We argue for a theme overarching the pivotal traits in cognitively complex lineages, providing a theoretical story to explain why and how they lead to cognitive complexity. We ground these findings in the free-energy principle and life-mind continuity and argue that they contribute to generating effective, isomorphic world models.

Life-mind continuity approaches argue that the principles governing the origin of cognition are the same as those giving rise to life (Godfrey-Smith, Reference Godfrey-Smith1996; Thompson, Reference Thompson2010). One such principle is the free-energy principle and the related framework, predictive processing (Parr & Friston, Reference Parr and Friston2019). This view portrays living organisms as entropy-resisting systems keeping homeostasis to maintain their organization across time and in the face of threats, such as temperature changes or metabolic demands. This requires specific and timely regulation of homeostatic drives, which is done in line with the good regulator theorem: “Every good regulator of a system must be a model of that system” (Conant & Ashby, Reference Conant and Ashby1970, p. 1). This imperative enables the sustenance of life and cognition, and it is argued that living organisms, even those not conventionally considered “cognitive,” adhere to it (Kirchhoff et al., Reference Kirchhoff, Parr, Palacios, Friston and Kiverstein2018). On a phylogenetic scale, species-relevant regularities about the environment are encoded in the physical characteristics of organisms (Rovelli, Reference Rovelli, Aguirre, Foster and Merali2018). On an ontogenetic scale, a nervous system can fulfill this imperative more flexibly by maintaining a (hierarchical generative) model of the environment, constantly updated to be aligned with the world (i.e., reduce prediction error; Bruineberg, Kiverstein, & Rietveld, Reference Bruineberg, Kiverstein and Rietveld2018). This model goes beyond a surface-level sketch of the world, including latent causes to enable informed prediction.

This simple mechanism of prediction-error minimization accommodates classically “human” psychological features such as beliefs, goals, or motivation (Clark, Reference Clark2015), as well as higher-level cognition; one critical feature of higher-level cognition is the ability to detach from sensory input such as when considering consequences of actions before performing them (Pezzulo & Castelfranchi, Reference Pezzulo and Castelfranchi2007). Forward models, originally evolved as efference copies, form the basis for detaching simulations until they are small-scale models (Grush, Reference Grush2004; Pezzulo & Castelfranchi, Reference Pezzulo and Castelfranchi2009). Beyond detaching models, higher-level cognition also exapts models: Ostensibly primitive, action- or perception-based abilities serve higher-level cognitive functions. This includes metaphoric mapping, where simulating concrete actions represents abstract concepts (e.g., walking for moving forward in time; Friedrich et al., Reference Friedrich, Fischer and Raab2025) or conceptual spaces where representations of space deploy to organize semantic knowledge (Bellmund et al., Reference Bellmund, Gärdenfors, Moser and Doeller2018). With these exaptations, life-mind continuity extends from simple bacteria to abstract human cognition (Friedrich & Fischer, Reference Friedrich and Fischer2025). Therefore, cognition is understood as an improvement on, and re-use of, more primitive world model strategies.

This framework further posits that mental representations are structural representations (Gładziejewski, Reference Gładziejewski2016; but see Ramstead et al., Reference Ramstead, Kirchhoff and Friston2020). This means that they are isomorphic to their referent by retaining the internal relational structure. Therefore, the term model does not imply direct mirroring but is rather like a key is a model of its lock (cf. Shepard, Reference Shepard1981). Similarly, the world model retains the relevant structure of the environment. This reading of life-mind continuity suggests a deep coherence in which the world, the nervous system, and the phenomenal experience are all isomorphic to one another (Friedrich & Fischer, Reference Friedrich and Fischer2025; Lyre, Reference Lyre2022). Therefore, effective interaction with the environment involves extracting organism-relevant structures from the physical world and retaining their isomorphism in neural and phenomenal media.

The target article provides evidence for these theoretical postulates. The pivotal traits which precipitate cognitive complexity 1) enable better world models, 2) enable the detached models critical for cognition, and 3) evidence the central role of isomorphism.

First, the pivotal traits for cognitive complexity are those that enable developing richer world models. The authors state that these are “united in terms of their involvement in active sensing” (p. 10). One example is flexible limbs. These accompanied further ways of interacting, perceiving, and representing the world, such as neuromuscular control systems, antennae, mouth parts, and so on. This consequently pressured the development of high-resolution eyes, which could extract salient object features by edge detection, increasing the resolution of the world model. Therefore, the pivotal traits, from limbs to antennae to eyes, share the attribute that they add novel ways of representing the world and increase its resolution. Other pivotal traits also enrich the world model, but for a different reason. Appendages with multiple degrees of freedom enabled grasping, which granted the capacity to manipulate objects. This ability to intervene with objects in the world is important because it allows generating counterfactual competence about underlying causes (Williams, Reference Williams2018). In human ontogenetic development, the importance of this is already well-evidenced (Goddu & Gopnik, Reference Goddu and Gopnik2024). Together, these aspects demonstrate that the pivotal traits all substantially enrich the world model by improving its modalities, resolution, and counterfactual competence.

Second, the important ability to detach models is evidenced. With limbs and motor skills, it became necessary to forecast to, for example, avoid obstacles. Furthermore, the ability to navigate along with visual information also necessitated being able to differentiate between the consequences of an organism’s own or others’ movements. Both of these generate selective pressures for efference copies, which, later in phylogeny, enabled detached models.

A third aspect of the target article is that it evidences the importance of isomorphism. It argues that processed information retains its “retinotopic register” (p. 14), meaning that it retains the structure of the previous “processing station” (p. 13). This provides evidence for isomorphism. In addition, beyond evidencing its existence, the development of cognitive complexity required retaining “retinotopic (or more generally, topographic) maps that preserve the spatial order” (p. 15). This therefore provides support for isomorphism’s presence as both a consequence and a necessary condition for cognitive complexity.

Concluding, three central features put forth in the target article provide evidence for central parts of life-mind continuity: First, the traits of cognitively complex lineages all allow for richer world models by more diverse modalities, finer resolution, and counterfactual competence. Second, motor skills in concurrence with visual sensing precipitated efference copies, critical for developing detached models. And lastly, isomorphism not only accompanied cognitive complexity but was necessary for its development.