Where is the boundary separating “me” from the “others” with whom I share the universe? Are we truly individuals? These aren’t merely philosophical riddles. They are part of an emerging understanding that patterns from the living world echo patterns found in mathematics.

The domains of neuroscience, evolutionary biology and computer science have been quietly converging for decades. At the heart of the Venn diagram where they intersect lies a humble but immensely powerful theoretical construct: the Markov blanket.

Andrey Andreyevich Markov (1856-1922) was a Russian mathematician best known for his research on stochastic processes. Stochastic applications are used in probability theory to model systems and phenomena that behave randomly. Whether measuring the growth of bacteria, fluctuations in electrical signals or the movement of individual molecules, stochastic models are ideal for describing the natural world.

In his seminal work, later described as a “Markov chain,” Andrey Andreyevich described a series of events in which the probability of each event depends on the final state, or output, of a previous event. Essentially, Markov’s work converted certain forms of randomness into predictable causality. This body of work provided a statistical foundation for the development of Bayesian networks, thermodynamics, signal processing, and simulations used in chemistry, economics and finance.

In 1988, the Israeli / American computer scientist Judea Pearl coined the term “Markov blanket” to describe a process for inferring the value of a single variable from a subset of input variables. The blanket is a boundary that excludes variables that are extraneous to solving the problem.

Building on Markov’s ideas, Pearl’s work created a probabilistic foundation for the machine learning tools used to build artificial intelligence models. In this Bayesian framing, software is used to model an inference engine, which can adapt to and thus learn from patterns in the data the model is fed. In 2011, Pearl received the Turing Award, the most distinguished accolade in computer science.

These constructs, inspired by the randomness of the natural world, became some of the most essential tools in modern computing. In the process, they have unmasked fascinating echoes between biology, physics and the tech industry.

Cells are a perfect example of Markov constructs in action. One of the factors that differentiates the three domains of life (archaea, bacteria, and eukaryotes) from one another are variations in the chemistry of cell walls. Change the conditions of these cellular Markov blankets and a fundamentally different biological system evolves inside.

As for what defines our physiological individuality, the most common answer would be the epidermis. Our skin is a Markov blanket that isolates the other cells in our bodies from routine interaction with the world outside. Notably, there are trillions of bacteria and viruses cohabitating inside that boundary, suggesting that we are indeed communal, rather than singular, organisms. Since pronoun edification is all the rage, instead of “I,” you really should be using “we” to describe the Markov blanket that defines your individuality.

Cloud computing, which shares data center hardware across thousands of individual clients, is a technological Markov construct. Each client receives a ration of computing power, and within a discrete container the client hosts the operating system and applications of their choice - segregated from the data of other clients by a Markov blanket made of software.

British neuroscientist Karl Friston is arguably the foremost pioneer of brain imaging techniques and the author of numerous MRI patents. He is also, academically speaking, one of the most cited scientists of all time. Friston is the key architect of an enigmatic concept called the “free energy principle” (FEP) that extends the implications of Markov constructs back into the natural world.

The FEP describes systems that are distinct from, but coupled to, another system. This can occur in technology, like multiple cloud computing containers running simultaneously on a common hardware platform. The interface between the two systems, which effectively determines the amount of freedom that the embedded system enjoys, is a Markov blanket.

The FEP also describes the biology of multicellular life, where specialized heart, muscle and brain cells – despite sharing the same genetic code – perform radically differing functions on behalf of the greater organism they reside within. Multicellular life is the logical result of nested layers of Markov blankets, each one mediating the exchange of information with the next layer.

Friston originally introduced this mathematical principle to explain perception/action loops in neuroscience, and in that context the framework is known as “active inference.” The FEP postulates that self-organizing biological systems like the brain strive to minimize the difference between their internal models of the world outside, and the feedback their senses and perceptions provide from the other side of that Markov blanket. This difference is variational free energy, and in healthy brains it is minimized by continuous correction of the internal world model, or by making the world more like the predictions of the system.

By actively changing the world to make it closer to the expected state, systems minimize the free energy of the system. Friston suggests this principle underlies all biological reactions and argues it applies equally to mental disorders and artificial intelligence.

Any random, dynamic system, if large enough, will display the kind of boundary that allows one to apply the free energy principle to its behavior. The mathematical underpinnings of this principle are beyond my ken. However, like the Markov blanket, its metaphorical utility in describing the world is virtually boundless.

We each carry a model in our minds that describes how the universe works. Particularly during our formative years, we continually adjust this model based on new information. We build libraries of Jungian personality categories as we strive to model other minds. And eventually we get better at avoiding surprise, actively preferring data that validates our existing model over data that requires changing the model.

In politics, populists seek to demonize the “others” who are not part of a tribal “us.” But “us” is just another Markov blanket. Zoom out. All political tribes are contained in the “we” called American. Zoom out again. All of us belong in the Markov blanket called human.