Review: Darwin Machines and the Nature of Knowledge

The central theme of this book is that biological adaptation and knowledge are the same thing. A bird migrating south for the winter might be doing so out of some genetic instinct, but Plotkin makes the case that instinct is just knowledge encoded into that bird's genetics. By framing knowledge in this way we can begin to study knowledge in a scientific way.

The first couple chapters are an overview of evolutionary thinking since Darwin. Darwin's theory of natural selection, first introduced in the late 19th-century, has been revised and updated with the advent of modern genetics, undergoing an evolution of its own. Nothing here should be too surprising for anyone with a basic knowledge of evolutionary theory.

Where the book gets interesting is in chapter 3 when Plotkin introduces what he calls Universal Darwinism. When Darwin introduced his theory of natural selection, he intended it to be an explanation of how species change and evolve over time. But on a closer inspection the processes of natural selection have more far-reaching implications about how individual organisms develop over time, and in fact can provide a framework for understanding every aspect of biological life, including the generation of knowledge. As a result, our thoughts themselves evolve in a manner analogous to how species evolve.

So what does it mean for thought a thought to evolve? Well Plotkin proposes that each thought that arises in our conscious experience is a product of the reproduction of previous thoughts. Whether this thought lives or dies depends on selective pressures in our mind such as whether the thought grabs our attention or whether that thought provides an accurate explanation of the environment in which we live.

Plotkin then describes the scientific enterprise--a systematic accumulation of human knowledge--as an evolutionary process of its own. Many scientific theories are put forth every day, but only the theories that can withstand scrutiny by peer-review and provide empirically testable explanations of physical reality are passed on the next generation of scientific theories. These barriers are in effect the selective forces which determine the fitness of our theories within the environment of scientific thought. It is because of the evolutionary nature of human thought that the scientific enterprise has had such success in explaining physical reality.

What the book doesn't delve into is how knowledge is physically embedded within our brain's neural structure. In my opinion, modern research in neural networks and natural language processing provide the clearest avenue to a scientific understanding of how knowledge is physically instantiated within the brain. It seems clear that we have yet to unravel the many mysteries of how the brain creates knowledge, but it also seems clear that the scientific process is powerful enough to unravel these mysteries in the coming decades.