Mind Everywhere: A Framework for Conceptualizing Goal-Directedness in Biology and Other Domains-Part Two

Before concluding our essay, we would like to address a few key objections to provide further support for our view and help the reader better understand it.

You are implying that scientists and philosophers should abandon current approaches to goal-directedness, which seem to do a pretty good job of explaining some important cases, such as homeostatic regulation and the evolution of biological adaptations. Why throw out the baby with the bathwater?

As we stressed in the second section of this article, we are not arguing that these other approaches should be rejected, but we are arguing they cannot adequately explain, and more importantly, extend, the experimental approaches described in the section on experimental results in Part One of this article and therefore a new approach, which more effectively drives research into the competencies of unconventional embodiments of mind (with its many practical applications), is needed. Current approaches could be used, depending on the type of goal-directedness that scientists are trying to understand. Some types of low-level goal-directedness, such as homeostasis, might be best explained in terms of causal roles and relationships with no mention of agency or intelligence. Biological adaptations, in many cases, could be explained in terms of the operation of natural selection over time and nothing more. We see no need to interfere with the use of explanatory frameworks that yield fruitful results. Since we accept multiple approaches to goal-directedness, our view could be characterized as pluralistic and inclusive, even though it represents a conceptual shift toward mentalism. We argue that, on the spectrum of agency, the goal should be to empirically determine the optimal position for a given system—not intentionally skew low (Morgan’s Canon (Morgan 1903) nor high (animism) but use experiments to determine which discipline (ranging across physics, cybernetics, behavioral science, psychoanalysis, and others) provides the most payoff in terms of prediction, control, and synthetic discovery of new vistas of research.

Your approach is incompatible with scientific materialism because it involves mentalism, which is too high a price to pay for explaining experimental results. Scientists and philosophers have worked hard since the 1950s to show how teleonomic language in biology provides materialistic explanations of goal-directedness, and your approach would undo all that.

As we have argued in Part One of this article, we think that our approach is compatible with scientific materialism, because we are not assuming the existence of metaphysically irreducible minds. We hold that mind, intelligence, agency, and related terms are ultimately compatible with the objects, processes, and properties belonging to physics and chemistry (Levin 2022; Levin and Dennett 2020). While cognitive scientists, neuroscientists, and philosophers have made tremendous progress toward understanding how the human mind functions over the last 50 years, their achievements are still decades away. We readily acknowledge that the mind–body problem remains one of science’s great mysteries (Chalmers 1997, 2010), but we argue that the way to make progress on this and other problems is to move away from the implicit dualism of artificial sharp boundaries between canonical mindful entities and their more primitive origins and instead develop principled research programs that seek to understand the scaling and transformation of diverse cognitive capacities on evolutionary and ontogenic timescales.

Using the word “intelligence” and other cognitive terms outside of their familiar context of brainy animals immediately raises questions: might not these terms be misused? Are not morphogenetic systems simply following the rules of chemistry—why anthropomorphize them?

First, in the modern age, we must accept that all cognitive systems—ourselves included—exhibit chemistry, not magic, when one drills down to examine the lower levels. Thus, there simply is no special human category that one can correctly anthropomorphize as somehow being beyond the laws of physics at its base. We argue that this word—“intelligence”—is in some ways an anachronism that needs to be reenvisioned in favor of an empirically grounded view, updated with the latest findings in causal information theory (Albantakis et al. 2017; Hoel and Levin 2020; Hoel 2017, 2018; Hoel et al. 2013, 2016; Moore et al. 2018), in which it is perfectly possible (in fact, unavoidable) for a system to both, be subject to chemistry, and also to possess additional levels of description and control whose recognition affords novel benefits. We offer two points in clarifying our use of this terminology (developed in detail in Levin 2022). An uncontroversial aspect of our view is that claims of intelligence (and other cognitive terms), like all others, must be based on rigorous experiment, supported or ruled out by the degree of objective benefit that a given framing affords in terms of a) prediction and control, and b) future discovery (and new research programs) it suggests. The latter is most significant because almost any paradigm can be rescued by enough epicycles; indeed, after one has discovered a new effect or reached a new capability, it is easy to drill down to the chemistry and, looking backward, claim that there is no intelligence here because it mechanically follows the laws of physics. The same is true for any act of a complex human brain–body system—if one insists on a view from the level of particles, it will always be there.

The key question is: does that level of perspective provide the most interesting platform from which to make the next discovery or develop the most effective control policy? The emphasis should be on novel capabilities, and new research programs facilitated (or suppressed) by a given perspective. Thus, we propose that attempts to mine the rich toolbox of behavioral science to understand and exploit capabilities of morphogenetic systems will continue to pay off in many (but no doubt, not all) cases. We have fleshed out the prior gains facilitated by this view, as well as the promises for regenerative medicine, elsewhere (Lagasse and Levin 2023; Mathews et al. 2023). The less conventional, and sometimes uncomfortable, aspect of our position is that the empirical utility of framings needs to be applied fearlessly and followed wherever it may lead: its empirical consequences must be taken seriously even when they contradict long-cherished a priori commitments to how non-intelligent a non-brainy system must be. In other words, if a specific framing, which uses tools usually reserved for brains, results in fruitful new research programs on bacterial biofilms (Martinez-Corral et al. 2019; Prindle et al. 2015; Yang et al. 2020), plant roots (Bassel 2018; Calvo et al. 2016; Ciszak et al. 2016; Davis et al. 2023; Johnston and Bassel 2018; Topham et al. 2017; Yokawa et al. 2018), the training of gene-regulatory networks (Biswas et al.2021 , 2022; Csermely et al. 2020; Gyurkó et al. 2013), or developmental/regenerative biology (Levin 2021), then the scientific approach requires that we consider those systems to be bona fide subjects of that corner of the natural world that is supposed to be described by the behavioral science of a spectrum of minds.

Aren’t you mistaking emergent complexity and unpredictability for intelligence? Is there really a research program here?

Explicitly, we hold that the reason to attribute intelligence to a process such as morphogenesis is not because it looks complex, unpredictable, agential, or anything else. The only reason to attribute a specific degree of intelligence (and claim specific capabilities as listed above) is to do perturbative experiments that reveal them. These competencies cannot be inferred from strictly observational data, and complexity alone is insufficient—different types of problem-solving capacities need to be explicitly demonstrated in functional experiments. Likewise, they cannot be ruled out based on philosophical commitments or classical categories and colloquial language use but must be argued against based on experimental evidence. The research program we propose is rigorously focused on empirically testing, not historically assuming, the boundaries between disciplines that have suppressed new discoveries by the gate-keeping threat of “category errors”. The categories must evolve with the science, not be maintained for historical reasons. Just as people are using tools of cognitive science to rigorously test the capacities of swarm intelligence (Couzin 2007, 2009; Dreyer et al. 2025) in collectives of animals (going up a level from conventional individuals), we emphasize the utility of testing these tools in novel substrates and at novel scales (Baluska et al. 2022; Baluska and Reber 2021 ; Calvo et al. 2017, 2020; Deisboeck and Couzin 2009; Levin 2023c; Mori and Koaze 2013; Reber and Baluska 2021 2021; Reid et al. 2013; Saigusa et al. 2008; Trewavas 2016; Vallverdu et al. 2018; Zakirov et al. 2021). In all of these cases, an operational and pragmatic view of cognitive terms suggests that when their application leads to progress and new discoveries, they are legitimately applicable.

On the contrary, taking evolution seriously requires that there be ancestral functions leading up to the development of the most obvious features of brains. Indeed, it is now known (Fields et al. 2020; Keijzer et al. 2013) that all of the mechanisms used by the brain to carry out its amazing functions—ion channels, electrical synapses, neurotransmitters—are evolutionarily ancient and are already doing brain-like signaling in bacterial biofilms. Not only the mechanisms, but many of the algorithms around information processing, active inference, and so on, are widely conserved beyond brains (Friston et al. 2015; Kuchling et al. 2020; Pezzulo and Levin 2015). Thus, if either mechanistic composition or functional behavior is taken as the kind of things that provide evidence for cognition, there should be no surprise to find it in the living material that was speed-optimized and augmented to become the central nervous system.

Returning to where we began in Part One, since Aristotle’s time, scientists have been trying to understand purposiveness in the living world. Attempts to banish teleological concepts from the life sciences have proven to be largely unsuccessful because (a) life exhibits myriad forms of goal-directedness that cannot be easily explained in causal-mechanical terms and (b) teleological concepts, such as goal, purpose, and function, are useful in hypothesis formation, model building, and experimental design in biology. In response to this predicament, scientists and philosophers have proposed theories of biological teleology that avoid the use of methodologically “dubious” mentalistic concepts and explain goal-directedness by referring to properties, processes, and objects that seem to be compatible with a mechanistic worldview, such as genetic program, causal role, adaptation, and field. While these non-mentalistic approaches to teleology can account for many phenomena in biology and other domains, they are not capable of adequately explaining ground-breaking experimental results in morphogenesis, synthetic biology, and other fields, as described in this article, which involve the manipulation of cells and tissues connected by bioelectric networks. In this article, we have defended a mentalistic approach to goal-directedness to account for these and other experimental results, but the view also applies to other biological systems that exhibit complex, goal-directed behaviors, such as ant colonies, economies, and genetic regulatory networks. Our view is motivated not by the desire to prove any philosophical points but by the practical necessity of developing concepts that can fruitfully guide inquiry in these fields of research and suggest testable hypotheses, explanatorily robust models, and theories.

Based on the arguments and conceptual clarifications we have made in this two-part article, and examples of real research programs that are not only consistent with, but actually driven by, the ideas we have defended, it is clear that a continuous, pluralistic, and pragmatic approach to agency, mind, and cognition beyond the brain is possible. The field of diverse intelligence offers numerous opportunities to move beyond the increasingly limiting teleophobia and animism of the past, toward developing principled theories that port tools and concepts across the increasingly blurring boundaries of disciplines that were established centuries ago. Truly transdisciplinary work, spanning science and philosophy, can now be done to positively impact clarity about the status of embodied minds and the biomedicine, engineering, and other practical advances that await our ability to optimally interact with intelligence in unfamiliar guises.