On the prospects of basal cognition research becoming fully evolutionary: promising avenues and cautionary notes

We began this paper by distinguishing two distinct manners of investigating the scope of biological cognition. The first of which, the definition-first approach, engages in the philosophical practice of constructing and amending definitions of cognition and it is primarily guided by a priori intuitions of what cognition is and where it is found in the biological world. Using an evolutionary-driven approach represents a second manner of investigating cognitive scope. While working definitions are a starting point, this kind of approach leads with what empirical investigations and phylogenetic contextualization reveal about the presence or absence of various cognitive capacities in the organisms investigated. Here the question of cognitive scope is not a question that reduces to one of strictly characterising and meeting a definitive ‘mark of the cognitive’ but is rather grounded in a question about how certain cognitive capacities that we recognise as such have evolved. The programme of basal cognition embodies this latter kind of approach and has the potential to avoid some of the stalemates and unending semantic battles–our familiar many-headed hydra–that have plagued definition-first approaches. In this paper, we have uncovered four parallels between the fields of Evo-Devo and basal cognition: (i) the search for shared, conserved toolkits and the importance of comparative causal-mechanistic research; (ii) panextensionalist framings to counteract the evolutionary limitations of oligoextensionalism about development and cognition; (iii) the implementation of phylogenetic thinking to uncover homologies and homoplasies (especially possible convergent routes of evolution); and (iv) the possibility to study losses, gains and uniqueness of traits, including complexification and simplification in particular lineages.

In its onset, Evo-Devo, like basal cognition today, was embroiled in controversy due to its departure from received views of development and its evolution. Evo-Devo eventually addressed many explananda regarding the (somewhat shared) causal-mechanistic underpinnings of morphogenesis and how development matters for accounting for evolutionary dynamics. In giving treatment to each of these four parallels, we have aimed to show that much of the current scepticism towards basal cognition and the idea of fruitfully investigating of cognition outside of the animal kingdom may also be unwarranted, or at least it should be tempered down. From each of the parallels identified, related cautionary notes for basal cognition have been drawn out (e.g., the need to complement investigation on evolutionary conservation with research on evolutionary divergence and differences, the pitfalls of trying to adjudicate between the antipodes of panextensionalism and oligoextensionalism about cognition in an empirically detached manner, the lack of attention paid to convergent evolution and the loss and gains of traits, notably secondary simplification, and assuming that ‘basal’ means ‘primitive’). These are cautionary notes that we hope will provide some guidance for this exciting new research programme on its way to becoming a fully evolutionary field.

Let us return to the examples of scepticism. As we saw, one form argues that if there is no strong evidence of similarity between what is being called “learning” or “memory” in non-neuronal organisms is and what cognitive scientists are interested in, there is no reason to deem those processes in non-neuronal organisms as such. To this, we respond: Firstly, there is no unified consensus within cognitive science regarding what cognition is, and the interests of cognitive scientists vary significantly (Cf. Sims, 2021). Secondly, even if there was consensus, using the interests of cognitive scientists as an exclusive metric is problematic given that the historical focus of this field has been on humans and non-human animals. Insisting that all instances of memory, learning, and decision making across taxa must resemble those of humans or of non-human animals is to be under the spell of oligoextensionalism. Moreover, just how similar a process must be to be considered similar enough to what cognitive scientists are interested in is also a vague matter. For instance, if P. polycephalum (slime mould) exhibits habituation in experimental conditions that rule out sensory fatigue (which it does), and habituation is widely acknowledged as a form of learning in animals, how much more similar must its habituation be to that in the rodents, corvids, or primates to count as relevant to the interests of cognitive science? As empirical research in basal cognition advances, some of this scepticism is likely to increasingly resemble the mid-20th-century resistance among developmental and evolutionary biologists to acknowledging developmental similarities across diverse taxa—a stance now seen as an artifact of pre-theoretical bias.

The second form of scepticism argues that even when basal cognition researchers successfully use feedback (cybernetic) models to represent adaptive and flexible behaviour in non-neuronal organisms, these models omit the phylogenetic and conceptual connections between such behaviours and those of organisms that are already recognised as cognitive. Our response is this: cybernetic models, like all scientific models (e.g., systems biology models), are abstract and idealized simplifications (for discussion, see, e.g., Andrews, 2021). By abstracting away from the complexity of target phenomena, models can serve as proof-of-principle tools for capturing cognitive principles in a, say, mathematically or empirically tractable manner. They do not provide evidence for cognition in a specific target system but buttress our understanding of possible shared mechanisms underlying observed behaviour. Thus, while Figdor (2024b) is correct that current basal cognition’s models usually omit phylogenetic and conceptual relationships, this omission is not a flaw on the part the basal cognition research programme or their cybernetic models, but rather a limitation inherent to the specific purpose of such models. The difficult task of elucidating such relationships instead falls within the domain of bioinformatics, comparative genomics, molecular phylogenetics and the models they deploy for cross-species comparison and phylogenetic reconstruction. A fully evolutionary approach, like the one basal cognition research could foster, would be able to complement comparative-mechanistic research with sound phylogenetic reconstructions and evolutionary-contextualized inferences. So basal cognition researchers could also challenge this line of scepticism in the future.

Importantly, since behaviour (cognitive or otherwise) cannot be reduced to gene activity, identifying phylogenetic and conceptual connections between behaviours of different taxa also requires taking reciprocal organism-environment dynamics into account (Baedke et al., 2021; Sims, 2024a). Considerations from Evo-Devo thus have a further significant role to play here in tandem with basal cognition research.

Finally, scepticism based on definitions of cognition, takes for granted that there is one agreed upon meaning (i.e., a natural kind) and views the job of empirical research of one of finding instances that satisfy this definition. Not only is there no such consensus, but, as our analysis in this paper has highlighted, putting all of one’s eggs in a rigid definitional basket might hinder and distract from hypothesis-driven investigation. As we have shown, investigating conserved toolkits, offsetting strict multicellularism, implementing phylogenetic thinking, and studying losses, gains and unique traits have yielded valuable insights in Evo-Devo. Similarly, an evolutionary approach to cognition can benefit from moving beyond using a priori intuition as an ends that prematurely constrain enquiry rather than as a means to further discovery. Although some degree of scepticism can foster scientific rigor and should always be weighed to evaluate scientific research programmes, scepticism based solely on (contested) definitions rarely contributes meaningfully to science in an empirically profitable manner.

Placing the scope question aside, we would like to end briefly with a few additional questions for future research that we believe may also be useful for basal cognition or any other approaches to cognition in which evolution is central.

How might exaptation (i.e., the co-option of an adaptive trait to play a different function than its original one) have played a role in the evolution of some cognitive capacities or cognitive mechanisms in particular linages that we see today and how can this be studied?

How can the practice of good taxon sampling (i.e., collecting data across major branches of the tree of life for comparisons at different degrees of taxonomic distance) be implemented in basal cognition to both avoid the tendency to overgeneralise and support specific generalisation across phylogenetic branches when evidenced? Better phylogenetic resolution has changed our understanding of neural evolution in animals (Hejnol & Lowe, 2015; Martín-Durán & Hejnol, 2021), and this is a necessary step for the advancement of basal cognition research as well.

How can we build robust frameworks for launching meaningful comparisons across (neuronal and non-neuronal) species that (might) exhibit cognition? A good pointer in this direction might be how reproduction, a widespread but radically divergent biological process across phylogeny, is studied comparatively (see Fusco & Minelli, 2019).

Can the kind of mesocosm experiments deployed in ecology be used in basal cognition research to investigate context-sensitive, cognitively-driven behaviours that might be otherwise supressed in largely artificial lab settings?

How can scientists study how cognition could affect patterns and dynamics of phenotypic evolution (similar to how development is taken to be important in Evo-Devo), complementing the studies on the evolution of specific cognitive repertoires and capacities?

Although much work still remains to be done, basal cognition shows great promise for developing into a fully-fledged field of evolutionary research.

We thank Carrie Figdor and Fred Keijzer for their helpful comments on an earlier version of this manuscript. We also acknowledge the feedback from attendees and participants at the BSPS 2024 Annual Conference’s symposium “The Many Implications of Basal Cognition for the Philosophy of Science” at the University of York (July 2024) and the CEFISES Seminar (MolDevBio series) at UCLouvain (December 2024). Likewise, we thank the two anonymous reviewers of this journal for their helpful feedback. MS acknowledges the generous support from the VolkswagenStiftung. AFT is grateful to The Ramsey Lab at KU Leuven and to the Research Foundation Flanders (FWO) for their financial support—in particular, for grant G070122N. MS is currently a part of the Major Transitions in Cognitive Evolution project, supported by grant no. TWCF-2020-20539 from the Templeton World Charity Foundation, awarded to A. B. Barron, M. Halina and C. Klein, all of whom he is incredibly grateful to.

Research Foundation Flanders (grant G070122N). Templeton World Charity Foundation (TWCF-2020-20539).