the context or frame within which we have to consider.

For example, following its laws, we predicted the particle should have kept flying north.

In reality another particle suddenly came in from the west and knocked it off course.

If we were modelling at the lower level, we'd have been assuming a de-facto horizon beyond which we weren't expecting to have to look. The new particle is an "act of God". Something we couldn't possibly have been expected to hypothesize or include within our model.

However a higher level of modelling (say economics) might have told us that the laboratory we were working in had been build cheaply and not to the standards that we were expecting in terms of shielding from external particles. At that level we could have predicted (at least statistically) the disruption to our experiment from particles further afield.

This is a kind of silly example, but it stands for examples that are more necessary to engage with. DanielDennett points out that the IntentionalStance, that takes into account my desire to go on holiday, can predict fairly accurately that the molecules in my body will be the other side of the world next week.While mere physics and chemistry models would be very unlikely to take into account causal interactions at such great distances.

Of course this still makes the higher level model a mere epistemic convenience. Making tractable reasoning about larger quantities of particles or wider contexts.

Perhaps even consciousness (OnConsciousness) might be one example of this situation.