The Science of “Superposition”: Towards an Interdisciplinary Perspective
This article introduces a perspective that perceives the various characteristics of DNA created in the process of evolution as a probability distribution.
This viewpoint of probability distribution is inspired by the superposition state of quantum mechanics. Usually, the superposition state is a concept applied to quanta. By redefining state space to DNA, we propose an abstract model that can be used to describe biological evolution using the superposition state.
Furthermore, I will explain how this perspective of superposition states can be extended to domains of intelligence and society, such as the chemical evolution of organic compounds, language, knowledge, academia, and culture.
Overview of Biological Evolution
The starting point of biological evolution is the creation of individuals with DNA different from their parents, due to mutations or breeding.
When the features of the new DNA match the environment and benefit survival and reproduction, the genes will be passed on and spread to the next generation. On the other hand, if the characteristics are disadvantageous for survival or reproduction, the likelihood of descendants carrying those features decreases, and that DNA eventually disappears.
It’s believed that organisms have evolved in this way, where new DNA emerges and those beneficial for survival remain.
In this evolutionary process, there’s a fascinating phenomenon: co-evolution.
Co-evolution: The Example of Flower Nectar and Bee’s Eyes
There are cases where two species establish a give-and-take relationship, benefiting each other’s survival and reproduction. This is co-evolution. The relationship between flowers and bees is a classic example.
Why did bees start collecting flower nectar, and why did flowers get bees to assist in pollination?
Let’s assume that flowers first offered nectar, attracting bees to collect it. If so, then the flowers evolved to ensure pollen sticks to bee’s legs more easily.
If that was the case, why did nectar-producing flowers have a better survival rate? It’s hard to explain.
Conversely, if bees first started pollinating by collecting pollen on their legs, flowers must have evolved to provide nectar to bees.
Why would bees approach flowers even if they didn’t get nectar? Why did bees evolve to be attracted to the color of flowers without any nectar rewards? Did something else incidentally develop that also benefited flower pollination?
Model of Co-evolution
I believe that understanding evolutionary models becomes easier by taking hints from the superposition state in quantum mechanics.
Let’s assume there were flowers that didn’t offer nectar and bees without eyes attracted to flowers.
Both flowers and bees reproduce.
Among a large number of individuals, their DNA characteristics are almost the same. This unchanging characteristic, from the perspective of superposition, can be seen as a definitive state.
However, due to DNA mixing or mutations, there are individual differences in DNA. From the viewpoint of superposition, overlapping all individual DNAs shows multiple states that are not yet determined.
This superposition state exists in both flowers and bees. Within it, there’s a slight chance of flowers offering nectar and bees being attracted to those flowers.
The overlap of these two states has a low probability but isn’t zero. When co-evolution actualizes a point of overlap that slightly boosts their survival chances, the probability distribution of this overlap increases in the next generation.
After stacking this for many generations, the probability distribution of the evolution of flowers and bees converges to this point.
Example of the Probability Distribution Concept
Let’s assume there were 10,000 evolutionary possibilities for flower DNA. Bees also had 10,000 patterns. Among them, let’s say there was just one pattern where the flower could offer a little nectar and one where the bee was attracted to the color of that flower.
By simple calculations, out of 10,000 individuals, one flower can offer nectar, and one bee is born attracted to that flower.
Even if the one bee and one flower were born, it would be meaningless if they didn’t meet. If one bee can only access 100 flowers, the chance of them meeting is 1%.
Even if they meet, the chances of them mutually benefiting in terms of survival and reproduction might be slim.
However, that’s just for one generation. Over time, these slight differences in probabilities accumulate. Initially, the 10,000 evolution patterns had an even probability distribution. But small differences in survival chances slowly skew this distribution.
As generations stack, this bias amplifies. After tens, hundreds, or thousands of generations, almost all flower individuals would produce nectar, and nearly all bee individuals would be attracted to the color of those flowers.
Superposition State: Quantum Mechanics and Genes
Some might feel uneasy about juxtaposing the concept of quantum superposition with that of genetic superposition. In quantum mechanics, superposition refers to the concept where states are genuinely overlapped, while in the case of genes, the entities are separated among different individuals, meaning the states aren’t truly superimposed.
However, I believe this issue boils down to how one defines the state space.
For quanta, it can mean considering the superposition of where a quantum might exist in a physical three-dimensional space or, for instance, the idea that an electron’s spin could be both up and down simultaneously.
In these examples, the state space could be considered as a continuous position in physical three-dimensional space or a discrete two-value spin state. From my perspective, I am defining the state space in terms of the evolutionary patterns of DNA.
While the definition of state space differs based on the subject, if one can represent a probability distribution under that definition, the rest follows the same abstract model.
Over time, one adds to the model the rules by which this probability distribution changes. Given enough time, this should allow for simulations predicting the evolution of the probability distribution.
Beyond the passage of time in a single state space, the approach of two hydrogen atoms and their transition to a bonded state can also be seen as a change in probability distribution. This operates on the same mechanism as the co-evolution of flowers and bees.
Differences and Commonalities between Quantum Mechanics and Genes
My argument is abstract: it looks at the existing proportion of differences in the DNA of a species’ entire population as a probability distribution. I’m not suggesting that the physical DNA is in a superposition state or discussing the superposition states of the atoms or electrons within that DNA.
The probability distributions of an entire population’s DNA and the states in quantum mechanics are different in nature and their presupposed state spaces. However, they share a mechanism: over time, their probability distributions change and tend to converge towards certain distributions, a purely mathematical property. By modeling as probability distributions, a common model has been extracted for both.
Of course, DNA is heavily influenced by its environment, so this simple model cannot capture all the nuances of real-world evolution. Yet, the repetitive change and convergence in the probability distribution can greatly impact DNA evolution, especially in specific patterns. Therefore, considering models based on disturbance-free environments can be meaningful.
In quantum mechanics, measurement or observation can collapse a superposition, resulting in a definitive state. Some might argue this doesn’t apply to DNA. However, how one defines observation and state determination matters.
I view the state where flowers produce nectar and bees being attracted to those flowers as equivalent to a collapse in the probability distribution, a definite state. In this scenario, when flowers and bees interact in the same space and influence each other’s DNA probability distribution, that’s an observation.
Application of Superposition States
Up to this point, we have abstracted the evolution of DNA as a probability distribution and proposed the idea of perceiving co-evolutionary phenomena as a convergence of states due to interactions between multiple probability distributions. This is similar to the way atoms bond, resulting in a convergence of the probability distribution to a specific state upon observation.
Though the abstraction discards certain external disturbances, making the model deviate from real-world phenomena, it can be argued that the convergence of states may play a more significant role than the discarded factors. Therefore, this model might still be capable of capturing general tendencies and directionalities in the evolution of DNA.
This abstract modeling is not limited to quantum mechanics or the evolution of DNA. It can be applied to the probability distribution of organic compound bonding in the chemical evolution prior to the origin of life, terms and grammar in language, the development of academia due to the interconnectedness of knowledge, and the progression of concepts in arts and culture.
When knowledge and ideas connect in our minds, we experience a phenomenon called “inspiration.” This might be similar to multiple atoms bonding or flowers and bees co-evolving. Multiple pieces of knowledge probabilistically overlap and interconnect at certain points, which then rapidly intensify in our minds, resulting in a confirmed state of excitement. This inspiration mechanism might be modeled similarly to quantum observation.
Superposition States in Neural Networks
While we discussed how the concept of superposition states can be applied to abstract ideas like knowledge, it can also be applied from a structural perspective in the brain.
Artificial intelligence technologies that emulate the movements of human and animal brain neurons are called neural networks. In neural networks, computations are made using multiple layers of neurons. The structure consists of an input layer, several intermediate layers, and an output layer.
In typical AI application systems, the neuron with the largest value in the output layer is considered the final output. However, the values in the intermediate layers can be thought of as probability distributions. Thus, each layer of the neural network can be said to possess a superposition state. As vast amounts of data are fed for training, the superposition states of each layer tend to converge to a certain probability distribution.
As previously explained, neural networks share a common mechanism with human and animal brains and AI. Thus, it can be postulated that both our brains and AI might store knowledge as superposition states.
In Conclusion
In this article, we introduced the perspective of superposition states in probability distributions to the evolution of life and discussed its applicability to co-evolution.
Furthermore, we presented the possibility of applying this concept of superposition state probability distributions to areas like the chemical evolution of organic compounds, knowledge, academia, culture, and the neural networks in AI.
Considering evolution and development across phenomena like life and intelligence with this model might offer new analytical perspectives across various domains. Additionally, cross-disciplinary applications of insights from this viewpoint could serve as an interdisciplinary approach.
