My baby and his beeping box

Today my baby (who is nearly six months old) was absorbed with a plastic box that makes different sounds when its buttons are pressed. He would stretch his hand out, press down on a button, stop, and listen to the sound produced. He would wait until the sound ended, and then he would reach his hand back out again. I was fascinated watching him so fascinated. It led me to thoughts about buttons.

A button is a perfectly designed cause-and-effect device. The box has been designed so that various physical and chemical interactions are channelled towards establishing clear, consistent and visible causal interactions. The cause-effect relationship is clear in the sense that it is not ambiguous or multifactorial, it is consistent in the sense that it is reliably reproduced every time, and visible in the sense that the entire set up is designed for a prominent experienced effect.

Four most of the universe’s 4.5 billion year history, and in most places in the universe today, such boxes do not exist even though their underlying components may be present in substantial numbers. This indicates that most combinations of the physical and chemical subcomponents does not lead to a clear, consistent and visible cause-effect pathway. In most cases, causes are interconnected and effects are variable, which is why scientific experiments need to be carefully designed and controlled to extract a predominant relation from its context. For example, some combinations of plastic and metal (say if they were dust) might instead leads to many little nonrepeatable causes, like forming a pile or getting pushed by the hand into some shape, or swept away by wind. In other cases, the causality might be invisible, such as would happen if there were slow chemical reactions happening at a microscope level. Even if the shape was retained by the device was broken, we can imagine the same parts participating in different kinds of causal interactions such as being used as a projectile or a bat. If the plastic were still oil it would obviously have a range of other causal possibilities.

Through interacting with a device with well-defined causal dynamics, my baby is learning to see the world mechanistically. As I define it here, a mechanism is a structure that has reduced the degrees of freedom of its components to ensure the reproducibility of certain (unlikely from a thermodynamic perspective) causal paths. We can see how this operates in the plastic box: all its components are organised to reproduce the relation ‘button pressed -> sound effect.’ This relation is maintained at the expense of other causal interactions through redundancy. For example, I could scrape away bits of plastic or wire, and the device would still function. This means that ‘extra’ matter is engaged is harnessed to buttress a specific causal relation rather than doing something else. This is another way of saying they have been highly constrained and put together in an unlikely way, i.e. that the system has reduced degrees of freedom.

As an environmental educator, I have both learned and instinctive suspicions of mechanistic thinking. It is a common critique of unecological thinking that it is ‘mechanistic and reductionistic’ (though I hope it is obvious that these are not the same thing, and in some tension. On reductionism in environmental education, here). One reason environmental educators are often suspicious of ‘mechanistic’ thinking is that one of the key insights, from ecology to relational pedagogies, is that things are ‘interconnected’. Another is that mechanistic thinking seems to suggest that everything is determined (by its structure or organisation), which goes against the freedom and variability we sense in the living world and that we may also see as crucial for ecological ethics or politics.

I will consider the first aspect of mechanistic thinking first. A mechanism actually does not teach that things are not interconnected. Instead it foregrounds and reinforces a small subset of interconnections. Consider how the button pressed and the sound produced are clearly interconnected. When my baby becomes a child, finds a screwdriver and opens the box, he will see that there are a number of components that are in very specific reciprocally dependent interactions that underlying the translation of button pressing into sound. But what mechanisms teach by virtue of foregrounding certain interconnections is that others do not exist or are not important. This happens because the mechanisms has redundant features, as discussed above, that ensure robustness against perturbation from other things, and a buffer against diverse causal interactions. So, precisely because a mechanism is able to show a causal pathway so consistently and clearly, it also obfuscates or diminishes other causal interactions. This is part of the meaning of reducing degrees of freedom.

Before turning to the second aspect of mechanistic thinking mentioned above, I want to point out that even freedom or indeterminacy, even the most robust mechanism teaches my baby something about ’emergent properties’. Emergent properties can be defined in many ways, and are a complicated philosophical topic with many facets (just like mechanism), but for the sake of this discussion I define it as features that unexpectedly arise when material components are configured in specific ways, and that are not aspects of of those components in isolation. In the case of the plastic box, my son is witnessing that the organisation of matter leads to things happening that are unlikely otherwise. This will become clear when he eventually breaks the toy and realises there are vastly *many* possible broken states compared to states that function.

The second feature of mechanisms mentioned is the sense that if something is mechanistic it is therefore deterministic. The case of the plastic box may be quite deterministic and therefore re-enforcing some particular notions about how things work in the world. But many mechanisms that he will encounter are not so fixed in what they do. While all mechanisms reduce causal pathways when compared to the disorganised bustle of underlying atoms, some mechanisms produce emergent properties that do still have indefinite possibilities. For example, a guitar is a mechanism constrained to produce sound but this is clearly an ‘enabling constraint’ that can generate countless kinds of sound in concert with a person’s skill development in drawing them out. Living organisms can be considered mechanisms insofar as the constraints of their anatomy and physiology create entities that behave unpredictably and learn. From this, one might then think a mechanism is something where the underlying components are ‘deterministic’ even if the macroproperties appear not to be, and that those properties are therefore an illusion. But mechanisms don’t even need to have fixed components to be mechanisms. Sometimes stochastic probabilities are sufficient (like how intrinsically disordered proteins provide an indefinite ‘range’ of possible structures that feed into what kinds of broader organismic structure and behaviour are possible).

One might think a kind of appraisal of good vs bad mechanisms is possible, and be wary of having a child exposed to too many of the bad kind for fear of what they may be tacitly learning from them. But it is tricky to define what those might be. For example, from the discussion you might be attracted to the idea that indeterministic mechanisms are to be favoured educationally because they reveal contingency and interconnectedness more clearly. A caution against this is that many indeterministic mechanisms are likely problematic too (for example scrolling Youtube channels is novely generating), while on the other hand it may be that acquaintance with clear and consistent mechanisms is actually helpful for emphasising how life is different from such typical machines. What seems crucial is the kinds of questions and discussions that surround different kinds of mechanisms so that the presumption that the world is ‘just’ or ‘ultimately’ a machine is challenged: either through realising cases where it is not, or by challenging presumptions about what it means to be a machine, or both.