Lecture 5: The Mechanistic Hypothesis (Turvey, 2019, Lectures on Perception)

In the previous lecture, Turvey spent a lot of time defending the idea that nonlocal causality is a legitimate option for a physical system. He did this by looking at quantum mechanics and the extraordinarily robust evidence it has provided for nonlocality. The reason he did this is because perception as a two-term relation (direct perception) seems to require at least some nonlocal causality in order to work without additional terms (epistemic mediators). 

This chapter is a discussion of properties and how to organise them - ontology. Part of Turvey's work is to lay bare the fact a lot of our science is still working within the 17th century, mechanistic framework - all psychologists have encountered the idea that there are primary and secondary properties of things, and that we are in the business of understanding how the secondary ones work. But, as Turvey is describing, science has moved on, and the possible types of legitimately physical (primary) properties has expanded far beyond what Galileo, etc thought possible. For example, in quantum mechanics, particles don't have properties such as 'position' or 'velocity' until they are placed into a relation with either a position or a velocity measuring device. The property 'position' is impredicative, both defined and actualised by the presence of the relation to a position measurer. Historically, properties that are both defined and actualised by the property holder is in the relation have been considered to be secondary properties; subjective, not objective properties. But quantum mechanics considers particle position to be perfectly real, and it has the extraordinary level of empirical success required to back that claim up. 

This chapter walks the reader through the development of the mechanistic framework that still drives how we classify properties, thus affecting how we study them. It ends by identifying that this taxonomy begins to creak as we use it at the scale of behaviour, and points to the success of quantum physics to suggest a more expansive and interesting taxonomy of real properties is both required and an option. 

The Mechanistic Hypothesis

Scientists want to study how and why things happen they way that they do. As a first step, we need to correctly describe what is happening. However, it turns out that how things appear and how things are actually unfolding can be different! The sun, for example, appears to rise in the east and set in the west. It took the application of reason to understand that this appearance was misleading, and that the earth orbits the sun instead. This breakthrough is a genuinely radical shift in thinking that changed how we do science. 

The main thing to come from this Copernican Revolution was the use of mathematics to guide reason as we identify which apparent properties of things were real, and which were merely appearance. If a candidate property of an object (say, the mass of a stone) could have it's role in chains of local causation analysed mathematically, it could be a primary, objective property of the stone - part of the reality of the stone. If not (say, the colour of a stone) it was demoted to a secondary, subjective property - mere appearance. 

This then leads to the 17th century mechanistic hypothesis, which will be the boogeyman for many of the upcoming chapters. Physical systems were to be analysed into primary properties who had their effects via chains of local causation. These primary properties make up the matter of a thing, and that matter is moved via the operation of machines. Machines cannot move themselves, and so when, say, the machine that is my body moves, it must be moving because of the operation of a non-physical system. This analysis of properties and how they work is inherently a Cartesian dualism. 

Problematic Properties

This analysis gets complicated, fairly quickly. For example, the appearance-reality distinction does not always help when explaining the behaviour of biological systems. Turvey describes a vine, Monstera gigantea, which finds trees to grow up by growing towards the darkest sector of the horizon. This works because in the vine's niche, that darkest sector is always the nearest tree. The plant doesn't know anything about the reality (trees), it just works with a specifying appearance (this is revealed by experiments that artificially separate trees from darkest sectors). Turvey notes here that growing towards the appearance isn't a mistake, from the point of view of the plant, because the appearance usually specifies the reality. 

The taxonomy of properties into primary and secondary is supposed to help us predicate properties to objects. We can predicate mass to a stone in a way that makes it a primary property. We can also predicate velocity to a stone, because it is a relation between two primary properties (the stone's motion and an environmental reference frame). A property like colour is predicated to an object as a secondary property, because the relation between object and observer that defines it also actualises it.

We can also predicate properties such as 'hammer' to a stone. This property is is both defined by and actualised by the presence of two primary properties (the stone and the person wielding it). This technically makes it a secondary property, but unlike colour the relata are two primary properties, which implies 'hammer' should be an objective property of the stone. 'Hammer' is a potential property of the stone, actualised by the presence of other potential properties of a person, all constituted by real, objective, primary properties. 

These kinds of potential properties of object are important to a science of behaviour. So how do we manage them, scientifically? Here we apply the lessons of quantum physics from the previous chapter; these properties are real, but do not exist until measured. To a scientist working with the mechanistic hypothesis, this makes no sense. But the extraordinary success of quantum physics has shown that such properties can be perfectly legitimate parts of a physical system. We are now working towards showing that such properties can also be legitimate parts of an ecologically scaled physical system (Turvey points to Chapter 22 here, 'Ontology at the Ecological Scale'). 

Reflections

It was again a real slog to find the core here; Turvey isn't making it easy. But he is doing a lot of important work. The ecological approach is going to require a very modern ontological structure, one that is not at all common in science outside of quantum physics. People tend to let quantum physics be weird on the premise that the weirdness seems to be restricted to the quantum scale. Turvey is laying the groundwork to say that while perception is not a quantum phenomenon, it does trade in the kind of properties that makes quantum physics so weird and annoying. This is going to be ok; weird and annoying are not grounds to deny something is real, and quantum physics works so well everyone has had to accept it is onto something. We have a long way to go before we get to weird and annoying at the ecological scale (we have more accounts of perception to discuss, for one), but we are on the way.

I'll just repeat here for the record; when Turvey points out the flaws in the mechanistic hypothesis, he is not addressing the mechanistic approach to explanation. He's referring to a specific type of physics, the kind we still sort of a appeal to for every day events. Mechanistic explanations are not nailed to that physics, and I've argued that approach is a great match for the ecological approach, weird properties and all.