People love to talk about improvement. Adding this improved the fertility of the soil. The yield of apples was improved with this pruning technique. This piece is a reflection on the limitations of this approach when managing complex systems and alternative mindsets that may prove more useful.
The first blind spot to consider is the problem of measurement. When we grow potatoes for example they serve a large number of uses. Their primary use is as a source of carbohydrates, but also provide protein, some minerals and vitamins, flavour and texture. Even these broad categories could be further broken down into more nuanced details. If you are managing a potato crop but only measuring the total yield of tubers then it is likely you will eventually compromise one of these many roles served by the potatoes. This easily happens if you choose the highest yielding variety. You will almost certainly select a starchy potato that is low in protein and high in water rather than a waxy variety. You could also select a variety that produces large numbers of tiny tubers that are difficult to harvest and process. If you were only selecting for the largest tubers then perhaps hollow ones would win out. There is an old adage that what gets measured gets managed. In the age of industrial vegetable production where the weight and appearance of a crop are the primary considerations it is not surprising that the flavour and nutritional value is sacrificed over time. With improved measurement techniques modern breeding is getting better at paying attention to the full range of factors, but illiterate peasant farmers have been successfully breeding crops for thousands of years using nothing more than their senses.
The next major pitfall is the simple link between more and better. A good example of this is soil carbon. After generations of industrial farming many soils have depleted soil carbon, so increasing it through various techniques will give a wide range of benefits. However the real relationship is more complex. As soil carbon levels increase eventually you reach something akin to a highly acidic pure peat soil, where minerals are so scarce and immobile that only specialist plants can grow slowly. Most relationships fall more into the shape like a bell curve with an optimum value somewhere in the middle. On the upslope, long before adding more of the input causes the output to fall, the return on extra added input starts to diminish. So for example if you gradually added more manure to poor soil the yields would initially increase dramatically, then only a little more as the plants needs are mostly met and other factors become limiting. With further added manure the plants would start to suffer excesses and maybe just be plain buried in the manure.
This idea of there always being a limiting resource in biological systems is called Liebig’s law of the minimum. Living things need many resources to grow and reproduce and any single limiting ingredient can limit overall output. Think of life like the recipe for a cake with different ingredients in fixed proportions. If you run out of eggs but still have everything else on hand you can’t make any more cakes. It is also sometimes explained as like a barrel where every slat in its side represents a different required input. Whichever stave is the lowest limits how much water the barrel can hold. There are a couple of limitations to this model (ironically perhaps?).
Firstly, there are complex interactions between all of the inputs, so they don’t act independently. The bioavailability of many key minerals for example depends on the relative abundance of a competing or complementary mineral. Calcium and magnesium can compete, as can copper and zinc, and with 25 elements essential for life there are a vast number of potential interactions. A similar situation occurs in the intake of essential amino acids in animals where an overabundance of one can limit utilisation of one that is scarce. Secondly while the basic composition and requirements of all organisms is similar they have all developed different adaptations to allow them to cope with varying imbalances in their environment. Different groups of plants are adapted to soils with unusual pH levels or an abundance or scarcity of specific elements. The bioavailability of soil nutrients depends critically on the microbial communities in the soil. Plants form specific relationships with different microbial communities to absorb all their requirements, including water and macronutrients. These are the main tool that allows species to adapt to different mineral balances but is highly species specific. This is one reason why the vision of a garden with every kind of vegetable growing in abundance is folly since their needs are often irreconcilable. The mineral balance of the underlying geology can barely be dented by human efforts to import fertility even with the full power of industrialism, possibly with the exception of a couple of trace elements like selenium and molybdenum that are needed in very tiny amounts.
There is another potential trap in the concept of optimisation. That simple graph with a single optimum might be a smaller part of a much more convoluted landscape. Rather than a single optimum there may in fact be many. If you only make small adjustments in the inputs of the system then you will remain trapped at whichever optimum you first encounter. This is kind of like tuning in on one radio station without realising that if you kept turning the dial you would eventually reach a station that you like even more. A good example of this occurs commonly in plant breeding. Modern dessert type bananas are the result of hybridising two wild species to produce various infertile clones. These are unable to produce seeds, resulting in fruit filled with sweet pulp rather than tooth shattering seeds. In a single step the crop is improved amazingly, but because it is infertile it cannot be further improved. It is a dead end from a breeding perspective, resulting in inevitable disease outbreaks and periodic industry collapse. Though modern biotechnological techniques are finding new ways forward it is also theoretically possible to return to the wild seedy species and go through a more prolonged breeding process to develop bananas that produce just a few seed, and potentially allowing even more productive or disease resistant forms to be continually produced.
The idea of an optimum between competing factors, like increasing yield and decreasing taste, can also be constructed but suffers the problem of needing to weight the relative importance of each factor. There is no universal way to do this as it depends on the relative importance of the various factors, a deeply personal choice. In the absence of anything else the financial impact of all these factors can be weighed up to maximise profitability, but this seems to mostly lead to farms that produce massive quantities of pretty but empty produce. The landscape has so many different dimensions to consider it would fry Einstein’s mind trying to visualise it all at once. Instead science tries to isolate a small number of factors at a time and optimise them before moving onto another factor. With a landscape littered with numerous local optimums this approach is still limited to finding the closest one. There are two types of evolution in a way. One occurs when organisms are pushed to their limits, resulting in them being honed to the nearest viable optimum. The other occurs when the restraints are lifted and the organisms are free to wander the landscape of possibilities without immediately being eliminated.
What is optimum in the short term is often detrimental in the long term. The logic of industrialisation is centred on this idea where present productivity is valued above sustainability. Adding fertiliser inputs or increasing irrigation can increase returns this year, but undermine the functioning of the system in a way that decreases yields over time. The most pernicious of these issues are ones that creep up gradually over many years or even over multiple generations. Grandpa on the verandah will tell everyone that the potatoes from that field are nowhere near as good as when he was a boy, but who would take him seriously given the frailty of memory? Even worse, issues like erosion or mineral depletion can take much more than a human lifetime to take effect. Can a peasant farmer feeding his family seriously consider changing their practices to reduce yields today in the hope that people in a thousand years will have more?
One final factor also needs to be considered when optimising a system. As the system approaches its optimum, balanced and strained between a wide variety of competing factors, the optimisation landscape can become increasingly steep and treacherous. Like a violin string that is tightened to produce a higher note it also gets closer to the breaking point, with outputs potentially fluctuating more wildly as inputs vary. Like a village where there is zero food waste and people are perfectly efficient and economical, there is no buffer if the crop yield decreases by 20 % one year. Efficiency and resilience are often opposites in complex systems.
Hopefully this meandering rant has given you some new mental tools to avoid some of the pitfalls of the modern optimisation mentality. Sometimes less is more in a multitude of surprising ways.
One thought on “Good, Better, Best”
I can see the reasoning of Fukuoka trough your writing, it is heart warming to see. Another good article.
I wish society will slowly transform once again from short-term gains to long term sustainability and most important of all resilience.