The evolutionary "pressure" formulated in England's theory, and the alignment mechanism described by Friston's principle, are applicable to all living organisms. However, as noted above, they are most vividly manifest in Homo sapiens – the only known species that possesses consciousness and reason. The primary distinction between an intelligent being and a merely living one, in terms of their interaction with the world, is not quantitative but qualitative – a nonlinear amplification of that interaction.
In systems science, nonlinearity is associated with sensitivity to initial conditions, rapid, significant amplification of small perturbations, and the emergence of new qualities that were not present at the level of the individual elements. Intelligence manifests in precisely this way: a thought or idea, originating in a single human brain, can trigger a cascade of changes far exceeding the scale of the initial impulse.
While the influence of a non-intelligent organism on its environment is roughly proportional to its mass and reproductive rate, an intelligent organism can amplify its impact by many orders of magnitude. A herd of antelope stampeding across the savanna might cause a brief tremor in the earth, but it remains a local event. In contrast, a discovery by a brilliant physicist or an algorithm created by a talented mathematician can alter the developmental trajectory of an entire civilization, unleash powerful energetic or computational processes, and ultimately, radically accelerate entropic transformations on the scale of not just the planet, but near space as well. As I wrote in “Cogito Man”, "imagine what just a few molecules of an aromatic substance can do to the world – the faint scent of perfume from a beautiful stranger, inspiring someone to a great accomplishment..."
This is exactly why it is "advantageous" for nature (in a statistical sense) to promote a higher degree of intelligence, thereby increasing the speed and magnitude of a system's intellectual engagement with the external world. The more advanced the intellect, the more efficiently it fulfills its dual function: accelerating energy dissipation and reducing information entropy.
But even more importantly, nonlinear effects are amplified manyfold when multiple intelligent agents unite into a coordinated community. What emerges is something we can call collective intelligence, whose capabilities far exceed the sum of its individual members. Systems theory suggests that networking is itself a nonlinear process, giving rise to new stable structures and mechanisms of self-organization. This intuitive notion is confirmed by formal approaches such as the theory of autocatalytic sets and the concept of cultural stability, which are discussed below.
The developmental process of collective intelligence can be mathematically described within the framework of complex systems theory – through the concept of autocatalytic sets [7-9], proposed by Stuart Kauffman. These are collections of elements capable of self-reproduction, self-maintenance, and self-amplification: reactions or actions within the set create positive feedback loops, that is, the conditions for new, significant reactions and actions to emerge. A classic example is a commune of specialists: a baker bakes bread for a blacksmith, the blacksmith makes tools for a carpenter, and the carpenter builds a new oven for the baker. No single individual is self-sufficient, but the collection of specialists forms a self-sustaining and productive system.
Why is this notion important? Because mathematics shows that as soon as a set of diverse "specialists" crosses a certain (very modest) threshold of complexity, it begins to grow explosively, generating ever more reactions and actions. The process is called autocatalysis, and it is obviously nonlinear in nature due to the aforementioned feedback loops: each new catalyst accelerates the emergence of subsequent ones. In chemistry, this manifests as sudden "phase transitions" from a set of simple compounds to complex ones (including organic molecules). When applied to intelligent systems, both individual and collective, autocatalysis means that ideas, discoveries, and inventions interact, forming a "network of cognitive achievements." In this network, new concepts serve as catalysts for the birth of even more complex and powerful ones. For example, the invention of machine learning algorithms spurred the development of new computer architectures, which in turn paved the way for progress in artificial intelligence – which is now inventing new algorithms itself. Each new node in such a network doesn't just add something to what already exists; it accelerates the entire process, increasing its nonlinearity.
Thus, human intelligence and the cognitive tools it has created form an autocatalytic structure that tends to accelerate its own evolution. In this sense, the development of civilization can be regarded as a statistically probable outcome of the dynamics of autocatalytic networks – "self-amplification machines" within the domain of knowledge, technology, and cultural practices.
The theory of autocatalytic sets perfectly illustrates how collective intelligence accelerates entropic processes. Evidently, it’s highly "advantageous" for nature to bring the development of life – particularly intelligent life – to the threshold where autocatalysis begins. Each new factor of nonlinearity (a new positive feedback loop) opens up additional opportunities for energy dissipation and efficient information compression, and this happens much faster than if the efforts of individual participants were merely summed.The quicker an autocatalytic set develops, generating increasingly powerful and diverse ideas, the more statistically probable it becomes. In this way, the universe itself is "pushing us forward," compelling us to accelerate progress.
[7] Kauffman, S.A. (1986). “Autocatalytic sets of proteins.” Journal of Theoretical Biology 119.1: 1–24.
[8] Hordijk, W., Kauffman, S., Steel, M. (2015). “Autocatalytic sets and boundaries.” Journal of Systems Chemistry 6.1: 1-10.
[9] Hordijk, W. (2022). “Autocatalytic sets: Complexity at the interface of chemistry and biology.” The Evolution Institute / John Templeton Foundation.
If autocatalytic sets explain how ideas amplify one another, then the concept of cultural stability shows why this process does not fade away or collapse into random fluctuations. For collective intelligence to preserve its accumulated achievements, a mechanism for long-term stabilization is necessary – and that mechanism is culture [10,11].
Culture, in a broad sense, is a system of symbols, knowledge, skills, and institutions passed down from generation to generation. From the perspective of information dynamics, culture reduces info-entropy: it filters and structures colossal flows of data, transforming them into relatively stable forms – languages, sciences, arts, and technologies. Moreover, each new cultural layer becomes a platform for the subsequent acceleration of innovation. As a result, the core principle of stability is achieved: the preservation of key patterns amid the continuous renewal of their content.
The fundamental property of cultural artifacts is that they outlive their creators. Culture is a mechanism for long-term preservation, allowing intelligent systems not only to evolve but also to retain their achievements in a changing environment. New generations do not need to start anew; from birth, they have access to libraries, the internet, and scientific theories developed before them. Cultural achievements form "networks of influence" that expand across the planet and through time – effectively creating an additional dimension for positive feedback loops that collectively accelerate autocatalysis.
КFurthermore, cultural heritage provides future generations with a more stable and predictable environment, reducing Friston's 'free energy' – the gap between expectations and reality. Thanks to the experience of our ancestors, we do not have to constantly re-adapt to the oddities of our surroundings. These freed cognitive resources can instead be directed toward higher-level challenges, expanding the autocatalytic set of ideas.
Thus, intelligent systems gain an evolutionary advantage not only from their cognitive power but also from cultural memory. The collective intelligence it supports becomes a stable product of the autocatalytic process, capable of outlasting both individual agents and entire civilizations. The universe statistically favors not only intelligence as a generator of entropy (and an absorber of info-entropy), but also intelligence as the creator of autonomous cultural forms that allow subsequent generations to fulfill its entropic demands even more effectively.
[10] Dean L., Vale G., Laland K., Flynn E., Kendal R. (2014). “Human cumulative culture: a comparative perspective.” Biol. Rev. 89, 284-301.
[11] Vale G., Coughlin C., Brosnan S. (2022). “The importance of thinking about the future in culture and cumulative cultural evolution.” Philos Trans R Soc Lond B Biol Sci., Dec 19;377(1866)