Allan Savory’s Challenge: Fractals
In a TED talk last spring, Allan Savory made a huge splash in the debates on sustainable food and climate change, triggering a cascade of contention and controversy among scientists the world over. When examined in depth, his methods, when applied, seem to work very well sometimes, and to fail at other times; so any ideologue desiring empirically-generated fuel gets it.
Although Savory and his proponents typically attribute the failed efforts to a lack of rigor, which is undoubtedly true in many cases, I think the inconsistency of results is largely because of something more subtle: The conditions for his methods to work with greater frequency are not sufficiently understood. The ideal conditions have not existed on this planet for thousands of years, whenever humans started herding animals and cultivating the soil for consumable vegetation. These ideal conditions include, but are not limited to,
- Plenty of healthy soil, clover, grass, etc. already on the ground, unaffected by humans
- Many predators to influence the migration patterns
- Allowance for many carcasses to die, feed lower creatures, and decompose into the soil
The migratory patterns, and the interplay of parameters involved, are far too complex for us to understand, at least at the moment, let alone reliably “mimic” them (which makes Savory’s successes even more remarkable.) Like the weather and the stock market, the movements in both space and time are fractal. In particular with living organisms the number of variables to consider, and number of feedback loops (think “butterfly effect”) is mind-boggling. The holy grail of modern complex mathematics is a modelling system that can predict fractal patterns. We don’t have one. (My personal opinion is that it is no more possible than a mathematical theory that explains consciousness in Psychology, or one that explains singularities in Physics. It is simply beyond what we can understand, given the types of tasks for which our brains evolved. Of course, Pride makes us think otherwise. Simple, but beside the point.)
Until we have such a modelling system, Allan Savory and his friends have two tools:
- approximate the patterns, and/or
- use intuition.
It’s hard to know how accurate Savory’s approximations are, nor how good his intuition is. Judging from the high rate of success in light of such a formidable problem, he seems to be very good at it, though. It is disturbing that opponents of his methods have so much money, are so ideologically motivated, and hold him to a standard of perfection. They require that the application be successful 100% of the time, or else they dismiss it altogether. This is a little myopic, kind of like saying that since meteorologists’ predictions are frequently off, their methods don’t work at all.
Authentic intuition (ie, that which is testable) is usually something that emerges in the human nervous system after a duration of making direct observations, in combination with an innate propensity which is not well understood. It seems that Temple Grandin has it. She made the observations, intuited, and communicated the patterns that were best for the cattle. However, this was in a feedlot, where the numbers were relatively small, and the geometry challenge was between mere right angles and different types of fairly simple curves – many orders of magnitude less complex than the open systems and sophisticated fractals that Allan Savory faces. It would be nice if some “Team of Temple Grandins” could come along and save us from desertification and the other unintended outcomes of the current human food system. In the meantime, I think Allan Savory deserves a little more credit for taking this as far as he has.
