Author: Lynn Chiu (University of Bordeaux/CNRS)

Cover Page

Title: Decoupling, Commingling, and the Evolutionary Significance of Experiential Niche Construction

Author: Lynn Chiu (University of Bordeaux/CNRS)

Address: CNRS UMR 5164 – Université de Bordeaux, Site de Carreire, Zone Nord, Bâtiment 1B, 146, rue Léo Saignat, 33076 Bordeaux Cedex – France

Tel:+33 (0) 557 579 246, Fax:+33 (0) 557 571 472, Email:

Abbreviated Title: Experiential Niche Construction

Acknowledgement: I would like to thank Kevin Laland and Tobias Uller for the opportunity to develop this paper at the EES workshop at the KLI inKlosterneuberg, Austria. Special thanks to André Ariew, Denis Walsh, Rasmus GrønfeldtWinther, Sonia Sultan, John Odling-Smee, Kevin Laland, and Tobias Ullerfor valuable feedback and comments on the various versions. Very early versions of the paperhave been presented atISHPSSB Montpellier 2013, the KLI (2013), 2nd Toronto Annual Roundtable Meeting 2015 (organized by Denis Walsh), and members of the ERC-IDEM group (led by Thomas Pradeu).Scott Dyson provided invaluable content and copyediting consultation at the beginning of this paper, and the detailed comments from the three reviewers greatly improved the orientation of the paper.Thanks to Greg Dupuy for stylistic comments.This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme- grant agreement № 637647 - IDEM.

1. Introduction

Evolutionary theory is dominated by models, theories, strategies, and metaphors that appeal to the environment to explain adaptive fit, or “good design”(Endler, 1986; Godfrey-Smith, 1996, 2001a; Gould, 1977, 2002; Lewontin, 2000; Mayr, 1982; Walsh, 2015; Williams, 1966).Externalist explanatory strategies presume that adaptations map onto environmental features. They recommend looking for environmental “problems” that traits are supposed to “solve” or adaptive traits that are supposed to “fill” an environmental niche (like keys fitted to a lock). Richard Lewontinargued that these externalist metaphors are “bad biology”(Levins & Lewontin, 1985; Lewontin, 1982, 1983, 2000).Organisms are not “adapted” to environments. Instead, they“construct every aspect of their environment themselves. They are not the passive objects of external forces, but the creators and modulators of these forces” (Levins & Lewontin, 1985, p. 104).The better metaphor for evolution is thus “construction,” not adaptation.

Lewontin’sconstructivism hasinspired a large and expanding literature on the causal significance of “niche construction” on evolution[1]. However, most followers reject Lewontin’sradical ontology of the environment. According to Lewontin, organisms do not just alter the world they occupy, they also change how the world is experienced.The niche constructed by the organism is thus not entirely made up by the external world, but by the experiences of the organismas well. Many admirers of Lewontinfindit difficult to comprehend or operationalize“experienced niches” asexternal environments and causes of evolution(Brandon & Antonovics, 1995; Godfrey-Smith, 1996, 2001a; Odling-Smee, 1988; Odling-Smee, Laland, & Feldman, 2003; Sterelny & Griffiths, 1999). Theythus opt to identifyconstructed environmentsas the intrinsic properties of the external world relevant to the organism and focus onthe evolutionary significance of these constructedenvironments.

In this chapter, I make the case for the evolutionary significance of “experiential niche construction” (coined by Sultan 2015).I start by arguing that recent analyses that draw on theagency and plasticity of organisms (Walsh 2015, Sultan 2015)can address a major objection against experiential niche construction (Godfrey-Smith 1996, 2001a). I then propose a way experiential niche construction is evolutionarily significant. Theories of niche constructionthat leave out the experiential variety tend to maintain an externalist characterization of natural selection and argue that niche construction feeds into the environmental causes of natural selection. Natural selection, however, does not adapta population to its environment when different organisms of a population construct and experience different environments.Instead, in these scenarios, the causes of selection are spread across varying organisms and their varying constructed environments. I argue that experiential niche constructionhelps maintain the spreadof selective causes across organism and environment interactions. It thus creates the conditions fora kind of natural selection that is not “externalist.”

My approach is pluralistic. Sometimes,natural selection can be heuristically approximated as environmental selection and niche construction contributes to selection’s environmental sources. In these scenarios, organisms and environments are “decoupled” causes.Other times, however, natural selection cannot be heuristically treated as environmental selection. This occurs when organisms and environments “commingle” and niche construction constitutes natural selection. I propose a decoupling/commingling framework that specifies when it is and is not appropriate to heuristically assume that natural selection explanations are externalist.

In the following, I will refer to the external world surrounding the organism, characterized by its intrinsic properties, as the external or physical world, environment, or surround. When I am referring to the environment experienced by the organism (which will be determined in part by the properties of the organism), I will qualify the world, environment, or surround with terms such as “experienced” or “experiential.”

2.Experiential Niche Construction and Discontents

According to Lewontin, mainstream evolutionary theory assumes that the environment presents well-defined problemsfor organisms to solve.

“The word ‘adaptation’ reflects this point of view, implying that the organism is molded and shaped to fit into a preexistent niche, given by the autonomous forces of the environment, just as a key is cut and filed to fit into a lock.” (Levins and Lewontin 1985, p. 98)

He argues that there are two problems with this problem-solution metaphor. The conceptual problem is that organisms do not “fit into a preexistent niche” as the nichesof a species come to existthroughinteractions between organisms and their environments. It is conceptually impossible, then, for a niche to preexist and select organisms. The empiricalproblem is that niches are not “given by the autonomous forces of the environment” as they are instead determined by the biology of the organism.Organisms create niches by determining what’s relevant, altering properties of the world, transducing external signals (into different types of signals),ortransforming environmental patterns (into different types of patterns) (Levins & Lewontin, 1985, pp. 98–106; Lewontin, 2000, pp. 55–68).Therefore, the properties of environments emerge from interactions with organisms.

Since organisms also gain properties through their interactions with the environment, Lewontin concludes that organisms and environments are not decoupled causes and effects with pre-existing intrinsic properties. The inter-relation between organism and environment that induces their properties is dynamic and dialectical. The properties of organism and environment emerge through interactions that also propel their future change.

Consider the way plant engineers attempt to improve crop yield by designing (through artificial selection or genetic engineering) leaf phenotypesoptimizedto a measured microenviroment, e.g., the temperature, light exposure, humidity, oxygen and carbon dioxide concentration around the plant. The problem is that the newly selected leaf morphologies tend to alter the humidity, light, carbon dioxide, etc., distributionand create a different, less optimal microenvironment. The plant engineers can intervene again, but only to have the plants change the environment once more. “The plant engineers are chasing not only a moving target but a target whose motion is impelled by their own activities,” states Lewontin, “this process is a model for a more realistic understanding of evolution by natural selection" (Lewontin, 2001, p. 57).In the wild, plants are constantly changing the environments as they develop and evolve in response to them, which in turn results in further change to their environments.

Lewontin’s Mediational Niche Construction

Lewontin’sexamples of niche construction fall under two main categories: physical and experiential niche construction.

Physical Niche Construction(also known as perturbational niche construction or habitat construction) is the causal manipulation of the external world by the organism, changing the environment’s intrinsic properties.

Experiential Niche Construction is changes in the environment experienced by the organism without changes to the intrinsic properties of the external world. There are two types of experiential niche construction:

Relocational Niche Construction(also habitat choice) determines which intrinsic properties of the external world surround an individual.

Mediational Niche Construction determines the relevance, impact, and significance of the external world for the organism. It determines howthe intrinsic properties of the environment is experienced by the organism.

The core question of this paper is whether mediational niche construction is a type of niche construction, and if so, whether it has any evolutionary significance qua niche construction. Even though relocational and mediational niche construction are both cases of experiential niche construction(they don’t change the intrinsic properties of the environment), I single out mediational niche construction as the central concern. Relocational and physical niche construction both determine which intrinsic properties surround an organism (the first by choosing an environment, the second by alteringan environment). However, mediational niche construction does not alter nor determine which intrinsic properties are around an organism. Instead, it changes the way an organism experiences them.[2]Is this niche construction?

What I coin “mediational niche construction” is currently an underexplored category[3]. Lewontin raises several examples of mediational niche construction. One exampleis the transduction of temperature into the biochemical signals of organisms such that a one-degree drop in the outside world is experienced as a smaller difference for one organism but a greater difference for another. Another example is the way an organism perceives its environment as resource-rich or poor. The perceived scarcity of an environment is relative to the organism’s level of fat storage. Yet another example is when the physiology of an organism incorporates rates of change of environmental factorsinto its experienced environment, thus perceiving and reacting to sudden changes instead of absolute levels.

Mediational niche construction occurs because the organism standsbetween itself and the world. Through this type of niche construction, states Lewontin, “the common external phenomena of the physical and biotic world pass through a transforming filter created by the peculiar biology of each species, and it is the output of this transformation that reaches the organism and is relevant to it” (Lewontin 2000, p. 64). It is as if the organism is residing insidea Plato’s cave of its own making, “determined by the shadows on the wall, passed through a transforming medium of its own creation” (Ibid). An organism in its self-created “bubble” is still affected by the physical world, but the effects of the world are distorted and transformed by organismal activities and physiology.

Mediational niche construction was dismissed by philosophers (Godfrey-Smith, 1996, 2001a) and left out by proponents of NCT (Odling-Smee, Laland, & Feldman, 1996; Odling-Smee et al., 2003), but kept aliveby close allies of Lewontin(Sultan, 2015; Walsh, 2015). In the following, I consider Godfrey-Smith’s objections and argue that new theories of mediational niche construction can addressmost of the concerns.

Godfrey-Smithagainst Mediational Niche Construction

It is not obvious thatmediational niche constructionis a type of niche construction. How caninternal changes within an organism count as changesto theenvironment?Internal changes are usually considered phenotypesunder selection, not determinates of selective pressures. That is why some argue that relocational niche construction are actually phenotypesfor habitat choice while mediational niche construction is a phenotypic response to environmental pressures (Brandon 1990; Godfrey-Smith 1996, 2001).

Godfrey-Smith suggests that mediational niche constructionas niche construction might make sense if we adoptone of Lewontin’s dialectical principles(found in the conclusion chapter of The Dialectical Biologist). Lewontin and colleague Richard Levins have long advocated for a dialectical biology against the “Cartesian” decoupling of causes and effects, parts and wholes, and insides and outsides (Levins & Lewontin, 1985; Lewontin & Levins, 2007). Organism and environment do not exist independently of each other as causes and effects, they argue, but “interpenetrate,” or “commingle.”Godfrey-Smith suggests that thedialectical principle that parts do not exist independently of each other can help make a case for mediational niche construction. If we assume that organisms and environments are two parts of a whole and parts do not independently exist,then a change to any part wouldlogically(not causally) entail a change in the other parts. The part-whole principle thus explains how an internal change to organisms wouldis a change in their environments.

Applying this dialectical principleto organisms and environments, however, seems to create an undesirable mix ofanti-realism andintractableholism. The “environment” or “niche” of an organism is not the objectively measurable environment, but something (in part) subjectively constructed by the organism. Furthermore, any change to the organism, the environment, or their relationwill count as niche construction, thus trivializing the concept. Anall-inclusivenotion of niche constructionfails to capturethe complex and varied relations between organisms and environments that are important for empirical study.

Godfrey-Smith proposes that it is fruitful to just acknowledgephysical niche construction as an evolutionarily relevant processand treatmediational niche construction as mere traits undergoing selection. To illustrate, consider two species of bacteria in an environment of toxic molecules. One evolves a different internal physiology such that the chemical is no longer toxic. For instance, the organisms may no longer have the receptors or signalling pathways that react to the chemical in a self-destructive way. The other evolves a mechanism that excretes toxin-degrading enzymes. These are two distinct evolutionary responses to an environmental challenge. The first is a selected internal accommodation to the environment, whereas the second is a selected trait that also alters the environment. The second trait changes the environment for future generations to come whereas the descendents of the first are still living with the toxin. Yet for Lewontin, both count as niche constructionasthe organisms all end up experiencing a non-toxic environment. This coarse-grained dialectical perspective glosses over important differences and outcomesof distinctevolutionary responses.

There are many advantages toa narrower scope of niche construction. First, it gives us a sharperboundary of what counts as niche construction and what does not. Not any change to organism or environment is niche construction. It is clear that the line is drawn at whether organisms alter the intrinsic properties of the external world. Second, we retain a commonsensical notion of the environment as an objective, physicalphenomenon instead of constructed experiences.Third, it distinguishes between mere adaptations from those that also alter the external, selective environment.

There is a final reason for discounting mediational niche construction. This phenomenon is not even neededto reject externalism. Godfrey-Smith argues that there are two types of externalist explanations: symmetric and asymmetric. The problem with adaptationist thinking is asymmetric externalism, that is, theposition that while the environment accounts for organismal evolution, organisms cannot account for changes in the selective environment. Externalism can also be of the symmetric kind, which permits the evolutionary effects of niche construction. As physical niche constructionalone rejects asymmetric externalism, leaving out experiential niche construction does notdiminish the Lewontinian challenge against what’s wrong with externalist thinking.

In sum, contrary to Lewontin’s call to replace adaptation with construction, Godfrey-Smithinsists that there is acomplementary coexistence between the two. “Rather than a replacement, there should be a supplementation. Both adaptation and construction are real relationships that organisms have, in particular instances, to environmental conditions” (Godfrey-Smith 2001, p. 263).Physical niche construction provides “constructivist” explanations that explain the intrinsic properties of the environment by properties of organisms, whereas natural selection supplies “externalist” explanations that explain the properties of organisms in terms of the intrinsic properties of the environment. Together, they give a fuller picture of how evolution works.

3. New Supportfor Mediational Niche Construction

Godfrey-Smith’s major concern was that mediational niche construction implies holistic intangibilityand anti-realism about the environment. I argue that recent developments, in particular,DenisWalsh’s affordance theory and Sonia Sultan’s theory of plastic cue and response systems, provide rich and testable characterizations of mediational niche construction that address almost each of the challenges. They offer clear-cut distinctions between the various ways affordances or experienced environments can change and explain why experienced environments are constructed environments.

Walsh’s Affordance Framework

Walsh (2012, 2014, 2015)argues that adaptive traits are not evolutionary responses to the external world per se, but responses to opportunities in the environmentthat appear to the organismas opportunities for action. Organisms, as purposeful agents, perceive and respond to opportunitiesthat appearbecause of what the organismcan do and what it aims to do[4].

An example from Walsh can help clarify how opportunitiesdepend on the goals and capacities of organisms. A stick does not become a tool just because sticks can solve a problem and there are sticks lying around. The problem-solving agent needs to have the capacity to perceive the stickasa usable tool that can potentially solve a problem, that is, to experience it as affording a particular type of action that can fulfill a goal. For a hominoid, opposable thumbs and “precision grip” are necessary for the manipulation of hand-held tools such as sticks. Without these hand structures, objects in the environment would not seem “grip-able” or “grasp-able” for use.Yet the evolution of precision grip is not the result of direct selection for tool use, but an evolutionary by-product of bipedalism. The structural changes in feet that enabled hominoids to stand up and run also altered the structures of hands. These new hand structures opened new possibilities: objects previously inconceivable as graspable are now within “reach.”The evolution of precision-grip is a nice example wherebyorganisms“can make a change in its own form, without affecting the environment, which in turn alters the affordances provided to the organism” (Walsh, 2015, p. 181-182).

According to Walsh, environmental opportunitiesare affordances[5]that appear only when there is a purposeful agent. An affordance exists when the environment is experienced by the agent as having the sorts of properties that can either enable or disrupt it from achieving its goals. An agent responds to these affordances by interacting with them in ways that help attain its goals, either through the exploitation of facilitators or the elimination of obstructions. Affordances are emergent properties of a purposeful agent and its environment. Agents and environment bothconstitute[6]an agent’s affordances, that is, their properties and relationsunderlie what an environment can afford to an agent.