Introduction
One of the most profound mysteries in biology is the persistent increase in complexity over evolutionary time. While the Second Law of Thermodynamics dictates that physical systems tend toward disorder, life on Earth has consistently transitioned toward higher levels of organization—from unicellular life to multicellular organisms, complex social structures, and symbolic language. This paper proposes a formal computation-theoretic framework to explain these "Major Evolutionary Transitions" (METs) as directional "Turing jumps" that resolve computational inconsistencies within self-modelling systems.
Tangled Hierarchies and Self-Modelling
The authors define biological systems as "Tangled Hierarchies" (TH), where macro-scale patterns exert causal influence over micro-scale components. A critical distinction is made between two types of hierarchies:
- Type I (TH-I): Systems with emergent behavior but no internal representation (e.g., ant colony pheromone trails).
- Type II (TH-II): Systems that possess an encoded, compressed representation of their own dynamics (e.g., the genotype-phenotype relationship).
In Type II systems, the organism effectively acts as a Universal Turing Machine (UTM) that processes its own "program" (genetic code) as data. This self-modelling capability allows for a sophisticated level of control but introduces a fundamental computational challenge.
Computational Undecidability and the "Turing Jump"
The core of the paper's argument lies in the tension created by self-modelling. When a system encodes its own rules, it eventually encounters the limits of its own logic. Because the space of possible phenotypes (expressions) is always larger than the fixed genome (referent), the system hits a wall of computational undecidability—problems that cannot be solved within the current level of complexity.
To resolve these "tensions," evolution must expand the system's problem-space. The authors liken this to a "Turing jump" in mathematical logic, where a system gains access to an "oracle" capable of solving previously undecidable problems. For instance, the transition from prokaryotes to eukaryotes provided a massive expansion in the biological "alphabet," allowing for the resolution of previous computational bottlenecks.
The Biological Arrow of Time
Unlike the thermodynamic arrow of time, which is defined by increasing entropy, the biological arrow of time is defined by increasing correlations and the open-ended generation of computational novelty. Evolution is viewed as a directional "spiral" where each resolution of a computational tension creates a new, more complex landscape, which in turn generates new future tensions.
This framework suggests that biological information is not a passive record but has "causative efficacy." Higher levels of the tangled hierarchy drive the selection process of lower levels, leading to a shift in causal structure where information gains direct, context-dependent control over matter.
Conclusion
By synthesizing recursion theory with biological dynamics, this paper provides a formal taxonomy for complexity and a predictive explanation for the open-ended nature of evolution. While the framework remains highly theoretical, it offers a unifying perspective that links assembly theory, stability, and information theory, suggesting that the increase in biological complexity is a necessary consequence of resolving self-referential computational inconsistencies.
Summary updated