In this speculative talk, I'm going to "think" adjacently with Stuart Kauffman's recent work on what he calls "the adjacent possible" in biological systems (Kauffman 2019). My aim is to articulate a way of thinking about the role of "environments" and behavior as the leading edge of evolutionary transitions (here: transitions in both individuality and inheritance, in a particular sense). Kauffman articulates an interesting thesis on what makes a living system: that it must be self-reproducing (in a particular sense) and carry out at least one "work cycle" (again in a particular sense). Kauffman muses that we lack mathematical theories to articulate what he considers the heart of the "problem" with the evolution of such systems -- because their operation changes their own configuration spaces, in a sense there can be no mathematical theory in the conventional sense of dynamical systems theory (which presupposes a fixed configuration space to get the math off the ground). I agree, this is a hard problem for a science of organized, living agents. I think there may be an adjacent "less hard" problem. Kauffman observes that for the kind of living organization he discusses "it takes work to make constraints and it takes constraints to make work." I speculate that the less hard problem can be formulated by considering the production of constraints through processes of ecological scaffolding. The problem is quite as open as Kauffman’s general problem of open “niches,” but it is less hard in the sense that there may be systematic ecological patterns of developmental scaffolding that allows us to study some highly limited problems of evolution into the adjacent eco-developmental possible. I speculate that these scaffolding interactions can lead to the development of configuration spaces, hence eco-developmental scaffolding introduces novelty into development from an environmental source. The talk will link this idea to what I have called “developmental reaction norms” in contrast to standard “ecological reaction norms” (Griesemer 2014). Differently put, step-wise nearby (adjacent) changes in configuration space can be studied by looking for systematic patterns of scaffolding constraints for these local phenomena of constraint production via scaffolding work, rather than by taking on Kauffman's completely general mathematical problem and challenge. I have no illusions that this will solve Kaufmann's hard problem, but there might be some clues to the kinds of mathematics we could be looking for and the kinds of phenomena that might present less-steep empirical challenges than developing a completely new mathematics for such completely general, open empirical problems as the origins of life or the evolution of the biosphere.

A popular account of evolutionary transitions in individuality (ETIs) postulates a crucial change in the nature of fitness during an ETI. Fitness at the collective level is supposedly “transferred” or “decoupled” during the process (Michod, 2005; Okasha, 2006). Recently, this view of ETIs has been challenged on the grounds that it may be better to focus directly on traits as opposed to fitness (Bourrat et al., 2021). In this paper, I will attempt to reconcile these two views by attending to their distinct conceptions of fitness. Following one account, fitness is considered a complex trait whose measurement involves summing over the totality of an entity’s phenotypic traits (Brandon, 1990, Bouchard & Rosenberg, 2004). Following the alternative account, fitness is the long-term reproductive output of an entity (Sober, 2001). If one adopts the former approach to fitness, I will argue that it becomes possible to understand how fitness changes during an ETI. When fitness is a complex trait composed of all the other traits of an entity, it is unsurprising that its nature changes during a transition. This follows from the fact that the selectively relevant traits of individuals often change during a transition. For instance, during the transitions from unicellular to multicellular organisms, trade-offs between different traits of the unicellular organism no longer apply when they become part of a collective. However, this does not correspond to a “transfer” of fitness. It is simply a change in the relationship between traits and their environment(s). Further, it seems that this change in the nature of fitness is not reconcilable with the competing view of fitness as the long-term number of descendants.

Evolutionary transitions in individuality (ETIs) are often conceptualized in a static rather than dynamical way. Abstractly, once an ETI is complete, the particles or lower-level entities (e.g., genes or cells) are regarded as the “bricks” constituting the “building” of the higher-level entities or collective (e.g., chromosomes or multicellular organisms). However, this static view underplays the dynamical nature of the collective—results of interactions between lower-level entities are only described in a phenomenological way. This is particularly detrimental when studying ETIs. We propose a different view in which both particles and collectives have their own developmental (internal) and ecological dynamics (external). One subtlety of this view is that the dynamics at both levels are intertwined in nested systems: a process described as the ecology of the cells is the development of the multicellular organism. Likewise, the development of the cells constrains their ecology and, ultimately, limits the space that can be explored by collectives. In this paper, we will focus on how regarding nested systems from the point of view of their life cycle, as a central object for the study of ETIs, can help clarify the problem of nested dynamics. First, we will show under what conditions a dynamical system can qualify as a life cycle by drawing from the Darwinian properties framework and, in particular, Godfrey-Smith’s (2009) Darwinian space. Second, we will show how focusing on life cycle rather than the entity at one point of the life cycle permits us to clarify the problems of entangled timescales and fuzzy boundaries, particularly in the context of the ecological scaffolding scenario of ETIs (Black et al., 2020; Doulcier et al., 2020). Third, we will contrast this view with some alternative formalizations of the ETIs. Finally, we will outline a research program detailing how this focus could be used to tackle outstanding questions in the field.

During an evolutionary transition in individuality (ETI), lower-level entities interact in such a way that they produce higher-level entities that become new units invoked in evolutionary explanation at this higher level (Michod, 2005; Okasha, 2006). In this paper, we will argue that to understand an ETI, it is crucial to first understand the type of ecological conditions under which the formation of higher-level entities can occur. For instance, from the perspective of a cell or, more abstractly, a particle, one will ask under what environmental conditions is it advantageous to become part of and possibly fully dependent on a larger entity such as a multicellular organism or, more abstractly, a collective? This leads to a view of ETIs in which being part of a larger entity is regarded as a potential strategy or phenotype from the point of view of the lower level. Starting from recent works in experimental and theoretical evolution (e.g., Black et al., 2020; Hammerschmidt et al., 2014, Bourrat et al., 2021), we will provide boundary conditions on the environment for ETIs to be possible. First, we will argue that the environment must be complex, where complexity is defined by the number of discrete states and an associated probability distribution, which can be connected to the notion of entropy in information theory. Second, we will argue that these states must have a sequential order that can be defined either spatially or temporally. Finally, we will argue that the difference between sequential states must be relatively smooth because large differences between sequential states would exceed the adaptive capacity of the entities undergoing the transition.