Coombs and Trestman (Reference Coombs and Trestman2025) propose to leverage what we know about the evolution of bodies to help explain the evolution of advanced cognition. (I also advocate this approach: Figdor, Reference Figdor2024, Reference Figdor2022). Their version improves over more abstract and braincentric proposals (e.g., Cisek, Reference Cisek2019; Barron, Halina, & Klein, Reference Barron, Halina and Klein2023) by focusing on a wider range of whole-organism traits and their known distributions in a widely accepted phylogeny of Animalia (similar to Ginsburg and Jablonka, Reference Ginsburg and Jablonka2019). But while their “pivotal” traits are intended to ground the “future evolvability of complex cognition” – including such charismatic abilities as mental time travel/future planning and theory of mind/mental state attribution (target article, Table 1) – it’s not clear how we are supposed to get from here to there. Building this bridge will require addressing theoretical as well as methodological challenges. I identify three particularly important theoretical challenges:

1. Individuating cognitive characters. C&T write only of “traits,” but phylogenetic approaches rely specifically on characters in the evolutionary-biological sense: inherited traits that are individuated and shared across species, and are the relata of homology (Wagner, Reference Wagner2000). They are helpfully contrasted with phenotypes, or traits of individual organisms that result from heredity, developmental processes, and environmental influences. Identifying and individuating characters are difficult biological problems, especially for characters above the gene level, since the same character in distinct species need not share superficial similarity in form or function and distinct characters can be manifested in superficially similar phenotypes. But the problems are exacerbated in psychology for a number of reasons.

First, the charismatic cognitive abilities we ultimately want to explain have long been considered uniquely human and defined by the phenotypes of a single species. Characters can be unique to a species or family (an autapomorphy), but not by definition. We will need to distinguish features of human phenotypes that are relevant to defining characters from the features unique to human phenotypes by expanding the sample of species we use to initially define the character. While the sample may be augmented later to revise the character definition, our choice should be guided by phylogeny, not serendipity.

Second, psychological properties are often thought of as abilities or states (or functions, in one sense of that term). Abilities don’t fossilize and are more difficult to individuate in general than structures like bones or genes. That makes tracing them through phylogenetic timescales difficult. Third, we must infer psychological abilities from behavior and other markers even in existing species, including humans (a.k.a. the problem of other minds). This source of epistemic uncertainty is by no means unique to psychology, but it can be turbocharged by the human uniqueness assumption mentioned above.

The bottom line: We need to define cognitive characters if we want to build a bridge to charismatic cognitive “traits” from “pivotal traits.” Otherwise, we can’t use a phylogenetic approach to help explain charismatic cognition.

2. Elaborating a taxonomic hierarchy of cognitive characters. C&T’s universality condition is a requirement for characters that are distributed broadly in Animalia, but not too broadly (e.g. olfaction). The condition is well-intentioned but doesn’t quite identify what we need, which are cognitive characters individuated at multiple grain sizes linked to phylogeny. For example, while Cnidaria (and possibly Ctenophora) are capable of simple forms of associative learning, more advanced forms (such as trace conditioning) may need a brain or at least neurons with synapses. Meanwhile, mindreading and mental time travel, among other charismatic abilities, are presumably not widely distributed at all and may not be present in two of their three “CC” lineages. If nature were strictly regular, we might hope for cognitive clades – groups of species with shared cognitive abilities that are also monophyletic. However, homoplasy is likely for at least some cognitive character relationships across species.

The bottom line: We can aim for a nested hierarchy of cognitive characters that can be mapped to the clades in phylogeny based on genetic and morphological characters, without requiring a precise match. This would reveal the evolutionary relationship of charismatic characters to the more widely distributed “pivotal” ones.

3. Clarifying non-accidental psychobiological relationships. C&T write that their analysis supports the hypothesis that “the brain architecture necessary for (a) processing three-dimensional, spatiotemporal properties of the visual world and (b) controlling object-oriented behaviors laid the neural groundwork for more advanced cognitive abilities.” This and other claims invoking necessity (and related modal concepts of possibility and counterfactuality) imply a move from character correlations – mappings of “CC” abilities to phylogeny – to non-accidental relationships between these characters. (They are not claiming that advanced cognitive abilities must be visually guided.) Non-accidental relationships play important roles in scientific explanation, inductive inference, and counterfactual support. It follows, for example, that an organism without a brain and high-resolution eyes couldn’t have these types of advanced abilities, and that computer models might reveal which evolvable but non-actual structures could underlie them and which counterfactual biological possibilities and biological impossibilities are possibilities in artificial agents.

However, identifying specific types of biological structures as necessary for a psychological ability is a direct assault on the dominant functionalist consensus by which psychological properties are functional roles that can be realized by any physically possible entity. This principle of medium (or substrate) independence does not restrict the realizers to what is biologically possible. The principle has been a boon to cognitive science. But biopsychological necessity claims expose the growing tension between this principle and the biological and behavioral research that enables and motivates phylogenetic approaches to psychology. For example, adherence to the principle threatens to impede our ability to distinguish cognitive homologies from homoplasies.

The bottom line: Phylogenetic approaches can help us identify non-accidental biopsychological relationships for “pivotal” and charismatic cognitive abilities alike (and everything in between). But understanding these relationships requires clarification of biological modality, as distinct from physical (and logical) modality.