Intelligence as Prediction
Intelligence As Prediction:
Our Knowledge of Reality
How We Come to Know the Structure of Our World: When you are born your cortex essentially doesn’t know anything. It doesn’t know your language, your culture, your home, your town, songs, the people you will grow up with, nothing. All this information, the structure of the world, has to be learned.
We experience the world as a sequence of patterns, and we store them, and we recall them. And when we recall them, we match them up against reality. We’re making predictions all the time.
— Jeff Hawkins
The solution of great problems requires the giving up of great prejudices.
— Paul Dirac
Every great advance in science has issued from a new audacity of the imagination.
— John Dewey
Prediction Must Be Centralized—It Leads to Self
Given that prediction is the ultimate and most pervasive of all brain functions, one may ask how this function is grounded so that there evolved only one predictive organ. …. For optimum efficiency it would seem that prediction must function to provide an unwavering residency and functional connectedness: it must somehow be centralized to the myriad interplays of the brain’s strategies of interaction with the external world. We know this centralization of prediction as the abstraction we call the “self.”
— Rudolpho Llinás
We’re into some very important stuff here. We’re into how the brain works and what the self is. We can also add the idea of multiple selves or sub-selves. First, I’d like to comment on Hawkins’ ideas. Hawkins is both an engineer and the developer of a new theory of brain function. In his book, Hawkins has explained in great detail how the neocortex works. Basically, He explains, “Discovering causes is the pinnacle of what brains do.” Those causes constitute not just our ideas about reality, they are the stuff with which we create and experience our reality.
Getting good at this “fancy pattern recognition” is something developing humans seem to do effortlessly, and computers only with immense labor. The neocortex is layered with nodes that make inferences about outside sensory data and it passes these hunches up a hierarchy of nodes until a consensus—a belief—evolves about the source of the data. The use of belief propagation techniques, says Hawkins, enables an entire system to reach the best overall consensus swiftly.
One central concept is the memory-prediction framework. It states that the neocortex matches sensory inputs to stored memory patterns and this process leads to predictions of what will happen in the immediate future when a pattern occurs. For example, if I start saying “a, b, c, d”, you’ll automatically predict that “e” will be the next thing I’ll say. You predict that because you know that pattern. It will be so obvious to you, you will not even notice that you’re predicting that. But, you would notice, if I said “l”. When events do not fit our unconscious predictions, we normally notice. That happens consciously. We events conform to our unconscious predictions, we do not notice. Hawkins has gives a good example. Suppose that, while you were out of your house for a few hours, someone came and somehow moved your front doorknob a few inches from its usual place. (No need to worry here about how he did it.) You come home. You’re talking on your cell phone say. You reach for the doorknob without thinking and, wham, you’re jolted out of your phone conversation with the failed prediction about where the doorknob “ought” to be.
The memory-prediction framework provides a useful way to think about the adaptive unconscious and its control of behavior. Everything you’ve ever experienced is stored in your memory. It is these memories that create what you perceive. But, because the predictions that they make happen so quickly, so continuously and are based on what has happened so many times, you only sense the results and not the process. As Llinás has said, “The brain is a virtual reality machine.” We live in that virtual reality. Luckily, for most of us, the machine remembers.
I had a friend who used to finish people’s sentences. He did it all the time and seemed to be functioning at something approaching light speed in predicting how someone going to complete every phrase or sentence. His ability to do this is a perfect example of Hawkins’ memory-prediction framework. Another example given by Hawkins is this: Once you know a song or a melody, you can sing along when you hear it sung or played by someone else. When you do this, you are continually predicting what will follow moment by moment. But of course, it’s almost impossible to sing a song you know backwards. Of course, you have to have heard the song, probably more than once. You have to practice to confirm and test your predictions. The sequence has to be in memory in order to predict, even in order to be “real”. And it’s like that with everything we experience!
Memory and prediction are how the brain works to create the realities we live in and the behaviors we use to meet those realities. That makes this of paramount importance to our work with people.
Before we talk about selves, let’s look at a diagrams I adapted from Hawkins talk.
Intellegence as prediction model
The neocortex is memorizing everything that comes in. If something is coming in which has been sensed before, the neocortex recognizes it and remembers what followed the previous times this sensory pattern appeared. This then allows for prediction. The prediction, based on recognition and memories, modifies the reflexive input-to-output patterns of the non-mammalian brain. The result is complex behavior based on experience. It is this ability that defines intelligence. The complex behavior demonstrates intelligence, but it does not define it.
Now, about the selves.
Here’s Hippolyte Adolph Taine’s description of the multiplicity of selves within a single individual. Writing in 1871, Taine says:
One can … compare the mind of a man to a theater of indefinite depth whose apron is very narrow but whose stage becomes larger away from the apron. On this lighted apron there is room for one actor only. He enters, gestures for a moment, and leaves; another arrives, then another, and so on…. Among the scenery and on the far-off backstage there are multitudes of obscure forms whom a summons can bring onto the stage… and unknown evolutions take place incessantly among this crowd of actors. [Taine, 1817, cited in Whyte, 1960, pp. 166-167]
So, it would seem these “actors” are embodiments of the “complex behavior patterns” that Hawkins is taking about. They are very complex, learned as adaptations to sets of experiences that originally created the memories, predictions and beliefs about what causes what. In other words, what is real. These memories are part of our astounding ability to recognize patterns. In the experiment, participants were shown 10,000 pictures for five seconds each. Two days later, they correctly identified 8300 of them.
Two more things. As Llinás tells us, (1) we need a self as “centralization of prediction” and (2) “prediction is the ultimate and most pervasive of all brain functions”. Llinás also tells us this self is an abstraction, in the same way “Uncle Sam” is. It is not an isolatable entity. Still, it has a definite function: prediction. In Buddhism it is said, “all is without a self”, meaning, a separate self, a non-isolatable self.
Within this functional self, there are many sub-selves. Recall that Taine’s “actors” come on stage one at a time. The appearance of each separate self is no doubt the result of the centralizing coordination of the use of them. Each can come on stage when a situation is recognized as a summarization of a set of memories. On the basis of that summary, the brain predicts what will happen next and organizes the appropriate complex behavior to deal with it. All of this happens quickly and automatically. William Calvin writes:
“Just as intelligence has been described as ‘What you use when you don’t know what to do,’ when no standard response will suffice, so too consciousness is prominently involved when the situation contains ambiguity or demands creative responses, ones that cannot be handled by a decision tree. Many mental activities can be handled by subroutines; consciousness helps to deal with the leftovers (and create new subroutines for the next time).”
— William Calvin
It is a tenet of this method — the Refined Hakomi Method — that mental, emotional and relational suffering happens due to automatic behaviors operating out of the “memory-prediction framework”. Often these automatic behaviors that are protective, over-generalized and outmoded. To change them, they must first be brought into consciousness. The method is designed to, first, recognize indications of such behaviors, and then to use mindfulness and experiments to bring them into consciousness. It coming into consciousness initiates a healing process, as it often does, the last phase is designed to support that process.
If we keep in mind that unconscious memories and predictions are responsible for the behaviors that result, we can design our experiments and follow up questions to evoke awareness of the predictive process itself. Techniques like having the client, while in a mindful state, notice what happens automatically when you offer a statement like, “If you (do such and such)….”, the result is likely to evoke an automatic reaction. Once that reaction takes place, the memory, the prediction or both will come into consciousness. These associated memories and beliefs start the healing process and can then be attended to in a way that creates new memories and changes the predictions, moving the healing process towards resolution and healing.