What were ancestral diet patterns? 🍠
Taking a view of the landscape we emerged from
Summary
The idea of the paleo diet has changed how a lot of people think about diets. The key insight is to be informed by how our ancestors ate in the palaeolithic period.
There are good reasons to think the main aspect of the ancestral diet was variety. If we want to preserve this we need to be creative in how we eat.
Emergence is a difficult to define process that roughly happens when more complex patterns emerge out of simpler patterns added together.
Many of the critical insights about emergence first started to be developed by playing with computer models that were the progenitors for today’s machine learning techniques.
What can we know about ancestral diets?
Among the many diet paradigms that have emerged in the last 20 years, the paleo diet has an interesting place within the fold. In recent years, as a specific protocol for what to eat, it has been eclipsed by the idea of keto diets. However, the theses that paleo promotes are increasingly important for how many people think about how they eat.
The core idea is that the majority of human evolution occurred during the palaeolithic (hence 'paleo') period from ~3 mya to ~10,000 years ago. For this entire time the food landscape remained more or less consistent. In the comparatively short period since then the environment has changed dramatically. First as a result of agriculture, then again in the modern era with the rise of ‘processed’ foods.
The argument goes that because the majority of our evolution occurred in the palaeolithic environment our bodies are suited for the kinds of foods that were available then. In practice this means we should stop eating modern processed foods, as well as anything derived from the grains and other cultivars that only became available to eat with the rise of agriculture.
There are meaningful questions that could be asked about this theory. In particular, around the gradualistic concept of evolution that it implicitly assumes. The punctuated equilibrium model of evolution would suggest that in fact most major evolutionary changes happen over time spans shorter than 10,000 years. If that were the case then the ways our diets have changed in the last 10,000 years could actually be more instructive of how we should eat then anything gleaned from a study of palaeolithic times.
That said there is other evidence to consider like the studies linking modern diets and disease that add to the picture. For now, let's take the principle of the paleo diet at face value. If we believe the way our ancestors ate is instructive for our diet the major question still remains, what exactly did our ancestors eat?
It may be easy to say that it was some combination of meat and vegetables but that doesn't meaningfully narrow down the possibilities. What proportion of meat and vegetables was it? Which meats and vegetables? How often did they eat? When they did eat did they cook the food, if so, in what way? Wait... What about fungi? Did they ever eat minerals like salt directly? If you start down the path of trying to survey the archaeological record and imagine the life of a fabled hunter-gatherer answering these questions becomes fraught with difficulty and most analysis ends up amounting to guesswork.
A different approach could be to turn the question on its head: Why did our ancestors choose to move towards agriculture and stop eating the way that they did [1]?
Presumably there was some downside to the old ways that our ancestors wished to rid themselves of, what was it? One of the more likely candidates seems to be food security. In a hunter-gatherer lifestyle you are always at the whims of animal migration patterns, geography, climate, and luck. You can easily end up in a situation without enough food to live. If you can get agriculture right, you might not solve this problem altogether but you can go a long way towards alleviating it.
If this is true then we have learned at least one thing: in a pre-agriculture paradigm without a stable food supply our ancestors would not have been picky. If you’re worried about being able to eat at all you take what you can get. There are a some other observations that encourage me in the importance of this point.
First, the way that babies eat. A baby will eat anything. Watch how a 2 year old navigates the world - if they can put it in their mouth then it's worth trying out. This kind of behaviour makes sense if you imagine they are preparing for a world where food is scarce. If you are always looking for your next meal you need to make sure you eat everything you are able to. In this paradigm a strategy of trying out all the options would be a good one.
Second, the variety of culinary science. Today the preparation of food is one of the most diverse human endeavours. We seem to constantly be trying to outdo ourselves to discover new ingredients, preparation methods, and configurations. Such exploration speaks to a level of creativity built in to the way humans operate with our sustenance.
Third, the array of archaeological evidence attesting to complex human behaviour. For almost any kind of food or preparation methodology there seems to be at least one study that has some (even if quite flimsy) evidence for its existence among Paleolithic humans. Such a wealth of material and possibilities indicates a real variety.
Fourth, one of the things we know for sure about humans from archaeological research is that by ~200,000 years ago we had migrated to essentially every corner of the globe. This means that at the very least we were adaptable enough that, prior to agriculture, we could survive on whatever the world was willing to offer us across it's surface.
Fifth, there are general principles of complex systems that make this variety seem likely as the default state. This may be the least empirically or deductively developed view but for me is probably the most compelling. If you accept the principles developed in Antifragile including the view that living systems are often best modelled as those things that ‘“gain from disorder“ then it becomes hard to not assume we are built from the ground up for chaotic environments.
All of this leads me to believe that the ancestral diet - something hard to imagine in specificity - may be more easily pursued through exploration rather than discipline. A paradigm for eating that is more attuned to creativity than analytical precision. Perhaps the thing for our diets that is more important than any one food, configuration of foods, or eating pattern is to seek out variety.
[1] There is a question here about human agency, how can we know our ancestors ‘chose’ this path rather than falling into it? In this case I think this is answered by the sheer improbability of ‘falling‘ into agriculture (especially when it seems to have happened in multiple places independently). However, apart from the colossal effort and extreme risk that the first agriculturalists must have faced and overcome - I am personally just biased to always assuming human agency as the primary factor in any major historical shift. Anything else is much harder to use empirically or convincingly as a consistent driver of behaviour. We are just too complex.
📘 Book Review: Emergence
I was first introduced to the concept of Emergence when reading Reinventing the Sacred by Stuart Kaufman. The general sense that I got from that book was that emergence was the key to new scientific paradigms and incorporating science into a holistic worldview. For this alone it was immediately compelling. I picked up this work by John Holland as another considered classic in this field to further my understanding.
After reading Holland, and since then being enrolled in ACS101 I’ve realised that while many thinkers have developed and relied on concepts of Emergence it remains difficult to define. Like the systems that it describes emergence itself seems to come to light through the interaction of sets of simple observations and repeated patterns rather than a top-down scientific description. The best way I've discovered to understand it is as the creation of new complex realities from simple underlying rules [2]. In this way a tree is emergent from a seed, the gaseous properties of nitrogen are emergent from N2 molecules, and the stock price of Gamestop is emergent from the individual behaviour of the individuals looking to buy and sell it.
To put it more precisely you can see emergence happen with physical systems where the simple behaviour of individual agents interacts to perform more complex patterns in aggregate. A canonical example of this is how the interactions of ants, each of whom have a very simple set of rules for what they will do leads to the creation of ant colonies that display many more complex patterns.
Emergence, more shockingly, can also happen across ontological levels. In physics there are a set of universal laws that govern the behaviour of atoms. The interactions of these laws gives rise in turn to the laws of chemistry and the interactions of molecules. Although the chemical laws are ultimately constrained by the underlying physics processes they nonetheless operate on a dimension all their own not reliant on an explanation at the level of physics.
Investigating this phenomenon in ‘Emergence: From Chaos to Order’ I learned two things, one intended and the other incidental.
The intended aim of the work was to lay the groundwork for a more formal definition of emergence. Specifically this meant developing a type of modelling that could be used to 'map' emergent processes. The belief is that if such a repeatable type of model could be produced then it would be possible to reflexively investigate that model to learn more about emergence in the abstract. Hopefully in that process we could take out some of the surrounding mystery.
For this approach Holland introduces the idea of a Constrained Generating Procedure. In honesty, I found the sections of the book where CGPs were given their proper formal treatment tough. To really develop the idea I imagine I would need to think through and build one from scratch with real numbers to grasp experientialy and intuitively how they work. That said the framework did engender insights at the high level about emergence. Many of which Holland himself called out. In no particular order these included:
Systems that exhibit emergence are commonly made up of many small sub-system, these could be described as 'agents' of one kind or another. These subsystems, although simple, lead to emergent behaviours through feedback loops where the outputs of one agent become the inputs for another. This can also be the source of the non-linearities that dominate chaotic systems.
Emergence is usually a persistent pattern formed by the repeat behaviour of the underlying systems. These persistent patterns can include rotating sub-units, as in the way that an organism replaces all the molecules in its body while retaining the same body.
These emergent patterns can be subject to phase changes that lead to new dominant, emergent, patterns within the same systems.
The macrolaws that can apply to the interactions of these emergent patterns can be difficult, maybe impossible, to predict a priori from the rules that govern the subsystems. For example, the rules that govern a gazelle's grazing habits would not allow you to ahead of time understand the lands the herd would choose to roam. These macrolaws are likewise subject to phase changes and 'shocks' that invalidate their paradigms.
The book is dense. These ideas alone are worth mulling on for many days, if not years a-piece. However, there is still much more in the book and in the concept of emergence in general to explore. I'm excited to re-read this with more context in future.
Incidentally, a surprising thing I learnt from the book was something of the history of neural networks and the origins of machine learning. The book was originally released in 1998 and in it Holland goes through many of the things that he and his colleagues had learned and continued to learn from studying neural networks and other machine learning tools starting in the 1950's!.
Even more shocking and perhaps prescient are Hollands predictions that the "winter" of development in these fields that he and his colleagues were experiencing would eventually pass and that new and amazing things would be done with these models. It gave me a new perspective on the history of development for these techniques and perhaps even more excited for the future possibilities they present.
[2] That said, this more formal treatment I read has also helped tremendously with my understanding.
Photo by Jisun Han on Unsplash
I like this photo for all the kinds of emergence it highlights. It has the natural forms of emergence like the palm trees and the bushes emerging out of the ground with complex patterns in the foreground. Behind that there is the emergence of particular architectural forms, in this case maybe with some inspiration from the natural ones. Behind that the emergence of city-scape forms as different architectures and building paradigms overlap on each other over time.