He calls it the omnigenic model.
the sciences where progress is hard are the ones that have what seem like an unfair number of tiny interacting causes that determine everything. We should go from trying to discover “the” cause, to trying to find which factors we need to create the best polycausal model. And we should go from seeking a flash of genius that helps sweep away the complexity, to figuring out how to manage complexity that cannot be swept away.
I prefer the term “causal density,” which James Manzi introduced in Uncontrolled. Many economic phenomena are characterized by causal density. Unfortunately, the mainstream approach is to “sweep away the complexity” by coming up with the simplest possible model that might explain some phenomenon.
Like Einstein is reputed to have said, models should be as simple as possible, but no simpler.
I prefer to think of it somewhat differently: models should be used with appropriate humility. That is, instead of being viewed as mirrors or maps to reality, one should see “models as a technology”, mean to achieve a particular purpose (even if just to aid understanding and illuminate relationships) and for there is a “safe operating window”, outside of which they are prone to failure.
So the canonical example is that a Newtonian view of space and time and gravity is “too simple” and “wrong”, but near perfectly good enough for just about any purpose until speeds or scales in which relativity starts to matter.
When causal density rises too high, there may exist no models able to accomplish the goal of providing actionable policy advice in terms of allowing a confident prediction that a particular outcome can be brought about.
That being said, one should be cautious about making the mistake of over-using the “casual density” criticism and being too eager to dismiss some simple models that really do have extraordinary explanatory model. Some things really do pass the “dominant single dimension” test in a factor analysis, despite emerging out of hopelessly complicated situations. Whatever IQ is, or whatever people think about the tests, it remains remarkable good at statistically predicting average life outcomes. Very simple evolutionary assumptions lead to clean mathematical models about the origin of disparate variability of sexual dimorphism, which we observe in nature. And despite all the possible things that one can imagine influences human psychology, personality, and behavior, a focus of social status ranking really does seem to be of principle importance.
It would be very easy and prima facie reasonable-seeming to throw a “these matters are too causally dense to permit such a simple model to be that good at explaining and predicting things” criticism at these phenomena. But it would also seem to be an error to do so.
FWIW, Newtonian physics also breaks down at very small scales, when quantum effects become important.
The pithy little saying that seems to go with your point is, all models are wrong but some are useful. As someone else said here, we need the wisdom to know the difference.
Reminds me of “physics envy”. Many of the results in physics are so simple and powerful that I think every scientist hopes an analog exists in their own field.
Newtonian physics owes its existence to planets, pendulums, and cannonballs. Those were the only three systems in the early modern world that were simple enough for Newton’s laws to be valid; everything else has complicating factors (friction, if nothing else).
Economics and other social sciences are more like protein folding than pendulums (pendula?). They’re too complicated for simple models to be useful. Greenspan knew as much about economics as anybody, but after 2008 he was forced to admit that there were unforeseen factors at work of sufficient magnitude to bankrupt millions.
Sounds like we need a complexity prayer to go with the serenity prayer. “And wisdom to know the difference.”
Another manifestation of this phenomena is the need to control the “narrative”, which highlights yet another incentive to prefer, if not to concoct, a simple (convenient) model of any reality. The incentives to do this are huge. This is why having “skin in the game” is so important in tempering the oversimplification that leads to the misunderstanding of reality.
Such as: the solution to health care is “Medicare for All.” Easy peasy. And the hurricane was caused by climate change. And gluten is responsible for all health problems.
Immediate thought is the applicability of this notion to psychology. There are people who say genetics “explains” schizophrenia (despite the fact that, for example, male schizophrenics are very unlikely to reproduce), and others who say trauma is unquestionably the only cause, which also seems absurdly reductionistic.
I think of the great jazz pianist Bill Evans, who had an abusive father. Bill became an addict, his brother a schizophrenic who ended up committing suicide. Bill’s suicide simply took longer. Perhaps trauma plus differing genes equals bad, but differing results.
There are two methods for dealing with causal density. Neither works.
One way is to simplify until you get a simple, mechanistic model. Applying that model to real, complex situations predictably fails.
The second way is the “big data” approach – add enough variables and try to model everything. That inevitably picks up enough spurious correlations that the model has little predictive power. Example: https://xkcd.com/1122/