On the Evolution of Mental Models

Our mind is the primary way in which we understand the world, but it has many limitations. The world is a big place and our minds are small, so generally the best we can do is create simplified theories that roughly match up with one of more aspects of reality. But any theory or model you create will fall short of reality given enough scrutiny. Rather than reach some absolute truth with our models, the best we can hope for is some sufficiently useful approximation to help us form a clearer picture of the world.

For example, throughout history people have continually been updating their mental models of the their position in the universe. To an illiterate pre-scientific hunter-gather, a reasonable model might have been that they live on a flat surface that extends outward infinitely in every direction. The Earth was the biggest object familiar to these early humans, therefore it must have been the biggest object in the universe. This might have been the dominant mental model until some clever person noticed that the sun as well as the stars in the night sky appear to rotate around the Earth. This might have led them to conclude that our planet is actually a sphere in the center of the universe. Then someone noticed that certain stars in the night sky, the planets, followed irregular orbits that were out of sync with the other stars. An initial mental model might have invoked some god or mysterious force to explain the strange movement, but eventually people created elaborate theories that described the paths of these irregular "stars" with respect to Earth with pinpoint accuracy. These complex theories, however, eventually fell apart, because someone realized that they would be much simpler if we just assumed that the sun was the center of the universe. This created a new framework within which the older theories could be subsumed and simplified. This paradigm shift was the basis of modern cosmology, and it allowed humans to create a more general discription of the universe at large. I could continue this story, but the point is that humans have always strived to create better models of our world. This is an on-going process, and even though it might seem like we live in a world where pretty much everything is understood, we still have a long way to go.

Knowing all this, the central question in science becomes: what makes a good theory? Generally we can judge a theory by how useful it is, whether that be in terms of how easy it is to understand, or how it accurately explains the complex in terms of the simple, or how well it matches up with experimental data. The most powerful models are the ones that convey the largest amount of information in the fewest words possible. Our theories might not always be right, but they can still be useful if they provide sufficient explanatory power.

When Darwin proposed his famous theory of natural selection, he was presenting a model of the world that compressed his twenty years of experience as a naturalist into a few simple arguments. His model was specifically about how species change over time due to competition for resources with other species. Animals deemed "fitter" were able to pass on more copies of their DNA to the next generation, hence the phrase "survival of the fittest". He called this process natural selection, and its tenets were brought under the umbrella of what today is called the theory of evolution.

Though Darwin's theory was simple, it provided a powerful explanation for the endlessly complex patterns we see in the natural world. It rediscovered a world which was previously understood mostly through folk wisdom and superstition, and allowed biologists to see that world through a clearer and more objective lens. The beauty of the theory was in its elegance--its ability to explain a huge and complex part of our world in a few short sentences.

Today evolution has been extensively studied under the light of genetic research, but its principles have remained unchanged. It seems to me, however, that there is growing evidence that the biological evolution is not merely a description of what happens in biology, but a subset of a larger universal pattern. Specifically, I think it describes a general principle of how the physical universe changes from one moment to the next.

For example, every gene in an organism will either help or harm that organism in terms of its ability pass on its genes to the next generation. A gene that causes a plant's leaves to turn toward the sun will probably survive longer than a gene that caused the leaves to turn away. We can describe the success of these genes in terms of fitness functions, which simply measure how well or poorly a gene will affect an organism's reproductive success. Because the success or failure of an organism also depends on its environment, the environment provides a fitness landscape within which that gene will either succeed or fail.

Lets says there are two variants of a gene, Gene_1 and Gene_2, and F(Gene_1) > F(Gene_2) for a given environment. This indicates that P(Gene1Survives) > P(Gene2Survives). Given enough time, assumed a fixed environment, Gene_2 would eventually disappear from the gene pool due to competition from Gene_1.

For the sake of illustration, we can formulate the probability that a gene survives to a given time in terms of a basic law of probability:

P(GeneSurvives_n) = \prod\limits_{i=1}^n P(GeneSurvives_i)

If we assume that in any given time, an organism with the first gene has a 90% chance of survival, and and organism with the second gene has a 80% chance of survival, we can see how quickly these numbers compound on themselves:

  Time Step

  Gene 1

  Gene 2



















By the sixth time step the P(Gene1Surives) is almost twice as much as the P(Gene2Surives), even though there is a relatively small difference in their respective fitnesses for that environment. When you're playing the long game of evolution, small advantages eventually win out.

But, you might say, genes are not independent of one another, they usually have interactives with groups with other genes. That's true, and you can resolve this problem by considering the gene assembly as an object with its own fitness function. In this light, the fitness function of an organism could simply be defined in terms of F(O) = \sum\limits_{i=1}^n{F(g_i)} where g_1, g_2, ..., g_n are each of the genes in that organism. Taking this model one step further, a species could have a fitness function in terms of F(S) = \sum\limits_{i=1}^n{F(o_i)}, where o_1, o_2, ..., o_n are all the individual organisms that make up that species. Again, all else being equal, one species will eventually out-compete another species if its collective fitness is greater than that other species. In ecological models, it is known that any given ecological niche can only be exploited by one species as a direct result of the way the environment evolves. Selection pressues will eventual force out competitor species, leaving a single dominant species. A lot of complexities of biodiversity are a result of environments paritioning themselves into different ecological niches, allowing different species to coexist side-by-side. These selection pressures also seem to be scale-invariant, meaning that they play themselves out at all organizational scales, whether that scale be in terms of individual genes, or individual animals, or individual species.

The crucial thing to realize here is that this principle of fitness landspaces can be applied arbitrarily to any level of organization in the physical universe, not just within natural selection. We don't need stay confined within the realm of observational biology to see the principles of ecology at work, because it turns out that ecology plays out within all aspects of human society.

The easiest way to see this might be to look at the world of economics. It's not a stretch to call the competition between different businesses for customers and money a survival of the fittest. We just have to redefine fitness as a general ability for an organization to stay organized. The fittest companies are those which can extract as much money as possible from the market, and they do that by exploting specific segments, or niches, of the market.

For example, consider a small rural village with no restaurants. One day a popular fast-food restaurant opens up along the side of the road. Because of the novelty of the new building and the people's appetite for burgers, that restaurant is able to tap into the market niche of people who like burgers. If a month later a rival fast-food burger chain opened up across the street, that first business has a problem. When once it could siphon that money freely from the burger-loving public, now it suddenly has a competitor trying to tap into the same market. If nothing else changes in the town, and that market was only big enough to support one restaurant, one of the two businesses will eventually close. A burger-chain executive might chalk this up to his restaurant not having "product market fit", which is essentially the same as saying the restaurant is not able to compete in that economic niche.

The reason we see so many different resaurant chains in big cities is because big cities create an endless variety of economic niches for businesses to exploit. A Vietnamese restaurant in downtown New York that opens up accross the street from McDonalds doesn't have to worry about tapping into the market of people-who-love-burgers, because there are plenty of other economic niches for that restaurant to exploit.

Just because we live in a society with a rich history and sophisticated technology does not mean that we can escape ecology. Take a look at one of the newest jobs many people are working today: running a YouTube channel. YouTube creators compete for viewers, subscribers, and ad revenue among the constant stream of other YouTube creators. The ones who create better content--at least in terms of the meme-laden fitness landscape of YouTube videos--are the ones who get the majority of attention. But because the internet is so vast, there is a huge variety of niches for content creators to exploit. You can become an overnight millionaire if you strike into some weird untapped Youtube video niche, and at the same time you will face a Sisyphean struggle if you try to break into a well-established niche.

On a broader scale you can also see evolution at work in language. Two thousand years ago the Roman empire spread across half the known world, and as a result of its might and influence, the Latin language was spoken far and wide. The need to communicate effectively created a fitness competition between different languages, and many local dialects were lost after the locals switched to speaking Latin. The authorities spoke Latin and the merchants spoke Latin so Latin simply had a higher fitness than the local languages. Eventually the empire fell and local variants developed. The language no longer the authority it once had and so it became less fit compared to local variants, like French, or Spanish, or Italian.

Crucially, this principle reaches beyond the biological world. Ever since matter in the universe started to differentiate, fitness functions have been present in some form or another. When you look at the universe at a marco-scale, you can see that it is largely homogeneous. The matter in the physical universe is composed of about 73% helium and 25% hydrogen. The remainder of elements, like Oxygen and Carbon, are formed inside the stars as a result of a process known as nuclear fission. Hydrogen, because of its simple and stable molecular structure, outcompetes Helium and Oxygen and other elements that are relatively more complex. This too, is a result of how fitness functions dictate the shape and composition of our universe. The structures in the universe that have a higher fitness are the ones we tend to see more often.

So if certain elements are fitter, why isn't the universe 100% hydrogen? Well because the universe is in a constant state of expansion, and because the distance between regions of the universe is so large, differences between regions tend to form. Niches for particles or planets or galaxies are formed and they in turn lead to even greater complexity. The reason a planet like Earth can exist is because the vast size of the universe has created different niches in which the earth could survive. If the sun was in a stellar region with many super-novas, the planet we live on probably wouldn't be able to exist as we know it. If the sun was bigger and hotter the earth might not have been able to survive. Therefore we can view the geological-biological collection of matter that is our planet as an object with its own unique fitness function.

The fitness of any object in the universe is a function of the fitness of the components of that object, going down all the way to fundamental sub-atomic particles. Evolution, both biological and chemical, is an emergent by-product of fitness functions playing out at a macro-scale. For its first few thousand years of its existence after the big bang, the universe was fundamentally different from what we see today. It was entirely made up of hot and dense radiation and it was even a little boring. It wasn't until the conditions became just right that matter began to form. The first sub-atomic particles that formed were quickly destroyed, but eventually the conditions became stable enough for larger particles to form. The particles that were more fit within the fitness landspace of this early universe were the ones that persisted into the future. Eventually matter started clumping and larger structures formed, such as elements and molecules, each of which were affected by and contributed to a fitness landscape.

Even stars, despite their magnitude, follow a somewhat predictable pattern of evolution. The universe is so old, that looking up in the sky you can see all kinds of stars in various phases of their lives. So we know, for example, that stars like ours live long lives and tend to become hotter and dimmer as they age, whereas other stars are hot and bright and burn up all of their energy in a short amount of time. Therefore it makes sense to speak of the fitness of stars, because stars have their own sphere of influence and can be disrupted by other stars. In general, the stars with higher fitness are the ones we see in our night sky today. A star that strays too close to a red giant might eventually have all its gas sucked up by its bigger neighbor. All that is saying, in this context, is that the red giant has a greater fitness than that smaller star. This is simply a result of ecological processes playing out at all levels of organization.

One more thing about mental models. Human survival throughout history has itself depended on our ability to create useful mental models of our environments. People with larger brains, and more importantly better mental models, were able to outcompete people and animals that lacked those assets. Forty thousand years ago Europe was covered in ice and inhabited by both humans and Neaderthals. Today there's only us--the Neanderthalers were outcompeted and went extinct some time before the start of recorded history. Because these two species occupied the same niche and exploited the same resources, it was inevitable that one would eventually out-compete the other. In this case, it turned out to be the species that had the better mentals models--the one that was better able to plan ahead, provision food, and stay alive in a harsh environment.

This is not the same as saying that humans always the correct mental models. Prehistoric humans believed in fictional deities and practiced human sacrifice. Clearly something was wrong. All it's saying that they had better mental models than the competition. It's not always important for an idea to be 100% correct, what's important is that it gets us somewhere closer to the truth. Science itself is an evolutionary process whereby ideas are born, developed and discarded as even bigger and better ideas are born. New ideas can only be born if people are willing to look outide the lines of previous models and risk being wrong. So take a chance and risk being wrong: scientific progress depends on it.

Further reading

Chaisson, Eric: Cosmic Evolution

Deutsch, David: The Beginning of Infinity

Kauffman, Stuart: The Origins of Order