Part one is here.
Consciousness as controlled hallucination
The word “hallucination” highlights that all perception is based on guesses about the sources of sensory information, and the word “controlled” means the perceptions are good guesses that drive adaptive action.
Seth’s notion of controlled hallucination is based in the predictive processing framework for perception. If you're not already familiar with this idea, I wrote a post about it here. Being You has an awesome review of all the technical details of this model, written in clear language, but if I summarized it all here, this review would be too long. But here’s a very brief summary of some basics:
Predictive processing inverts the common sense idea that our perceptions and conscious experiences are mirrors of the outside world, happening through a passive process whereby sensory information arrives at the brain from outside-in and bottom-up.
Instead, conscious perception is far more active, top-down, and inside-out than we might imagine, involving predictions, guesswork, probabilistic thinking, and models of the world based on past experience.
And here’s Seth’s description of three basic ingredients in the controlled hallucination view:
First, the brain is constantly making predictions about the causes of its sensory signals, predictions which cascade in a top-down direction through the brain’s perceptual hierarchies … If you happen to be looking at a coffee cup, your visual cortex will be formulating predictions about the causes of the sensory signals that originate from this coffee cup.
Second, sensory signals—which stream into the brain from the bottom up, or outside in—keep these perceptual predictions tied in useful ways to their causes: in this case, a coffee cup. These signals serve as prediction errors registering the difference between what the brain expects and what it gets at every level of processing. By adjusting top-down predictions so as to suppress bottom-up prediction errors, the brain’s perceptual best guesses maintain their grip on their causes in the world. In this view, perception happens through a continual process of prediction error minimization.
The third and most important ingredient in the controlled hallucination view is the claim that perceptual experience—in this case the subjective experience of “seeing a coffee cup”—is determined by the content of the (top-down) predictions, and not by the (bottom-up) sensory signals.
Seth provides an interesting intellectual history of the idea of perception as inference:
The first glimmers of a top-down theory of perception emerge in ancient Greece, with Plato’s Allegory of the Cave ...
[A] thousand years ago, the Arab scholar Ibn al Haytham wrote that perception, in the here and now, depends on processes of “judgment and inference” rather than providing direct access to an objective reality. Hundreds of years later again, Immanuel Kant realized that the chaos of unrestricted sensory data would always remain meaningless without being given structure by preexisting conceptions.
In the late nineteenth century, [the German physicist and physiologist Hermann von Helmholtz] proposed the idea of perception as a process of “unconscious inference.” The contents of perception, he argued, are not given by sensory signals themselves but have to be inferred by combining these signals with the brain’s expectations or beliefs about their causes.
The brain’s expectations and beliefs about the causes of sensory experience are called a generative model. Here's an example of how models and expectations affect perception. Take a look at this picture and note what you perceive. For most people it will look like a bunch of random blobs.
Now look at the picture below, and then go back to the first to see if your perception changes.
Seeing the second picture transforms the way you perceive the first picture, because you updated your model of the world to include a high likelihood that you would see a women kissing a horse, and therefore that's what you perceive. Put differently, you were unable to perceive the woman kissing the horse because you did not predict it.
Here’s another interesting thing about your ability to see the woman and the horse in the first picture – they aren't really there, and your brain fills in all sorts of missing details in order to see them. Your perception of the woman and horse is a controlled hallucination.
Our experience of color provides another example. There are no colors “out there” in the world, but we perceive them to keep track of objects as the light changes.
Color is not a definite property of things-in-themselves. Rather, color is a useful device that evolution has hit upon so that the brain can recognize and keep track of objects in changing lighting conditions. When I have the subjective experience of seeing the red chair in the corner of the room, this doesn’t mean that the chair actually is red—because what could it even mean for a chair to possess a phenomenological property like redness? Chairs aren’t red just as they aren’t ugly or old-fashioned or avant-garde. Instead, the surface of the chair has a particular property, the way-in-which-it-reflects-light, that my brain keeps track of through its mechanisms of perception. Redness is the subjective, phenomenological aspect of this process.
So we hallucinate color because it helps us take appropriate action. It’s like color-coding folders of documents, so you can tell them apart easier.
We perceive the world around us in order to act effectively within it, to achieve our goals, and—in the long run—to promote our prospects of survival. We don’t perceive the world as it is, we perceive it as it is useful for us to do so.
Thinking about our perceptions this way can get sort of mind boggling when applied to other aspects our consciousness, such as the three-dimensionality of objects and our sense of self.
Deconstructing Objects and Selves
The visual data you get from a coffee cup is almost identical to that from a two-dimensional picture of the same coffee cup, but our perceptions about them are different in an important respect:
When I look at the coffee cup on my desk, in some sense I perceive its back, even though I cannot directly see this part of it. The cup appears to me to occupy a definite volume, whereas a coffee cup in a photograph or drawing does not. This is the phenomenology of “objecthood.”
Perception of volume is there because it affords action. It allows you to predict what you would see if you rotated the coffee cup to get a look at the other side of it. Like color, the perception of three-dimensionality is constructed as a way to facilitate potentially useful action.
The sense of self is also constructed:
It may seem as though the self—your self—is the “thing” that does the perceiving. But this is not how things are. The self is another perception, another controlled hallucination.
This idea has been around for a while:
Kant, in his Critique of Pure Reason, argued that the concept of the self as a “simple substance” is wrong, and Hume talked about the self as a “bundle” of perceptions. … Buddhists have long argued that there is no such thing as a permanent self.
We know that the self is a construction because it can be deconstructed. This might happen after meditation, drug experiences, mental illness, or brain damage. Neurologists have documented many ways that the self can disintegrate after brain injury. And split brain patients raise the possibility that one self could become two.
So if the self is a construction, what functional purpose does it serve? There are many, and each gives rise to a slightly different perception. These tend to get all bundled together, but we know that sometimes one will disappear while others remain. Here is a list of separate perceptions about selfhood, along with some examples of what happens when they go away for some reason.
Bodily identification. Perception of self is partly about determining which objects in the world are part of us, and which are not. In rare cases this perception can be highly irregular, such as phantom limb syndrome or alien limb syndrome. It can be disturbed by the rubber hand illusion. And the sense of bodily identification can be additive as well - consider the feeling of using a tool that you can operate with a high level of skill.
First person perspective. We experience the self as having a location behind our eyes. This perception is disrupted when people have “out of body experiences”, which often occur in near death situations or with epileptic seizures. These experiences can also be created with virtual reality experiments.
Agency: We experience our selves as agents who can make choices and effect changes in the world. Thus, self-concept is a way to organize beliefs about volition, intention, responsibility and basic cause and effect. “We perceive ourselves to control ourselves.” This can be disturbed in schizophrenia, where people misread internally generated neural signals as having an external source, so they feel controlled by outside forces, or they hear “voices” in their head.
Free will. We all perceive that we have free will, even if we have a deterministic scientific perspective. This is because free will is a useful perception: “our sense of free will is very much about feeling we ‘could have done differently.’ … [This] does not mean that I actually could have done differently. Rather, the phenomenology of alternative possibilities is useful because in a future similar, but not identical, situation I might indeed do differently.”
Personal identity: Selfhood involves a narrative about life history that explains who we are and where we come from. These can be altered by pychotherapy or disappear in cases of dementia or amnesia. Seth tells the story of a man who woke from a coma with his ability to form new memories destroyed. So every twenty seconds he re-experienced the feeling of having just woken up, because that’s where his life narrative ended and never continued.
Social self: This forms a model of how others perceive us, and is related to emotions like shame, guilt and pride.
All of the above aspects of self form powerful models that generate predictions, which affect perception and action. These models tend to be stable over time, but they are modifiable:
Compared to perceptual experiences of the outside world, self-related experiences are remarkably stable. Our perceptions of the world are always changing, objects and scenes coming and going in a continual flux of events. Self-related experiences seem to change much less.
However:
Our perceptions of self are continually changing—you are a slightly different person now than when you started reading this chapter—but this does not mean that we perceive these changes. This subjective blindness to the changing self has consequences. For one thing, it fosters the false intuition that the self is an immutable entity, rather than a bundle of perceptions.
Part of the purpose of mediation and mindfulness practices is to notice that change is always happening in small amounts all the time. Noticing this can perhaps change your self-image and then maybe your self.
Interoception, mood and emotion
We have so far been talking mostly about perception about the outside world, but we also perceive the inside world of the body. This is called interoception:
Interoceptive signals report things like heartbeats, blood pressure levels, various low-level aspects of blood chemistry, degrees of gastric tension, how breathing is going, and so on.
The key property of interoceptive signals is that they reflect, in one way or another, how well physiological regulation of the body is going. In other words, how good a job the brain is doing of keeping its body alive.
The brain can only guess at how well the body is doing:
Just as the brain has no direct access to the causes of exteroceptive sensory signals like vision, which are out there in the world, it also lacks direct access to the causes of interoceptive sensory signals, which lie inside the body. … Interoception is therefore also best understood as a process of Bayesian best guessing, just like exteroceptive perception.
Our best guesses about how well the body is doing form the basis for mood and emotion. In fact, some would say those are one and the same thing. This idea has been around for a while:
Back in 1884, William James and—independently—Carl Lange argued that emotions … are perceptions of changes in bodily state. … For James, the perception of bodily changes as they occur is the emotion: “We feel sorry because we cry, angry because we strike, afraid because we tremble, and not that we cry, strike, or tremble, because we are sorry, angry, or fearful.”
Seth wants to update this idea by looking at emotions as controlled hallucinations about the body’s physiology:
In the same way that “redness” is the subjective aspect of brain-based predictions about how some surfaces reflect light, emotions and moods are the subjective aspects of predictions about the causes of interoceptive signals. They are internally driven forms of controlled hallucination.
Here's an interesting observation about the nature of an emotional experience: Unlike perceiving objects in the world, emotions don’t have shapes, colors, locations, or dimensions. For example, fear may be a perception of the heart beating fast, but it not a subjective experience about the location, color or shape of the heart or any “thing” at all. Why not?
The answer is that interoception serves a different purpose than exteroception. The reason we perceive our physiology is to control it. A thermostat cannot control the temperature in a room without knowing the current temperature, and our brains cannot control all the internal conditions required for life and health without perceiving them. So interoception is all about control, while exteroception may be more about just finding things out. This helps explains the difference in their phenomenology:
Experiences of fear, jealousy, joy, and pride are very different, but they are more similar to each other than any one of them is to a visual experience, or to an auditory experience. Why is this? The nature of a perceptual experience depends not only on the target of the corresponding prediction—perhaps a coffee cup on the table, or a racing heart—but also on the type of prediction being made. Predictions geared toward finding out things will have a very different phenomenology from those geared toward controlling things. When I look at a coffee cup on my desk, there is the strong perceptual impression of a three-dimensional object that exists independently of me.
…
Anxiety doesn’t have a back, sadness doesn’t have sides, and happiness is not rectangular. The perceptions of the body “from within” on which affective experiences are built do not deliver experiences of the shape and location of my various internal organs—my spleen here, my kidneys over there. There is no phenomenology of objecthood … instead of experiencing emotions as objects, we experience how well or badly our overall situation is going, and is likely to go.
This perspective on emotions and moods ties them even more closely to our flesh-and-blood nature. These forms of self-perception are not merely about registering the state of the body, whether from the outside or from the inside. They are intimately and causally bound up with how well we are doing, and how well we are likely to do in the future, at the business of staying alive.
Well I am out of time and space again. I will need to write a third part, and this will be mercifully shorter I think.
I love this book but haven’t found much traction in translating these ideas to patient care (besides for my own understanding of the patient’s experience). Have you been able to get patients on board with predictive processing?