An introduction to Darwinian physics

Darwinism's application domain expanded in the 20th century from its traditional area - of the evolution of DNA-based organisms. Despite vociferous objections from some anthropologists and philosophers, it is now widely recognized to cover a wide variety of cultural phenomena - including science, technology, language and religion.

Paralleling the rise of Darwinian cultural evolution, Darwinism was also applied to the development of organisms. First, evolution within the immune system was observed - and then various evolutionary models of other developmental stages were found to be useful. Multi-cellular organisms can be usefully regarded as populations of cells which themselves evolve over the lifetime of the organism they are part of. Darwinian models of psychological evolution were also developed. Skinner's model of learning, for example, was explicitly Darwinian. Gerald Edelman was involved in both the immune and neural breakthroughs.

These expansions of the application domain of Darwinian evolution led to a back-to-basics study of Darwinism and its limits. These studies suggest that Darwinism has applications well beyond biology.

Physics has long laid claim to observation selection effects - though the main students of selection are those studying evolutionary biology. However, it was additionally found that many simple physical systems can be modeled using traditional Darwinian models - based on copying with variation and selection. It turns out that the tree-shaped patterns found in electrical discharges, propagating cracks, and fractal drainage systems are all composed of family trees - representing patterns of descent which are subject to variation and selection. This can be seen in slow motion images of these phenomena.

One of the key concepts in understanding these systems is positional inheritance - or, more generally, spatio-temporal inheritance. When entities split, the offspring often inherit a variety of properties from their parents. One set of properties that are regularly inherited with high fidelity are spatio-temporal coordinates. Other properties are also regularly inherited. Velocity, charge, temperature and chemical composition are other examples.

In practice, many variables correlate with position. A splitting pebble inherits its parent's position, but it also inherits its acidity level, humidity, temperature, and many other parameters from its parent's ecosystem. This is due to the smoothness of nature. Nature's smoothness helps to ensure that fitnesses are inherited - which is one of the requirements for adaptive evolution.

The upshot of all this is that we can use Darwinian models to study simple physical systems, and the concepts of adaptive fit and fitness are applicable. The resulting field of study is known as Darwinian physics, which is considered to be part of Universal Darwinism.

Physicists have long realized that physical laws might cover these types of complex systems - and various models of them have been built. Often these models have then been exported to biology. Fractals, reaction diffusion systems and diffusion limited aggregation are examples of physics being applied to biology. However, physicists seem to have been reluctant to look to the fundamental principles of biology and see how they apply to physics. The thinking seems to have been that Darwinian principles only apply to living systems - and that the high fidelity copying required is rare elsewhere. On this line of thinking, biologists have Darwinism covered.

In fact copying with variation and selection are ubiquitous in nature. Copying takes place whenever information at one location spreads to multiple locations. This can be formalized in terms of Shannon's concept of mutual information. The process is a common one. It happens when waves radiate. It happens when sunlight hits dust. Selection is similarly ubiquitous. In its simplest form, selection involves choosing a subset of items from a set. That happens all the time - and isn't confined to the death of living things or to choosing mates. When rocks erode, some of them collapse into the water, while others do not. Some of the Sun's photons hit planets, while others do not. These can usefully be interpreted as cases of natural selection.

Many simple physical systems behave as goal-directed fitness optimizers - in the same way as biological systems do. They can be used to solve optimization problems - much as genetic algorithms can. Examples of this are the way a lightning strike finds a short path to the ground or a stream in a mountainous region finds the fastest path to the sea. Physicists have noticed these optimization capabilities and have developed their own models of them - most notably maximum entropy production principle. These thermodynamic models and the Darwinian models cover similar territory and mutually illuminate each other.

Another area where Darwinian models have found to be applicable involves observation selection effects. Spencer's survival of the fittest can be usefully be generalized to observation of the observable - bringing observers into a central location within Darwinism. Classical Darwinian models of selection apply to observers in the exact same way that they apply to other entities.

One of the more studied areas of Darwinism within physics involves quantum physics. According to the many worlds interpretation, the world is constantly splitting. Some branches reproduce faster than other ones, and so become more numerous and thus more likely to be observed. Other interpretations propose a selection effect acting on these worlds - known as "wave function collapse". Whichever interpretation is closer to the truth, this looks as though this will be an interesting area for Darwinian or quasi-Darwinian models.

It has also been speculatively proposed that the visible universe is the product of Darwinian evolution - and thus has a lineage of ancestors which existed before the big bang. If so, we might reasonably expect to find some clues relating to this evolutionary heritage. We can see that the visible universe had a birth date - but other evidence of an evolutionary history currently remains elusive. We should do some further research on this possibility.

Darwinism is now over 150 years old. However, according to the picture here, the Darwinian revolution is only part way through. There's a lot of remaining revolution to go - and the process involves a lot more than just mopping up some creationists.

For references not hyperlinked to above, see the references here and here.