Seth provides a timely and compelling set of arguments for why life could be crucial for consciousness. However, his account does not clarify what exactly in life makes the difference – remaining agnostic whether this should be found in intrinsic, causal, or functional properties. Could a process that “looks like” life suffice, or does it need to be “life all the way down”?

It is nearly uncontroversial that consciousness operates on a larger scale than the constituent elements of its substrate. Thus, deliberating on the role of the substrate for consciousness requires a careful treatment of how different scales interact in the natural world and therefore the notion of emergence (Anderson, Reference Anderson1972). In the section Emergence and the separation of scales, Seth speculates that biological systems, by their very nature, cannot support emergent phenomena. To evaluate this claim, one needs to first articulate what emergence might mean in a world like ours.

To this end, we introduce the notion of a compatibilist emergence – akin to compatibilism in the philosophy of free will (Dennett, Reference Dennett1984). This position grants the potential for macroscopic scales to possess novel dynamical properties while simultaneously asserting that such properties obey supervenience. In other words, emergent features can be both unique irreducible processes in their own right while being contingent on their constituting factors, thus being at the same time substrate independent and dependent in distinct ways. Importantly, this view emphasizes not just the hierarchical spatial scales inherent to emergence but places explanatory weight on the way in which the system unfolds over time.

Under this theoretical purview, the scientific study of emergence has delivered multiple distinct frameworks for understanding emergent phenomena over dynamical processes (Barnett & Seth, Reference Barnett and Seth2023; Hoel, Reference Hoel2025; Hoel, Albantakis, & Tononi, Reference Hoel, Albantakis and Tononi2013; Rosas et al., Reference Rosas, Mediano, Jensen, Seth, Barrett, Carhart-Harris and Bor2020, Reference Rosas, Geiger, Luppi, Seth, Polani, Gastpar and Mediano2024; Seth, Reference Seth2010). Empirical applications of these frameworks are revealing that emergent phenomena are pervasive in the brain – and, what’s more, that neurobiology has no monopoly over them (Hoel & Levin, Reference Hoel and Levin2020; Klein & Hoel, Reference Klein and Hoel2020; Klein et al., Reference Klein, Hoel, Swain, Griebenow and Levin2021; Luppi et al., Reference Luppi, Mediano, Rosas, Holland, Fryer, O’Brien, Rowe, Menon, Bor and Stamatakis2022; Mediano et al., Reference Mediano, Rosas, Luppi, Carhart-Harris, Bor, Seth and Barrett2021; Milinkovic et al., Reference Milinkovic, Seth, Barnett, Carter and Andrillon2025; Pigozzi, Goldstein, & Levin, Reference Pigozzi, Goldstein and Levin2025; Tolle et al., Reference Tolle, Luppi, Seth and Mediano2024; Varley, Reference Varley2020, Reference Varley2023). Interestingly, these frameworks possess significant differences, coalescing into a collection of distinct facets of emergence:

• Emergence through macroscopic sufficiency – when a macro contains all the information relevant to itself (Barnett & Seth, Reference Barnett and Seth2023).

• Emergence through macroscopic causal dominance – when it is most efficient to control the system from the macro (Hoel et al., Reference Hoel, Albantakis and Tononi2013).

• Emergence through mereological constraints – when the whole carries more information than the sum of its parts (Mediano et al., Reference Mediano, Rosas, Luppi, Carhart-Harris, Bor, Seth and Barrett2021; Rosas et al., Reference Rosas, Mediano, Jensen, Seth, Barrett, Carhart-Harris and Bor2020).

This taxonomy of facets of emergence thus forces us to ask: if consciousness is an emergent process, then what type of emergence could it rely upon?

Seth claims that the generative entrenchment inherent in evolutionary design and the autopoietic imperative of organisms to regenerate their own basis may prevent biological systems from demonstrating macroscopic sufficiency. However, it is important to note that while the argument Seth provides may apply to one kind of emergence (macroscopic sufficiency), it does not undermine the relevance of others. On the contrary, autopoiesis implicitly invokes the latter forms of emergence, as it raises the question of which higher scales have been endowed with the appropriate horizontal and downward causal powers to execute these functions.

Seen from this perspective, the link that Seth proposes between self-organization, self-reproduction, and consciousness is compelling from both theoretical and empirical standpoints. Theoretically, at the highest scale, this may be a functional reason for why brains bother to furnish particular mental states with experience at all, since feeling carries the quality of intrinsic value and care that cognition-as-usual cannot offer in service of self-maintaining and directed activity (Cleeremans & Tallon-Baudry, Reference Cleeremans and Tallon-Baudry2022). Empirically, while the human brain appears to exhibit macroscopic sufficiency, this may hinder rather than help consciousness (Milinkovic et al., Reference Milinkovic, Seth, Barnett, Carter and Andrillon2025). In contrast, considering consciousness as emerging through mereological constraints, not only does evidence suggest that brains can orchestrate this kind of emergence across scales (McSharry et al., Reference McSharry, Kaplanis, Rosas and Mediano2024), but indeed it seems to be attenuated in states of reduced consciousness.

This link between life, consciousness, and mereological constraints is supported by the fact that our nature as living organisms is one of nested interdependent parts and wholes, rather than consistently autonomous macroscales (Levin, Reference Levin2019). In effect, one way to distinguish machines from organisms is that while for both the whole is made of and depends on the parts, the parts of a machine are oblivious of the whole, while the parts of an organism critically depend on the whole for their subsistence (Nicholson, Reference Nicholson2013). In a similar way, consciousness may require the integrity of the whole to allow unified percepts but also the ability of that whole to constrain its parts – for example, to allow perceptual filling-in.

In sum, consciousness as a biological process is more plausibly realized not through separation but through interdependence, feedback, and synergy. Thus, contrary to Seth’s claim, it may be because of – rather than in spite of – emergence that the distinction between software and hardware may not be the right way of understanding creatures like us. In this way, we reiterate Seth’s remark that the properties of life have a significant purchase in the prediction and control of features of consciousness. However, we believe that this is, at its core, true because life has found itself to be a potent means of orchestrating the kind of emergence crucial for consciousness.

Lastly, we emphasize that although emergence may be crucial for consciousness, it is likely to be a necessary but not sufficient condition for it. This implies that the ability to display emergent properties does not necessarily result in synthetic phenomenology. Emergence is the medium, not the message – and as extolled here, it has many dimensions. Accordingly, optimizing artificial systems involves choosing a specific facet of emergence. For instance, naively pursuing macroscopic sufficiency alone is likely to lead to systems with runaway scales, lacking relevant bottom-up constraints, and possessing aberrant structure. Despite these limitations, measures of emergence may still serve as a useful indicator of whether a particular system is worthy of further inspection for the presence of consciousness.

Figure 1.

Facets of emergence. Top: Emergence through macroscopic sufficiency occurs if the microscale history adds no predictive power to the macroscale’s own self-prediction. Middle: Emergence through macroscopic causal dominance occurs if the coarse-grained macroscale causally constrains its own possible past and future states more effectively than the microscale does for its respective states. Bottom: Emergence through mereological constraints occurs if the whole has irreducible predictive power over itself or individual parts.

Figure 1. Long description

Panel A: A diagram illustrating macroscopic sufficiency. It shows a sequence of events over time, with macro-level interactions at the top and micro-level interactions at the bottom. Arrows indicate the flow of time from left to right. The macro-level interactions are represented by colored nodes and connections, while the micro-level interactions are represented by clusters of smaller nodes. The diagram shows that changes at the macro level can influence the micro level, but not vice versa. Panel B: A diagram illustrating macroscopic causal dominance. Similar to Panel A, it shows macro-level interactions at the top and micro-level interactions at the bottom. However, in this case, a perturbation affects the macro level, which then influences the micro level. The arrows indicate the flow of time from left to right, showing how the macro level dominates the micro level in response to perturbations. Panel C: A diagram illustrating mereological constraints. It shows the relationship between a whole and its parts over time. The whole is represented by an oval at the top, and the parts are represented by smaller ovals at the bottom. Arrows indicate the flow of time from left to right, showing how constraints on the whole can influence the parts and vice versa.