Review of "Being You" by Anil Seth (Part One)
An interesting new book on the science of consciousness
This post is part one of a summary of my notes on Being You, a new book about the scientific study of consciousness by the philosopher Anil Seth. Seth is a world-class thinker on this subject and a major proponent of the predictive processing framework.
His book is surprisingly accessible given the confusing nature of the subject matter. It deals with some very puzzling questions such as:
Why does consciousness even exist? In other words, why is it "like something" to be a human?
If consciousness evolved, when exactly did the “lights come on” and for what purpose?
Does consciousness have something to do with information processing or complexity?
If so, would a sufficiently sophisticated artificial intelligence be conscious?
Can these questions be answered scientifically?
Or is there something about consciousness that makes it an eternal mystery, or a “hard problem” that can’t be solved?
Seth thinks these questions are best addressed by looking at consciousness as a “controlled hallucination” that helps us navigate the world, find food, and evade predators. These hallucinations are created through a constant process of modeling the world, making predictions, evaluating sensory information, and correcting errors. If you are wondering: “why in the world should making predictions cause subjective experience”, Seth recommends holding your questions until you get a clear idea of how predictive processing and controlled hallucinations work. That's what most of the book is about.
Consciousness defined
People use the word consciousness to refer to different things, even when those people are sophisticated philosophers and scientists trying to have clear discussions. The word is used alternately to mean being awake, having free will, being self-aware, or behaving with intelligence.
Seth’s definition is simple and broad: consciousness means “any kind of subjective experience whatsoever.” Under this definition, you can be conscious while dreaming, when you have no control over your actions, or when you have no concept of being a self. All that is required is that you have phenomenal experience, a first-person perspective on the world, and that it is “like something” to be you.
Philosophical ideas about mind and matter
The relationship between mind and matter has been a philosophical puzzle for thousands of years. For most of history, the proposed answer was that physical objects owe their existence to an immaterial spirit or God. This is one version of the doctrine of idealism, which holds that some form of intelligence or consciousness is the ultimate ground of reality, and that matter emerges from mind.
This order was reversed by advances in science, especially evolution. It is now taken for granted by most scientists that mind emerges from matter, and that consciousness depends on physical events in the brain. This philosophical position is called physicalism, explained by Seth as follows:
[Physicalism] is the idea that the universe is made of physical stuff, and that conscious states are either identical to, or somehow emerge from, particular arrangements of this physical stuff. Some philosophers use the term materialism instead of physicalism, but for our purposes they can be treated synonymously.
Physicalism can be distinguished from dualism, which is a kind of middle ground between idealism and physicalism, positing that mind and matter are fundamentally separate forms of reality. This raises the difficult question of how these separate realms can interact, and therefore few philosophers or scientists would admit to being dualists.
Physicalism is therefore a near universal assumption in any credible scientific study of the brain. A related doctrine called functionalism is also very popular, but more controversial:
Functionalism is the idea that consciousness does not depend on what a system is made of (its physical constitution), but only on what the system does, on the functions it performs, on how it transforms inputs into outputs. The intuition driving functionalism is that mind and consciousness are forms of information processing which can be implemented by brains, but for which biological brains are not strictly necessary. Taking functionalism at face value, as many do, carries the striking implication that consciousness is something that can be simulated on a computer.
Seth explains that he is a committed physicalist, but not sure about functionalism. In other words, he is sure that human consciousness depends on the physical stuff in our brains, but he's not sure that computers would be conscious if they started acting like humans.
Another minority viewpoint worth mentioning: Panpsychism is the idea that everything is at least a little bit conscious (even rocks!), but in different amounts depending on their complexity.
The hard problem, the easy problem, and the real problem
The philosopher David Chalmers is famous for distinguishing the “hard problem” of consciousness from the “easy problem.”
The easy problem is explaining how physical systems like brains can give rise to intelligent behaviors associated with consciousness, like reacting to stimuli or figuring out math problems. The hard problem is explaining why the brain activity associated with these behaviors produces conscious experience.
Chalmers argues that even if we solve all the easy problems, we will never have an intuitively satisfying answer to the hard problem, because this would simply be inconceivable. No matter how much we learned about the brain, consciousness will always remain a mystery.
Seth disagrees. He argues that it is possible that when we learn more about the brain, we will think it’s not weird that the it creates consciousness. Consider the progress we have made in biology:
Not so long ago, life seemed as mysterious as consciousness does today. Scientists and philosophers of the day doubted that physical or chemical mechanisms could ever explain the property of being alive. The difference between the living and the nonliving, between the animate and the inanimate, appeared so fundamental that it was considered implausible that it could ever be bridged by mechanistic explanations of any sort. …
The science of life was able to move beyond the myopia of vitalism, thanks to a focus on practical progress—to an emphasis on the “real problems” of what being alive means … biologists got on with the job of describing the properties of living systems, and then explaining (also predicting and controlling) each of these properties in terms of physical and chemical mechanisms.
Seth also points to progress in physics, which involves stuff that is pretty weird and mysterious.
Physicists have made enormous strides in unraveling the secrets of the universe—in explaining, predicting, and controlling its properties—but are still flummoxed when it comes to figuring out what the universe is made of or why it exists. In just the same way, consciousness science can make great progress in shedding light on the properties and nature of conscious experiences without it being necessary to explain how or why they happen to be part of the universe in which we live.
In both physics and biology, scientists made progress by trying to explain, predict and control the properties in question. This is exactly what should be done in a science of consciousness, which Seth calls solving real problems:
According to the real problem, the primary goals of consciousness science are to explain, predict, and control the phenomenological properties of conscious experience. … This means explaining why a particular conscious experience is the way it is—why it has the phenomenological properties that it has—in terms of physical mechanisms and processes in the brain and body. These explanations should enable us to predict when specific subjective experiences will occur, and enable their control through intervening in the underlying mechanisms.
The rest of the book sets forth our current progress in solving real problems related to consciousness. One key strategy is recognizing that consciousness, like life, is not one irreducible phenomenon or essence, but many related events. Thus, it can deconstructed into separate phenomena which can be studied separately. For example, there is a different between the contents of consciousness, and the level of consciousness.
Levels of consciousness
The content of your consciousness is made up of sights, sounds, thoughts and emotions. The level of consciousness is about the degree of detail in the contents.
Consciousness level can be conceived as existing on a spectrum from zero (e.g. brain death or coma) to very little (dreamless sleep) to very high (e.g. an unforgettable experience).
Here’s an example from my own experience where I reached an abnormally high level of consciousness. I was in Costa Rica and getting ready to lay my cheek down on a pillow when I noticed a scorpion just a few inches from my cheek. I backed off and looked at the scorpion thinking: “this is real. That is a real scorpion.” Everything that was happening felt very vivid and almost hyperreal.
Consciousness level is not the same thing as wakefulness. They are often correlated, but you can be asleep and have highly detailed dreams filled with interesting information. Or you could be wide awake and zoned out with your mind almost blank.
Seth identifies two basic factors that make consciousness more “high-level”: information and integration. Information describes how much detail is available. If your conscious experience was a computer screen, information would measure the number of pixels. Integration measures the extent to which the bits of information relate to each other in meaningful patterns. Using the computer screen analogy again, integration means the pixels form identifiable pictures or symbols, as opposed to random noise.
This balance between information and integration is a sweet spot between order and chaos that creates creates the maximum degree of complexity. Here’s how this balance plays out in terms of neural activity:
In a maximally information-rich brain, all neurons would behave independently, firing randomly as if they were completely disconnected … But this brain—with lots of information but no integration—would not support any conscious states. At the other extreme, a maximally ordered brain would have all neurons doing exactly the same thing, perhaps firing in lockstep together, somewhat like what happens during global epileptic seizures. Algorithmic complexity here would be very low. This brain would also lack consciousness, but for a different reason: lots of integration, no information.
With these ideas in mind, scientists are looking to record and quantify the kinds of brain activity and neural complexity that might be associated with consciousness level.
One popular measure involves calculation of “integrated information” and another is “algorithmic complexity.” So far, scientist have been able to use these and other measurements to accurately “predict” the level of consciousness in the following cases:
Psychedelic use, where consciousness level is high.
Dreamless sleep and general anesthesia, where the values are low.
Neurological damage, where consciousness level would be expected to vary with the severity of the damage.
One of the most interesting results: Italian researchers were able to show in 2013 that measures of brain activity could be used to prove the existence of conscious experience in people with locked in syndrome, which is a condition where subjects are unable to move their bodies in any way, but have normal conscious experience.
Seth points out that all of this work is very preliminary and speculative, but indicates forward progress.
Part two coming later this week
This review ended up longer than expected, so I'm breaking it into two parts. I hope to write the next part later this week. It will cover controlled hallucinations, perception of the self, interoception, and the basics of predictive processing.
I think this book is a very good read, very well written, and in most part very helpful, but I think the phrase "controlled hallucination" is unfortunate. I fully understand and agree that all perceptions are indirect, via a model, but calling this a hallucination gives the wrong impression. We know that our brains build a model, and that our perception works by prediction, but the model is all we can be aware of, so it is not a hallucination at all, it is reality to us. Seth cites evidence supporting his claim with examples of how the model can be fooled by well-known illusions, but to me this is simply evidence of the nature of the model, and helps us understand how it is built and updated.
He then goes a lot further and says that the model the brain builds of the self is also a hallucination, and gives similar reasoning about how the model of the self can be fooled or disassembled. Again, I take this as good evidence that we perceive ourselves via a model that the brain builds of our self, and also gives some clues about how it is built and updated.
Seth refers to Thomas Metzinger’s book "Being No One", calling it a brilliant book and saying that it is "a powerful deconstruction of the singular self"”. He goes on to point out, as others have done, that Buddhists philosophy argues "that there is no such thing as a permanent self", and almost seems to be using this fact to support his conclusion. I find Metzinger’s arguments totally unconvincing, and therefore I also disagree with Seth on these points. I am my model of my self, and I am most definitely not a hallucination.
I have been developing a set of proposals over the last 8 years that I have documented on a new website hierarchicalbrain.com
Yes, and that hyperarousal might prompt some extreme mobility away from that scorpion!