Interactionism and Nomic Complexity

Interactionist dualism is thought by many to be implausible because it requires the existence of physical laws that look rather different from what we currently know about laws of physics. This objection can take several different shapes. Here are four.

First, if interactionism is true, then the movement of physical particles can't be determined solely by the distribution of fundamental properties like mass, charge, spin, etc. Conscious properties must exhibit some influence over the behavior of quarks, atoms, etc. Conscious properties are not like physical properties, and so laws involving conscious properties would require broadening the possible relata of physical laws.

Second, in addition to the mere fact that we would need to expand the relata of physical laws, interaction requires  a connections between different kinds of properties that some have thought to be mysterious. How could something non-physical effect something physical?

Third, assuming that panpsychism is false, such laws would have an unusually limited domain of applicability. It is a bit strange that there should be laws of physics that only manifest themselves inside the skulls of living animals. We don't expect to find exceptions to existing physical laws inside our brains -- for instance, we don't expect gravity to exert a different force on neurons than on other things. So why hold that there might be other forces that only have their effect on neurons?

Fourth, given what we know about how the brain functions, interactionist laws would need to look very very strange. The strangeness of these laws comes from their massive complexity. This makes them very different from existing laws in a way that really seems to matter.

I don't think that the first three possible concerns are very concerning. There is possibly something problematic about postulating additional kinds of things, but parsimony doesn't seem like a super strong theoretical consideration. If we have no evidence for something more than physical stuff, then of course we shouldn't posit anything more. But if we do have some evidence, we shouldn't let parsimony stand in our way. Laws of nature look fairly arbitrary as is, so I don't see any major problem in postulating laws between conscious properties and physical properties. And finally, if non-physical consciousness exists, it is not unreasonable to think that it requires fairly special physical circumstances, and that will be where the ordinary physical laws need supplementation.


The fourth concern however, provides more than enough reason to reject interactionism.

The brain  interacts with itself and controls the body through the signaling of neurons. Neurons communicate with each other using potentials that travel down their dendrites/axons which leads to chemical and electrical signaling across synapses. Any interaction between the mind and the brain would have to result from an influence of conscious events over one of the following:

• local electric fields (directly)
• gates (to generate electric fields)
• neurotransmitter release

And of course in order to influence any macrophysical structures like nerve cells (or Na+ channels), the mind would have to influence physical objects at lower levels: individual atoms, the quarks that they are made up of, etc.

There is nothing intrinsically odd about this. Why not allow that the occurrence of sadness or the intention to move one's arm might produce or modulate some movement of atoms?

What is odd, however, is the way in which conscious experiences would need to have a very specific influence in order to exert precise control over the brain.

Consider how an intention to move one's arm might lead to an arm actually going up. Let's suppose that the mind works directly on the motor cortex, rather than at some connected area. In order to get the arm to move, the mind would have to cause the nerve cells responsible for motor control of the relevant muscles to fire. Such nerve cells are tightly packed in alongside each other in the topographic maps of the motor cortex. To avoid causing the fingers to twitch or the shoulder to jump, the mind would need to only tickle the nerves corresponding to the muscles of the arm.

 How does this happen? The neurons that are responsible for arms are not much different from the neurons responsible for other parts of the body. Their main difference is not in their neural structure, the kinds of neural signals they employ, the strength of field they generate, or sorts of gates or neurontransmitters that they use.  Their difference lies just in their location in the topographic map, and crucially, the parts of the peripheral nervous system that they connect up with. 

Having a precise physiological effect like moving the relevant arm requires that the conscious experience directly tickle parts of the brain characterized in terms of their downstream connections rather than their local physiology. This is what makes the process so strange. What determines which neurons are influenced would need to take into account very very fine-grained facts about the neurophysiology of the creature at large: whether the intention causes a particular channel to open depends on the structure of the neuron it is part of, the structure of the neurons it is connected to, and perhaps its eventual effect on the peripheral nervous system.

Spelling this complex property out in physical terms at a fundamental level would be extremely difficult. 

The problem is compounded by the fact that conscious states seem to be simple. Even if we allow that conscious states reside spacially in the brain, we are still faced with the problem of explaining how a mental state that presumably isn't local to the neurons that it influences is able to influence those neurons and not the vast majority of other nearly-identical neurons. The only way to do this is to build hyperspecific laws. But such hyperspecific laws look very different from the laws we are accustomed to.