“Learning is often thought of as the alternative to instinct, which is the information passed genetically from one generation to the next. Most of us think the ability to learn is the hallmark of intelligence. The difference between learning and instinct is said to distinguish human beings from “lower” animals such as insects. Introspection, that deceptively convincing authority, leads one to conclude that learning, unlike instinct, usually involves conscious decisions concerning when and what to learn.
“Work done in the past few decades has shown that such a sharp distinction between instinct and learning—and between the guiding forces underlying human and animal behavior—cannot be made. For example, it has been found that many insects are prodigious learners. Conversely, we now know that the process of learning in higher animals, as well as in insects, is often innately guided, that is, guided by information inherent in the genetic makeup of the animal. In other words, the process of learning itself is often controlled by instinct.”
Source: Gould, James L. and Peter Marler. Learning by instinct. Scientific American, January 1987. Reprinted in William S.-Y. Wang, ed.. 1991. The emergence of language. Development and evolution. New York: Freeman. 88–103.
What we have here is a learning/memory process that is shared by all biological systems. Nothing is really “hard wired” in biological systems. Even DNA is a pattern for growth and not a firm design. Even after DNA is expanded during cellular growth into a “final” adult design stage, cells grow and adapt to changing conditions. Hard wired is a term from computers that really doesn’t apply to biological systems, unless maybe that “learning” itself, is hard wired into all living systems.
Learning is a process that functions at all levels of biological systems. Looking at Life as an information processing system, reveals learning as a primary function of the system. DNA is the system wide knowledge base, stored in such a way to guarantee maximum stability and robustness. Every piece of code has a built in sundown clause. It is constantly being replaced with new code that has been thoroughly tested under real life conditions.
The life cycle of all living things is a two part process that simultaneously tests its DNA in the real world and learns about how the local environment is changing. Successful species continue to build on the knowledge base, adding and adapting to their experience.
The best designer in the world would not be able to design a mobile agent that could function successfully in a new environment without some knowledge of what that required. The system that designs agents must work from inside with intimate knowledge of the outside. The outside is constantly changing, so the inside must adapt.
Even though the design team doesn’t have to redesign the whole agent every time, it needs to be able to set the appropriate control switches, like environmental triggers for imprinting and such. Each species has a set of specific triggers it expects from the environment in order to initiate specific learning patterns. Even though many systems and subsystems arrive fully functional regardless of the species, like basic metabolism, control and sensory systems, they still must adapt to resource availability during their development. They also grow and reinforce what is successful in real life tests. The design team will need to know the available food sources, their nutritional values and such, before they can set the necessary triggers.
Of course, there is no design team as such. I use the term to personify an observed function, for ease of discussion. It is a very handy tool, if the imagination is flexible enough not to get caught up in the analogy.
Whether the process is driven by dumb luck or an intelligence of unknown origin, the results are some very fine subsystems, of which the human animal is one. We seem to be very smart, however, our very best design teams produce nothing, but crude examples showing only primitive engineering savvy, when compared to the simplest life forms. Flatworms are primitive according to evolutionary standards, yet their design is extremely efficient, even elegant. From a system viewpoint, the flatworm design, represented by its DNA, is extremely robust, stable and well dispersed in the environment. It has been a very successful design. It is part of our exploration to identify the design process, wherever we find it.
In building the LifeOS model, i was satisfied to have a system that would monitor reality and project solutions to problems, but idea that the system is constantly projecting an intended path is much more “intelligent”. Rather than have to identify “problems”, which could be a challenge, the system attempts to maintain a status quo. It treats any deviation as suspect. Seems to me that shows that the system is dealing pro-actively with the future.
In our culture, most people simply react to what comes their way. Our most intelligent citizens are people who don’t wait to simply react to events, but take control of their own future. These people set goals and work to reach them. Biological systems in general, are not just responding as would a person of average intelligence, but in the manner of one of superior intelligence; not by simply reacting, but by forming expectations(hypotheses) and testing them in the real world. This intelligent action is going on at the cellular level.
This is not in a long drawn out process, but a rapid firing cycle of cellular metabolism. With every cycle, the present input is compared to predictions. This process becomes a flowing wave of expectations, met or not, evaluated by waves of dopamine.
Back in Free Will i described the behavior of critters like cats and that bossy fly that quit playing by my rules. This is the same pattern of behavior that we find in dopamine neurons. These neurons are firing a steady stream of expectations that cause no reaction when they are met. Any deviation causes immediate adjustment to expectations. This stream by a single neuron becomes a wave when all the neurons are firing. This wave is constantly being compared to the waves of the past.
When an agent meets a new object in the environment, these neurons fire away, sending a flood of dopamine carrying info about the mystery object. It appears that the dopamine acts like a reward for the agent, making the exploration of new objects, “feel good”. It isn’t a simple on/off reward, but contains levels of description, evaluation and judgment. It is like emotion, with a wide range of feelings about the object. The simple act of observation begins an elaborate process of carving away the mystery to reveal the reality of the novel object or event.
When the new object is “figured out”, the neurons quit firing, the dopamine stops flowing, and interest in the object fades. Novelty and boredom controlled by the flow of dopamine. When a kitten is learning a new game, the dopamine is flowing. Once they learn the game, they lose interest. Back in Free Will, i attributed that behavior to the ego. The kitten and the fly just wanted to have their way. They just like to “boss us around”. From this new information, i would say that the reason for that behavior was because they learned the trick and their neurons quit firing. The dopamine rewards them for learning new tricks. Well, human beings are the bossiest of all, and our ego seems to be fueled by the same kind of neurons.
So, who gets bored? Who gets the reward that dopamine provides? Who is giving out this reward? When the fly or kitten turns its back on me and refuses to acknowledge my play cues, the animal is communicating to me that they are no longer getting the internal reward they need to stay interested. Who is initiating that communication? I still say, even the smallest creature has an ego, a sense of self that relates to its environment with intent, expectations and strategies.
From a system viewpoint, adaptation must be a sought after goal of some protocol or other. Adaptation is accomplished by learning at all levels of biological systems. Adaptation is a fundamental attribute of Life, part of the protocols of its operating system.