Aspects of Causality

Prompted by reading The Oxford Handbook of Causation

I’m not going to talk too much about the book itself, except to say that it was way too short. Although it’s possibly strange to say that about an 800+ page book, causation is a huge topic and each approach to it is deeply embedded in the thought and technology of its time. Reflecting this, it’s best to consider each chapter as a quick overview accompanied by a useful set of pointers into more comprehensive literature (and each chapter does have pointers to further reading.

In general, the book reemphasized the extent to which our expectations are predicated by our interaction with the world. Which to use Brian Cantwell Smith’s phrasing in On the Origin of Objects  occur at a middle distance — our everyday experience of interacting with the world is with “human sized things.” Things that we can sense and manipulate with our bodies. These experiences are what ground our intuitions, when we move to wildly different  scales, they tend to lose their grounding requiring constant effort on our part to adjust our expectations accordingly.

This became apparent to me when reading the discussion of Hobbes and the causality of motion

Hobbes (1994: 7) is clear that ‘motion produceth nothing but motion’, but again produce is itself a causal concept. Such insights offer no explanation of how motion produces motion.”

Which just strikes the modern reader as being “off” aka what about the whole F=ma thing?  and then you realize the Hobbes was writing in 1631, while Pricipa was in 1687 — it’s hard to think in frameworks that haven’t yet been discovered, and those frameworks that we have available inevitably deeply color what constitutes a “good explanation”.

Along those lines, it’s ironic to read about those writing in the “classical physics era” who wanted to reductively ground causality in “fundamental physics” aka the “fundamental physics of their time.” — their desiderata bear little relationship to the “fundamental physics of the early 21st century” which has become untethered from the billiard ball causality they desired (as we realized that the this deterministic billiard ball causality is essentially an ensemble effect of the fact that billiard balls contain an unimaginably large number of fundamental physical particles). I can only wonder what parts of our desiderata will be considered ironic to the readers of the 25th century.

As I’ve mentioned,  I’m a big fan of the manipulationist approach and believe that our desire and ability to manipulate the world has a large impact on what we deem to be acceptable explanations of causation (likely reflecting my pragmatic/engineering background). 

My take is that there are two aesthetic axes in the examination of causality: the pragmatic and the Platonic

Pragmatists use manipulability/impact upon the outcome to evaluate the adequacy of a causal explanation, e.g., can you change something that changes the outcome?  If not, you don’t really have a handle on the cause. Now the cause may not be apparent at any level (think Alzheimer’s), but the goal of understanding the cause is to influence the outcome. Similarly, it might not be possible in practice to implement the change that would impact the outcome (the classic example here being plate tectonics), but the ability to model the process and make changes to the model that match results in the world will suffice.

In my characterization, Platonists encompass both reductionists who wish to reduce the explanation to first principles (quoting Michael Tooley’s chapter) 

“did the other British empiricists, Locke and Berkeley—that while some concepts can be analysed in terms of other concepts, in the end analysis must terminate in ideas that apply to things in virtue of objects having properties and standing in relations that can be immediately given in experience”

and also those who search for a detailed understanding of the processes though which things occur. This search wouldn’t have a manipulationist goal but would be searching for the first principles appropriate for the matter under consideration. These first principles need not have a material realization. Canonical examples here would be deists who look for how the mind (soul) controls the body, or the structure of free will in the presence of an infinitely powerful god.

 I don’t think there’s any hard boundaries between these two approaches; recall my discussion of Object Oriented Ontology  which ended up referencing corrosion simulations. The same problem can be addressed by either camp, it’s just that what’s a successful explanation in one camp will often be characterized as completely inadequate in the other.