Expanding Causality: Constraints Within a Relational Ontology

How Constraints Shape the Possibility Landscape of Complex Systems

In this article I will try to expand our common notions about causality. Typically, we tend to interpret causality in a mechanistic manner, meaning that only direct and forceful interactions are considered causes, which grounds itself in Newtonian principles of causation. This type of causality is known as the law of cause and effect (i.e., A causes B). However, there are areas in which another story is more fitting than a linear deterministic relationship between cause and effect. I want to emphasise that my intention is not to undermine the importance of efficient causes, but rather to complement the toolset by which we can understand the world. This complementary story speaks about constraints as factors that limit or enable possibilities.

Why Complexity Applies

Juarrero’s work is interesting for a variety of different reasons. In the last article we discussed issues about identity and how that relates to the paradox of the One and the Many. What makes something one or many? Another way of talking about this is in terms of wholes and parts. Especially important are the relations between wholes and parts. Reductionism holds the position that we can reduce wholes to their parts and that is true as long as we don’t deal with emergent phenomena. The general assumption is that “because wholeness is nothing but particles and their secondary interactions, any seemingly novel properties of purportedly coherent wholes can be derived, in principle, if not in fact, from laws pertaining to its constitutive parts.”¹ The relevant point for this discussion is the belief reductionist hold, namely that everything that is important can be explained from bottom-up processes, meaning that the causal power resides only within the parts and not within the interconnected whole.

When you deal with linear relationships within controlled environments best described by discrete mechanisms, the aforementioned approach is suited best, but as soon as we expect interdependencies between parts that create feedforward or feedback loops or nonlinear dynamics, reductionism hits its limitations.

Paradoxically, in the face of the primordiality of relationality, the attempt to particularise and individualise a phenomenon, or in other words, the attempt to reduce a whole to its parts in the hopes of understanding it, inverses what one tries to achieve. In another context Filler writes, “the more one becomes individualised, the less one is”. I would argue, this is the mistake reductionism, also known as the “nothing but-ism” makes. By reducing wholes to their parts, it narrows our vision and we start to lose sight of the connections that contributed to the generation of the wholes in the first place by cutting off the strings of relation.

What are Constraints anyway?

“Constraints” can be understood as “entities, processes, events, relations, or conditions that raise of lower barriers to energy flow without directly transferring kinetic energy”² Juarrero explains. But what does that mean? With the notion of constraints she wants to emphasise, among other things, the importance of context dependence. The context in which something happens, influences what happens - not through mechanical impacts, but by shaping the probability landscape of what is possible. The easiest example I can think of is road traffic. The cars velocity is dependent on the road and the cars in front of them. A highway affords higher speeds than a back road in the countryside. What greatly determines the car’s velocity from the standpoint of save driving in road traffic is not only the type of car or the person steering the wheel but also the shape of the street and the traffic conditions. In the end, every car has to slow down at a sharp turn and everything else gains less importance.

Different Types of Constraints

If we imagine a universe at thermal equilibrium, in which everything is equally probable, context-independent constraints take such a universe away from equiprobability Juarrero argues. Because these constraints create inhomogeneities in energy flow we see the probability landscape changing. She explains that “context-independent constraints take conditions away from equilibrium. They render conditions, events, and processes that were equally likely no longer equiprobable.” Furthermore, she asserts that they “turn the space of possibilities in which a system’s events and processes play out nonuniform or inhomogeneous.”³

Context-dependent constraints, on the other hand, are a little bit different. Whereas, context-independent constraints make all events upon which they have an effect unequally likely, context-dependent constraints start to connect certain events that were previously independent of another. This makes them mutually conditional on one another and, in doing so, begin to the shape possibility space.⁴

A good example to illustrate context-dependent constraints and the emergent properties created by coordination dynamics is the tragedy of the commons. This refers to the dynamics created by a situation where individuals begin to overuse a finite resource, because they act in their own self-interest. For example, if fisherman prioritise maximising their own catch without keeping in mind the finite nature of the resource, it will lead to a depletion of fish stocks, ultimately harming everyone who relies on the resource. While fishing in the ocean is an independent activity as long as enough fish are available, it quickly becomes mutually conditional when the resource is depleted. Individual fishermen are then forced to limit their catch if they want the resource to recover, but they are often unable to do so because they rely on profit.

Another fascinating and often overlooked type of constraints are temporal and spatial constraints. Juarrero gives a bunch of examples in which timing as a temporal constraint matters. For example, if a child wants use a swing, it must not kick randomly but at the appropriate time to swing higher. What matters is not the force alone, but rather the force combined with the proper timing.⁵ Something similar applies to seesaws, but this time spatial constraints are at the center (although timing matters as well). A child cannot seesaw, if it sits too far towards the tipping point. Seesawing only works because there are background constraints that establish where the child must sit given its weight.⁶

Top-Down Causation within Biological Systems

Before we can understand top-down causation, we need to delve deeper into what makes it possible in the first place. Juarrero defines enabling constraints as “context-dependent constraints that irreversibly link and couple previously separate” parts “by lowering barriers to energy, matter and information flow”. They make the probability of events that have been independent before conditional on each other.⁷ Imagine you are in a classical concert and applaud for the orchestra. All of a sudden, you begin to notice a rhythm forming and you decide to join in and clap according to the rhythm. This is a instance of a enabling context-dependent constraint. Juarreo explains that they are “nature’s mechanism for coherence making, generalisation and emergence.”⁸ The individual claps turned from independent asynchronous claps to dependent synchronous claps. Believe it or not, this exact thing happened to me a while ago - it was remarkable.

But what keeps the rhythm in place? Juarrero argues that the possibility space is stabilised by governing constraints. The beat of the rhythm affords the appropriate timing of all future claps - by this, governing constraints limit possibility space, by biasing the likelihood of future events. I want to stress the point that all emergent phenomenon depend on the metastability that governing constraints afford. They regulate the interdependent processes between the parts, by controlling the dynamic’s synchronisation and coordination. If this metastability begins to break down, for example, if enough people lose sight of the rhythm, the whole dynamic falls apart.

Since all of that might seem abstract and mere philosophical wish-wash that leads to nowhere I want to emphasise that this different way of thinking has already practical applications. It is always important within theoretical speculation to think about how ideas can shape new research programs to test their merit.

Micheal Levin, an American developmental and synthetic biologist, does just that and works on various different highly interesting research topics. He argues that although reductionist models in biology are successful by assuming that causation flows from the parts to the whole (i.e., bottom-up causation), a complementary strategy may be helpful to understand biological phenomena, which recognises causation as also occurring from the top-down.⁹ This top-down causality, as described here, should be understood as governing constraints. What Levin tries to achieve within developmental biology is “the programming of shape by specifying organs and their topological relationships, instead of attempting to micromanage the construction at the ‘machine language’ level”.¹⁰ One pathway towards realising this hypothesised possibility towards growing organs is through bioelectricity, since we know that bioelectric signal influence cell growth, migration and differentiation. This has important implications for future medicine and human health.

Conclusion

This concludes this short exploration of constraints. In reality, this article only scratches the surface. Take it that constraints are abundant and all around us. We rarely notice them because they are often accepted as background constants (e.g., gravity). I hope to have showed why classical substance ontology isn’t enough to account for emergent phenomena within complex systems. Constraints don’t fit the classic picture, because they aren’t any concrete thing per se, like substance ontology would suggest. Within a relational ontology, constraints realise themselves within relations and therein have actual effects on possibility space, by either limiting or enabling possibilities.

1. Juarrero, A. (2023). Context Changes Everything: How Constraints Create Coherence. The MIT Press, p. 11

2. ibid, p. 40

3. ibid, p. 49

4. ibid, p. 67

5. ibid, p. 45

6. ibid, p. 41

7. ibid, p. 41

8. ibid, p. 71

9. ibid, p. 71

10. Pezzulo, G., & Levin, M.. (2016). Top-down models in biology: explanation and control of complex living systems above the molecular level. Journal of the Royal Society Interface, 13(124), 20160555. https://doi.org/10.1098/rsif.2016.0555