I read this book as a follow up to the article “Is reality real? How evolution blinds us to the truth about the world” by Donald Hoffman in New Scientist. In this article Donald Hoffman asserts that our subjective reality does not have a straightforward relationship with objective reality.
Most people, myself included, would probably assume that what we see and hear has a direct relationship with the objective world. We might not see, hear or smell as well as some animals but what we can perceive is at least real.
Donald Hoffman asserts that this is not the case. He makes the case that evolution optimises for fitness, not truth and that therefore what we perceive does not have a simple mapping to reality. He then takes this further to say that both spacetime and the neurons in our brains are just part of our subjective reality and not objectively real.
According to Hoffman, attempting to work our how human consciousness arises from our neurons is the wrong problem to tackle because our neurons themselves are not real, just another item in our subjective reality.
Hoffman takes the analogy of the graphical desktop interface. The things we experience, such as a car driving on a road towards us, are not objectively real but icons on our desktop of consciousness that represent something. The car may not be objectively real as we experience it but stepping out in front of it still has consequences in the same way as dragging a document to the trashcan on our computer desktop has consequences.
Hoffman became interested in the problem of how consciousness arises from the neural network in our brains. No current theory, Hoffman asserts, that starts with neurons and neural activity has so far been able to explain the origin and nature of conscious experience and its correlation with neural activity. This led him to wonder if neurons were real in the first place.
The first three chapters are introductory and set the scene.
In chapter 4, describes the Fitness-Beats-Truth (FBT) theorem. Consider two sensory strategies, each capable of distinguishing N distinct perceptions. One strategy attempts to render reality with the best fidelity. The other sees none of reality but provides perceptions tuned to fitness payoffs (ie indicates to the organism whether the thing perceived is helpful or harmful). The Fitness-Beats-Truth theorem states that (perceiving) fitness drives (perceiving) truth to extinction with probability at least (N-3)/(N-1). Hoffman claims to have conducted many simulations that verify this result.
Chapter 5 goes into more detail about the illusory nature of our perceptions and the desktop metaphor. This quote sums it up:
“…, there is no need to posit any physical object, or a spacetime, that exists when no one observes. Space and time themselves are simply the format of our interface, and physical objects are icons that we create on the fly as we attend to different options for collecting fitness payoffs. Objects are not pre-existing entities that force themselves on our senses. They are solutions to the problem of reaping more payoffs than the competition…”
Spacetime then is not real and the only reason we reach a consensus about the nature of the world is that we construct our subjective illusions in the same way as our fellow humans.
So far this is fascinating and seems plausible. The obvious question is that if we don’t perceive objective reality, what is objective reality like. What is out there? Sadly, Hoffman does not (yet) have answers to this.
However, where I have trouble is the assertion that since we don’t see reality then neurons are not real and so it does not even make sense to try and understand how neurons give rise to consciousness.
Assuming Hoffman is correct and our senses evolved to show us a view of the world that maximised fitness; then that view would be tailored for a hunter-gatherer existence in which there would be no concepts of neuroscience. It seems inconceivable then that our senses have a specialised presentation for neurons. This leads me to think that if we see neurons our representation for them must borrow from representations of other objects that would be common in a hunter-gatherer’s world. In other words, rather than constructing neurons out of nothing there is probably something there. To my mind, this means that while the Interface Theory of Perception (ITP) may be true, it does not necessarily follow that neurons don’t exist.
In summary, while I’m not convinced by all his conclusions this is a fascinating and thought-provoking book and worth reading by anyone with an interest in physics or neuroscience. I would love to hear how this work progresses and whether anyone more knowledgeable than me can spot flaws in it.