Organism-Environment Codetermination: the Biological Roots of Enactivism

Home , Grubbing, Ron McClamrock, Tim Lewens

Organism-Environment Codetermination: The Biological Roots of Enactivism

A dissertation submitted to the Graduate School of the University of Cincinnati in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

in the Department of Philosophy of the College of Arts and Sciences by

Amanda B Corris June 2020 M.A. University of Cincinnati M.A. University of Sussex

Committee Chairs: A. Chemero, Ph.D., A. Potochnik, Ph.D.

Abstract

Traditional approaches to cognition take it to be a fundamentally brain-based phenomenon. On

this view, the brain functions as a type of information processing center, making cognition a

matter of computational processing and representational symbol manipulation. In contrast,

embodied, enactive approaches to cognition emphasize the role of the body in cognition and non-

representational perception-action dynamics. While the embodied and enactive paradigm has

been gaining in popularity, it has yet to adequately engage with complementary approaches in biology that aim to define the organizational structure of organismal life. In this dissertation, I argue that an enactive approach to cognition in nature can be enriched by incorporating the central tenets of both developmental systems theory and extended interpretations of evolutionary biology. This framework, which I term biological enactivism, defines organisms as cognizing systems structured by both their internal dynamics and their dynamic relations with environmental features corresponding to their sensorimotor capacities, developed as a result of their coupled interactions with their environments over both developmental and, on a population scale, evolutionary time.

Amanda B Corris

“Go, thou sluggard, learn arts and industry from the bee, and from the ant!

Go, proud reasoner, and call the worm thy sister!”

Erasmus Darwin, Zoonomia

iii

Acknowledgments

I have been met with surprisingly little resistance with regard to my philosophical pursuits, and for that I am thankful. Here are the people who are responsible for this, and of whom I am wholly appreciative. Ron McClamrock instilled in me the confidence needed to stay afloat in the raging sea that is academic philosophy. Thank you, Ron. Tony Chemero made me feel like I am famous since the day I met him. I am not sure why, especially considering he is the famous one. But it is a very nice ego boost. Angela Potochnik, nearly singlehandedly, made me into the professional philosopher I am (slowly becoming) today. Any errors I have made along the way were due to me not asking her for advice first. Tom Polger inspired me to work hard now to save myself from extensive scrutiny later- in a good way, of course! Nate Morehouse welcomed me into a stimulating, resourceful, and happy environment, and deeply enriched my understanding of biology along the way (especially the weird stuff). Bob Richardson made me love philosophy even more. Rob Skipper drove me to do philosophy well. The philosophy faculty at UC are excellent people. The Morehouse Lab regularly entertained my ‘armchair biology’, and showed me that it is possible to be an academic and have fun at the same time. Imagine that! My family never gave me the “So when are you going to get a real job?”, which is appreciated. Brad and Lisa put up with endless musings on what could possibly be going on in the mind of every animal (and sometimes plant) that I have come across. Thank you for your patience, friendship, and love. And I think it is only natural to acknowledge the forms of life that have inspired me along the way, for even the “minirobots” enact a world of their very own.

A portion of Chapter 4: Making Sense of the Environment will be published as Corris A. (2020). Defining the Environment in Organism–Environment Systems. Frontiers in Psychology, 11:1285. doi: 10.3389/fpsyg.2020.01285.

This research was supported in part by the Charles Phelps Taft Research Center.

Table of Contents

Abstract ...... i

Acknowledgments ...... iv

Table of Contents ...... vi

List of Figures ...... ix

Chapter 1: Introduction ...... 1

1.1 The path that leads here ...... 1

1.2 A sketch of the project ...... 3

Chapter 2: A Biological Approach to Cognition ...... 6

Introduction ...... 6

2.1 Life and mind ...... 7

2.1.1 Bringing forth a world through living ...... 8

2.1.2 Varieties of enactivism ...... 10

2.2 Development and evolution on the enactive view ...... 18

2.2.1 Evolution as natural drift ...... 18

2.2.2 Enactive evolution ...... 21

2.3 Enaction in a developmental system ...... 24

2.3.1 Biological enactivism: a synthesis and expansion ...... 26

2.3.2 Integrating DST with the enactive approach ...... 29

Chapter 3: The Role of Cognition in Evolution ...... 32

Introduction ...... 32

3.1 An alternative synthesis ...... 36

3.1.1 Extending evolutionary theory ...... 39

3.1.2 Interdisciplinary opportunities from the EES ...... 43

3.2 Parallels in cognitive science ...... 46

3.2.1 The epigenesis of behavior ...... 46

3.2.2 Causal directionality and information flow ...... 48

3.2.3 Shared commitments ...... 52

3.3 Cognition as evolutionary ‘bootstrapping’ ...... 57

3.3.1 Inheritance beyond the gene ...... 59

3.3.2 Plasticity and the shaping of evolutionary trajectories ...... 64

3.3.3 Speeding up evolution...... 69

Chapter 4: The Organization of Cognizing Systems ...... 74

Introduction ...... 74

4.1 The organization of living systems ...... 76

4.1.1 Self-organization as autopoiesis ...... 78

4.1.2 Order from chaos: autocatalysis and dissipative structures ...... 80

4.1.3 Self-organization over ontogenetic time ...... 83

4.2 Plasticity as structural flexibility ...... 88

4.2.1 An array of possibilities ...... 89

4.2.2 Relations between traits: integration and modularity ...... 93

4.2.3 Controls and constraints ...... 97

4.3 Acting on flexibility ...... 100

4.3.1 Modulating coupling with the environment ...... 102

vii

4.3.2 Assessment of environmental conditions...... 105

4.3.3 Structural modification as cognitive strategy ...... 108

Chapter 5: Making Sense of the Environment ...... 113

Introduction ...... 113

5.1 Making meaning through sense-making ...... 115

5.1.1 Intrinsic purposiveness in living systems ...... 117

5.1.2 Sense-making as projective teleology ...... 119

5.2 Constructing an enacted world ...... 122

5.2.1 Coherence and reliable reconstruction of the life cycle ...... 123

5.2.2 Specifying the cognitive domain ...... 127

5.3 The environment as developmental niche ...... 130

5.3.1 Multiple senses of environment ...... 131

5.3.2 The developmental niche ...... 135

Chapter 6: Conclusion ...... 141

6.1 Summary ...... 141

6.2 Future directions ...... 144

Bibliography ...... 153

viii

List of Figures

Figure 1. Gottlieb's diagram of the bidirectionality hypothesis. Processes at different scales of biological organization form dynamic interactions, the result of which shapes an organism's developmental trajectory...... 49

Figure 2. Laland et al.’s (2011) depiction of the causal pathways involved in the evolution and development of traits. Reciprocal causation between components is emphasized, illustrating feedback mechanisms from the environment to developmental processes...... 51

Figure 3. Waddington's developmental landscape diagram, with modifications from Noble (2015) to illustrate how development can follow alternative favored routes, with different phenotypes resulting from the same genotype...... 66

Figure 4. Waddington's epigenetic landscape, representing complex interactions between genes (pegs at the bottom) and environmental conditions...... 101

Figure 5. Bird song as an example of an integrative approach to Tinbergen’s ‘four questions’. (A) details bird-song learning mechanisms. (B) details current utility of song function. (C) details developmental mechanisms for singing capacity. (D) details the evolutionary history of songbird populations. (From Bateson and Laland 2013) ...... 147

Chapter 1: Introduction

1.1 The path that leads here

Natural selection is not a consequence of how well the organism solves a set of fixed problems posed by the environment; on the contrary, the environment and the organism actively codetermine each other. The internal and the external factors, genes and environment, act upon each other through the medium of the organism. Just as the organism is the nexus of internal and external factors, it is also the locus of their interaction. The organism cannot be regarded as simply the passive object of autonomous internal and external forces; it is also the subject of its own evolution. (Levins and Lewontin 1985, 89)

Although this dissertation is ultimately concerned with the nature of cognizing systems, it is the above quote that serves as a guiding principle for the work being done here. It forms a link between two fields – biology and cognitive science – that I hope to draw closer together. It does

so by offering a criticism that while directed at one field in fact applies to received views in both

fields. The criticism is that organisms are not passive receivers of environmental input, as

conventional theory claims, but instead form a dynamic, co-actional relationship with their

environments. Although the above quote offers this criticism in the context of evolutionary

biology, it is complementary to one made in cognitive science, by theoretical frameworks

intended to serve as alternatives to conventional theory in that field. These frameworks, namely

those found in 4E (embodied, embedded, enactive, and extended) approaches to cognition, also

reject the view of the perceiving organism as a passive receiver of environmental information in

favor of a view of the organism as actively shaping its dynamic interactions with its

environment. This reciprocal, dynamic view of organisms coupled with their environments is

thus one that has relevance for both biology and cognitive science.

The primary goal of this dissertation is to offer a framework for understanding cognition

as a biological phenomenon – one that emerges as a result of the dynamic interaction between

organisms and their environments. I do this by working through theory from both biology and

cognitive science to show how organism-environment interaction is central to understanding cognition. From biology, I highlight the strengths of developmental systems theory (e.g., Oyama

2000 [first edition 1985], Oyama, Griffiths and Gray 2001) as a way of understanding organismal life, and the related calls for extension of standard evolutionary theory (e.g., Laland

et al. 2015; Müller 2017; Pigliucci 2007) that have recently been gaining traction in evolutionary

biology. From cognitive science, I focus on the enactive approach to cognition found in work by

Francisco Varela, Evan Thompson, and Ezequiel di Paolo (e.g., Varela, Thompson and Rosch

1991; Thompson 2007; Di Paolo, Buhrmann, and Barandiaran 2017). These views all have much

in common, and their synthesis, I suggest, has significant implications for both fields. By bridging the gap between them, my hope is to set the foundations for a novel framework for understanding biological cognition. Interpreting new findings using this approach could potentially result in the solving of some old problems in cognitive studies (see Chapter 6, Section

2: “Future directions”) as well as suggest innovative trajectories for future research (see Chapter

3, Section 3: “Cognition as evolutionary ‘bootstrapping’”). I term this framework biological enactivism to signal that it is a branching off of the enactive approach, one that traverses deeper into the biological realm to incorporate some new theoretical and conceptual claims, and thus one that deviates in some key foci.

The arguments put forth in the following chapters are intended to primarily contribute to

this larger positive project; it is a secondary goal for the arguments to serve as reevaluations for extant theory. This is not to suggest there are no opponents, nor that I do not take issue with

received views in both biology and cognitive science. Rather, the entire body of work should be

taken as an alternative standpoint to many issues in both fields, and I have tried to highlight the

shortcomings of mainstream claims as they arise in the discussion. Both developmental systems

theory and the enactive approach to cognition were developed in response to perceived faults

with extant views in their respective fields, and this work should be seen as carrying on that

tradition. Overall, my hope is for the project to serve as a way of bridging the gap between two

theories that have much in common, and to offer a way forward for thinking about biological

cognition as a complex, dynamic phenomenon that requires integration across numerous cognate

fields. This requires rejecting many ‘textbook’ views, either in part or wholesale, along the way, from computationalism to standard evolutionary theory. In this way, this project is radical – it suggests a significant overhaul of how we think about life and the mind, and the relationship between them.

1.2 A sketch of the project

The dissertation begins with an account of the origins of the enactive approach and the varieties

of enactivism that have since emerged from that view. I argue that what is termed ‘autopoietic

enactivism’ offers the best way forward for understanding cognition in nature and fits most

naturally with claims from developmental systems theory. Autopoietic enactivism specifies

organisms as self-organizing (autopoietic), autonomous systems, in that they are endowed with

the ability to maintain their internal dynamics through self-regulation and accordingly sustain

themselves as a unity. They also engage in sense-making, in that they undergo positively- or negatively-valenced interactions with their environment that in turn require regulatory activity for the sake of their viability. I argue that an expansion of the theory’s framework to incorporate concepts from developmental systems theory results in a richer approach to the study of cognition, namely by presenting a detailed account of the structural coupling relation between organisms and their environments.

After providing a call for the synthesis of autopoietic enactivism and developmental systems theory, I investigate the related call for an Extended Evolutionary Synthesis (EES) in

biology with the goal of showing how insights from the EES research program are relevant for

the study of the role of cognition in evolution. The EES’s endorsement of the evolutionary

impact of extra-genetic inheritance and reciprocal causation between organisms and their

environments (namely via niche construction) creates a space to discuss cognitive structures,

such as learning and experience, as elements that can impact a species’ evolutionary trajectory. I

argue that cognition can function as a driver of evolution through mechanisms of non-genetic

inheritance, thereby expanding upon Jablonka and Lamb (2005). I draw on research in both

molecular genetics and behavioral ecology to support my argument for the incorporation of

enactive cognitive science into the EES research program’s broader scope.

The second half of the dissertation is comprised of a detailed analysis of the concepts

central to the biological enactive approach that frames the dissertation. According to biological

enactivism, organisms are self-organizing, operationally closed, plastic systems embedded within

a larger developmental system. In virtue of this organizational makeup, they have the ability to

actively assess and modulate their coupling with their environment. In Chapter 4, I argue that

this organizational makeup serves as the precondition for cognition. ‘Sense-making’ is a key

enactive concept that refers to an organism’s ability to cognitively engage with the world in such a way that creates significance for that organism. On many enactivist interpretations, sense- making simply is cognition. In Chapter 5, I argue that reformulating sense-making as a coherence relation between organisms and their environments results in a more natural fit for the integration of the enactive approach with developmental systems theory. Additionally, this reformulation has value for biology more generally, as it can shed light on how to conceptualize an organism’s enacted world as its developmental niche.

Both the enactive approach to cognition and shared perspectives in developmental systems theory and extended interpretations of evolutionary theory have much to offer their respective fields. This dissertation attempts to bridge the gap between these areas in the hopes of providing a refined and comprehensive framework for understanding the phenomenon of cognition as it is found in nature. The overarching goal is to motivate a picture of biological cognition that emphasizes the strongly dynamic relationship between embodied organisms and the complex environments in which they live.

Chapter 2: A Biological Approach to Cognition

Introduction

What kinds of things are cognizing things? In virtue of what can certain things be said to be cognizers? One way forward for addressing these fundamental questions about cognition is to investigate the continuity between life and cognition. The kinds of things we typically take to enjoy cognitive capacities are, in fact, living things, so grounding the study of cognition in a biological framework may help to shed light on cognition as a biological phenomenon.

Taking this kind of approach means looking not just at individual aspects of cognition, such as neurological functioning, but cognition on a global scale, as a process that a living being actively engages in over time. This view of cognition takes into consideration contingent relations such as historicity, feedback mechanisms, and inter- and intra-systemic interaction. In other words, it suggests a holistic methodology to the study of cognition, one that requires insights across multiple levels of analysis.

One such approach is the enactive view of cognition. According to this view, cognition emerges as a result of dynamic interactions between a living being and its environmental milieu.

A living being can be said to ‘enact’ or bring forth a world through these dynamic interactions, and it is this enaction that is central to cognition. Therefore, on the enactive approach, understanding cognition requires investigating the relationship between a living being and its environment, how that relationship is formed, and what differentiates a living being from its environment.

In this chapter, I will discuss the enactive approach to cognition as a biologically based theory of cognition and as one that is offered as an alternative to traditional ways of thinking about cognition. In Section 2.1, I provide a historical overview of enactive theorizing found in works by Maturana and Varela (1987) before describing how this theorizing is expanded upon in the enactive approach to cognition found in Varela, Thompson and Rosch (1991) and Thompson

(2007). I then provide a brief account of enactivist theories that have evolved from these prior views. My primary goal in this chapter is to draw attention to the biological claims at the heart of the enactive approach for the sake of taking these lines of argument further by incorporating complementary claims from developmental systems theory, a perspective in developmental biology, as well as extensions into other biological fields including evolutionary biology. With that goal in mind, in Section 2.2 I look at enactive claims about development and evolution, and I end in Section 2.3 by offering a first pass on the integration of developmental systems theory and related views with the enactive approach, the details of which I take up in the remainder of the dissertation.

2.1 Life and mind

Work by biologists Humberto Maturana and Fransisco Varela laid the theoretical foundation for the enactive approach to cognition. Their theory of cognition centered around the organizational pattern of living beings. On their view, living beings are cognitive beings in virtue of this organizational pattern, and so an account of biological cognition must start with a description and explanation of the way that living beings are organized. As each enactive trajectory that subsequently developed out of Maturana and Varela’s original approach shares some

fundamental characteristics with it, it will be helpful to first outline the basic tenets of that

theory.

2.1.1 Bringing forth a world through living

Maturana and Varela claim, in their 1987 book The Tree of Knowledge, that what defines living

beings is their continual self-producing nature (Maturana and Varela 1987, 43); the

organizational pattern that they exhibit is one of self-production. They refer to this as

‘autopoietic’ (literally, self-producing) organization. A unity is autopoietic when its processes

form a network of continual, dynamic interactions and in doing so produces its own components.

Cell metabolism serves as an example of autopoietic organization. The chemical transformations

that take place during cell metabolism constitute a dynamic, self-referential network of interactions that, in their activity, create a boundary for that network. An autopoietic unity is thus one that undergoes dynamic, self-referential interactions (such as metabolism) while creating a boundary (such as the cell membrane) (46). The same organization is meant to apply to living beings beyond cells. Autopoiesis is central to an enactive understanding of life as it specifies the kind of organizational pattern that living beings exhibit.

In addition to being autopoietic unities, Maturana and Varela describe living beings as being autonomous, or self-governing, unities. They are autonomous in virtue of their autopoietic organization. They produce themselves, and so are unities that exist predominantly as a result of their own internal processes. Maturana and Varela state that autopoiesis is the mechanism that makes living beings autonomous systems (48). We can thus think of living beings as autopoietic, autonomous unities.

Living beings are also historical beings. Multicellular organisms and individual cells

alike originate from an existing entity producing another entity of the same class – “it gives

origin to another unity that an observer can recognize as possessing the same organization as the

original one” (57). Reproduction presupposes an entity that does the reproducing, so in the act of

reproduction a history is enacted. Entities or unities originating from other reproducing unities

inherit elements from that unity; Maturana and Varela state that “[t]hose aspects of the initial

structure of the new unity which we evaluate as identical to the original unity are called heredity”

(68, original emphasis). A unity’s heredity is an important part of its history, because it creates a

link between it and the unity which gave rise to it. Reproductive variation arises when aspects of

the initial structure of the new unity are different from those of the original unity.

Structural change in living things also occurs in the unity “either as a change triggered by

interactions coming from the environment in which it exists or as a result of its internal

dynamics” (74). A coupling relation arises between autopoietic unities and their environments

with recurrent interactions occurring between those two systems. These recurrent interactions

can bring about structural changes over time, resulting in a “history of mutual congruent

structural changes” (75). The notion of structural coupling captures this relation and the shared

history that results from it. Organisms and their environments are structurally coupled with each

other and thus enjoy a shared history.

Structural coupling is an important aspect to the enactive picture because it is relevant to an understanding of both organismal change over time and the foundations of cognition. Changes in structural couplings result in what Maturana and Varela refer to as ‘natural drift’. As will be discussed in Section 2.1 below, on an enactive view, we can conceive of evolution as being a

process of natural drift. Structural coupling is thus relevant to understanding the evolution and

diversity of life. It is on this picture of living beings structurally coupled with their environments

that emergent enactive frameworks build. In the next section, I will discuss three varieties of

enactivism, highlighting the ways in which their explanatory tools and foci differ from that of the

original approach put forth by Maturana and Varela.

2.1.2 Varieties of enactivism

The term ‘enactivism’ refers to several closely related theories that have come out of the enactive

approach to cognition. Broadly speaking, enactive views are those that conceive of cognition as a

matter of organism-environment dynamics. Also central is the claim that the sort of world an

organism enacts is dependent upon its sensorimotor capacities. The term ‘enaction’, first

introduced in The Embodied Mind, refers more explicitly to the construction of a world of

relevance for an organism – a ‘bringing forth’ of a world for that organism on the basis of embodied action. For example, the enacted world of an ant corresponds to the ant’s physiological, sensorimotor makeup – certain environmental features will be relevant to it that would be irrelevant for an organism with a different sensorimotor makeup.

Enaction is linked up with an organism’s history of structural coupling with its environment. Certain sensorimotor structures arise in an organism due to recurrent interactions with environment features that correspond to those structures. A honeybee’s capacity for perceiving ultraviolet light is coupled with the behavior of seeking out floral structures that exploit that capacity. For Varela, Thompson and Rosch, embodied, perceptually guided action is central to the enactive approach. They hold that there are two central points to the enactive approach: that “(1) perception consists of perceptually guided action and (2) cognitive structures

emerge from the recurrent sensorimotor patterns that enable action to be perceptually guided”

(Varela, Thompson and Rosch 1991, 173). On this view, we can begin to understand cognition by looking at perception-action dynamics arising out of the recurrent interactions in structurally coupled systems (i.e., organism-environment systems).

On the received view in cognitive science, cognition is a matter of brain-based information processing, with the brain responsible for creating representations of the external world. The enactive view rejects this understanding of cognition in favor of a theory of cognition that emphasizes the dynamic relationship between perception and action as well as the dynamic relationship between an organism and its environment. Cognition should be understood not as representational information processing but rather as enaction. As Varela, Thompson and Rosch suggest in defining their approach, “[w]e propose as a name the term enactive to emphasize the growing conviction that cognition is not the representation of a pregiven world by a pregiven mind but is rather the enactment of a world and a mind on the basis of a history of the variety of actions that a being in the world performs” (9). The enactive view is thus offered as an alternative to representational, computational perspectives on cognition.

Where Maturana and Varela’s early work begins with biological claims and works its way up to claims about cognition and human understanding, Varela, Thompson and Rosch’s The

Embodied Mind (1991) is primarily grounded in contemporary cognitive science. Their goal is to offer the enactive approach as an alternative to traditional paradigms in cognitive science, namely cognitivism and computationalism. They provide a synthesis of ideas from phenomenological philosophers such as Merleau-Ponty and Heidegger with enactive theorizing and Buddhist insights. Thus The Embodied Mind is an extension of some of the arguments made

in The Tree of Knowledge to engage more directly with contemporary cognitive science and

phenomenological approaches to the study of the mind.

Thompson’s Mind in Life (2007) argues for the continuity between life and mind by

furthering the synthesis made in The Embodied Mind between enactive biology and

phenomenology. On Thompson’s view, there is a deep continuity between life and mind due to

the fact that “[l]ife and mind share a core set of formal or organizational properties, and the formal or organizational properties distinctive of mind are an enriched version of those fundamental to life” (Thompson 2007, ix). According to Thompson, a framework that addresses

this continuity can serve as a way of bridging the explanatory gap between life and mind, so it is

a worthwhile project for the sake of naturalizing the mind and conscious experience. As in The

Embodied Mind, Mind in Life emphasizes the role of the body in perception and looks to

perception-action dynamics for an understanding of cognition.

The enactive framework at the core of The Embodied Mind and Mind in Life has been

referred to as ‘autopoietic’ enactivism, though it is more commonly referred to as simply the enactive approach to cognition. Several themes are carried through this framework that draw on as well as expand Maturana and Varela’s original approach. Autonomous, autopoietic organization features prominently, but additional concepts have been incorporated. Autopoietic enactivism offers an account of the organization of living systems in virtue of their features that make them cognizing systems. Organisms are structurally organized such that they are

“internally self-constructive in such a way as to regulate actively their interactions with their environments” (Thompson and Stapleton 2009, 24). In other words, organisms are endowed with the ability to maintain their internal dynamics through self-regulation. Because the internal, self-

regulatory dynamics of the autonomous system are sufficient for its persisting as a unity, it can

be said to be operationally closed. Importantly, as Thompson and Stapleton point out,

“operational closure does not imply that conditions not belonging to the system cannot also be

necessary” (24). Autonomous systems are thermodynamically open, in that they undergo

processes to regulate the flow of energy both between them (from the environment into the

system) and within them (as regulatory processes internal to the system). The properties of

autonomy and operational closure can sometimes define a spatial boundary to a system as well1.

Living systems are autonomous systems that are structured by their own internal, operationally

closed regulatory dynamics as well as their thermodynamically open regulatory dynamics with

the environment.

Organisms engage in energy transfer from environment, but they do not do so entirely

passively. Environmental features have a valence for individual organisms. To use Jakob von

Uexküll’s example, a tick stationed on top of a blade of grass is spurred into action by a specific

environmental feature – a passing mammal’s butyric acid. Organisms engage not in passive

behavioral repertoires but in positively- or negatively-valenced interactions with the environment. This process is referred to as sense-making: it is “behavior or conduct in relation to environmental significance or valence, which the organism itself enacts or brings forth on the basis of its autonomy” (Thompson and Stapleton 2009, 25). Sense-making, then, is a way of relating to the world, and responding to environmental stimuli for the sake of enabling further actions. On the enactive picture, organisms are autonomous systems that engage in sense-

1 The concept of autopoiesis is often discussed in the enactive literature, but its invocation suggests further claims to which enactivists are not firmly committed. Autonomous systems are not necessarily autopoietic systems, because autonomous systems do not need to be spatially bounded for their self-regulation. Thompson and Stapleton offer the example of an animal or human social group as an autonomous system that is not spatially bound and therefore not autopoietic. 13

making, or regulation of behavior in relation to environmental valence. Sense-making occurs in virtue of a norm of self-regulation, but as Thompson and Stapleton note, this is an ‘all-or- nothing’ norm rather than the sort of ‘graded norms of vitality’ that correspond to an organism’s behavior. Di Paolo (2005; 2009) argues that the concept of adaptivity is needed in order to make sense of the graded norms of vitality seen in organismal life. Sense-making is thus a matter of adhering both to a norm of self-continuance and a norm of self-interestedness. And it is identified as cognition because it serves as a way for the organism to relate to its enacted world on the basis of its own viability and vitality.

Di Paolo, Rohde, and De Jaeger identify five core ideas of the enactive approach: autonomy, sense-making, emergence, embodiment, and experience (Di Paolo, Rohde, and De

Jaeger 2010, 37). As detailed above, autonomy and sense-making are central because they explain how cognizing systems are organized and how they establish a perspective on their enacted world via structural coupling with the environment. As Di Paolo, Rohde and De Jaeger explain, emergentist claims which view cognition as a process resulting from the interaction of various processes have a natural fit within the enactive framework, though they note that more work is needed to clarify an enactive notion of emergence. Embodiment is central to the enactive approach because “for the enactivist the body is the ultimate source of significance; embodiment means that mind is inherent in the precarious, active, normative, and worldful process of animation” (42). It is through movement and sensorimotor engagement with the world that we understand it. Lastly, experience is relevant on the enactive approach because it is “a guiding force in a dialog between phenomenology and science … [and] is itself a skillful aspect of embodied activity” (43). Enactive researchers suggest that taking accounts of experience seriously illuminates the study of cognitive activity and ability.

A second variety of enactivism, sensorimotor enactivism, developed from the views expressed in O’Regan and Noë (2001). Sensorimotor enactivism seeks to provide an account of perception and perceptual experience that is complementary to some of the central tenets in generalized embodied cognition as well as accounts of the body from the phenomenological tradition. O’Regan and Noë (2001) emphasize an organism’s sensorimotor contingencies in their enactive framework. They summarize their view as being a theory of vision that describes it as

“a mode of exploration of the world that is mediated by knowledge of … sensorimotor contingencies” (O’Regan and Noë 2001, 940). Sensorimotor contingencies refer to the “structure of the rules governing the sensory changes produced by various motor actions” (941).

Sensorimotor enactivism is primarily a theory of vision and visual experience, and therefore makes claims about the nature of visual perception more so than about cognition across the board. That said, it is complementary to other enactivist views in its rejection of mental representations in visual experience and its emphasis on the role of the body in perception. These commitments make it such that it falls in line with general enactivist thinking, but with a closer focus on vision and visual experience. Subsequent work from O’Regan and Noë gives treatment to topics such as the relationship between the physiological aspects of consciousness and visual perception (O’Regan 2011) and how bodily engagement with the world (which is fundamentally intertwined with vision) impacts our understanding (Noë 2012).

While the enactive approach to cognition is fundamentally non-representational, Hutto and Myin (2012) offer a more pointed argument as to why representationalism fails, denoting their variety of enactivism as radical enactivism. In Hutto (2017), the radical enactivism program is further clarified: the goal is not to put forth radical enactivism as its own variety of enactivism set against competing varieties, but rather to suggest it as a tool “designed to cleanse, purify,

strengthen and unify a whole set of existing anti-representational offerings. [Radical enactivism]’s aim is to radicalize existing versions of enactivism and related explanatory accounts through a process of philosophical clarification” (Hutto 2017, 379). Indeed, as Ward et al. note, “REC should be seen as an attempt to improve and unify anti- representationalist approaches to cognition rather than as competing with autopoietic or sensorimotor enactivism” (Ward et al. 2017). It can be thought of as a method by which to draw attention to the particular lines of argument in enactivist thinking that are most conducive to an anti-cognitivist, anti-representationalist approach to cognition.

The radical enactivist approach begins by rejecting some of the central tenets of the cognitivist program. For example, it denies any explanatory role to the notion of ‘informational content’ (Hutto 2017, 380). Radical enactivists see it as a fundamental flaw of cognitivism that it places emphasis on the representational concept of informational content while at the same failing to “supply a naturalistic account” (380) of it. With attention drawn to the problematic nature of this cognitivist tenet, radical enactivism seeks to provide an alternative account of information sensitivity that is in line with generalized enactivist thinking.

The term ‘enactivism’ may be used more broadly to refer to argumentative leanings that follow the central tenets of the enactive framework though they may follow a different trajectory in terms of their philosophical aims. These types of views can thus be said to remain committed to the enactive approach generally speaking, while not necessarily making arguments in the same domain as the enactive approach to cognition nor contributing to it as a philosophical perspective. Autopoietic enactivism makes some claims about vision and representations, but sensorimotor enactivism is silent on biological organization, and radical enactivism does not take

it as a given that representations are unwarranted. These three “varieties” of enactivism, autopoietic, sensorimotor, and radical, can each be more accurately described as their own enactive trajectories rather than as an alternative version to one another.

One way to frame a ‘taxonomy’ of enactive views is to draw attention to the research goals of each view. On this understanding, enactivism itself can be thought of as a generalized research program, and each view can be said to focus on different research goals with different tools at its disposal, yet at the same time committed to the same overarching philosophical goals of the enactive perspective. Gallagher (2017) raises a similar point in noting that enactive embodied cognition is a ‘holistic’ approach to cognition that incorporates both theoretical considerations and empirical data across numerous disciplines and areas of philosophical analysis (Gallagher 2017). He suggests that enactivism might be thought of as a ‘philosophy of nature’ (Godfrey-Smith 2001) rather than a scientific research program. In this context, a philosophy of nature “takes seriously the results of science, and its claims remain consistent with them, but it can reframe those results to integrate them with results from many sciences”

(Gallagher 2017, 22). Enactivism broadly conceived may be thought of as a philosophy of nature in that it is a holistic approach to questions about cognition, and one that makes use of theory from fields as diverse as phenomenology and dynamical systems theory.

While each variety of enactivism shares some fundamental theoretical claims about cognition, they differ in their foci and explanatory tools. This differentiation invites a question of whether one variety is more suitable for the incorporation of insights from developmental and evolutionary biology that bear on our understanding of the nature of cognition. In the next section, I will look more closely at some of the biological claims made in the autopoietic variety

of enactivism. My goal in drawing attention to these claims is to show how autopoietic enactivism converges with views in developmental systems theory and related perspectives in such a way that lends support to the integration of their larger theoretical frameworks.

2.2 Development and evolution on the enactive view

The enactive approach is a naturalistic theory of cognition that draws on biological insights about the organization of living beings to understand how cognition emerges in an organism’s enactment of a world. Though a variety of enactive perspectives have developed out of the original autopoietic view, the enactive approach, generally speaking, is grounded in organism- environment dynamics. One consequence of this view is that it suggests an alternative way of conceptualizing some key ideas from developmental and evolutionary biology. In the following section, I outline how the evolution and development of living beings is conceptualized on an enactive reading, drawing from work in both The Embodied Mind and Mind in Life. My goal in highlighting these claims is to provide a set up for thinking about the integration of enactive thinking with complementary views in developmental and evolutionary biology.

2.2.1 Evolution as natural drift

In order to illustrate the trajectory of biological insights through the enactive literature, I begin by focusing on an overview of the enactive concept of ‘evolution as natural drift’ as discussed in

The Embodied Mind. A commonly held view on cognition and evolution is that widespread cognitive capacities, such as color perception, are selected for due to their survival value. The structural coupling between an organism and its environment are the result of an organism’s fit to its environment. So the cognitive capacities that evolve in a history of structural coupling must

be in some way optimal. The cognitive capacities that are selected for are those that reflect optimality in some way or other.

Varela, Thompson and Rosch take issue with the notion of optimization and thus an adaptationist perspective as an explanatory tool for the evolution of cognitive capacities (and indeed, for organismal life). They argue that adaptationism is problematic for the same reasons that representational views of cognition are problematic, so “any evidence that weakens the adaptationist viewpoint ipso facto provides difficulties for the representationalist approach to cognition” (194). Recall that a central line of argument in the enactive approach is that representationalism is an inadequate basis for cognitive science. If arguments against adaptationism have force against the views of cognition that the enactive approach rejects, then they are worth considering further.

Instead of thinking as selection operating on a logic of optimality, they suggest that

“selection discards what is not compatible with survival and reproduction. Organisms and the population offer variety; natural selection guarantees only that what ensues satisfies the two basic constraints of survival and reproduction” (195). This view of selection leaves room for suboptimal traits and traits that seemingly make no contribution to survival or reproductive value. They suggest that we can then “analyze the evolutionary process as satisficing (taking a suboptimal solution that is satisfactory) rather than optimizing: here selection operates as a broad survival filter that admits any structure that has sufficient integrity to persist” (196). Structural coupling between an organism and its environment does not ascertain that the organism’s traits are optimal but rather the ‘organismic pattern’ is satisfactory for the viability of that coupling.

Natural selection thus becomes a matter of allowing for whatever can persist and pruning

whatever fails to remain viable. This view of selection, they suggest, fits more comfortably with

the considerable diversity of organismal life.

Varela, Thompson and Rosch label this view ‘evolution by natural drift’. They state that

it has four basic points:

1. The unit of evolution (at any level) is a network capable of a rich repertoire of self-organizing configurations.

2. Under structural coupling with a medium, these configurations generate selection, an ongoing process of satisficing that triggers (but does not specify) change in the form of viable trajectories.

3. The specific (nonunique) trajectory or mode of change of the unit of selection is the interwoven (nonoptimal) result of multiple levels of subnetworks of selected self-organized repertoires.

4. The opposition between inner and outer causal factors is replaced by a coimplicative relation, since organism and medium mutually specify each other.

The concept of evolution by natural drift is meant to capture how recurrent interactions in organism-environment coupling results in transformations over time. As they explain, traits or networks are selected solely on the basis of their viability, and as conditions and elements change in each pattern of coupling, the organism must change in response in order to remain viable. So long as it does so, it will persist. Therefore we can think of evolution as not being a powerful weeding out of everything except for optimal patterns but rather a less restrictive process that provides general constraints to viability over time. As the recurrent interactions in a coupled system will necessarily change with instances of variation and perturbations, the structural patterns of each system will naturally drift over time.

Varela, Thompson and Rosch emphasize the relevance of this view of evolution to the enactive perspective on cognitive science. They state, “[e]volution as natural drift is the biological counterpart of cognition as embodied action, and therefore also provides a more embracing theoretical context for the study of cognition as a biological phenomenon” (Varela,

Thompson and Rosch 1991, 188). A conception of evolution as adhering to a logic of satisficing rather than optimizing fits more naturally with a conception of cognition as embodied action and enaction of a world. These points suggest that a rethinking of standard evolutionary theory is necessary in the same way that a rethinking of standard cognitive theory is necessary. This is an important point for our purposes – if standard evolutionary theory is incompatible with an enactive approach to cognition, then an alternative theory must be worked out in its place. I will return to this point in Section 2.3 of this chapter, but it will also be the main focus of Chapter 3.

2.2.2 Enactive evolution

Thompson’s (2007) exploration of enactive perspectives on development and evolution begins by drawing on the historical contingencies of living beings that I discussed in Section 2.1.

Organisms have both a developmental and an evolutionary history, and thus, Thompson notes, life can be conceived of as a historical phenomenon. This entails that understanding organismal life means looking not just at an individual organism at a particular time but at the full historical pattern of that organism. According to Thompson, the living body, “seen as a temporally extended lifeline, is a developmental process or life cycle that initiates new life cycles through reproduction. Living beings are constituted by historical networks of life cycles, and the units of evolution are developmental systems comprising organism and environment” (167).

Developmental and evolutionary timescales are thus just as relevant to understanding living

beings as cognizers as behavioral timescales are.

As in The Embodied Mind, Thompson begins with a critique of the received view of evolution. He argues that the adaptationist perspective suggested by an emphasis on optimality in understanding natural selection is problematic for the same reasons that Varela, Thompson and

Rosch offer, but he also notes that the adaptationist thesis is also linked up with a genocentric perspective that invites further problems. On such a view, “organisms evolve as elaborate contraptions … constructed and controlled by genes” (173). Genes are involved in information

processing in the same way that the computationalist view suggests the brain is involved in

information processing. Naturally this view of evolution presents an issue for the enactive

framework, as it firmly rejects the computationalist view of cognitive activity as information

processing. Besides this, Thompson works through several issues with the genocentric approach

to evolutionary biology, highlighting its inadequacy as a biological perspective. Though I will

not go into the details of those criticisms here, the upshot is that there are at least two good

reasons for rejecting the received view of evolution: there are internal inconsistencies, and it is

incompatible with the enactive approach to life and cognition. If we are convinced that the

enactive approach is a viable way forward, an alternative conception of evolution must be

supplied.

One way forward, Thompson suggests, is to consider evolution from the perspective of

developmental systems theory (DST). According to DST, evolution can be defined in terms of

change in organism-environment systems rather than as change in gene frequencies. Moving the

focus away from changes in genes to changes in organism-environment systems dissolves the dichotomies between genes and environment (and relatedly, between nature and nurture) that

DST researchers have suggested are inadequate conceptual bases for understanding developmental and evolutionary change. Instead of looking at genes and the factors that bring about genetic variants, DST suggests that we look at the processes of interaction between elements in an organism-environment system. Importantly, this system is processual and dynamic, with change and variation arising over various timescales. It can thus be conceived of as a developmental system. Individuals are conceived of as the process of a life cycle that is embedded within a larger developmental system. It is this developmental system that is the unit of evolution, Thompson claims, rather than genes. In emphasizing the reciprocal relationship between organism and environment as well as the structural coupling that forms as a result of recurrent interactions between organism and environment, DST shares a theoretical basis with the enactive approach and is therefore better suited as a biological counterpart to the enactive approach than the genocentric traditional approach.

Together, the autopoietic view of living systems that the enactive approach espouses and the DST perspective on organism-environment systems suggest an alternative view of development and evolution. As Thompson explains:

Autopoietic systems (and autonomous systems generally) are unified networks of many interdependent processes. Organisms are accordingly not the sort of systems that have atomistic traits as their proper parts; such traits are the products of theoretical abstraction. Similarly, from the viewpoint of developmental systems theory, the adaptationist notion of the organism as an array of traits on which selection acts obscures development. In development, there are no static traits, but rather integrated developmental processes. (203)

Thus, both theories emphasize the dynamic interaction between organisms and their environments and the co-constructive nature of each system. The environment is not merely an

external force that acts upon particular genetic traits. Understanding development and

evolutionary change, on a shared enactive-DST view, is thus a matter of investigating the changes in structural pattern that arise as a result of processual interactions within a developmental system, with both organism and environment initiating alterations. Thus the environment is not conceived of as an external, independent force that exerts pressures on an organism and poses problems which the organism must solve in order to stay alive, but rather as a setting that both constructs and is constructed by the organism structurally coupled with it. This mutual specification is central to both enactive thinking and DST perspectives on development and evolution.

Thompson reiterates the enactive claim that selection operates on a logic of satisficing rather than optimizing, which he labels as the ‘conservation of adaptation’, and states that this claim along with the emphasis on structural coupling in organism-environment systems makes

up a theory of ‘enactive evolution’. As he notes, “[a]ccording to this perspective, all living

beings are adapted as long as they are alive. Reproductive success – the measure of fitness – is

not determined by isolated traits, but by the entire life cycle” (205). On a shared enactive-DST

view, enaction refers to the playing out of a life cycle as a process embedded in a developmental

system. Organism and environment co-determine one another in their structural coupling. The

unit of evolution is taken to be the whole developmental system rather than isolated traits.

2.3 Enaction in a developmental system

Thompson asserts that “a natural kinship exists between developmental systems theory and the

theory of autopoiesis” (193) that grounds the enactive approach to cognition. He notes:

…beyond their criticisms of neo-Darwinism, the two theories support each other. The theory of autopoiesis is needed to describe the self-producing organization of living things on the basis of which development and evolution proceed, and developmental systems theory is needed to give a nondichotomous account of the structural coupling of organism and environment in ontogeny and phylogeny. (193- 194)

Autopoietic enactivism thus fits naturally with insights from developmental systems theory.

What we gain is an enriched view of the organism and its enacted world over developmental, behavioral, and evolutionary time. DST further specifies the organization of living beings as developmental processes embedded within a developmental system and coupled with a developmental niche. Recall that my goal in this dissertation is to investigate more carefully the biological elements of autopoietic enactivism and show how they fit with DST and related perspectives for an integrated framework for understanding biological cognition. In this section, I will sketch out how this integration might look, which will then be expanded upon in the remaining chapters.

According to autopoietic enactivism, cognizing organisms are adaptive, autonomous

systems with the capacity for sense-making, and these features entail that they exist as self-

regulating, individuated unities in constant dynamic relations with the environment. However,

more can be said about how cognitive functioning has played a role in an organism’s

developmental and evolutionary history. While autopoietic enactivism provides a sketch of the

architecture of cognitive systems, further elaboration that incorporate findings across

developmental, behavioral, and evolutionary timescales can aid in understanding the various

roles that cognition plays in ontogeny and phylogeny. This work can be done by incorporating

theoretical considerations from, for example, evolutionary developmental biology, which

emphasizes looking at the organism across both scales in order to understand why certain

features have evolved, and from developmental systems theory, which emphasizes the role of

organism-environment interaction in explaining the existence of certain traits. Such a framework would also be suitable for discussion of the development and evolution of cognition itself.

2.3.1 Biological enactivism: a synthesis and expansion

The coevolution of organisms and environmental features is a phenomenon central to an enactive understanding of biological cognition, yet there exist only a few discussions on the topic, such as Thompson, Palacios and Varela (1992), which focuses on comparative color vision. In order to serve as a robust theory of biological cognition, autopoietic enactivism can be bolstered by additional scaffolding from recent developments across cognate fields in biology.

This biologically-informed enactive approach can thus be termed biological enactivism, and can

be thought of as an extension of autopoietic enactivism intended to serve not only as an account

of cognition that focuses on the organization of cognizing systems but also the roles that

cognition plays in development, behavior, and evolution.

Biological enactivism, as a theory of cognition as it is found in nature, makes the

following central claim:

Organisms are cognizing systems structured by both their internal dynamics and their dynamic relations with environmental features corresponding to their sensorimotor capacities, developed as a result of their coupled interactions with their environments over both developmental and, on a population scale, evolutionary time.

Note that this claim is firmly consistent with Thompson’s enactive theory, especially regarding

the concept of enactive evolution and the relationship between the enactive approach and DST.

A biological enactive approach should thus be viewed as an expansion of Thompson’s insights

rather than as an alternative version of them. This expansion entails analyzing research in

biology through an enactive lens in order to bolster these claims. It is a synthetic project in that it

brings together the enactive approach with complementary areas of research in biology both

including and beyond those discussed by Thompson. It elaborates and expands upon the

developmental and evolutionary claims Thompson makes, in turn forging more and stronger

links between the enactive approach and DST, as well as related perspectives in biology such as

extended interpretations of evolutionary theory (which will be the focus of the following

chapter).

In particular, biological enactivism investigates the underlying mechanisms of structural

coupling that result in the emergence of coupled systems over developmental and evolutionary

time. The idea is that the notion of structural coupling can be incorporated into a developmental

and evolutionary framework to make it clear how cognizing systems develop and evolve in

nature. Biological enactivism takes the enactive approach a step further by specifying how these

dynamic relations are instantiated both by individual organisms and historically as an effect of

evolution as selective pressures both on species and on the environment. The hope is that this

framework is explanatorily robust as a theory of cognition for both the fields of cognitive science

and biology. It can also serve as a useful tool for current debates in a number of research domains in biology. The next chapter, for example, investigates how it might be fruitfully

applied to debates regarding the nature of evolutionary processes.

Biological enactivism thus seeks to describe and explain cognition at both the organism

scale and the population scale. Enactive approaches are generally progressive in their focus of

the whole organism as the right unit of analysis for understanding cognition rather than a

reductionist approach. However, this view has both implications and relevant results at the

population scale as well. If the goal is to understand the phenomenon of cognition, giving an

account of its properties at the population scale will be relevant to that understanding. The

framework would thus benefit from the inclusion of theoretical and conceptual tools that have

application across developmental, behavioral, and evolutionary timescales.

In order to make sense of the central claim outlined above, it is necessary to specify how

it is that organisms (1) maintain internal dynamics and (2) engage in dynamic relations with their

environments corresponding to their sensorimotor capacities. This requires both a developmental

account and an evolutionary account. Organisms are not simply placed in the environment as

static entities, nor is the environment one which itself is static and independent from forces that

might bring about change. Understanding the flower-seeking behavior of honeybees requires understanding why honeybees have the visual systems that they do, and undergo sense-making with regard to angiosperms in the way that they do. If cognition is ultimately a matter of action- perception dynamics, autopoietic enactivism can benefit from a richer account of coevolutionary dynamics in order to make sense of why certain sensorimotor systems are found in nature. It has the advantage of specifying the organizational features internal to the organism, but there is more to be said on the impact that self-organization and sense-making have on the organism itself throughout its life cycle and evolutionary effects on later generations. Therefore, a robust account of cognition should be one that has the resources to be able to explain it diachronically –

not only across the developmental and behavioral timescale of the individual organism, but

across the evolutionary timescale of the species as well.

2.3.2 Integrating DST with the enactive approach

To elaborate on Thompson’s claims, where DST becomes relevant to enactivism is in its

identification of coevolutionary dynamic structures between organisms and their environments.

Regarding self-organization, which captures both the notions of autonomy and adaptivity, DST

can provide a biological backdrop to how organisms are structured through development as self-

organizing systems. The language used in discussing structural coupling in autopoietic

enactivism and coevolutionary dynamics in DST is essentially the same; both concepts serve to

draw attention to the factors relevant to an organism developing the sort of traits that it does.

Sense-making arguably requires an extension into evolutionary theory, as it is concerned with behavior in relation to environmental valence, a phenomenon that impacts not just the individual organism but the organism’s lineage as well. DST also helps to identify constraints on possibility spaces for the evolution of certain forms. Recall that “DST views both development and evolution as processes of construction and reconstruction in which heterogeneous resources are

contingently but more or less reliably reassembled for each life cycle” (Oyama, Griffiths and

Gray 2001, 1). This point is relevant to the types of behavioral strategies that correspond with

particular biological structures. Digit formation, for example, places constraints on certain

behavioral strategies. Thus, cognitive science can benefit from the insights of DST by way of

specifying the processes involved in the instantiation of certain cognitive traits. If the picture of

embodied cognition is convincing, then providing a story of how organismal bodies develop and

evolve through time can reveal valuable insights into the structure of cognition itself.

The framework of DST can be incorporated into the enactive approach to cognition in

order to provide an account of how organisms as cognizing systems develop and evolve in

relation to their environments. DST also has the tools to explain changes in the structures of

environments as coevolving with organisms, thus treating neither the organism nor its

environment as static and discrete. It provides a biological basis for how the structure of

cognizing systems has developed over time. Cognition can be thought of as an emergent property

of these systems.

Within the biological enactive framework, organisms can be thought of as cognizing

systems that have the following properties. They exhibit self-organization, through operating as autonomous, adaptive systems that typically meet the requirements for being autopoietic systems as well. They engage in sense-making, in that they have the capacity for augmenting their behavioral and structural traits in relation to environmental valence. Lastly, their relation to the environment is one of structural coupling, where organisms develop and maintain the structures necessary to engage in sense-making, have the capacity for adaptivity, and maintain autonomy.

Cognition can be thought of as an emergent property of systems structured in this way, and can thus be defined as follows:

Cognition is an emergent property of cognizing systems, where cognizing systems are those systems which have as their fundamental properties self- organization, sense-making, and structural coupling.

This chapter constructs an enactive framework for understanding biological cognition through investigating core concepts in extant varieties of enactivism, such as autopoiesis and autonomy,

sense-making, and evolution as natural drift. I then suggest expanding these concepts by

incorporating complementary theoretical claims from developmental and evolutionary biology.

In the following chapter, it will be argued that this framework is complementary to the current push in evolutionary biology for an extended synthesis of evolutionary theory, where the goal is to incorporate new concepts into evolutionary theory in order to shed light on recent findings that go beyond concepts in the standard framework. As central tenets from developmental systems theory are woven into biological enactivism and developmental systems theory is incompatible with the genocentric view put forth in standard evolutionary theory, (a claim elaborated upon in the next chapter), some conceptual faults with standard evolutionary theory will be investigated and an argument for a reinterpretation offered.

Chapter 3: The Role of Cognition in Evolution

Introduction

In the previous chapter, I discussed the enactive approach to cognition, which holds that the dynamic relationship between an organism and its environment is central to understanding cognition. I argued that a philosophical perspective stemming from the enactive approach, autopoietic enactivism, finds a natural fit within a broader theoretical framework on living systems offered by developmental systems theory as well as extended interpretations of evolutionary theory. Bridging the gap between these two areas by integrating the autopoietic view with these biological perspectives results in a novel approach to understanding cognition in nature. Specifically, it results in a richer theory of cognition that details the structural coupling between organisms and their environments.

In this chapter, I consider how this novel enactive theory parallels and provides resources for evolutionary biology. I suggest that exploring how extended interpretations of evolutionary theory invite new ways of thinking about what factors shape a species’ evolutionary trajectories can in turn invite further dialogue about the role of cognition in evolution. Drawing out similarities between theoretical insights from enactivism and arguments for an expansion of current evolutionary theory sheds light on the explanatory advantages of both accounts.

Innovative research on ways in which cognition can facilitate the inheritance of traits motivates new theoretical considerations and encourages a novel perspective on the processes underlying evolution.

Within the past two decades, there has been a call for revision to the received view of evolutionary theory. Contemporary evolutionary theory is structured by the central tenets of population genetics as a theory of heredity and Darwinian evolution – a bridging of areas referred to as the ‘Modern Synthesis’ (MS). Some researchers hold that this framework is inadequate for addressing numerous questions and concerns in evolutionary biology, namely those raised by new findings in related fields such as evolutionary developmental biology (evo- devo) and evolutionary ecology. A substantial revision to the underlying structure of evolutionary theory, they argue, is necessary (Gould 1983; Pigliucci 2007, 2009; Laland et al.

2015). This suggestion has prompted significant debate among both philosophers of biology and evolutionary biologists (Laland et al. 2014). Some of those in favor of revision claim that an alternative theoretical framework, what they call an ‘Extended Evolutionary Synthesis’ (EES), is a richer framework that is better suited for addressing the diversity of topics increasingly relevant to evolutionary biology, and thus should be adopted as representing standard evolutionary theory.

The EES has not been without scrutiny itself. Proponents of traditional evolutionary theory argue that such significant revision is not necessary, as the MS has the theoretical and conceptual tools to tackle new problems in evolutionary biology. What’s more, they hold that some of the central theoretical claims that the EES makes currently lack robust empirical backing, and so there is no urgency to their inclusion into standard evolutionary theory. The

EES, it is suggested, is an unnecessary overhaul of the conceptual architecture of evolutionary biology, and there is little empirical support for the need of such an overhaul (Mesoudi et al.

2013).

Further, while proponents of both sides of the debate agree that research from cognate fields, including fields as diverse as genomics, ecology, and social science (Laland et al. 2014), should be incorporated into the broader framework of evolutionary theory, they disagree on more philosophical points about how results from those fields should impact the structure of biological knowledge. One key concept on which the two frameworks differ is non-genetic inheritance, or the idea that information can be transferred generationally via mechanisms other than changes to

DNA sequences. Mechanisms of non-genetic inheritance are detailed in work in evolutionary ecology (Bonduriansky and Day 2009), behavioral ecology (Head et al. 2016), and even psychiatry (Toth 2015). More broadly, differing views on non-genetic inheritance can have an impact on how evolutionary processes are conceptualized. Accepting that non-genetic inheritance occurs and emphasizing its causal role in bringing about evolutionary change paints a broader picture of the processes underlying evolution, while views that dismiss the causal importance of non-genetic inheritance narrow the scope of the processes relevant to evolution.

The EES argues that non-genetic inheritance plays a prominent role in evolutionary change, while the MS focuses on Mendelian inheritance as the primary mechanism for inheritance, with selection as the driver of evolution (Plutynski 2014; Gould 2002).

On an extended interpretation of evolutionary theory, there are several distinct systems of non-genetic inheritance (Jablonka and Lamb 1995, 2005; Danchin et al. 2011). Two systems of particular relevance for this chapter are ecological inheritance, whereby organisms modify their environments through their behavior and in turn alter the inherited niche of subsequent generations, and cultural inheritance, where phenotypic variation arises as a result of social

learning2. Recent research at the intersection of biology and cognitive science illustrates how

cognitive phenomena, such as behavior, trial-and-error learning, experience, and culture via

social learning, have the potential to facilitate these forms of non-genetic inheritance. On this view, key features of both ecological inheritance and cultural inheritance are thus cognitive in nature. If it is the case that non-genetic inheritance can initiate and shape evolutionary trajectories, as some researchers hold, then cognition, as a facilitator of non-genetic inheritance, can potentially shape not only an individual’s phenotypes but the evolution of its population as well. Therefore, looking at ways in which cognition shows up as part of the causal story in systems of non-genetic inheritance can shed light on one possible role of cognition in evolution.

This chapter proposes a novel framework at the intersection of evolutionary biology and cognitive science that supports the claim that one role that cognition can play in nature is as a driving force for evolutionary change, namely via mechanisms of non-genetic inheritance. This framework requires theoretical commitments to both the role of non-genetic inheritance in evolution and to the explanatory relevance of the dynamic relationship between organisms as cognizing systems and their environments. I hold that the EES is particularly well-suited for such a project due to its emphasis on reciprocal causation between organisms and their environments, which can be thought of as a structural feature of non-genetic inheritance. What’s more, a view of cognition that is centered on the dynamic relationship between organisms and their environments, such as that offered by biological enactivism, the view I am espousing in this dissertation, provides the theoretical and conceptual tools to explain the role of cognition in both

2 Avital and Jablonka (2000) also describe behavioral inheritance as a system of non-genetic inheritance. There is overlap between behavioral, ecological, and cultural inheritance, though less between the latter two, and so for the purposes of this chapter I will focus on mechanisms of non-genetic inheritance that arise across all three systems. Epigenetic inheritance is an additional system of non-genetic inheritance detailed in the literature, but the connection between cognition and epigenetics is beyond the scope of this chapter (v. Marshall and Bredy 2016). 35

development and evolution. Taken together, the EES research program and biological enactivism

result in a rich conceptual architecture that can support the claim that cognition can serve as a

driving force of evolutionary change.

Section 3.1 provides an overview of the call for an extended synthesis and explains how

the framework builds on that of the MS in such a way that encourages the incorporation of

research from cognate fields and takes seriously how new findings might impact the structure of

biological knowledge. Section 3.2 draws parallels between debates concerning conflicting views

in both cognitive science and evolutionary biology, suggesting that these parallels reveal

valuable insights about both debates. I suggest that what is at the heart of both debates is a

defining claim regarding causal directionality, and drawing attention to the claims made by each

side reveals similarities in underlying commitments that impact the conceptual architectures of

both fields. Section 3.3 focuses in on the topic of ecological and cultural inheritance as biological

phenomena that illustrate the richness of what I will refer to as a bidirectional view of causality.

There, I discuss empirical findings that exemplify how an individual’s cognitive capacities can

shape its development and evolution. This discussion helps to set up the picture of biological

cognition that I argue for in the remainder of this dissertation. Looking at how cognition

facilitates ecological and cultural inheritance (as forms of non-genetic inheritance) is just one kind of role that cognition can play in nature, and the following chapters will address other ways in which cognitive abilities can directly impact organismal form and function.

3.1 An alternative synthesis

A historical account of the Modern Synthesis begins with Darwin’s conception of common

descent and natural selection (Darwin 1859). With the advent of work by Fisher (1918), Wright

(1932), and Haldane (1932) that would come to form the field of population genetics, Darwin’s

theory of natural selection would be supplemented by a theory of heredity, thus making it viable

as a comprehensive theory of evolution. The synthesis of these two fields, Darwinian

evolutionary theory and population-statistical genetics, would come to be known historically as

the ‘Modern Synthesis’ (MS), and would then serve as the conceptual and theoretical

architecture for the growing field of evolutionary biology.

Karl Popper was among the first to propose that significant revisions to the MS were

necessary, suggesting that the theory is “strictly a theory of genes, yet the phenomenon that has

to be explained in evolution is that of the transmutation of form” (Pigliucci 2009; quoted in

Platnik & Rosen 1987). As Pigliucci notes, some have argued that while standard evolutionary

theory does indeed provide a suitable theory of genes, it has yet to adequately deal with

questions about the evolution of organismal forms (Pigliucci 2007, 2009; Robert 2004). Gilbert

and Epel (2009) argue that new research on developmental plasticity “is causing a need to reflect

on and critique our gene-based models of biology. Phenotype is not merely the unrolling of

genotype” (Gilbert and Epel 2009, 395). The initial concern, then, is that the theoretical structure

of the MS is insufficient on its own, because little significance is given to concepts deemed by

critics to be explanatorily relevant for understanding the evolution of traits, such as

developmental bias and shaping influences from the environment. Further concerns include expanding what counts as a system of inheritance and an emphasis on the dynamic relationship between organisms and their environments, specifically the role the organism plays in its own development and the evolution of its species (Jablonka and Lamb 1995; Levins and Lewontin

1985). These concerns have led some researchers to propose an alternative to the MS, an

‘Extended Evolutionary Synthesis’, which is built on a framework that, in their view,

significantly revises, expands, and elaborates on the framework for the MS. While the name

suggests that the EES merely builds on the MS, proponents of the EES argue that the elements it

introduces and the emphasis on their impact require reconsideration of what counts as the central

theoretical and conceptual tenets of evolutionary biology.

Evolutionary developmental biology, or evo-devo, investigates how the study of

development has theoretical implications for evolutionary theory, arguably motivating the need

for an extension of the MS. Müller (2007) suggests that evo-devo (and relatedly, eco-evo-devo, which incorporates ecology into the evo-devo framework) presents a view of phenotypic change in evolution that “differs significantly from the prevailing focus in the standard theory of evolution” (Müller 2007, 945), thus inviting questions regarding the need for reconceptualization of evolutionary change beyond change in gene frequencies. Similarly, Carroll (2008)’s analysis of empirical work in molecular developmental biology motivates an argument for the incorporation of evo-devo into standard evolutionary theory. Other researchers have drawn attention to phenomena at the intersection of evolutionary and developmental biology for which the framework of the MS, they argue, fails to adequately account. Gilbert et al. (1996) suggest that macroevolution, homology, and morphogenetic fields are significant problem areas for current evolutionary theory, and it is an open question as to whether empirical findings in those areas necessitate a reshaping of the theoretical framework. Danchin et al. (2011) offer a novel approach to quantifying the interaction of mechanisms of inclusive or non-genetic inheritance, urging that “the theory of inheritance that currently prevails also needs to be extended in order to incorporate all non-genetic inheritance as participating to the development and inheritance of the phenotype” (Danchin et al. 2011, 481). Developed out of these and similar concerns, arguments for an extended synthesis center on a necessary restructuring of evolutionary thought as to allow

for the incorporation of new fields of interest as well as to do work toward the wider epistemic

goals of evolutionary biology, such as understanding the generation of novel phenotypes and the

mechanisms responsible for speciation.

3.1.1 Extending evolutionary theory

One of the central claims of the EES is that research on evo-devo, developmental plasticity,

inclusive inheritance, and niche construction invites an interpretation that “views the same

findings as challenging important assumptions of the MS” (Laland et al. 2015, 2). These four

areas each look at how phenomena arguably underexplored in standard evolutionary theory can

have a significant impact on an understanding of the evolution of traits. Evo-devo research

“provides a causal-mechanistic understanding of evolution by using comparative and

experimental biology to identify the developmental principles that underlie phenotypic

differences within and between populations, species and higher taxa” (3). One of the key areas of

evo-devo research is developmental bias, which suggests that variation in phenotypes can be

brought about via influences during development. Developmental plasticity is defined as an

organism’s ability to change phenotypes in response to certain environmental effects. For

example, in the presence of chemical cues emitted by its natural predators, the water flea

(Daphnia) develops an elongated ‘helmet’ and tail spine which serve as protective elements

(Laforsch and Tollrian 2004; Nettle and Bateson 2015). The EES emphasizes the role of

development in the generation of phenotypic traits – such traits are not merely determined by an

organism’s genotype, but can arise (or fail to arise) as a result of certain environmental cues.

Laland et al. argue that developmental plasticity can be thought of as a “cause, and not just a

consequence, of phenotypic evolution” (3). Again, the goal is to show how this phenomenon has

relevant explanatory implications for the evolution of phenotypes. Inclusive inheritance refers to the ways in which traits can be inherited by offspring through means other than transmission of

DNA. Non-genetic inheritance systems, such as epigenetic, behavioral, and cultural inheritance

(Jablonka and Lamb 1995, 2005), “can bias the expression and retention of environmentally induced phenotypes, thereby influencing the rate and direction of evolution” (Laland et al. 2015,

4). Lastly, niche construction theory suggests that organisms coevolve with their environments in such a way that makes organism-led environmental modifications a factor in selection. Laland et al. argue that niche construction itself ought to be thought of as an evolutionary process due to its significant impact on selection.

Two focal conceptual points can be derived from research in evo-devo, developmental plasticity, inclusive inheritance, and niche construction, and thus are central to the EES as an alternative interpretation of evolutionary theory. These two concepts are (1) constructive development and (2) reciprocal causation. Together, they capture what is distinct about the four areas of research listed above: they illustrate how processes beyond selection can play a role in evolution and do so in a way that extends relevant causal significance to factors other than genes.

Constructive development “refers to the ability of an organism to shape its own developmental trajectory by constantly responding to, and altering, internal and external states” (6). While proponents of the received view of evolutionary theory would not deny that constructive development occurs, the EES specifies it in such a way that stresses the complex interplays between genotype and phenotype rather than describing their interaction as an asymmetric relation where genes are the primary factor in guiding development. Reciprocal causation, notably exemplified in niche construction theory, emphasizes the dynamic relationship between organisms and their environments. This view of causation suggests that “developing organisms

are not solely products, but are also causes, of evolution” (6). Selective pressures are not independent forces acting on an organism, but rather are themselves shaped by organisms over developmental and evolutionary timescales.

The bidirectionality of both developmental and evolutionary causal influences and their integration within one another is seen to be a rejection of the distinction between proximate and ultimate causes (Mayr 1961; West-Eberhard 2003; Mesoudi et al. 2013). The concept of reciprocal causation suggests that there is no strict separation to be made: evolutionary processes are fundamentally intertwined with developmental processes, and vice versa. Variation cannot be understood on solely an evolutionary timescale. The role the individual organism plays in its own development, and for its species’ evolutionary trajectory, is relevant: “[t]he EES is thus characterized by the central role of the organism in the evolutionary process, and by the view that the direction of evolution does not depend on selection alone, and need not start with mutation” (Laland et al. 2015, 8). An individual organism’s characteristics can be part of the causal story as to why a particular phenotype arises for a species.

Some critics of the EES are skeptical that the areas of research the EES focuses on are robust enough to necessitate incorporation into standard evolutionary theory. In a response piece,

Wray et al. assert that “none of the phenomena championed by Laland and colleagues are neglected in evolutionary biology. Like all ideas, however, they need to prove their value in the marketplace of rigorous theory, empirical results and critical discussion. The prominence that these four phenomena command in the discourse of contemporary evolutionary theory reflects their proven explanatory power, not a lack of attention” (Wray et al. in Laland et al. 2014, 163).

Their argument is not that there is an a priori reason for the exclusion of areas of research such as developmental bias in evolutionary biology, but rather that the results of that research must first

be scrutinized and found to have significant bearing on evolutionary theory. They stress that until more robust research is offered, there is no impetus to integrate the EES’s key phenomena into evolutionary theory. For example, they claim that “there is little evidence for the role of inherited epigenetic modification (part of what was termed ‘inclusive inheritance’) in adaptation: we know of no case in which a new trait has been shown to have a strictly epigenetic basis divorced from gene sequence. On both topics, further research will be valuable” (164). While it is undoubtedly the case that new research will help to guide the debate further, any theoretical framework that incorporates non-genetic inheritance in the way discussed in this chapter would hypothetically be a suitable framework for understanding the role of cognition in evolution. Therefore, the argument presented here does not rest upon an argument for the EES as a viable research program, though as discussed later on, incorporating non-genetic inheritance into the framework of the MS has historically proved challenging.

Some philosophical concerns raised by both sides of the debate center on what biological phenomena get incorporated into evolutionary theory. Wallace (1986) asserts that questions about development are fundamentally distinct from questions about evolution, and so empirical and theoretical work in fields such as embryology have no bearing on the structure of evolutionary theory. Indeed, burgeoning work on developmental biology was historically left out of the architecture of the MS (Gilbert et al. 1996), with development being seen as “not only asking the wrong questions, but it was structurally incapable of asking the right ones” (Gilbert and Epel 2009, 444). While the field of embryology would in time be subject to much further refinement, any potential for impact on evolutionary theory was dismissed during the development of the MS.

3.1.2 Interdisciplinary opportunities from the EES

The extended interpretation of evolutionary theory “is distinctive for its emphasis on organismal causes of development, inheritance and differential fitness, the role of constructive processes in development and evolution, and reciprocal representations of causation” (Laland et al. 2015, 2-

3). Proponents of this alternative interpretation stress that is not meant to function as a wholesale rejection of the MS, nor is it meant to be suggestive of fundamental errors in its theoretical claims. Rather, the goal is reconceptualize the factors and processes responsible for evolutionary change and move away from an emphasis on genes as the sole explanatory vehicle for that change. Phenomena previously considered consequences of evolution rather than causes of it, such as non-genetic inheritance and niche construction, can be incorporated into general evolutionary theory with the goal of enriching concepts defined by the MS. Contemporary evolutionary biology is continuously shaped by new research that both clarifies and expands upon previous findings. There is no question that the field itself evolves as scientific research programs do over time. The proposed restructuring brought about by the EES’s alternative interpretation holds onto the central tenets of the MS, but provides a picture of evolution that builds upon those tenets in the hopes of a richer framework for the structure of biological knowledge.

At the heart of the debate between proponents of the EES and standard evolutionary theory is a question about what it means for the field of evolutionary biology to undergo theory change. One way to frame this suggestion might be to label it a paradigm shift (Kuhn 1970), which serves as a common way to conceptualize theory change in the sciences. However,

Pigliucci (2009) suggests that a move from the MS to the EES does not represent a paradigm shift, as such would entail a wholesale rejection of the currently existing conceptual structure of

evolutionary biology, and the EES does not propose such a rejection. The rethinking that the

EES does suggest, Pigliucci holds, is more aligned with Gould’s (2002) “model of a continuously expanding conceptual tree” (Pigliucci 2009, 219). On this view, as new research is done in evolutionary biology, its foundational architecture increases in scope and complexity, incorporating new claims about fundamental processes on which new research has shed light.

Rather than proving existing claims false, new findings enrich those claims, thereby guiding further research and filling in gaps in biological knowledge.

Another problem area that arises concerns a differentiation between questions regarding the content and the structure of evolutionary theory (Love 2010, 2017). Assessing proposed changes to the structure of evolutionary theory, rather than the content, may help to clarify what the epistemic implications of those changes are. Fábregas-Tejeda and Vergara-Silva (2018) suggest representing the structure of the EES alongside that of the MS in a vast, interwoven network representing the conceptual architecture of evolutionary biology. This network, made up of “evolutionary models, practices, and representation systems” (Fábregas-Tejeda and Vergara-

Silva 2018, 173), could ‘revolve around the epistemic goals of evolutionary biology’, with a central epistemic goal being “an understanding and acceptance of the constructive roles organisms play in shaping their own development and in changing their environments in evolutionarily meaningful ways” (181). They argue that framing the conceptual architecture of evolutionary biology as a vast network or set of interwoven networks illustrates a possible way to incorporate research and concepts from evo-devo especially. This alternative philosophical approach can help to draw out what the relevant epistemic units and epistemic goals are of evolutionary biology as a pluralistic and dynamic field of inquiry that is being continually shaped

and reshaped by research across numerous related dimensions. Their concluding point is worth

emphasizing:

Finally, we want to stress that, beyond worn-out Kuhnian or Lakatosian schemes, different philosophical insights are needed to better understand the interdisciplinary reconfiguration of evolutionary theories that are also opening to different fields of inquiry in the social and human sciences … (181).

Representing the conceptual architecture of evolutionary biology in the way Fábregas-Tejeda and Vergara-Silva suggest creates the opportunity to incorporate research from other related fields either previously considered irrelevant or simply yet to be considered robustly. This restructuring creates a possible avenue by which research from fields such as psychology and cognitive science more broadly can be woven into the framework for a pluralistic understanding of evolutionary biology, and can in turn help to shape its epistemic goals in a fruitful manner.

Some have suggested that an overlap between evo-devo and cognitive science (Gottlieb

2001; Stotz 2014) may be one avenue to begin exploring the role of cognitive science in a

pluralistic approach to evolutionary biology. The advantage to an approach that leaves room for

the impact of research from other fields is that it may do significant work toward the epistemic

goals of evolutionary biology. This philosophical framing can help clarify how taking on the

EES as an alternative research program for the field of evolutionary biology would impact the

structure of biological knowledge in such a way that allows for it to grow and develop. One such

effect, I want to suggest, is a new perspective on how to understand the role of cognition in

development and evolution.

3.2 Parallels in cognitive science

As discussed in the previous chapter, one of the central lines of argument against cognitivist and

computationalist approaches to the mind is that they posit the cognizer as a passive receiver of

sensory input that is processed internally via neurological mechanisms, effectively bracketing off

or even ‘black-boxing’ numerous important causal factors and influences. In this section, I want

to suggest that criticisms against the MS parallel those made against cognitivism; critics argue

that the MS, too, posits organisms as passive receivers of environmental stimuli, a view arguably

too asymmetrical to capture other relevant explanatory processes. In this way, the arguments put

forth against cognitivism and against the MS enjoy similarities, and highlighting the ways in

which these two debates share points of criticism can help to shed light on the explanatory

advantages to both enactivism (generalized here as 4E cognition3) and the EES, as well as how the two theories may complement one another. Illustrating these parallels requires drawing out related conceptual points, and a brief account of Gilbert Gottlieb’s research in behavioral development helps to guide the argument.

3.2.1 The epigenesis of behavior

While undergoing research in behavioral embryology, Gottlieb (1970) determined that there were two approaches to understanding seemingly innate behavioral development: either “the behavioral sequence is predetermined by factors of neural growth and differentiation

(maturation)” (Gottlieb 1970, 112) or “the sequence and outcome of individual behavioral

3 The term ‘4E cognition’ captures a view of cognition as embedded, embodied, enacted, and extended. This view holds that cognitive processes extend outside of the brain to include the body and, on some accounts, features of the environment, such as artifacts or other people. 4E views on cognition contrast with cognitivist or computationalist views, which assert that the brain is the proper domain of cognitive processes (See Newen, de Bruin and Gallagher 2018). 46

development is probable (with respect to norms) rather than certain” (Gottlieb 2001, 43). On this dichotomy, the epigenesis of behavior is either predetermined, making behavior “an epiphenomenon of neural maturation” and not a contributing factor to development itself (42), or it is probabilistic, viewing behavior as potentially facilitating certain developmental traits depending upon certain initial conditions. In duckling development, for example, vocalization behavior emerges as a species-typical fixed action pattern. On the predetermined epigenesis view, this behavior develops as an innate trait that arises independently of any environmental input. On the probabilistic view, on the other hand, it is contingent upon numerous sensory stimulative factors, namely the prenatal auditory experience of the duckling’s own vocalizations as well as its siblings’ vocalizations. Gottlieb’s research on the manifestation of vocalization behavior (Gottlieb and Vandenbergh 1968) supported the probabilistic view: ducklings who were deprived of their ability to experience their own or sib vocalizations “could not distinguish the mallard maternal call from the chicken maternal call” (Gottlieb 2001, 45; Gottlieb 1971 141-

142). These results showed that traits previously thought to be innate, central to work done by

Konrad Lorenz and Niko Tinbergen, had in fact a complex developmental story underlying them, and one that illustrated the causal role of both the environment and of individual experience in seemingly predetermined traits.

On Gottlieb’s probabilistic epigenesis view of behavioral development, structure (in the above example, vocalization behavior) and function (auditory experience) form a dynamic relationship – specifically, they form a bidirectional relationship. Predetermined epigenesis assumes a unidirectional relationship between structure and function – “structure is supposed to determine function in an essentially nonreciprocal relationship” (43). Alternatively, a bidirectional account of the structure asserts that function “can significantly modify the

development of the peripheral and central structures involved in these events” (Gottlieb 1970,

123). In the case of behavioral development, a probabilistic epigenesis account of behavior necessitated a bidirectional relationship between structure and function, as prenatal stimulative factors had been shown to impact the development of certain behavioral traits previously assumed to be predetermined.

3.2.2 Causal directionality and information flow

The bidirectionality hypothesis has important implications beyond behavioral development.

Received views of causal directionality in molecular biology assert that “‘information’ flows in only one direction from the genes to the structure of the proteins that the genes bring about through the formula DNA → RNA → Protein” (Gottlieb 2001, 47; see also Gottlieb 2002). This view is strictly unidirectional, making it the case that the genome functions as a distinct subsystem of the organism, effectively closed off from external influences. Gottlieb argues that there are problems with the unidirectionality of this account, asserting that “[g]enes express themselves appropriately only in responding to internally and externally generated stimulation”

(47). His definitive picture is one of a multi-level, non-reductive account of the reciprocal causal influences on organismal development that is at every stage bidirectional in terms of the flow of information.

This view, which cuts across developmental biology, molecular biology, and psychology, describes the structural development of the organism as being reciprocally related to functional factors such as experience. It is opposed to the view that even behaviors that appear strongly innate are predetermined as well as to the view that information from DNA flows in only one direction.

The bidirectional hypothesis motivates a holistic explanatory account of seemingly innate behavioral traits. The predetermined epigenesis view arguably black-boxes any extra-genetic factors that may impact genetic function, effectively restricting its domain to only the genetic level. Gottlieb’s probabilistic epigenesis view, in comparison, incorporates levels of biological organization beyond that of the gene in such a way that provides a richer account of the explanatory factors relevant to an understanding of behavioral development. Such an account undoubtedly suggests a much more complex picture of behavioral development. However, as an effect it also provides a way to synthesize findings across several cognate fields – namely developmental biology, molecular biology, and psychology – for a more detailed explanation of

target phenomena. Provided the bidirectional hypothesis is bolstered by robust empirical

findings, it can do significant work toward the epistemic goals of biology more generally,

namely explaining the mechanisms for species-typical behaviors, but also potentially

contributing toward an understanding of the generation of novel behaviors.

With an account of the bidirectionality of development outlined, two key interrelated concepts can be drawn out with the goal of illustrating the parallels between the two debates

discussed in this chapter and in the previous chapter. These two concepts are (1) causal

directionality and (2) information flow. Causal directionality refers to the relationship between

processes in organismal development. The bidirectional hypothesis states that these processes are

strongly reciprocal. The received view, in some areas in behavioral development and, according

to Gottlieb, in molecular biology, states that some subsystems of processes, such as genomic

activity, are effectively closed off from systems at the organismal scale and are distinctly

unidirectional. Information flow, as an application of causal directionality, refers to the way in

which “information” (in various forms) can influence certain processes and causal factors. For

example, on the received view, genetic information is said to flow in the direction toward protein

development, but not vice versa – information flow at this level is thus unidirectional.

The EES framework posits that the bidirectionality of influences across all levels is critical to understanding both development and evolution. This is illustrated best by the notion of reciprocal causation, a key theme of the EES. Recall that the concept of reciprocal causation illustrates how organisms can play an important role in their own development and evolution.

As Gottlieb argues, structure does not merely determine function in a unidirectional manner, but rather function can reciprocally determine structure – a duckling embryo’s exposure to its own vocalizations, stimulated by various mechanoreceptive and sensory influences, can aid in the development of the vocalization behavior itself. Information flows both ways and as a result can exert an influence on the development of behavioral traits. Under the theoretical framework of the MS, what is ultimately explanatorily relevant is the flow of information from the genome outwards. Phenomena such as niche construction may play a role in exposing organisms to certain environmental stimuli, or in guiding certain behaviors, but in terms of understanding the development and evolution of traits, the genes are doing the causal work. Standard evolutionary theory, then, can be seen as supporting a unidirectional hypothesis regarding the causal directionality of genomic activity lending to organismal development and species evolution, as it

places no strong explanatory value on potential causal factors beyond genes. They have causal primacy.

3.2.3 Shared commitments

The bidirectionality hypothesis shows up in cognitive science as well. Cognitivism and related views suggest that cognition can be understood locally: as operating laterally on a sublevel distinct from other levels (or, perhaps more accurately, biological scales) within the organism.

Cognitive function occurs unidirectionally within that level. For example, Adams and Aizawa

(2001) refer to cognitive processing as being ‘intracranial’ (in contrast to transcranial), suggesting that cognitive work is done within the bounds of the brain. This is not to say that input in the form of information from the environment has no causal bearing on the cognitive system, but rather that the source, mode of transmission, and variety of that information plays no significant role in explaining the phenomenon of cognition itself. What is relevant for the sake of understanding cognition, it is argued, are the mechanisms for the processing of that information.

On this view, information flows unidirectionally and in a feed-forward system: it is received as sensory input, processed according to a set of predetermined rules, and effectively transcribed into behavioral patterns.

Enactivism and related 4E cognition views, in contrast, strongly assert the bidirectionality of causal factors and of the flow of information. Feedback loops, illustrated most clearly by dynamical systems theory thinking on cognition, ascertain that information initially received by sensorimotor mechanisms is further modified by a variety of receptor types

(proprioceptive, interoceptive, and other sensory modalities) before being fed back into the sensorimotor system. Information can then be said to not only flow in loops but is also being

constantly modified through its interaction with other parts of the system. Additionally, 4E views

assert that cognition must be understood globally, in a holistic, non-reductive manner, due to the

explanatory relevance of factors at scales beyond the neuronal. Cognition is thus highly bidirectional across multiple scales and the emergence of certain phenomena can be understood as arising on a probabilistic basis rather than a predetermined, in line with Gottlieb’s proposed

dichotomy. Whether or not an organism exhibits a particular behavior will not be solely

dependent upon existent mechanisms in some manner inherited genetically but rather on whether

or not all of the factors that could potentially contribute to its instantiation are in place, including

environmental and developmental factors.

Despite their shared fundamental commitments to causal directionality and information flow, the relationship between the EES and 4E approaches to cognition remains underexplored.

Stotz (2014) touches on 4E cognition briefly in discussing varieties of evolutionary psychology,

suggesting that “an extended evolutionary theory is reciprocally related to the view of a cognitive

system as embodied, embedded, enacted, and extended, promoted recently by many proponents

in cognitive science” (Stotz 2014, 10). On Stotz’s view, 4E cognition perspectives are

complementary to EES approaches, as they both emphasize the role of the organism in its own

evolution in addition to its development. In contrast, cognitivist approaches that view cognition

as operating in a manner distinct from other physiological processes in the organism, and MS

approaches that posit the organism as a passive receiver of environmental input, are seen to be

complementary due to their shared asymmetry. Taking the whole organism as the unit of

analysis, an account of cognition that incorporates the organism’s physiological development

would need to identify a reciprocal relation between organism and environment. On this view

(and in line with developmental systems theory thinking), organisms do not follow a

predetermined developmental plan causally separated from the external environment, but rather

develop certain traits as a result of their environmental interactions, including those possible

through cognitive capacities such as experience and learning.

While both proponents of cognitivism and of traditional evolutionary biology would not deny that organismal behavior can have an impact on development at some stages and in some degrees, proponents of the 4E approach to cognition and of the EES place much greater emphasis on the causal role of these capacities at both the developmental and the evolutionary scale.

Incorporating the causal effects of learning behavior, for example, into the developmental and the evolutionary story for an organism and its species reveals valuable insights about the phenomenon of learning itself, how certain traits may be linked to learning behavior, and how learning can have a significant development and evolutionary impact on the organism.

Recall from chapter 1 that biological enactivism, as a theory of cognition, claims that organisms are cognizing systems structured by both their internal dynamics and their dynamic relations with environmental features corresponding to their sensorimotor capacities, developed as a result of their coupled interactions with their environments over both developmental and, on a population scale, evolutionary time. Built into the definition of biological enactivism, then, is the idea that the organism plays a role in its own development and evolution – and they do so in virtue of their structure as a cognizing system. As Stotz notes, “[t]he relationship between an active developing system with exploring cognitive abilities and its capacity to construct its own living environment—and that for its descendants—becomes immediately obvious. Since the organism-environment relationship creates selection pressures that will have an important feedback on the phenotype of future generations, the organism indirectly and partially controls its own evolution” (Stotz 2014, 11). On this view, an organism that can modify its environment

through behavioral means, or transmit information to others via communicative signaling,

transforms not only itself but also its selective pressures both individually and on a population

scale.

In Gottlieb’s duckling example, it is the ducking’s sensorimotor experience that helps to

shape its developmental trajectory. Stotz suggests that incorporating a concept of experience into

the study of developmental and evolutionary change “may help to bridge the gap between

learning and development by including all aspects of environmental stimuli that lead to long-

term adaptive changes of behavior” (5). In the sense used here, experience is defined as “the

contribution to development of the effects of stimulation from all available sources (external and

internal), including their functional trace effects surviving from earlier development” (5;

Schneirla 1957; see also Lehrman 1970). Stotz’s goal in introducing the concept of experience is

to highlight the inadequacies of a distinction between instinctual traits and learned traits as being

the only ways to interpret behavioral development.

Traditionally, innate behaviors are thought of as preprogrammed, automatic responses to

environmental stimuli. They are instantiated through a stimulus-response mechanism, with no mediating control input from the organism. Stotz suggests a different picture, where an organism’s individual experience plays a role in generating what seem like automatic responses.

Duckling vocalization behavior emerges in an early developmental stage and seemingly as a result of innate factors. However, as shown by Gottlieb and colleagues’ empirical results, ducklings whose abilities to hear their own vocalizations as well as their siblings’ vocalizations were removed while in the embryonic stage of development did not develop normal vocalization behavior. In other words, their early experience of vocalizations is an important factor in the normal development of their own vocalization behavior. The active exploration of the embryonic

environment via sensory capacities (in this case hearing) guided the development of what, under

competing paradigms, is considered a preprogrammed response, structured by the ‘reading off’

of the genotype alone. In this instance, a view of the organism as an active, cognizing explorer of

its environment, even in the earliest of developmental stages, can draw out new insights into

behavioral phenomena traditionally explained by appealing to automatic mechanisms.

Drawing attention to the distinction between innate and learned behaviors may help to shed light on why a mechanistic approach to animal intelligence has proliferated. Chittka (2017) argues that instinctual behaviors, namely in the case of honeybees and other eusocial insects, may not be as straightforward as previously thought. He notes that “very few behavioural routines are fully hardwired and even comb construction skills have to be partially learnt by honeybees. There are interactions between instinct and learned behaviour at multiple levels, and complex instincts can facilitate advanced learning behaviour” (Chittka 2017, R1050). Prior work by Lehrman (2001) on instinctual behaviors in rats and chicks corroborates a similar view in the case of development. Lehrman states, “analysis of the developmental process involved shows that the behavior patterns concerned are not unitary, autonomously developing things, but rather that they emerge ontogenetically in complex ways from the previously developed organization of the organism in a given setting” (Lehrman 2001, 29). Danchin et al. (2011) offer a related example that supports this point:

Ironically, social learning has been demonstrated to affect behaviours that biologists often consider as being under strong, if not exclusive, genetic control. For instance, the tendency of cockroaches to flee light and head towards darkness is usually considered to be genetically encoded because of its anti- predator selective advantages. However, a recent study showed that this presumed innate tendency may be at least partly acquired socially: cockroach-

like robots that were programmed to head towards a light shelter were sufficient to lead groups of cockroaches to often choose the light shelter when given the choice between a light and a dark shelter. (Danchin et al. 2011, 479)

This kind of research shows that a structure of complex interaction between several factors, including environmental stimuli, developmental processes, experience, behavior, and learning, is responsible for the appearance of certain behavioral traits. Seemingly innate behaviors may be in some sense ‘built from the ground up’ in the case of each organism, but are facilitated by a complex ontogenetic history, without which the trait in question may not be able to develop or may not develop normally.

This section identifies two interrelated concepts that help to draw out similarities between the MS and EES debates in evolutionary biology and the cognitivism and 4E cognition debates in cognitive science. These two concepts, causal directionality and information flow, illustrate how the EES and 4E cognition share a commitment to the bidirectionality of causal influences and the flow of information. In contrast, the MS and cognitivist views can be said to subscribe to a unidirectional hypothesis, where the organizational structure dictates that the flow of information does not involve causally relevant feedback mechanisms at the appropriate scale.

3.3 Cognition as evolutionary ‘bootstrapping’

So far, the focus in this chapter has been on more general claims about what qualifies as an evolutionary process and the causal roles that these processes play both in terms of development and evolution and in terms of the relations between them. This picture helps to set out some preliminary commitments which can now be looked at in more detail to investigate how cognition can play a role in development and evolution. This section argues for two claims about

cognition: first, that cognition facilitates some forms of non-genetic inheritance, making it a relevant factor in the development and evolution of certain traits, and second, that capacities that are cognitive in nature, such as behavior, learning, and experience, can work as evolutionary

‘bootstrapping’, initiating a causal trajectory that can potentially lead to genetic change and in some instances, what is referred to as the ‘genetic assimilation’ of phenotypic traits. In this way, cognition can be thought of as a driving force in evolutionary change.

While other roles for cognition in nature will be discussed in the remainder of this dissertation, I want to suggest that inheritance is a good place to start if we take as our

explanandum what it is that a capacity for cognition offers to organisms. While this is a much

larger question than this section, or indeed, this project, can take on, my hope is to provide a

starting point for addressing this question in the context of the intersection between evolutionary

biology and cognitive science. Godfrey-Smith (1996) suggests that the function of cognition is to allow organisms to deal with environmental complexity; my goal (here briefly but with more treatment in later chapters) is to build on this account by looking at how the flexibility of cognitive abilities initiates certain developmental and evolutionary trajectories in such a way that it can serve as a type of ‘fast and loose’ evolutionary process, with genetic change refining traits over a longer timescale. Research on factors that affect the speed of evolutionary change has been a topic of interest in the past several years, with cognitive abilities such as learning playing a prominent explanatory role (Aguilar, Bennati, and Helbing 2019; Sznajder et al. 2012).

Therefore, addressing how cognition can potentially initiate evolutionary trajectories can shed light on explanatory mechanisms across evolutionary timescales.

3.3.1 Inheritance beyond the gene

Recall from Section 3.1 that the EES posits mechanisms of non-genetic inheritance as evolutionary processes alongside genetic inheritance, as they have the potential to impact the evolution of certain traits. Mesoudi et al. (2013) define non-genetic inheritance as “the transmission of information across multiple generations of individuals through a mechanism other than DNA replication, such as cultural inheritance via social learning (e.g., imitation or language), epigenetic inheritance via epigenetic marks (e.g., methylation patterns of genes), or ecological/niche inheritance via the environment” (Mesoudi et al. 2013, 1). While accounts of each of these mechanisms serve to broaden understanding of the impact of inheritance systems beyond genes, ecological and cultural inheritance are especially worth investigating in the context specified in this chapter. This is because cognition, I want to suggest, can facilitate these forms of non-genetic inheritance through an organism’s capacity for cognitive abilities, such as behavior, learning, and experience.

Mechanisms of ecological inheritance arise through an organism’s interacting with its environment in such a way that the effects of their actions can have both developmental and evolutionary consequences for subsequent generations. One oft-cited example is the modification of soil constitution by populations of earthworms (Odling-Smee and Laland 2011; Jones et al.

1994; Odling-Smee et al. 2003; Darwin 1881[1966]). These activities generate an ecological niche that alters the evolutionary trajectory of not only the organisms modifying the niche, but of other organisms as well. In this way, an organism’s behavior can be said to be a contributing factor to the generation of ecological inheritances. As Danchin et al. (2011) note, earthworm- induced modification “has not only increased plant growth, as a result of the nutrient enrichment, but has also affected earthworm physiology. The well-mixed soil that results from earthworms’

burrowing activities makes it easier to absorb water and has allowed them to retain their

ancestral freshwater kidneys, rather than evolve novel adaptations to a terrestrial environment”

(Danchin et al. 2011, 478). The earthworm’s movement through the soil creates a specific kind

of environment that future generations then continue to modify with their own behavior. Each

generation inherits a niche modified by the previous generation, and those modifications can

bring about changes at both the developmental and evolutionary timescale (of multiple species,

as well).

The active process of an organism’s physical modification of its ecological niche is

defined by proponents of the EES as niche construction. They argue that niche construction is a

process that can lead to ecological inheritance (Odling-Smee and Laland 2011; Odling-Smee et

al. 2003). Individual organisms modify their environments through their metabolism, activities,

and choices (Odling-Smee and Laland 2011, 220). Their offspring then inherit those modified

environments and continue the chain of modification through their own actions. The modified

environment exerts modified selective pressures on future generations, thus leading to

inheritance via ecological conditions. Niche construction, as a mechanism for ecological

inheritance, “allows acquired characteristics to play a role in the evolutionary process, in a non-

Lamarkian fashion, by their influence on selective environments through niche construction”

(Laland, Odling-Smee, and Feldman 2001, 122). Dam building, for example, generates a novel environment into which future generations are born, in turn potentially modifying the selection pressures on that generation.

Characteristics with an evolutionary impact may be acquired through strictly behavioral means, as in the earthworm example, but also by learning and experience. The Galápogos woodpecker finch, for example, engages in grubbing behavior as true woodpeckers do, but

instead of using their bill to grub for insects, they use a tool: a cactus spine (Laland, Odling-

Smee, and Feldman 2001; Alcock 1972). There is an important difference between these two traits, although they achieve the same result. The woodpecker’s pointed bill is an adaptation, shaped by natural selection. The woodpecker finch’s tool use is not – “[r]ather, the finch, like

countless other species, exploits a more general and flexible adaptation, namely, the capacity to

learn, to develop the skills necessary to grub” (Laland, Odling-Smee, and Feldman 2001, 122).

Part of what is learned, in this example, is how to make use of environmental resources. This

activity has the potential to considerably impact the evolutionary trajectory of the species. As

Laland, Odling-Smee, and Feldman suggest, “[t]his behavior probably created a stable selection

pressure favoring a bill able to manipulate tools rather than the sharp, pointed bill and long

tongue characteristic of woodpeckers” (122). Traits associated with niche constructing abilities

may thus be stabilized through the regular behavioral activity of the finches to transmit cactus

spine use “knowledge” to subsequent generations.

Characteristics with an evolutionary impact can be acquired through more complex forms

of social learning and cultural transmission, even across different species. Another classic

example in this context is the learned behavior of opening milk bottles by different species of

British tits (Parus major, P. caerulus, and P. ater) (Fisher and Hinde 1949). The behavior was

first recorded in Southampton, UK in the mid-20th century, but over a relatively short period of

time had spread throughout the country and was even seen in Wales, Scotland, and Ireland.

Fisher and Hinde describe the behavior in detail:

The method of opening employed varies greatly. When the milk bottle is closed by a cap of metal foil the bird usually first punctures the cap by hammering with its beak and then tears off the metal in thin strips. Sometimes the whole cap is removed, sometimes only a small hole is made in it. Cardboard caps may be

treated in a variety of ways. The whole top may be removed, or only the press- in centre, or the cardboard may be torn off layer by layer until it is thin enough for a small hole to be made in it; the milk may be taken through this hole or the bird may insert its beak in the hole and flick off the remainder of the top. The records show that several different methods may be used in any one district, and that more than one method may be employed by one individual. (Fisher and Hinde 1949, 354)

Fisher and Hinde conclude that “[t]he distribution of the records is consistent with the view that this source of food was actually discovered de novo by only a small population of the tit population, and was then passed on in some way to other individuals” (351). The gradual increase in the geographical distribution of the milk bottle opening behavior supports the view that “when the habit has been acquired by one tit, it can then be spread through the population by some form of imitation or learning” (353). The variation in technique outlined above suggests that individuals were not merely imitating an initial observed technique but were engaging in active exploration of possible techniques, illustrating a significant learning ability available not only to conspecifics but to members of other species as well. Given the relatively rapid spread of milk bottle opening behavior, social learning seems a more likely explanation than numerous instances of individual ingenuity4. In terms of the possible developmental and evolutionary impact, Fisher and Hinde note that the behavior is seen more commonly in the winter months

(354), and they hypothesize that this is partly due to an increased need for fats during the winter.

A potential regular increase in nutrition can lend toward greater survival rates of milk drinkers, and can in turn increase the probability of successful offspring. Milk bottle opening behavior,

4 Lefebvre (1995) suggests that quantitative models support the hypothesis that bottle opening behavior arose several times independently rather than from a single originator, but does note that social influences then accelerated the spread of the behavior. 62

induced through social learning, can be seen as an effect of cultural transmission that can then lead to cultural inheritance, which has, in turn, relevant evolutionary consequences.

While there is some overlap of influences in cultural inheritance and ecological inheritance, the two processes function through different mechanisms and can therefore be said to be distinct forms of inheritance. Odling-Smee and Laland (2011) partition out ecological inheritance and cultural inheritance as being actualized through different forms of niche construction, namely in the case of humans. They suggest that cultural niche construction be thought of as a “subset of niche construction that is the expression of culturally learned and transmitted knowledge” (226). The two systems may differ, then, in the type of information that is transmitted both between and within generations. Ecological inheritance supposes the transmission of information about environmental cues, such as soil composition or resource utilization, while cultural inheritance supposes the transmission of socially attained information, such as social behavioral repertoires like mating rituals.

Despite their distinctions, cultural and ecological inheritance have an important factor in common: they both employ cognitive capacities in the transmission of information. Cultural niche construction requires social learning and experience, with language as a canonical example in the case of humans. In ecological niche construction, information is transmitted through the environment rather than socially, but that information is produced through the behavior of organisms. There is a sense in which both mechanisms may ultimately collapse into each other, as it is undoubtedly the case that other organisms are indeed part of an organism’s environment, and are therefore relevant parts of its niche. So the differentiation does not necessarily carve up distinct modes of interacting with the world, but rather helps researchers to identify particular mechanisms for transmission of particular kinds of information. The earthworm’s movement

through the soil invites more research on the chemical and molecular relationship between earthworm bodies and soil composition, while the tits’ bottle opening behavior invites more research on the spread of information through observation and behavior (i.e., perception and action) as processes of social learning.

3.3.2 Plasticity and the shaping of evolutionary trajectories

Processes of non-genetic inheritance, such as niche construction and cultural transmission, regularly involve the utilization of cognitive abilities, namely behavior, learning, experience, and cultural knowledge. And in cases where certain behaviors seem innate, such as in the duckling vocalization example, the active exploration by an organism of its environment can seemingly play an important role in generating what is typically conceived of as innate behavioral patterns.

These phenomena invite the question of how exactly cognition itself can serve as a driver of developmental and evolutionary change. If organisms regularly exercise their cognitive abilities in such a way that has both a considerable developmental and evolutionary impact, then the link between cognition and evolution is worth investigating further. One way to approach this question is to look at the relationship between modifications via cognitive capacities and genetic change over time. In order to make sense of this relationship, it is first necessary to look at how genetic change can possibly be brought about via epigenetic means.

The idea that genetic change can be brought about through means other than selection and mutation has a complex history within evolutionary biology. James Mark Baldwin, a comparative developmental psychologist, suggested in his 1902 book Development and

Evolution that individual differences in adaptability could potentially lead to genetic change.

Some organisms are more successful than others at adapting to a changing environment. Baldwin

held that “adaptations become congenital by further progress and refinement of variation in the

same lines of function as those which their acquisition by the individual called into play”

(Baldwin 1902, 98). The hypothesis that individual differences in adaptability can influence the

evolutionary trajectory of subsequent generations came to be known as the Baldwin effect.

Conrad Waddington, an evolutionary biologist working in the mid-20th century,

achieved a novel and intriguing experimental result that arguably lent credibility to the Baldwin

effect. Waddington exposed Drosophila pupae to heat shock during an early developmental

stage; this caused them to develop wings with few or no crossveins (Gottlieb 2002). He then bred

the crossvein-less individuals, repeating the heat shock to each new generation during early

development. After fourteen generations, some individuals displayed the crossveinless trait despite not yet being exposed to the heat shock. The crossveinless phenotype no longer required the environmental stimuli, the heat shock, in order to appear - it had been genetically assimilated.

Waddington’s results provided support for the Baldwin effect, that individual phenotypes have the potential to become reconstructed genetically.

Interpretations of the work that were in line with evolutionary thinking at the time proved challenging, however. Noble (2015a) explains the difficulty in assessing the results on a gene-

centric view:

What was happening at the gene level in Waddington’s experiments? A standard Neo-Darwinist explanation might be that some mutations occurred. That is possible, but extremely unlikely on the time scale of the experiment, which was only a few generations. Moreover, random mutations would occur in individuals, not in a whole group. Single small mutations would have taken very many generations to spread through whole populations, and many such mutations would have been required. (Noble 2015a, 816)

It was clear that a change had occurred, but the causal story proved more complex than standard evolutionary biology thinking at the time would suggest. Rather than relying upon random mutation, it picked out a feature of organismal structure that had yet to be fully explored: the potential of underlying developmental plasticity. As Noble suggests, Waddington’s experiments

“exploited plasticity that is already present in the population. That strongly suggests that all the alleles (gene variants) necessary for the inheritance of the characteristic were already present in the population, but not initially in any particular individuals in the correct combination. The experiment simply brings them together” (816). Importantly, genetic assimilation does not entail that phenotypes are transmitted generationally without any genetic change. The trait in question must exhibit plasticity – there must be an underlying genetic structure in place that, under the right conditions, allows for that trait to develop.5

5 As Gottlieb (2002) points out, this is why experiments by August Weismann arguably failed to discredit the idea that individual differences could be inherited. Weismann cut off the tails of mice and bred them for several generations. No individual in any generation was born with a shorter tail, and Weismann concluded that the acquired trait (a short/absence of tail) could not be genetically inherited. The problem is that the shortened tail does not qualify as a trait subject to plasticity in the way required for genetic assimilation, so will not be heritable under that mechanism. 66

Waddington’s findings, and the notion of genetic assimilation, did not make their way into the

MS, presumably due to their reliance upon a notion of non-genetic inheritance, which was at

odds with the MS’s emphasis on genetic causal primacy.

As Gottlieb (2002) points out, this historical detail is complicated further by Ivan

Ivanovich Schmalhausen’s role as a contributor to the MS6. Schmalhausen conceptualized

evolution “as a process wherein favorable adaptabilities instigated by the environment eventually

became genetically assimilated adaptations, in the sense of moving from external dependency to

complete internal control” (Gottlieb 2002, 126). What was crucial for this view was the role of

individual development in bringing about adaptations. As individuals face dynamic influences

and stressors from the environment, they are shaped in ways that can potentially have a genetic

impact via the activation of dormant or ‘hidden’ gene variants responsible for phenotypic plasticity.

As seen in the case of Waddington’s Drosophila, the crossvein-less trait acquired by

individual organisms as a result of experiencing environmental heat shock transformed from

being instantiated via environmental factors to being genetically instantiated in a matter of

fourteen generations. The genetic assimilation of the crossvein-less phenotype, originally a

characteristic acquired during development and brought on by environmental forces, meant the

trait was now an adaptation (albeit a maladaptive one)7 rather than simply an acquired

phenotype.

6 Historically, Waddington’s experiments came only a few years after Schmalhausen’s work, yet neither biologist’s research impacted the foundations of the MS. Presumably this was due to difficulties in incorporating the notion of genetic assimilation into the strongly gene-centric framework of the developing MS, but the question remains as to whether there is an underlying fundamental incompatibility or if there is room in the MS for genetic assimilation. 7 Gottlieb (2002) references another experiment by Waddington showing the development of an adaptive trait. Drosophila exposed to salty environments developed over several generations enlarged anal papillae which aid in

Arguments for an extended interpretation of biological theory give a prominent role to individual development in shaping adaptations, and thus to the possibility for genetic assimilation of phenotypes. Laland et al. (2015) hypothesize: “[t]he causal description of an evolutionary change may, for instance, begin with developmental plasticity or niche construction, with genetic change following. The resulting network of processes provides a considerably more complex account of evolutionary mechanisms than traditionally recognized”

(8). Rather than starting at the place where genetic change occurs through selection, this approach suggests working with an account of evolution that starts at the structure of interactions and conditions that bring about certain individual differences. These factors, it is argued, are relevant to the causal story, and investigation of them can shed new light on how and why certain traits evolve.

Empirical work on how cognitive capacities give rise to genetic assimilation of traits is limited. Indeed, the concept of genetic assimilation of phenotypic traits has only recently been given empirical treatment; Aguilar, Bennati, and Helbing (2019) note that “[o]nly from the mid-

2000s did the Baldwin effect start taking ground in the field of Evolutionary Biology” (Aguilar,

Bennati, and Helbing 2019, 2). The aim of the present project is to motivate a need for further research in the hopes of shedding light on the possible role of cognition in bringing about genetic change.

One way forward is a focus on how traits developed via learning mechanisms can lead to genetic change. Insights from recent research can also show how the capacity for learning itself evolved and continues to evolve across species. Avital and Jablonka (2000) have suggested that

processing excess salt. This trait was eventually genetically assimilated, no longer requiring exposure to the salty environment to appear. 68

particular structures of interactions involving learning have potentially led to the evolution of advanced learning behaviors. Their assimilate-stretch principle supports the claim that “part of a behavorial sequence that formerly depended heavily on learning is genetically assimilated, and this allows a new learned element to be added” (Jablonka and Lamb 2005, 291). This strategy has the potential to lead to increased fitness, as organisms can effectively ‘offload’ behavioral work onto a genetic mechanism for the same element, thus ‘freeing up’ cognitive space for other adaptive behaviors or skills to develop. As more parts of the behavioral sequence are offloaded genetically, the capacity for learning is increased, as fewer sequences have to be effectively built from the ground up via complex cognitive and behavioral abilities, and organisms can spend more learning-oriented resources on novel behaviors that aid them in dealing with environmental stressors and changes.

One oft-cited example illustrating evolutionary change brought on by human culture is the evolution of lactase enzyme production into adulthood in milk drinking populations

(Ranciaro et al. 2014). Individuals in cultures that historically practiced cattle domestication for milk production have retained the ability to process lactose as a genetically determined phenotype, whereas individuals in cultures without cattle domestication do not retain this ability.

In this instance, the cultural practice of cattle domestication (and milk drinking into adulthood) led to the formation of a corresponding genotype.

3.3.3 Speeding up evolution

The empirical and theoretical work outlined in the above section supports the claim that phenotypic plasticity can initiate certain evolutionary changes. But some researchers have also considered how phenotypic plasticity might influence the speed of evolution in addition to its

trajectories. On this suggestion, the development of some phenotypic traits might allow organisms to respond more quickly to dynamic environmental pressures in such a way that results in the speeding up of evolutionary change in comparison to the speed at which change is brought on by mutation or selection. For example, Danchin et al. (2011) note that formal models

“suggest that culture can increase an organism’s fitness (that is, its chances of survival and reproduction) under a wide range of conditions by allowing individuals to acquire and transmit adaptive behaviour that they could not have acquired by costly ‘trial-and-error’ learning and by providing a faster means of adapting to rapid environmental change than genetic inheritance alone” (Danchin et al. 2011, 479). In this way, phenotypic plasticity - here, in the form of social learning - may allow an organism to respond more effectively to environmental changes, ensuring that subsequent generations maintain an appropriate degree of fitness in a novel environment.

As an effect, learning, as a form of phenotypic plasticity, can initiate ‘fast-and-loose’ evolutionary trajectories, that are then potentially ‘fine-tuned’ via genetically assimilation, provided the phenotype has been maintained for a number of generations. Laland et al. suggest this framing invites the thesis that “much of adaptive evolutionary change may have its origin in plastic responses to novel environments, later followed by genetic changes that stabilize and fine-tune those phenotypes, rather than the other way around” (Laland et al. 2012, 4). In line with

Godfrey-Smith’s environmental complexity hypothesis, learning serves as a way for organisms to deal with unstable, changing environments, but in a way that goes beyond behavioral modification – it allows them to manage their degree of fitness for future generations by rapidly adapting to those environments.

Several conclusions can be drawn from work on genetic assimilation and phenotypic plasticity. The first is that the underlying potential for developmental plasticity and the influence that resultant traits have on an evolutionary scale is important for a story of genetic change.

Avoiding a charge of Lamarckianism is possible on this account, as the causal story begins with the activation of potential ‘hidden’ or dormant genetic traits, rather than relying upon inheriting phenotypes directly from the previous generations’ phenotypes. As Jablonka and Lamb (2005) stress, genetic assimilation experiments “show how, when faced with an environmental challenge, induced developmental changes unmask already existing genetic variation, which can then be captured by natural selection. Short-term evolution does not depend on new mutations, but it does depend on epigenetic changes that unveil the genetic variants already present in the population” (Jablonka and Lamb 2005, 274)8.

The second conclusion is that in order to explain the origin of some characteristics, a comprehensive causal story of phenotypic change may entail incorporating individual developmental history in addition to genetic change. Drosophila experiencing a hot environment prompted a change in phenotype, eventually leading to genetic change. An account that leaves out the initial influences of the change in genotype confines evolutionary biology to the realm of the gene, effectively black-boxing important causal factors, the investigation of which can shed light on the generation of novel traits.

Lastly, the broadest conclusion but the one with the greatest impact on the work being done in this chapter, is that in many instances, the fact that an individual organism’s significant interaction with particular environmental features is what helps to shape particular genotypes

8 Another part of the story here is ancestral genetic potentialities. Gottlieb (2002) draws attention to an experiment done on chick embryos resulting in “phenotypes of enameled dentition – otherwise known as teeth!” (154) – possible only because of underlying genetic potential resulting from reptilian ancestry. 71

means that its ability to interact with the world plays a vital role in its developmental and evolutionary story. What cognitive capacities do is allow the organism to be meaningfully influenced by its environment, and meaningfully influence it in return. Behavior, experience, and learning (both social learning and trial-and-error learning), as the canonical examples of cognition employed here, are all modes through which an organism can interact with its world.

Ultimately, I want to suggest, an account of cognition that emphasizes the dynamic relationship between organism and environment is an appropriate approach to take when considering how cognition facilitates the evolution of phenotypes and the causal factors of genetic change.

Certainly, further research is needed on the link between cognitive capacities and genetic assimilation of traits, but it is an area rich with potential for both evolutionary biology and cognitive science, and the integration of the two fields.

One common factor among some systems of non-genetic inheritance is that they are, in many cases, highly dependent upon the behavior or experience of individual organisms. Niche construction is exemplified by the actual niche constructing behaviors of organisms, such as the physical act of building a dam by beavers. Cultural transmission occurs through the learned behaviors of organisms that either experience others behaving in a certain manner or figure out for themselves through trial and error how to attain some resource, with both possibilities illustrated in the milk bottle opening case. Notably, relevant behaviors are not merely innate traits, nor are they simple stimulus and response reactions. Rather they are complex chains of behaviors, often deliberate, that result from a web of interactions with environmental resources, conspecifics, and sometimes members of other species, usually within a specific period in the organism’s life.

Despite the dependence upon these highly context-specific, complex behaviors, there is little talk of cognition itself playing a role in the theory behind systems of non-genetic inheritance. One reason for this may be due to the received view of cognition being one that investigates cognitive capacities locally – often at the neuronal scale, or through mechanistic metaphors intended to extend no further than the brain. Understanding an organism’s cognitive processing might not matter so much as understanding its outward behavior, so a behaviorist approach may seem to suffice in terms of how deep into psychology evolutionary biologists need to go. Enactivism and related 4E views, however, offer a picture of cognition that cuts across several scales of biological organization and considers reciprocal effects at both developmental and evolutionary timescales. This picture gives evolutionary biologists more tools at their disposal – they can look to cognition for relevant influences in understanding the causal history of certain evolutionary trajectories. Therefore, the integration of work from 4E approaches to cognition into evolutionary biology can reveal valuable insights about how cognition, as a facet of organismal life, can drive evolutionary change.

Chapter 4: The Organization of Cognizing Systems

Introduction

In chapter 1, I suggested that incorporating key concepts from developmental systems theory into

an enactive approach to cognition results in a richer framework that is well suited for understanding biological cognition. Specifically, while the autopoietic variety of enactivism has the conceptual tools to explain how living systems are organized, insights from developmental systems theory can show how organisms, as living systems, develop not merely as a result of genetic unfolding but rather through a web of complex interactive processes between organisms and their environments. According to autopoietic enactivism, cognition emerges as a fundamental feature of living systems structurally coupled with their environments. Because organisms are always situated in an environment and are dependent upon environmental resources to sustain them, they must engage in some form of activity in order to make use of those resources and maintain themselves as autopoietic systems. On an enactive account, this engaged activity constitutes the beginnings of cognition, which can then be fleshed out by specifying the ways in which organisms undergo such activity (for example, by looking at behavioral traits in organisms with a particular structural makeup).

Developmental systems theory (henceforth DST) follows a similar path in describing the organization of living systems and in explaining how those systems arise in nature, albeit without the goal of providing an account of cognition. DST defines organisms as developmental

processes, emphasizing the reciprocal nature of organism and environment9. It views the organism-environment system as not merely being a dyad consisting of two discrete entities, but rather as constituting a developmental system comprised of processes that extend over many interwoven biological scales. This conceptualization of living systems thus rejects the traditional view of organisms as passive, discrete entities on which the environment acts, and additionally invites the view that the organism both actively shapes and is shaped by its environment.

In this chapter, I pick up on the project proposed in chapter 1 by further refining a specification of the organization of living systems, drawing on theory from developmental biology and enactive cognitive science in order to show how concepts from both areas constitute a rich account of what a cognizing system is, how it arises, and how it functions. In section 4.1, I describe scientific and philosophical research showing that organisms are self-organizing, autocatalytic, dissipative developmental systems. This characterization of organisms requires that they be thought of as developmental processes contingent upon interaction with environmental resources rather than as individual entities on which the environment acts (Weber and Depew 2001). In section 4.2, I suggest that this characterization makes it clear how organisms are fundamentally plastic. As open systems situated in dynamic environments, organisms employ a tactic of flexibility in order to maintain themselves in the face of perturbations and to persist over time. Some authors (Griffiths and Gray 2005) have suggested that DST and evolutionary developmental biology, or evo-devo, share important theoretical commitments, and can in turn inform each other. I suggest that one area of integration concerns the mechanisms that act on plasticity as being under varying degrees of control. In section 4.3, I

9 Recall from Chapter 3 the importance of reciprocal causation for both DST and enactivism; on both accounts the dynamic interaction between organisms and their environments is a central part of understanding living systems. 75

argue that plasticity is acted upon through structural modification. Modulation and assessment,

as forms of structural modification10, can be thought of as forming the basis for cognitive acts.

Modulating coupling with an environment – with variations in modulated responses possible due to plasticity – is a capacity that serves as a precondition for cognition, as it requires exploration of and active response to the larger developmental system. Instances of structural modification result in organismal variation, as each organism’s responses will be dependent upon its own developmental life history, especially its learning experiences and access to resources.

New variants in developmental resources require flexibility in response; those organisms that are able to exploit those resources may discover adaptive responses. Variation therefore arises as a result of modifications to the coupling between individual organisms and their environments.

Thus, this chapter focuses on describing the organization of living systems and how cognition plays a role in modulating responses to variables within those systems, along the way illustrating how cognition can guide variation by initiating modifications in structural coupling.

4.1 The organization of living systems

It is standard practice to characterize living beings by the sort of things that they do – they take

in resources from the environment, convert those resources into energy, release waste products,

produce new cells, move around in and respond to their environment, and reproduce by various

means. This characterization tells us what activities all living things engage in. In studying living

things, we are also interested in how it is that they are able to engage in these activities. In other

words, we might want to know: in virtue of what can living things take in resources, make use of

10 Behavioral modification is an important aspect of cognition as well. The focus in this chapter will be on structural modification while the next chapter will deal more directly with behavioral modification. 76

them, and release them as waste? This is a question about the organization of living things. In order to understand what distinguishes living things from non-living things, we can investigate how they are organized so as to be able to do the sort of things that they do. In this section, I make use of the enactive concepts of autopoiesis as self-organization and autonomy as operational closure to describe the organizational makeup of living things. I then bring the enactive concept of structural coupling together with the DST concept of a developmental system to show how systems organized in this way develop over individual life cycles.

Autopoietic enactivists (Varela, Maturana and Uribe 1974; Maturana and Varela 1987;

Varela, Thompson and Rosch 1991; Thompson 2007) have suggested that a particular kind of self-organization is a defining feature of living things. On their view, the first step to building a living being is putting together a system that is capable of maintaining itself over time. For both single-celled and multicellular life, living beings are organized in such a way that allows them to persist as an entity. In this section, I spell out how to think about this type of organization in terms of organismal life. I suggest that some additional features addressed in the literature on self-organization and complexity science, namely autocatalysis and dissipative processes, are key to understanding the self-organizing nature of living systems. From there, it is possible to look at how these systems function on a developmental timescale. Organizational language from developmental systems theory can then be tied in to make it clear how living beings, as particular kinds of self-organizing systems, undergo change and variation11 over both spatial and temporal scales.

11 I partition these out in accordance with Oyama (2000). 77

4.1.1 Self-organization as autopoiesis

Early enactive work (Varela, Maturana and Uribe 1974; Maturana and Varela 1987) provided the foundations for understanding the organization of living beings. On this view, a defining feature of living beings is that they are continually self-producing – they are structured such that they are able to maintain themselves as a material entity that persists over time. Maturana and Varela refer to this feature as autopoiesis (from Greek auto-, self, and poiesis, production). Autopoietic systems are specified as networks of processes with certain relations that enable the system to persist. If these relations fail to hold, the system will necessarily disintegrate (Varela, Maturana and Uribe 1974). The canonical example in this body of work is the cell. A cell can be conceived of as autopoietic system due to the way in which the relationship between its internal processes enables the system to persist through time as a unity:

It is a network of chemical reactions which produce molecules such that (i) through their interactions generate and participate recursively in the same network of reactions which produced them, and (ii) realize the cell as a material entity. Thus the cell as a physical unity, topographically and operationally separable from the background, remains as such only insofar as this organization is continuously realized under permanent turnover of matter, regardless of its changes in form and specificity of its constitutive chemical reactions. (Varela, Maturana and Uribe 1974, 188)

Here the environment is specified as merely the background in which the physical unity that is the cell is contrasted. The cell, as an autopoietic system, is ‘operationally separable’ in that it undergoes a particular set of reactions that effectively form an operationally closed network.

Additionally, the cell’s membrane constitutes a boundary that distinguishes it as an entity from

its environmental setting. The cell may change in form and constitution, for example by taking in new resources and expelling old ones, but its organizational makeup remains the same12.

This organizational pattern can generally be found at the scale of larger organisms such as animals. Although these organisms may vary in structural form (both between and within individuals), they are organized in the same self-producing manner, in that they are “internally self-constructive in such a way as to regulate actively their interactions with their environments”

(Thompson and Stapleton 2009, 24). In other words, organisms are endowed with the ability to self-regulate so as to maintain their internal dynamics in a state of homeostatic equilibrium in the face of both internal and external perturbations.

The relation between self-regulatory processes, which are necessary in order for the system to persist, is captured in the enactive term ‘operational closure’. Operational closure describes the “network of processes whose activity produces and sustains the very elements that constitute the network” (Di Paolo et al. 2017, 112). An artifact of this organizational property is that it specifies an environmental state-space as well, described as the features which are operationally external to the organism’s self-regulatory capacities such that they are not necessary for the operational closure of the organism as a living system, though they may be necessary for its persisting over time.

A distinction can thus be drawn between the organism as an operationally closed entity and its environment. An organism as a self-organizing, autopoietic living system is set against the backdrop of an environment, yet the organism exists as a unity distinguishable from the

12 In this context autopoiesis captures metabolic self-production – it specifies the type of chemical reactions necessary for a living entity to maintain itself over time. For Maturana and Varela, metabolic processes are central to a conceptualization of life, as they constitute the ‘dynamic transformations’ that enable a living system to persist (Maturana and Varela 1987, 46), and as a result form a membrane that serves as a spatial boundary for an individual cell. 79

environment. This property of the organism is referred to as its autonomy13, because it specifies

the organism as a system that is “composed of processes that generate and sustain that system as

a unity” (24). Because the internal, self-regulatory dynamics of the autonomous system are

necessary for its persisting as a unity, it can be said to be operationally closed in the same

manner that cells are. Importantly, as Thompson and Stapleton point out, “operational closure

does not imply that conditions not belonging to the system cannot also be necessary” (24).

Living systems are thermodynamically open, such that they undergo processes to regulate the

flow of energy both between them (from the environment into the system) and within them (as

regulatory processes internal to the system). Thus, individual organisms, as living systems, are a

particular kind of self-organizing system – the autopoietic kind – and they exhibit autonomy as a

result of their organizational makeup and regulatory (namely, energetic) needs.14

4.1.2 Order from chaos: autocatalysis and dissipative structures

Living beings need sources of energy to function. By being thermodynamically open, they can take in resources from the environment and convert them into energy, which can then be used to maintain their integrity as an entity. Through conversion of energy resources, living systems produce waste, which can then be utilized by other systems. This exchange of materials highlights how living systems are in constant flux, with energy stores depleting and being restored continuously throughout their life cycles. This feature has been captured in a variety of sets of terms, such as Bertanlanffy’s distinction between ‘flow’ and ‘balance’ and Prigogine’s

13 This is not to suggest that organisms have autonomy in the way that rational agents have autonomy but simply that autonomous agents operate in accordance with their own self-generated processes. 14 For ease of reading, I will mostly refer to autopoietic, autonomous systems as self-organizing living systems, taking autopoiesis and autonomy to be defining features of living systems. Di Paolo et al. (2017) describe agents as being self-individuating, but considering some potential agents may be made up of multiple individuals (i.e., ant colonies), this term may raise some issues. 80

distinction between ‘dissipative’ and ‘structure’ (Bertanlanffy 1950; Prigogine 1967; see also

Capra 1996). These coexisting tendencies together form a description of the nature of open

systems – they highlight how open systems engage in material exchange in a bidirectional

manner as a feature that is distinct to them.

Prigogine’s analysis of living systems as moving between dissipative states and structural

states is particularly salient in the context here. A system is in a dissipative state if it is

destabilizing in the absence of new energy resources flowing into the system. Conversely, a

system is in a structural state if it is regenerating as a result of an influx of energy resources. A

system at a state of thermodynamic equilibrium is stable, with no further exchange of energy.

Living beings are, as highlighted above, in constant flux, are chemically unstable, and in self-

maintenance are in continuous exchange of resources in internal processes and external ones.

Therefore, they cannot, operationally, exist at thermodynamic equilibrium. As Capra notes:

A living organism is characterized by continual flow and change in its metabolism, involving thousands of chemical reactions. Chemical and thermal equilibrium exists when all these processes come to a halt. In other words, an organism in equilibrium is a dead organism. Living organisms continually maintain themselves in a state far from equilibrium, which is the state of life. (Capra 1996, 181)

Organisms are thus continually in motion, oscillating between dissipative and structural states in a cyclical manner, as new resources are taken in and used resources are expelled as waste. And it is through this cyclical motion that order in the sense of organismal form is created. Being far from thermodynamic equilibrium, dissipative states occur regularly in living systems as resources are used up through normal activity. A dissipating system is a failing system, however,

and so living systems must continually work to replenish resources and keep from breaking

down completely.

Prigogine thus refers to organisms as “dissipative structures” (Prigogine 1980). One

unique feature of organisms as dissipative structures is that they employ a certain method for

achieving efficiency in the face of dissipation. Autocatalysis describes a type of chemical

reaction that results from reaction products being utilized in the same catalytic reaction. Weber

and Depew (2001) provide a detailed description of how organisms make use of autocatalysis in

virtue of their dissipative structure:

The inherent tendency of dissipative structures to increase, even to maximize, their dissipative rate is linked with their ability to build better dissipative pathways, in the form of more efficient internal structures, which enable them to make other entities pay their entropic debts. The most effective way of building such structures in chemical systems is by means of autocatalysis. A chemical reaction that produces a substance that can facilitate the production of more of the original reactant will show a rapid amplification of the concentration of that substance. This is called an autocatalytic cycle. Organisms are chock full of autocatalytic cycling – indeed, within the environments to which they are deeply coupled, they simply are autocatalytic systems of a certain sort. (Weber and Depew 2001, 242-43, original italics)

Weber and Depew note that this organizational pattern of chemical activity is tightly linked with the same system exhibiting a dissipative structure, and that it follows from the notion of self-

organization as a kind of self-production that autocatalysis and dissipative states should be seen

in living systems as self-organizing systems. Put more concisely, if organisms are self-

producing, whatever mechanisms they employ must enable them to maintain themselves as a

unity far from thermodynamic equilibrium (where the system will no longer run). Naturally they

are dissipative structures, because of the fact that they are thermodynamically open, with energy regularly flowing out of the system. Autocatalysis serves as a way to efficiently maintain this organization.

As dissipative, autocatalytic systems, living beings are in a constant state of flux. They maintain themselves as a unity in virtue of their organizational makeup, which can be conceived of as a complex network of interacting processes, with enabling relations between certain processes essential for the system’s integrity. They are operationally closed in virtue of their capacity for generating and maintaining their own internal processes. At the same time, they are thermodynamically open, taking in resources from the environment and releasing consumed resources into the environment. These processes are repeated throughout the life cycle of the system, until a point when they can no longer function properly, and the system dissipates fully.

4.1.3 Self-organization over ontogenetic time

Naturally, organisms change form as they develop. They undergo cell differentiation and maturation as a result of interaction between developmental resources over time. Organisms show variation as well, with individual differences arising in members of the same species.

Developmental systems theory (DST) (Oyama 2000 [first edition 1985]; Oyama, Griffiths and

Gray 2001) was formulated as a response to commonplace views on development that were seen as failing to capture the complexities of ontogenetic processes. Instead of viewing the organism as a template onto which genetic information is read out, DST conceives of organisms as developmental processes in themselves. As Weber and Depew note, DST’s proponents describe organisms as “self-organizing processes rather than as discrete, hard entities on which ‘forces’ impinge” (Weber and Depew 2001, 242). The larger web of interactions, in which the organism

(as a developmental process) is embedded, is defined as a developmental system. A

developmental system is comprised of a matrix of resources, processes, and interactions between

developmental features. All organisms have a unique ontogenetic history that describes how they

came to take on the form that they do, as a result of the utilization of developmental resources

both internal and external to the organism itself. Here I suggest that, despite their structural

differences, what is common to them is their organizational makeup – as self-organizing systems

– and the fact that they are necessarily embedded within a developmental system.

In their early book The Tree of Knowledge, Maturana and Varela distinguish between

structure as the individual formation of a living system and organization as specifying the types

of relations necessary for a system to be classified as a particular kind of system (Maturana and

Varela 1987, 47). For example, a keyboard’s organization is defined by its capacity for typing,

but whether the keyboard is made of plastic, metal, projected onto a surface, and so on, does not

matter for its organization. Individual keyboards may differ in their structural (here, material)

makeup, but fall under the same classification of ‘keyboard’. In this sense, structure realizes

organization; individual systems or entities serve as realizations of kinds of organization15.

Organisms, as living systems, share their organizational makeup as self-organizing systems that are thermodynamically open but operationally closed. But they realize this organization in many different ways. In other words, they are all unique in their structural makeup, while being similar in their organizational makeup. Why organisms vary structurally –

15 This distinction may invite more problems than it solves, namely in the case of invoking a realization relation between systems. It may also be counterintuitive, with it seeming more natural to imagine the organizational makeup of a system changing in individual circumstances rather than its overarching structure. For our purposes, the distinction serves to merely point out how living systems across the board share a fundamental architectural makeup while differing in their formation. The goal is to pick out what a human and a bee have in common qua living systems; for Maturana and Varela it is their organization specifically as autopoietic organization. 84

even between individuals – is a question that follows naturally from an explanation of how they

are the same organizationally. So is a question of how they undergo often dramatic changes in

form, from conception up until maturation (and indeed, beyond to old age). Self-organization is a

constant feature throughout each stage in an organism’s ontogeny, though alterations to form

may be extensive, such as in metamorphic shifts.

Differentiation in structure, on both a DST account and an enactive account, is due to the

intimate coupling between organisms as living systems and their larger developmental systems.

Maturana and Varela define structural coupling as “a history of recurrent interactions leading to

the structural congruence between two (or more) systems” (Maturana and Varela 1987, 75). An

organism’s ontogenetic history can be thought of as mapping onto a unique pattern of

interactions between processes in a developmental system. Some of these interactions may be

highly conserved, occurring with regular frequency. Maturana and Varela offer the example of

the regular active transport of certain ions through a cell membrane such that the cells can in turn

interact regularly with these ions (76). Thus, some coupled interactions may occur frequently in

nature, across a variety of structural forms.

Structural coupling, importantly, scales up to the whole organism as well, and here I

suggest that the concept can help to make sense of what both enactivists and proponents of DST

mean when they talk about organisms and their environments being tightly interwoven. A

species-typical ontogeny requires structural coupling between the individual organism as a

subsystem and its developmental system comprised of developmental resources reliably

recurrent for that species16. Variation occurs when there are shifts in structural coupling, as a

16 A large body of research shows how temperature plays a direct role in species-typical development. One classic example is how temperature variation affects normal eye development in Drosophila (Foster and Suzuki 1970). 85

result of new resources entering the system, extant ones being removed, and differentiation in

how individuals interact with resources, with other sources of variation being possible as well.

Significant variation reoccurring in multiple generations results in evolution; similarly, new

patterns of coupling (such as a new or drastic shift in ecological niche) can result in speciation

events. Conceiving of both organismal and environmental variation as being effects of shifts in

structural coupling between those two systems can help to make sense of the relationships

between processes in each system. If a process in one is linked to a process in another, it may be

possible to explain variation in one by pointing to variation in the other.

One way in which the enactive concept of structural coupling can shed light on ontogeny

is in specifying what constitutes an individual life cycle. An individual life cycle can be thought

of as a temporal pattern in structural coupling between processes in a developmental system.

These processes form a self-organizing network on a larger scale, with constant, dynamic

interactions between developmental resources resulting in the formation of a complex, spatially-

bound cluster of operationally closed interactions – an individual organism. This

conceptualization may seem overly abstract on its face, but it highlights the fundamental

processes that are at play in an ontogenetic cycle. Within a complex web of dynamic interactions

between variable resources, an individual organism materializes as a locus of order, maintaining

homeostasis at a state far from thermal equilibrium and itself as a unity over time, and

dissipating only when met with significant perturbation from interacting forces within the

developmental system17. These are thus characteristics of an individual entity as a self- organizing living system.

17 A more natural way to put this would be to differentiate between environmental, external perturbation (e.g., loss of energy resources) and internal perturbation (e.g., organ failure), but this suggests a bifurcation between organism

Depew and Weber (1998; see also Weber and Depew 1996) suggest that thinking of organisms as autopoietic, autocatalytic, dissipative structures that are embedded in a developmental system can shed light on the phenomenon of natural selection. They highlight computer modeling work by Kauffman and Wimsatt (Kauffman 1993; Wimsatt 1986) that shows that “incredibly complex internal order in the genetic regulatory programs that govern ontogeny is to a large extent a product of self-organizational processes, which produces modular entities among which natural selection selects, locks in the results of selection, and constrains the paths along which it can run” (Weber and Depew 1996, 38). The patterns that emerge as a result of interaction within the larger, self-organizing developmental system can thus be specified in reference to a species population in addition to individual organisms, allowing for further analysis of the factors involved in bringing about change and variation at the population scale, over evolutionary time.

This section outlines an important first step in understanding cognizing systems by describing the organizational makeup of a living system. Maintaining integrity through self- organization is fundamental to all living systems. The enactive approach, as a naturalistic theory of living beings, defines autopoiesis as a type of self-organization that captures how living systems maintain themselves as a unity over time. Self-organizing systems are structurally coupled with their environments, such that the organism and its environment together form a complex web of interactions. DST’s notion of a developmental system describes how this web of interactions is made up of an organism as a developmental process and the developmental resources which construct its life cycle. With a picture in place of how living systems are

and environment, and both DST and enactivism stress there being no necessary distinction between the two – they are ‘bound up’ with each other in their structural coupling. I have not explicitly avoided using this language, but do wish to draw attention to this technicality. 87

organized, I next address how living systems undergo alterations in structure in virtue of this organizational pattern.

4.2 Plasticity as structural flexibility

In the previous section, I suggested that organisms, as self-organizing, operationally closed systems, undergo changes in structure by making use of different developmental resources within their larger developmental system. In this way, they are structurally flexible, with changes being initiated by numerous interactions within the developmental system. In this section, I suggest that this flexibility in structure describes how living systems are plastic in addition to being self-organizing and operationally closed. Plasticity enables organisms to undergo change and variation in structure, in turn giving rise to both variability and stability of form and of traits.

I argue that integrating concepts from evolutionary developmental biology, or evo-devo, into

DST, and the broader biological enactive framework developed here, can fill out some conceptual gaps in our understanding of structural flexibility (and rigidity).

Insights from complex systems theory can help to show why living systems are structurally plastic. Prigogine suggested that indeterminacy is a key feature of systems far from thermodynamic equilibrium. In contrast to linear systems, which follow a highly conserved path to equilibrium, non-linear systems have multiple possible trajectories at any point in time. At individual bifurcation points, the system may follow an entirely new trajectory, depending on a number of factors including the previous history of the system. Therefore, non-linear systems exhibit a kind of indeterminacy, because these bifurcation points can arise at any point in time and may initiate a new trajectory for the system. This indeterminacy of states suggests that there are points at which the system can effectively move between two (or more) possible outcomes.

In this way, they can be thought of as structurally flexible, or plastic. To be plastic is to possess the ability to undergo some kind of alteration or modification, typically in response to a change in conditions. So to exhibit plasticity is for there to be some kind of difference from a previous state, whether in form, kind, or degree of some condition. These points capture what is generally meant by the term plasticity, though its usage and definition varies.

West-Eberhard (2003) describes plasticity as environmental responsiveness, and notes

similar definitions used throughout the literature, including “modifiability of morphology during

development” (Wilson 1894), “environmentally sensitive behavior” (Baldwin 1902), adaptive

action on the basis of experience (Wheeler 1910), and intra-individual variation (Evans 1953;

Williams 1992). While a distinction can be made between adaptive and nonadaptive plasticity,

West-Eberhard suggests a fitness-neutral description of plasticity is most practical (noting

Schlichting and Pigliucci 1998). How fine-grained a definition is necessary will be context- dependent. Phenotypic traits that are plastic in nature, captured by the notion of phenotypic plasticity, enable an organism to respond to environmental inputs.

4.2.1 An array of possibilities

A plastic, non-linear system can undergo a multitude of structural alterations. Thinking of living systems as plastic opens up ways to discuss how such systems undergo both change and variation. In critiquing Konrad Lorenz’s analysis of innate and learned behaviors as being both ultimately determined by innate genetic mechanisms, Oyama suggests an alternative analysis that describes possible outcomes as operating out of an “array of possibilities” (Oyama 2000, 68). It is important to note here that the proper locus of the array of possibilities is the developmental system, comprised of both the organism and its environment. On Lorenz’s view, regardless of

whether stimuli resulting in an adaptive response result from the organism or its environment, the organism’s genome must be structured in such a way to allow for a response. Only what is genetically possible can arise; whether or not a particular response is possible (and is actualized) will depend on an organism’s genomic structure. Oyama stresses the importance of mutual selectivity between factors within the developmental system, noting that

The selection from the array of possibilities cannot be an exclusive function of the array itself. Nor, and this is an important point, can the array, the “potential”, be strictly defined apart from the array of selecting influences. The fact that, in normal development, a multitude of selections and influences occur inside a skin does not change their mutual dependence. A structured system selects its stimulus – indeed, defines it and sometimes produces it (the state of the system determines the kind and magnitude of stimulus that will be effective, and intrasystemic interactions may trigger further change) – and the stimulus selects the outcome (the system responds in one way rather than another, depending on the impinging influence). (Oyama 2000, 68)

I think Lorenz is right to the extent that an organism’s genomic structure places constraints on what is developmentally possible, and so ontogenetic outcomes are in some sense contingent upon genetic architecture. But what Oyama takes issue with regarding Lorenz’s position is not that genes constrain development, but rather that it is the genome that solely specifies the array of possibilities in ontogeny. Selective influences originate from the larger developmental system; the array of possibilities spans both organism and environment, and developmental resources throughout both subsystems. The reciprocal selectivity between both illustrates the importance of the mutuality between influences in a developmental system. Selection of an influence is a function of the interactions within the entirety of the system.

The upshot here is that movement through a dynamic array of possibilities is possible due to plasticity, and this shows how plasticity is fundamental to generating both organismal change and variation. An individual trait can undergo structural alterations over developmental time, such as morphological alterations, as a result of movement through an array of possibilities within the ‘matrix of developmental resources’ (Griffiths and

Hochman 2015). Variation can likewise arise over evolutionary time with selective influences altering species’ trajectories, occasionally resulting in speciation or extinction.

And centrally, individual development is presumably outright impossible without change in form over a life cycle. Likewise, individual variation arises out of new experiences and learned traits. A focus on plasticity as movement through the array of possibilities emphasizes how developmental systems are in constant flux, with multifarious influences morphing the structural landscape over time. A rich understanding of organismal change and variation therefore requires an account of what allows for such change and variation, and plasticity is well suited for the role.

Importantly, this picture of flexibility is not to suggest that there are no “fixed” elements in organismal development, nor that such “fixed” elements cannot arise. Indeed, rigidity of certain traits, captured by Waddington’s notion of canalization, is a key aspect of developmental regulation, defined as “the tendency for similar results to appear despite some variation in developmental conditions” (Oyama 2000, 109). The canalization of such traits illustrates a frequency measure of them, but not that they are predetermined. Here rigidity of canalized traits reflects their developmental regularity. They are subject to constraints as any traits are, but are identified by their tendency to produce the same or a similar outcome despite perturbations from these constraints. Thus the system itself

remains flexible, in that with enough pull from developmental influences they may be altered, but high frequency, canalized traits should be expected to develop under typical conditions. They are ‘resistant to variation’ (Oyama 2000, 110) but not impervious to it.

Another way to conceptualize resistance to variation is to consider it in the context of what Griffiths and Gray (1994) refer to as ‘transgenerational stability’ of factors interacting within a developmental system. Organisms, as developmental processes themselves, make use of resources across each resource domain depending on a shared history of interaction and with regard to individual needs. What is relevant for the sake of providing an evolutionary explanation is identifying recurrent interactions between resources and multiple generations of organisms with the capacity to at the very least access those resources via sensory modalities.

Variation arises when resources are utilized in a new manner, when new resources are introduced or existent resources are removed, the relationship with resources is altered, timing of interaction with resources is altered, among other possibilities. In this way, the developmental system is comprised of not solely an individual organism interacting with an environment over the course of its life cycle, but rather it extends over both the organism as a developmental process and the developmental resources with which it is coupled such that interactions with those resources constitute a species-typical life cycle. Phenotypic traits that emerge on a consistent basis obtain stability through the recurrent interactions that are responsible for bringing them about.

Importantly, transgenerational stability is dependent upon reliably consistent recurrent interactions; the dissipation of such a relationship would likely result in the dissipation, in turn, of the phenotype reliant upon that interaction18.

18 One example is the degeneration of eye structures in subterranean organisms (see Chapter 5 for discussion). 92

4.2.2 Relations between traits: integration and modularity

Under DST, phenotypes arise as a result of a dynamic array of interactions between features of

the developmental system. On its own, DST serves as a starting point for reinterpreting key

claims about biological systems. Its emphasis on organism-environment dynamics and multiscale

explanatory mechanisms offer a more complex, nuanced view of organismal development and

evolution. It also points the way forward for how to structure new research questions and

programs, including support for the integration of subfields (such as ‘ecological developmental

biology’ – see Gilbert and Epel 2009). Griffiths and Gray (2005) urge that DST, though

promising as a way of moving past nature-nurture debates in biology, is in need of a further theory of phenotypic integration and modular evolution (Griffiths and Gray 2005, 423). These

two concepts build on DST by describing the relationship between the emergence of traits. By

integrating theoretical claims about phenotypic integration and modular evolution into DST’s

framework, more can be said about how phenotypes, especially complex phenotypes, arise.

Griffiths and Gray suggest that evolutionary developmental biology, or evo-devo, offers a viable

way forward for studying the factors involved in generating phenotypic variation and thus new

phenotypes.

I am in agreement with Griffiths and Gray that theoretical claims from evo-devo can find

a natural home under the broader conceptual scheme of DST. While there has been discussion of

exactly the manner in which the two fields fit together (see Griffiths and Gray 2005) 19, evo-devo

can profit from concepts in DST that are harmonious with its theoretical framework. And

equally, DST can profit from work produced under the evo-devo research program as empirical

19 Griffiths and Gray (2005) note that evo-devo researcher Brian Hall has explicitly rejected the DST claim that “elements outside the genome are part of an evolving developmental system” (Griffiths and Gray 2005, 422), or the idea that non-genetic factors constitute an inheritance system as equally relevant as genetic inheritance. 93

evidence in support of its claims. In what follows, I will offer a potential way forward for

integrating evo-devo theory and DST. This integrated theory can in turn help to fill out some

conceptual gaps in the biological enactive framework I am putting forth; namely, it can provide

an enriched account of structural variation and stability in living systems as self-organizing,

operationally closed, plastic systems.

We can think of the imported evo-devo theory encompassing the notions of phenotypic

integration and modular evolution as specifying more closely the sort of interactions described

by DST. Phenotypic integration refers to the mechanisms by which phenotypic traits covary as a

result of their being functionally related (Pigliucci 2003, 265). As discussed in the previous

section, phenotypic plasticity enables differentiation between traits in response to perturbations or opportunities. But it is possible to track covarying changes to multiple traits, such that a domain can be identified within which integrated traits can be studied. One clear example of phenotypic integration is the coevolutionary relationship between phenotypes of pollinating insects and floral structure. Studying how these phenotypic traits correlate with each other over evolutionary time can shed light on the factors involved in the manifestation of both conserved traits and novel traits. As DST emphasizes the multitude of connections within the web of interactions occurrent in a developmental system, empirical research on the concept of phenotypic integration would lend support to DST as a broad theory of the organization of

developmental systems. Pigliucci (2003) describes phenotypic plasticity as a ‘twin’ issue to

phenotypic integration, suggesting that research on plasticity may aid in guiding new research on

integration. As novel research methods arise out of the integrated field of evo-devo, conceptual

claims about phenotypic plasticity and integration can serve as a way to tie DST into the

theoretical framework of evo-devo.

Indeed, providing treatment to the issue of phenotypic integration may be necessary in order to clarify some aspects of DST. In reference to a key passage by Lewontin (1978), Griffiths and Gray (2004) point out that traits cannot be entirely integrated with each other if they are to be selected for under adaptive evolution. Deep integration between traits would mean that a change in one trait would bring about change in every trait it is integrated with, with the potential for “deleterious changes in other aspects of the organism” (Griffiths and Gray 2004, 424).

Selection acting on one trait without affecting other traits thus requires that trait to be quasi- independent, or modular. This suggests that the organization of development may be described more generally as modular, and given the contingency of developmental factors with evolutionary factors (and vice versa) central to DST (and evo-devo), modularity may be a key facet of evolution as well.

At first glance modular evolution may appear to pose an issue for the integration of DST with evo-devo, considering DST’s emphasis on the intimacy of interwoven features within the developmental system. But as discussed in the previous section, DST leaves room for stability by specifying it as occurring in virtue of instances of reliably consistent recurrent interactions. And indeed, stability is a critical aspect of development, so any developmental theory must provide the conceptual tools to engage with it. While species-typical developmental factors may be contingent upon reliably consistent interactions, the fact that they appear on a consistent basis points to a need for an explanatory mechanism for how some traits remain relatively rigid and others are more malleable, both in individuals and in populations. And this is precisely what evo- devo offers in the notion of modular evolution (and modularity of developmental organization).

As Griffiths and Gray note:

The modularity issue is at the heart of the new discipline of developmental evolutionary biology (EDB), so it is striking that DST has had little to say about the issue of modularity, or the older, related issue of the extent of developmental constraints (Maynard Smith et al 1985, for an exception to this remark, see Oyama 1992). What, one might ask, is developmental about developmental systems theory? The answer is that the term ‘developmental’ in this context stands in contrast to ‘innate’. To think developmentally is to focus on the many factors that must be present for a fertilized egg to give rise to a normal life cycle and on the process – development – in which those factors interact. (Griffiths and Gray 2004, 425)

In thinking about the developmental process as a series of changes that are reliably recreated in each life cycle as well as a process that varies with each iteration, stability is as relevant as is mutability. DST thus requires a viable theory of both in order to serve as a theory of how organismal form arises. The concept of modular evolution, as discussed in the evo-devo literature, can provide valuable theoretical insights for DST, namely, it can explain the variation in degrees of plasticity seen across phenotypic traits. Some traits are more rigid in that they arise even when developmental resources vary in their availability, whereas others are more malleable in that even a slight variation in resource availability can cause phenotypic differentiation. What an enriched version of DST can offer, then, is a theory that explains not a distinction between rigid and malleable traits (as this falls into yet another unwarranted dichotomy that parallels the distinction between innate and learned traits) but instead a continuum in terms of constraints on plasticity.

4.2.3 Controls and constraints

So far I have gestured at the importance of constraints as an effect on both developmental and

evolutionary change and variation. While both phenotypic integration and modular evolution

make reference to constraints as integral elements in understanding the factors that control and

allow for alterations to be made, more must be said about how these constraints function within a

developmental system. Both DST and evo-devo offer accounts of what ‘does the molding’ in the molding of developmental and evolutionary elements, and it is an important part of the integration project to bring these accounts together, as evidenced by the quote from Griffiths and

Gray highlighted above. I begin by discussing the account of control that Oyama offers in The

Ontogeny of Information before discussing the relationship between adaptation and constraints from the evo-devo literature.

Oyama outlines three ways in which control emerges as a result of the organization of developmental systems: through interaction, hierarchical levels, and time. Rather than thinking of genes as ultimately controlling developmental outcomes, DST specifies the interaction

between elements across the developmental system, noting the context of the interaction, as

being the locus of control. Oyama states:

Every developmental interaction, and therefore the entire norm of reaction, is jointly determined. To question the notion of genetic constraint or potential is not to deny limits to variability, flexibility, or adaptation. It is simply to give all developmental interactants comparable theoretical status, to recognize the importance of levels of structure above (and below) that of the genes, to insist that the significance of an interactant must be discovered by investigating the roles it plays in ongoing processes, and to point out that phenotypic potential cannot be said to be limited in any practical sense until we know what the limits

of genotypic and environmental variation and interactions are. (Oyama 2000, 132, original italics)

Control also emerges as a result of the relationship between various interactions within sub- structures of the organism. Processes at the scale of the organs, for example, will be affected by processes at the scale of the cells as well as processes at the scale of the whole organism. The extent to which certain processes maintain control over others will vary, with feedback between outcomes from interactions potentially generating greater variation. Lastly, developmental systems have a history; life cycles are played out in a temporal context such that events preceding the enactment of a life cycle will impact elements of that life cycle. Phylogenies illustrate the temporal aspect to developmental systems by situating them historically. These histories are contingent upon factors across all scales, not merely major events. Small events may have generative effects if reconstructed reliably in each life cycle (Wimsatt 2001) just as large, constant events (exposure to elements that comprise the external, species-neutral environment, for example) place conserved constraints on both development and evolution.

Discussion of control elements also arises in the evo-devo literature, particularly with regard to questions about phenotypic integration. Pigliucci suggests that a question that naturally follows from the careful formation of a definition of integration is whether integration ought to be considered an adaptation or a constraint (Pigliucci 2003, 266). He notes that historically, integration has been thought of in adaptationist terms, with selection acting on integrated traits to move toward increased fitness in the environment. But recent work has suggested that patterns of covariance can serve “as an indication of constraints on future evolution” (Pigliucci 2003, 266).

Rather than thinking of adaptation and constraints as a dichotomy, Pigliucci suggests (among others, see Schwenk and Wagner 2004) that instead

we need to think of adaptation and constraints as players engaged in a continuous dialectic throughout evolutionary history. On the one hand patterns of phenotypic integration can surely be modified by natural selection in order to improve adaptation to the external environment and/or maintain the coherence of the internal developmental system. On the other hand, obviously whatever functional, developmental, or genetic relationships among traits are now present in an organism will limit and channel at least its short and mid-term future evolutionary trajectory. (Pigliucci 2003, 266)

Both adaptation and constraints act as forces on complex integrated phenotypes, with patterns emerging as a result of both selective pressures and structural capacities. The image of an emergent pattern of alterations dependent upon adaptive and constraining factors within the developmental system fits neatly with DST’s conceptualization of plasticity as well as transgenerational stability. Organisms are shaped by both adaptation and constraints in a dynamic sense, with some elements imposing greater or lesser force at different times depending upon other states of the system. In this way, development can be conceptualized as a pattern of movement through a dynamic landscape comprised of interacting elements.

On this view, as Oyama notes, studying differentiation as a function of variation in interactions over time dissolves concerns and debates over whether genetics or the environment is responsible for a given trait, and “[u]nderstanding ontogeny thus becomes partly a matter of charting the shifts from one source of change (including intraorganismic processes) to another, as one interaction alters the developmental system in a way that provides transition to the next”

(Oyama 2000, 161). This conceptualization emphasizes not genes and environment as separate forces acting in distinct ways, nor simply the interaction between the two. Instead, the focus is on dynamic movement through the developmental state-space. This focus allows for comparison

between developmental states at differing times, which can provide a clearer picture of how organisms undergo both change and variation over time.

In this section, I have discussed how plasticity is an important feature of living systems.

As non-linear, dynamic systems, living systems undergo both change and variation by coming to

“bifurcation points” in their systemic functioning and proceeding to maintain themselves as a unity. Phenotypic traits are subject to change as a result of complex interactions between features of the developmental system, which encompasses both the organism as a developmental process and the resources within its niche that shape and are shaped by its development over time.

Theoretical considerations imported from evo-devo help to further build on the framework proposed by DST by discussing how phenotypic integration, modular evolution, and adaptation and constraints capture the mechanisms by which plasticity is enacted. In the next section, I suggest that the kinds of changes that are possible through phenotypic and developmental plasticity can serve as cognitive strategies for responding and adapting to fluctuations in both the internal and external environment.

4.3 Acting on flexibility

Through plasticity, an organism can undergo a multitude of alterations in response to changes in conditions. This modification occurs over developmental, behavioral, and, for populations, evolutionary time. Individual organisms progress through life stages, experience and learn about their environments (and themselves) through their sensorimotor capacities, while undergoing adaptive changes in response to selective pressures and influences. But in order to understand how such flexibility plays out, it is necessary to look at how “choices” are made – how individuals and populations follow one trajectory over other potential trajectories through each

100

timescale. The language of ‘choice between options’ here is metaphorical; to mean it literally would be to violate DST by following the conventional language of the ‘cognitive-causal gene’

(Oyama 2000). Of course, there are instances where an individual organism must choose between two (or more) options, behaviorally speaking. But it is important to be mindful of what is being specified as doing the ‘choosing’, and what the ‘choices’ are. Instead of importing the cognitive-causal language, it is helpful to think of change and variation as arising as a result of the pulling or slackening of strings in Waddington’s epigenetic landscape.

Figure 4. Waddington's epigenetic landscape, representing complex interactions between genes (pegs at the bottom) and environmental conditions.

Interactions within the developmental system, between organism and environment, result in this pulling and slackening, with a pattern of alterations emerging over time. These alterations can be thought of more generally as structural modification of the developmental system.

101

In this section, I discuss structural modification via two concepts: modulation and

assessment. Modulation emerges as a feature of self-organizing, operationally closed systems structurally coupled with an environment in such a way that makes up a developmental system.

A related concept, albeit one with a slightly different focus, is the concept of assessment, which refers to “adaptive switching between phenotypes” (West-Eberhard 2003, 440). Both of these concepts, modulation and assessment, can be thought of as types of structural modification20. I

then argue that one way in which organisms cognitively engage with the world is through

modulation and assessment. These forms of modification are cognitive strategies for dealing with

fluctuations in both the external and internal environment.

4.3.1 Modulating coupling with the environment

Recall from Section 4.1 that living systems can be thought of as self-producing systems that are

operationally closed while remaining thermodynamically open. In this way, living systems

maintain their organizational makeup while making adjustments in response to the flow of

resources both into the system and out of the system. As Di Paolo et al. (2017) note, however,

the system does not merely passively exchange resources in the form of energy and materials,

but rather makes certain alterations depending upon its status at a time and its metabolic needs.

In other words, the system has a particular manner of behaving such that it exhibits a type of

agency – it acts as to produce a particular effect. For Di Paolo et al., the goal in describing living

systems in this way is to get to the heart of what makes this type of system an agent21. In order to

20 Modulation is a more abstract concept that describes an architectural feature of living systems, while assessment is a more concrete concept that describes individual actions that such systems carry out. 21 While agency is not my focus here, the discussion is helpful for understanding the factors involved in the materialization of certain behavioral traits. 102

describe how it is that living systems exhibit active evaluation of their conditions, it is necessary to describe the sort of organizational features they have that allow for agential behavior.

On Di Paolo et al.’s account, organism and environment stand in a symmetrical relation to one another as one is co-determined by the other as a result of their structural coupling.

Thinking of the organism as an agential system, however, highlights a fundamental asymmetry between organism and environment – agents behave in response to a valenced aspect of the environment; organisms act on the environment in a self-interested manner, whereas the environment acts on organisms according to constraints but without this aspect of self- interestedness. Di Paolo et al. spell this distinction out further:

…acts have the property of being asymmetrical in terms of the relation between agent and environment. Because the idea of coupling is symmetrical, one way to introduce an asymmetry in it is to go one level up and propose that an agent is sometimes able to modulate its coupling with the environment (i.e., to modify the way its own processes and those of the environment relate). This condition of interactional asymmetry is the second requirement for agency (Di Paolo et al. 2017, 117, original italics) with the first requirement for agency being active self-production and self-distinction. What Di

Paolo et al. point to is an organism’s capacity for modulation22 as a way of relating to the environment in an active, responsive, self-interested manner. They argue that this feature of agency is not sufficiently addressed in merely thinking of living systems as autonomous systems, because that characterization fails to capture the interactional asymmetry seen in organism- environment coupling.

22 ‘Modulation’ should not be confused with ‘modularity’, discussed in Section 4.2. 103

Modulation, then, is the way in which an organism modifies the relationship between its

internal processes and environmental processes, or in other words, its coupling with its

environment. By making these kinds of adjustments, organisms can undergo structural

modification in response to changing conditions both internally and externally. Modulation can

thus be thought of as an architectural feature of living systems, as they actively respond to

changing conditions within the developmental matrix in order to maintain themselves as a unity.

As Di Paolo et al. note, not all modulation will result from organismal initiation, as some

alterations will be necessitated by the environment. But as a more abstract feature of self-

organizing, developmental systems, modulation is necessary in order to respond to

environmental changes across each timescale.

Another sense in which modulation is necessary is in order to maintain homeostasis in

the face of environmental perturbations. Organisms with homeostatic physiology can maintain

their internal environment in response to fluctuations within that environment and external forces

responsible for variable effects on the internal environment (Nijhout 2001, 138). Temperature

regulation is one such example, with various mechanisms in place, for some species, for

adjusting to external temperature variation to maintain a constant internal temperature.

Thermogenic strategies may be behavioral, such as shivering, or morphological, with increases

of fur and plumage for insulation during certain parts of the year (Blix 2016). Modulation of

effects on thermoregulation processes can serve as an important adaptive strategy, namely in

cases of survival in a harsh environment such as those found in polar regions.

Structural modification in the form of modulatory behavior (at various scales of

biological organization) is a critical aspect of living systems conceived of as self-organizing, operationally closed, plastic systems. Without the means by which to effectively keep the system

104

running when met with internal and external variation, the system would disintegrate over time.

And unless some external force is able to make the appropriate adjustments on behalf of the system, it is, in some sense, up to the systems’ internal mechanisms alone to make these adjustments.

4.3.2 Assessment of environmental conditions

West-Eberhard describes assessment as the evaluation of environmental circumstances

(including the evaluation of conspecifics) that “occurs whenever a particular response correlates consistently with some environmental variable” (West-Eberhard 2003, 442). Environmental features, in the form of resources, conspecifics, predators, and so on, vary, and organisms that can engage in assessment of those features will vary in their responses to them. Organisms assess the relevant conditions and effectively choose between alternative trajectories, be they behavioral, developmental, or over a longer timeframe (for populations), evolutionary. Earlier I issued a disclaimer around the word ‘choice’, and West-Eberhard too notes the term is being used here in a very broad sense, and that debate has ensued as a result of its use. In one sense, some organisms do make choices, such as choosing between potential mates or choosing to flee rather than fight a predator. In a looser sense, choices can be said to be made simply when there is a “differential response to stimulus differences associated with the alternatives” (442). The evaluation of these stimulus differences, captured in the notion of assessment, is a way of initiating both structural (and behavioral) modification.

Acting on structural flexibility can be a matter of modifying what is possible to modify when it is appropriate to do so, but also sometimes simply when possible. The French biologist

François Jacob conceived of evolution as ‘bricolage’ (Jacob 1977), highlighting how

105

evolutionary changes seem to sometimes occur merely because it was possible for them to do so.

So-called evolutionary tinkering can potentially have adaptive outcomes, depending on other

contingent factors. Plasticity allows for such tinkering, and thus evolutionary innovation. Degree

of plasticity can also be relevant for assessing fitness. In the context of flexibility, generalist and

specialist strategies can be mapped onto degree of flexibility, with generalist strategies

illustrating more flexibility and specialist strategies illustrating less flexibility. Environmental

heterogeneity (or lack thereof) can affect which strategy organisms opt for, thus providing a

place for analyzing the relationship between environmental setting and degree of plasticity (van

Tienderen 1997). A capacity for assessing environmental conditions is utilized in order to

employ a particular strategy for adjusting to those conditions.

Making adjustments to traits in response to environmental conditions is an additional

quality of agents that Di Paolo et al. draw attention to. They use the term ‘normativity’ to refer to

a system’s active modulation of interactions with respect to norms (Di Paolo et al. 2017, 120).

They do not offer a precise definition of ‘norm’ in this context, but direct attention to writings

from Merleau-Ponty (1963) suggesting that organisms have an appropriate way of being and

sustaining themselves in their environmental niche. Some actions are, in a sense, purposeful23,

such as searching through a patchy environment for bits of food or singing to draw the attention

of potential mates. According to Di Paolo et al., these actions have a normative dimension to

them, and it is this normativity that fills out their account of agency24.

On West-Eberhard’s account of assessment, normativity appears in the notion of

adaptive switching between phenotypes. She notes that “[t]he fact that nonhuman organisms are

23 This teleonomic approach will be given further treatment in Chapter 5. 24 The ‘joint conditions’ of self-individuation, asymmetry, and normativity make up their full account of agency (Di Paolo et al. 2017, 120). 106

known to adaptively switch between conditional alternatives implies that they somehow use criteria of advantageousness to assess the conditions surrounding their decisions” (West-

Eberhard 2003, 443, original italics). The selection of criteria in making a decision may occur with respect to individual organismal norms. Daphnia that evaluate predator-indicating chemical cues and modify their morphological structure to be able to defend themselves against potential predators make an adaptive switch in phenotypes. In some species of fishes, changes in morphology can bring about adaptive outcomes. A bluehead wrasse (Thalassoma bifasciatum) male, rather than facing competition with other territorial males, can take on a smaller, female- like form, allowing them to intrude on the occupied territories unnoticed (West-Eberhard 2003,

458). Making an adaptive switch in morphological traits requires both the capacity for assessing environmental conditions and the capacity for initiating changes in form in response to those conditions.

The capacity to assess conditions and modify traits with respect to organismal norms is thus, in most instances, key to ensuring fit especially in a dynamic environment. As environmental conditions vary and new pressures arise, being able to adaptively respond is vital for survival and reproductive success. This is true for both morphological modification and behavioral modification. West-Eberhard offers a plethora of behavioral examples to illustrate this point, with varying complexity. One example shows how assessment can occur not just of environmental resources but of potential consequences of behavioral actions. She notes, “[e]arly observation of chimpanzees in the wild, for example, showed that when food is abundant, rich sources of fruit cause noisy excitement, which attracts others to the site, whereas in times of food scarcity chimps search in small groups and exploit their finds quietly” (Reynolds 1970; West-

Eberhard 2003, 445). The assessment of complex situations invites complex behavior in return,

107

and instances of planning behavior may be structured by assessing longer-term conditions in

order to ascertain a successful outcome. Variation in traits arising from assessment-initiated

modification can, over time, be selected for, leading to new evolutionary trajectories.

4.3.3 Structural modification as cognitive strategy

I am now in a position to suggest that modulation and assessment, as forms of structural

modification, are cognitive strategies organisms undertake in response to environmental

fluctuations. They employ these strategies according to organismal norms, which factor in a

valenced aspect to encounters with resources. Recall that a central focus of enactivism is on the

notion that organisms enact a world through sensorimotor engagement with the environment, and

cognition emerges out of this enaction. My claim here is that enaction happens (partly) through

these forms of modification, which are possible due to the organizational makeup of living

beings. At this stage, I am not making any claims about what cognition itself is, but rather I am

examining the properties of living systems that enable them to be cognizing systems. I began by

working through the organizational features of living systems, namely self-production, operational closure, and plasticity. I then discussed the kinds of mechanisms that plasticity makes use of to bring about both variability and stability. This work puts us in a good place to discuss specific forms of structural modification as strategies that are cognitive in nature.

To be sure, it is reasonable to consider some cognitive strategies as being more complex than others. Growing thicker fur to withstand colder temperatures merely involves passive regulation of physiology initiated primarily by sensory information. Learning from observing conspecifics involves much more. But both can be said to be cognitive strategies, in that both

108

entail the use of assessment and modulation to effectively ‘keep up’ with the world. The kind of strategies I detail here can thus be thought of as falling along a continuum of cognitive abilities.

Naturally, cognition itself can be difficult to define. A recent discussion in Current

Biology illustrates the many different approaches biologists have to what counts as cognition and what doesn’t (Bayne et al. 2019), with a general consensus being that there is oftentimes no hard line to draw. Rather, it makes sense to speak of some abilities as being in some way involved with cognition – such as flexibility in response (R612) – and putting aside questions about where on the continuum of cognitive complexity certain organisms should be placed. First on the agenda is explaining that there is such a continuum; more complicated questions about the function of cognition can be dealt with afterward.

I suggest that the capacities and abilities discussed in this chapter serve as preconditions for cognition, and thus they identify what kinds of systems have the capacity to be cognizing systems. Systems built in this way have the capacity to engage in cognitive activities in virtue of their organizational makeup. They serve as a place to look if we are interested in finding cognition in nature. The differences in capacity for modification can offer a starting point for drawing out distinctions in capacity for cognitive functioning. Being more structurally flexible and able to assess more conditions (perhaps by way of more sensory modalities, or multimodal perceptual capacity) may lead to greater cognitive ability, as there will be more factors to which the organism can respond. At the same time, being embedded within a richer developmental system, and endowed with a greater capacity to make use of developmental resources, may require greater responsiveness, and thus more complex cognitive functioning. Some eusocial insects, such as social bees and ants, provide a strong example here, illustrating not only complex

109

polymorphism but also a complex environmental setting (that they themselves create25). Despite

being merely small insects, social bees and ants are commonly studied for their complex

behavior and seemingly quite complex cognitive capacities26. But their complexity can be

investigated as a function of the contingent relations between the organism and its larger

developmental system. That they meet the conditions for identification as cognizing systems, in

virtue of their organizational makeup, is the conclusion to draw from the work being done here.

Organisms undergo structural modification predominantly via modulation and

assessment as ways of responding to environmental (both internal and external) variation. As Di

Paolo et al. note, this type of behavior is essential to understanding agential features of living

systems. In other words, in thinking about organisms as individual agents, highlighting the ways

in which they are self-producing, operationally closed systems with the capacity for responsive

structural modification helps us to pick out what is fundamental to organisms as living beings,

and what distinguishes them as agents. In addition to picking out these features as being facets of

agency, a twin claim can be made about them also being facets of cognition. The concepts of

modulation and assessment are arguably key to understanding cognition, as they are ways in

which a system can dynamically respond to variable conditions. If cognition is at its heart a

matter of active regulation and response, self-organization/operational closure/plasticity and

structural (and behavioral) modification pick out both the organization and the structure (sensu

Maturana and Varela) of cognizing systems. We can thus think of these two foci in this chapter

as being ontological features of living systems qua cognizing systems. Cognition emerges as a

25 These points are taken up further in chapter 5. 26 The honeybee ‘waggle dance’ is perhaps the most well-studied invertebrate signaling behavior. Signaling is thought to require complex cognition, as it involves representing environmental elements. 110

result of the architectural makeup of these types of systems. It is enacted in the playing out of each life cycle.

The set of developmental resources accessible within an individual life cycle is dynamic over each timescale, with modifications arising out of influences from both biotic and abiotic sources. Thus the organism’s developmental niche is rarely static for long at any particular level.

In order to cope with an environment in fairly constant flux, organisms make use of their structural flexibility by assessing and modifying their structure in response these changes.

Because organisms qua developmental processes are completely embedded in their larger developmental systems, the changes they make can initiate additional changes in that system; feedback loops between components of the system warrant that a change in one place can potentially initiate changes at many other places. To put the relationship more concretely, looking at any example of niche construction illustrates how this structural coupling can be enacted even just over an individual’s lifetime. The developmental “information” available to ant larvae, for example, initiates morphological changes depending upon the current status of the larger organismal system.

The question here thus becomes one of how this structural coupling, arising out of organismal plasticity over time, organizes individual lives. As discussed, the relationship holds over a behavioral timescale in addition to developmental and evolutionary timescales. Answering how organisms cope with changes in real time, and how this coping organizes their lives, is a matter of looking more closely at the organism-environment relationship. Individual choices and decision making can be made sense of through investigating an organism’s ability to maintain self-organization and structural coupling through its behavior.

111

With a picture in place of how organisms assess conditions for the sake of modulating between plastic traits, more can be said about how such activity plays out in organismal life. In the next chapter, I look at the enactive concept of ‘sense-making’ in order to clarify how individual organisms behave as agents in a dynamic environment. Sense-making can aid in guiding understanding of both species-typical organismal behavior as well as the factors responsible for generating novel behavioral traits. The capacity for environmental responsiveness, captured in the discussion of plasticity above, is necessary for living systems to persist in their environmental settings. But for a richer understanding of the lives of individual organisms as intelligent agents in a dynamic world, we must consider why they behave in the ways that they do.

112

Chapter 5: Making Sense of the Environment

Introduction

So far in the dissertation, I have discussed ways of enriching the enactive approach to cognition

by situating it in a contemporary biological context. However, I do not wish to adopt the enactive

approach wholesale. In this chapter, I provide an alternative interpretation of the enactive

concept of sense-making, which refers to an organism’s ability to engage with the world in such

a way that creates significance for that organism. On autopoietic enactivist interpretations (e.g.

Varela, Thompson and Rosch 1991; Thompson 2007; Di Paolo et al. 2017) cognition is grounded

in sense-making. Therefore, the notion is central to understanding an enactive conceptualization

of cognition. Sense-making is typically introduced as a term to capture how organisms relate to the worlds that they enact. It is described as the way in which organisms create a world of meaning and significance.

In this chapter, I suggest that to fit more naturally with the biological enactive framework

I am proposing in this dissertation, sense-making may be reconceptualized in developmental terms. Specifically, I argue for a notion of sense-making as generating coherence between an organism and its environment as coupled systems. This is a different way forward for specifying how organisms ‘enact’ a world - one which has the advantage of clarifying enactive thinking about the coevolutionary relationship between organisms and the biotic features in their environments.

113

In section 5.1, I describe how the enactive concept of sense-making is meant to show

how environmental features have a particular valence depending upon the status of an organism,

and that organisms engage in sense-making when they respond adaptively to changing

conditions. Through sense-making, meaning and significance are generated, because in an

organism’s valenced interactions with the world certain elements and resources become relevant

to the organism in ways specific to its individual needs. It is in virtue of the capacity for sense-

making that organisms enact a world rich with significance. In section 5.2, I offer my

reformulation of sense-making as a principle of coherence within a developmental system. My

motivation for doing so is to define sense-making in such a way that makes reference not to the

generation of meaning but instead to the generation of coherence. This conceptualization fits

more naturally with a developmental perspective on organismal life. I work through this

alternative conceptualization of how organisms construct their enacted worlds by appealing to

insights from developmental biology. I argue that rather than focusing on meaning as arising out

of the structural coupling between an organism and its environment, a developmental perspective

means focusing on the emergence of coherence between features of these systems. Thus, it is

according to a principle of coherence that sense-making occurs. In section 5.3, I address how to conceive of the enacted world of organisms engaged in sense-making activity. This requires defining the environment to which organisms are coupled. I argue that conceiving of an organism’s enacted world – its cognitive domain – as a developmental niche helps to clarify the structural coupling that occurs between an organism and its environment.

114

5.1 Making meaning through sense-making

Enactivism stresses that the dynamic coupling between organisms and their environments is

central to understanding cognition. This coupling arises out of an organism’s sensorimotor

engagement with salient environmental features. As discussed in the previous chapter, this

interaction is importantly asymmetrical, in that organisms do not merely passively respond to

environmental stimuli, but rather engage in active exploration of their environments according to

their individual biological needs. Therefore, environmental features have a particular valence

depending upon the status of the organism. For example, a starving animal may seek out more

risky or novel food sources than a sated animal, or a previous negative experience with a

potential food source may deter that animal from attempting to eat that item again. Certain

resources take on significance in relation to the organism’s need for maintaining its viability,

namely with regard to both metabolic needs and survival and reproductive success (i.e., fitness).

Naturally, maintaining viability in the face of perturbations (both internal and external) is

a fundamental part of keeping a living system alive. But the enactive approach goes further in

specifying the normative dimension that arises out of organism-environment coupling. On the enactive approach, it is not merely valence that is generated in this coupling, but meaning according to ‘graded norms of vitality’ (Thompson and Stapleton 2009, 25). To a tick, a mammal’s fur carries a certain meaning in virtue of it being a critical part of the tick’s enacted world – it is an indicator of a potential food source. There is thus an intentional aspect to the relationship between a tick and mammalian fur – the tick’s behavior is directed toward the fur as an indicator of a food source. Mammalian fur signifies in a way that other environmental settings do not. In its sensorimotor engagement, the tick’s enacted world is a particular world that is imbued with meaning relative to it.

115

‘Sense-making’ is the enactive term that refers to this ability to generate significance and

meaning and thus an enacted world. Sense-making is grounded in the organizational makeup of organisms as autopoietic, autonomous systems. It is therefore something that all living systems engage in, in virtue of their organizational makeup. Though the degree to which organisms engage in sense-making may vary, it is expressed in the enactive framework as being central to understanding cognitive behavior because it guides meaningful action with respect to individual organismic needs. Environmental features have significance because they stand in a functional relation to organismal needs. Sugar, for example, exists in a bacterium’s enacted world in order to satisfy its metabolic needs. As Thompson and Stapleton explain,

[s]ugar is significant to these organisms and more of it is better than less because of the way their metabolism chemically realizes their autonomous organization. The significance and valence of sugar are not intrinsic to the sugar molecules; they are relationship features, tied to the bacteria as autonomous unities. Sugar has significance as food, but only in the milieu that the organism itself enacts through its autonomous dynamics. (24-25)

If bacteria were unable to digest sugar, or otherwise had no use for it, it would cease to have such significance; indeed, we could say that it would not even ‘show up’ in bacteria’s enacted worlds.

Sensorimotor engagement with the world, then, creates a world of significance and meaning for an organism. Bacteria value nearby sucrose gradients, ticks value mammalian fur, and humans value a fragrant roast. This valuation suffices to explain organisms’ basic cognitive behavior as behavior that is in accordance with their organismal norms.

116

5.1.1 Intrinsic purposiveness in living systems

The intentional aspect of sense-making presupposes a teleological basis for sense-making activity. The bacterium’s behavior is directed at its enacted world in a context-sensitive fashion.

In traveling up a sucrose gradient, its aim is to collect and metabolize the sugar. However, early enactive work by Maturana and Varela (1980; Varela 1979) explicitly rejected the view that autopoietic systems had purposes, aims, or goals. Varela asserted that these notions were

“unnecessary for the definition of the living organization, and … belong to a descriptive domain distinct from and independent of the domain in which the living system’s operations are described” (Varela 1979, 63-64; quoted in Thompson 2007, 144). As Thompson notes, however,

Varela later revised this view27, ultimately suggesting instead that autopoietic systems maintain a

certain kind of purposiveness, and this notion plays an important explanatory role in addition to a

descriptive role. Following Kant’s writings on self-organization, mechanism, and teleology,

Weber and Varela (2002) describe living systems as being intrinsically purposive, or roughly, generating their own purposiveness in virtue of their self-organizational makeup28. On this reading, intrinsic (or immanent as Thompson suggests) purposiveness is unique to life and is what distinguishes living systems from mere machines.

There are two modes to purposiveness that explain a distinctive pattern of activity in living systems. The first mode, identity, highlights how an autopoietic system maintains a particular identity over time. The second mode is sense-making. In Thompson’s words:

27 It is worth pointing out that the term ‘sense-making’ does not appear in The Embodied Mind, though talk of ‘enacting significance’ does appear briefly in discussing how cognition is foremost a matter of embodied action (Varela, Thompson and Rosch 1991, 175). 28 There is a larger discussion and body of literature here that draws on Kant’s discussion of self-organization and teleology in the Critique of Judgement. In essence, what came out of the investigation of Kant’s work in this context is a specific kind of purposiveness – intrinsic, or immanent as Thompson suggests – that self-organizing systems appear to possess in virtue of the constitutive enabling relations that make up the autopoietic network. 117

An autopoietic system always has to make sense of the world so as to remain viable. Sense-making changes the physicochemical world into an environment of significance and valence, creating an Umwelt for the system. Sense-making, Varela maintains, is none other than intentionality in its minimal and original biological form (Thompson 2007 146-147; Thompson 2004; Varela 1997).

A living system’s purposiveness, then, is partly actualized through the active creation of meaning in making sense of its surroundings. Through the constant regenerating of itself as a self-

organizing unity, it creates its own identity, while at the same time creating an organism-relative

environmental milieu, an Umwelt or lived environment, populated by resources that have

significance for it for the sake of its active regeneration and viability.

According to Varela, these two modes of purposiveness, identity and sense-making,

constitute cognition. He explains: “the term cognitive has two constitutive dimensions: first its

coupling dimension, that is, a link with its environment allowing for its continuity as individual

entity; second its interpretative dimension, that is, the surplus of significance a physical

interaction acquires due to the perspective provided by the global action of the organism”

(Varela 1997, 81). On an enactive reading, the intentional relation29 an organism has to its

enacted world is thus, in part, central to a conceptualization of cognition.

29 For historical reasons, it is worth noting Varela’s justification of the import of intentional language here. He states: “it should be clear that the constitution of a cognitive domain links organisms and their worlds in a way that is the very essence of intentionality as used in modern cognitive science, and as it was originally introduced in phenomenology. My proposal makes explicit the process through which intentionality arises: it amounts to an explicit hypothesis about how to transform this philosophical notion of intentionality into a principle for natural science. The use of the term cognitive here is thus justified because it is at the very base of how intentionality arises in nature” (Varela 1997, 80-81). One reason for weaving intentionality into the enactive account, where before it was rejected, may be as part of the larger enactive project of naturalizing phenomenology. This is an intriguing route to pursue, but I do not have the space to do so here. 118

5.1.2 Sense-making as projective teleology

Di Paolo (2005) builds on Varela’s account of sense-making by drawing a distinction between intrinsic teleology and projective teleology. Autopoiesis succeeds at capturing intrinsic teleology, which refers to an organism’s ongoing self-production. The foundational normative dimension to sense-making is fulfilling metabolic needs as the “norm of self-continuance”

(Thompson 2007, 147). However, Di Paolo suggests that another concept is needed in order to capture the normative dimension of living systems that “admits of graded notions such as lacks and breakdowns and articulates in detail how signification is generated” (Di Paolo 2005, 437). Di

Paolo refers to this as adaptivity, or “the capacity of an organism to regulate itself with respect to the boundaries of its own viability” (430). Where autopoiesis captures a living system’s self- producing nature, adaptivity captures its self-interested nature. Drawing from work by Weber and Varela (2002), Di Paolo explains how this self-interestedness denotes a second sense of teleology – a projective teleology. He states: “this projective teleology, or sense-making (see also

Varela, 1991, 1997), would seem to follow naturally from the language of concernful self- affirmation, for an entity for whom its own continuation is an issue would immediately project this concern onto its surroundings” (Di Paolo 2005, 433). Sense-making thus arises not merely out of autopoietic continuance, but also out of respect to an organism’s graded norms of vitality and self-concern. Sense-making necessarily involves the active creation of significance through the projection of organismal needs onto the environment.

Autopoiesis Adaptivity

119

Intrinsic teleology: purposiveness is generated Projective teleology (sense-making):

in virtue of the system’s self-organizing purposiveness is generated in virtue of

makeup. system’s graded norms of vitality.

Norm of self-continuance: system operates Norm of self-interestedness: system operates

according to normative dimension according to normative dimension

corresponding to self-organizational corresponding to individual organismal needs.

mechanisms.

Table 1. Generation of normative dimensions based on autopoietic and adaptive features of the organism.

Table 1 shows the distinction between the normative dimensions generated by autopoiesis and

adaptivity. Where autopoiesis captures the normative dimension concerned with self-

continuance, adaptivity captures the normative dimension concerned with self-interestedness.

Behavior with regard to metabolic needs falls under the norm of self-continuance, for example, while behavior with regard to needs that cover a wider scope, such as survivability and reproductive success, falls under the norm of self-interestedness. It is behavior according to the normative dimension of self-interestedness that, on Di Paolo’s account, exemplifies sense- making, and thus cognition. The intrinsic teleology suggested by autopoiesis is not enough on its own to account for cognition. Cognizing systems must be both autopoietic and adaptive systems.

Sense-making generates a point of view, an attitude toward the world that is derived from an organism’s self-interested nature. A point of view is grounded in an organism’s needful concern for its vitality and viability. And it is this point of view that serves as a starting point for mindedness and lived experience. Therefore the notion of sense-making is central to

120

understanding not only the nature of cognition but also how mind30 emerges in living systems qua autopoietic, adaptive systems.

As Maiese (2015) notes, organisms are guided by non-metabolic values as well as

metabolic values. She draws attention to the fact that non-metabolic values, namely cultural or

social values, may even conflict with and override metabolic values (for example, in religious or spiritual asceticism). While adaptivity is meant to capture values beyond those generated by metabolic needs, norms of self-interestedness are grounded in concern for vitality, and there is a question of if an additional normative dimension is needed for social and cultural norms. Much of what we talk about when we talk about action being guided by norms and values, for humans, is in relation to our complex sociocultural environment31. Indeed, as Maiese stresses, the

question of how to make sense of non-metabolic values “suggests that the account of sense- making presented so far needs to be developed further in order to show how the dynamics of living organisms allows for more complex forms of meaning and value to emerge” (21). While it is not my goal to elaborate on this concern here, one way forward may be to specify a sociocultural normative dimension that identifies a norm of group-level or collaborative interests. To

30 One of the central goals of the enactive approach is to explain a continuity between life and mind – that there is a similarity relation between the organization of living systems and the organization of things with minds. So in describing how the autopoietic, adaptive organization of living beings necessarily results in the enactment of a value-laden world onto which organismal concerns are ‘projected’, enactivists hope to fill in the explanatory gap between life and mind. Weber and Varela appeal to the work of phenomenologist Hans Jonas (1973) in order to show how it is through our own lived experience that we come to understand the teleological nature of life. They explain: “[i]t is actually by experience of our teleology – our wish to exist further on as a subject, not our imputation of purposes on objects – that teleology becomes a real rather than an intellectual principle. Thus causality, as it is perceived by us as sentient beings, may be subsumed under the more general principle of life” (Weber and Varela 2002, 110). The grounding of intrinsic purposiveness in autopoiesis is meant to lay a biological foundation for phenomenological theorizing that in turn serves to bridge the gap between life and mind. The appeal to intentionality and teleology is thus central to the enactive goal of illuminating the strong continuity between life and mind. 31 I think food provides a very salient example here. Human beings go to great lengths to procure not just things that qualify as food but food we enjoy (for gustatory, but also psychological reasons). For humans foodstuffs do not have value merely as sustenance but have a rich, complex value-laden nature that is heavily influenced by sociocultural norms. 121

be sure, more work is needed to elaborate on the enactive picture of sense-making as being grounded in these normative dimensions.

5.2 Constructing an enacted world

Sense-making is a kind of minimal, biological intentionality. It captures the function of certain

self-affirming processes, such as searching for food, evading predators, or calling for mates. On

the enactive account, these behaviors are examples of undergoing sense-making, as their function is in reference to an organism’s self-interestedness and to the enacted world as one that is rich with significance for that organism. Meaning is generated through sense-making, in that certain resources become food for an organism, certain organisms become potential mates and others predators, and so on. An organism’s enacted world is imbued with meaning through sense- making.

As detailed above, according to autopoietic enactivists, the intentional nature of sense- making is not merely descriptive but is explanatory as well. We can explain why organisms behave the way they do in virtue of their enactive adaptivity and their projective teleology.

Behaviors are expressed in order to self-regulate over the individual’s lifetime. This picture is intuitive on its face; much of what organisms (of all capacities) do is in self-interestedness and directed toward basic viability. Sense-making applies to all cognizing systems qua living systems, and indeed all living systems in virtue of the deep continuity between life and mind

(Thompson 2007). This means that even the “simplest” system, for example a bacterium (and the bacterium moving along a sucrose gradient is the canonical example in the literature) engages in sense-making, and lives in an enacted world rich with meaning. Both Di Paolo and Thompson note that minimal autopoietic systems – those that lack an internal metabolic network that allows

122

for flexibility – are incapable of being adaptive systems, however, and thus do not undergo

sense-making.

A question that naturally follows from the claim that organisms enact a world through

sense-making is exactly how organism-specific worlds are enacted and what norms govern what elements populate them. In this section, I will work through an alternative conceptualization of how organisms construct their enacted worlds by appealing to insights from developmental biology. I argue that rather than thinking of meaning as arising out of the structural coupling between an organism and its environment, from a developmental perspective, we can think of coherence emerging as a result of interactions between features of these systems. Thus, it is according to a principle of coherence that sense-making occurs. I will begin by discussing the enactive view on organismal fitness relative to their ecological niches as this will set up the biological framing in which sense-making as coherence can be investigated.

5.2.1 Coherence and reliable reconstruction of the life cycle

Adaptations refer to characteristics that allow for a better fit between the organism and its environment and thus an organism that is better adapted can be said to be maximizing its fitness.

Behavioral traits contribute to an organism’s fitness when they aid in the survival of that organism and in its reproductive success. The fitness of an organism is context-dependent; an organism may be particularly well-suited for one environmental setting and poorly suited for a different environmental setting. Evolution by natural selection requires that variation between individuals is heritable, such that some lineages may have a greater degree of reproductive success (and therefore fitness) than others.

123

Talk of maximization of fitness implies that there is an ideal way for the organism to be in order to be well-suited for its particular environment (hence the phrase “survival of the fittest”).

However, autopoietic enactivists have proposed a different perspective on which “all living beings are adapted as long as they are alive” (Thompson 2007, 205). The rationale for this perspective is based on the notion of structural coupling between an organism and its environment. As Thompson explains:

Like two partners in a dance who bring forth each other’s movements, organism and environment enact each other through their structural coupling. Given this view of organism-environment co-determination, it follows that evolution should not be described as a process whereby organisms get better and better at adapting to the design problems posed by an independent environment … As long as a living being does not disintegrate, but maintains its autonomous integrity, it is adapted because it is able to carry on its structural coupling with its environment. (Thompson 2007, 204; Maturana and Varela 1987)

On Thompson’s conception of a theory of ‘enactive evolution’, the developmental system, as a network comprised of the organism as a developmental process (as discussed in Chapter 4) and resources or environmental elements as the developmental niche (which will be discussed further in the following section), can be analyzed in virtue of its viability rather than its fit or optimization. This view requires, first, that what is analyzable in an evolutionary context is not singular traits or genes but the developmental system as a whole. This point is important because there is no strict separation to be made between a trait and its developmental and evolutionary setting; the manifestation of certain traits is always context-sensitive. Second, developmental systems are evaluated on the basis of their viability because systems that remain viable in each reconstruction (i.e., each instantiation of a life cycle) are the ones that persist over evolutionary time.

124

Varela, Thompson and Rosch (1991) urge that “we must ask how, within the general

constraints of the laws of nature, organisms have constructed environments that are the

conditions for their further evolution and reconstruction of nature into new environments”

(Varela, Thompson and Rosch 1991, 202). One way to answer this question in the enactive

framework is to suggest that it is through sense-making that organisms construct environmental settings with the appropriate conditions for their viability. With respect to the normative dimension of self-interestedness, organisms change their environments to suit their needs – they build shelters or modify existing ones, they develop warning systems to alert conspecifics of predators, they ensure reliable access to food sources by regular, transgenerational behavioral means32, and so on. As stressed in the quote that began the dissertation, organisms both shape

and are shaped by their environments.

According to the enactive concept of sense-making, meaning is created through this

structural coupling. Sense-making tells us that there is meaning in virtue of structural coupling,

but more can be said about the contingent relations that result in the coupling of two systems.

Sensorimotor contingencies between an organism’s body and the resources it can interact with

might give us a starting point. Ultraviolet light means something to bees in the way that it does

not for humans. From a developmental perspective, however, it is possible to work through a

richer account of these contingencies, how they are formed, and how they result in the sort of life

forms that they do.

I want to suggest that, from a developmental perspective, the process of sense-making

results in coherence between features in a developmental system. In striving toward its own

32 I am predominantly referring to agriculture here, which it is worth noting is not limited to humans. Some ants have been known to “farm” aphids by protecting them from invaders and harvesting an energy rich honeydew from them (Stadler and Dixon 2005). 125

viability, the organism enacts enabling relations between processes within the developmental

system. It is the organism as the developmental process that ties the whole system together

through its projective concern for viability. What is ‘made sense of’ is whatever is within the

network; such making sense is possible due to the creation and construction of coherence

between processes. The ability to perceive ultraviolet light and the existence of ultraviolet

colored objects in the world are two phenomena that cohere with one another in a particular way

for a particular organism. Importantly, they cohere due to a history of structural coupling

between the organism and its environment which in turn is due to the sensory information in the

environment and the organism’s sensorimotor capacities to exploit that information.

Coherence creates links between processes that can become species-typical behaviors if

reliably reconstructed in each life cycle. According to DST, an individual organism is construed

as the enactment of “one cycle of a complete developmental process – a life cycle” (Griffiths and

Gray 2001, 209). The developmental niche, which I will discuss in detail in the next section, is

comprised of resources needed for the reliable reconstruction of a life cycle. For example, the

nests and shelters that animals build to house their offspring are structures that are inherited in

the developmental niche and aid in species-typical development. The coupling between nest shape and egg shape, for example, is an example of how those two elements cohere with one another. On this perspective, what is enacted in the building of a particular style of nest is coherence rather than significance – an animal builds a nest as a way of cohering with other relevant elements in its niche, not necessarily because the nest has a particular meaning to it.

Making sense of the environment amounts to behaving coherently in a developmental niche

rather than acting on meaningful information.

126

Drawing on Griffith and Gray’s example, hermit crabs that experience a dearth of

suitable discarded shells in which to live face significant selection pressure, but “[v]ariants with

a beneficial set of behaviors or a beneficial habitat association that allowed them to continue to

reliably reestablish their relationships to shells would be favored by selection” (208). The hermit

crabs do not assign meaning to a novel shell resource but instead create a new coherence relation

between them and the novel shells. So long as the coherence relation continues to be

reconstructed in each life cycle (and is not maladaptive), it may result in a species-typical behavioral trait.

5.2.2 Specifying the cognitive domain

I am not using the term ‘cognitive domain’ here in any specific technical sense; it is simply meant to refer to an organism-specific environmental state-space in which cognitive activity takes place. We can think of a bee’s cognitive domain, for example, as the environmental state- space containing whatever is potentially perceivable and actionable by the bee. Whatever those elements are will depend upon the bee’s sensorimotor capacities and its history of structural coupling. On an enactive reading, the bee’s enacted world is its cognitive domain, in virtue of its autopoietic, adaptive configuration.

Determining what elements populate an organism’s cognitive domain requires careful investigation of its sensorimotor capacities and its coupled history with its environment (the details of which matter, and will be addressed in what follows). For example, eyes are ubiquitous throughout the natural world. Yet even in locations where vision is seemingly no longer worth investing resources in, such as in subterranean habitats, eye structures, though reduced, persist

(Nevo 1979; Nikitina et al. 2004). Normal development for the naked mole rat (Heterocephalus

127

glaber), for example, results in a reduced eye structure, but an ocular phenotype nevertheless.

Researchers have asked why this trait still develops despite at least 25 million years of subterranean evolutionary pressures (Nikitina et al. 2004; Bennett and Faulkes 2000). Just like any organ, eyes have metabolic costs, and so individuals who direct those resources elsewhere may be better off. Nikitina et al. suggest, however, that a closer inspection of the environmental stimuli and the mole rats’ regular activities can provide clues as to why the phenotype has not been completely selected against:

… retaining the capacity for light‐dark discrimination is important for the survival of these animals. The soil‐removal activity of the naked mole rats results in their direct exposure to sunlight, as the animals kick soil out of an open mound. The open mound poses a further threat of exposure to aboveground predators (Sherman et al. 1991). An ability to detect light and dark and sudden transitions associated with the arrival of a predator at a well‐lit burrow entrance may confer a survival advantage and hence be maintained by natural selection. (Nikitina et al. 2004, 331)

While the species-typical habitat of the naked mole rat is categorized as being distinctly subterranean, brief instances of direct exposure to sunlight is enough to serve as an environmental pressure necessitating the retaining of an eye structure. The payoff of these reduced eye structures is a discounted metabolic cost along with a sufficient capacity for light- dark discrimination. These findings support an account of perception that stresses how perception is actually used – while there are no scientific findings that suggest that the mole rats use visual information in the way animals with fully-developed eyes typically do, their eyes still make use of environmental information – specifically, transitions in brightness. They may not be able to ‘establish perceptual contact’ with objects in the world (including conspecifics, which

128

they identify through olfactory and tactile cues [Nikitina et al. 2004, 331]), but, in a sense, light

still affords them seeing, albeit an unconventional mode of seeing.

This example is meant to show how careful investigation of the specifics of the

relationship between organisms and their environments matters to how we think about cognition.

The mole rats’ habitat is not merely a subterranean one, and the particular way they interact with

that environment, even if in brief moments, can end up impacting their evolutionary trajectory.

As we saw, occasional surfacing in the activity of burrow building has generated enough

selective pressure to retain minimal eye structures. It may not be a conventional way of using

eyes, but it works for the mole rats.

Persisting both at the developmental and at the evolutionary scale is often a matter of

getting by on what works rather than maximizing potential. Indeed, Varela, Thompson and

Rosch note this biological fact in their reference to evolution as natural drift33, stating a call for recasting selective pressures as “broad constraints to be satisfied” (Varela, Thompson and Rosch

1991, 198). They refer to this satisficing principle in describing the enactive notion of mutual specification, or the Lewontin-inspired notion of codetermination:

The key point, then, is that the species brings forth and specifies its own domain of problems to be solved by satisficing; this domain does not exist “out there” in an environment that acts as a landing pad for organisms that somehow drop or parachute into the world. Instead, living beings and their environments stand in relation to each other through mutual specification or codetermination. Thus what we describe as environmental regularities are not external features that have been internalized, as representationism and adaptationism both assume. Environmental regularities are the result of a conjoint history, a congruence that

33 See Chapter 2 for more detailed discussion on the enactive concept of ‘evolution as natural drift’. 129

unfolds from a long history of codetermination. In Lewontin’s words, the organism is both the subject and the object of evolution. (198-199, original italics)

This coupled history matters, particularly in instances where interacting features are both

organisms. Pollinators and angiosperms are typically thought to have a mutualistic relationship,

with one organism relying upon the other for its survival and reproductive needs. But an

established coupled history between these organisms matters for their viability. Aizen et al.

(2014) note that seemingly mutualistic relationships between a native organism and an invasive

organism may result in detrimental effects to the native organism. There is a lack of a coupled,

shared history between the two species, resulting in an imbalance in costs and benefits to each.

An established relationship matters for how environmental features modify organisms over

developmental, behavioral, and evolutionary timescales – features may be exploited (or not) with

a variety of effects over these timescales. And importantly, organisms in turn modify these

features which results in the generation of new developmental and evolutionary effects, as

discussed in the literature on niche construction (e.g. Odling-Smee, Laland, and Feldman 2003).

So the details of interaction matter, especially for a biological approach to cognition. If the goal

is to understand cognition as a biological phenomenon, we need to look at how it is actually

used, down to the individual differences and peculiarities such as those seen in the naked mole

rat. One approach to investigating the details of interaction is to consider the environment at the

scale of the individual; this requires specifying the environment in a more fine-grained manner.

5.3 The environment as developmental niche

Sense-making as coherence captures how the contingent relations in organism-environment structural coupling shape an organism’s enacted world. It also helps to pick out an organism’s 130

cognitive domain and the elements that populate it. In this section, I discuss different ways of

specifying the environmental setting that is the organism’s enacted world before ultimately

arguing that we ought to think of the organism’s cognitive domain as mapping onto its

ontogenetic or developmental niche. It is through active exploration of the developmental niche

that organisms make sense of their environments, and as a developmental niche is partially

constructed by the organism itself, it exists relative to a particular organism, thus serving as the

unique enacted world of that organism.

5.3.1 Multiple senses of environment

Brandon and Antonovics (1996) suggest that, in the field of population biology, there are three

ways to distinguish between conceptions of the environment for the sake of understanding

organism-environment coevolutionary dynamics. These conceptions differ based on what sets of environmental factors are taken to be relevant in the generation of selective pressures. The first sense of environment is purely external – the environment is constituted by a set of factors independent of an organism of interest, and these factors are measured independently from an organism of interest. Brandon and Antonovics suggest that a fundamental problem with conceptualizing the environment in this manner, however, is that because these factors are measured entirely independently from the organism, they may turn out to be irrelevant to an organism’s fitness, and thus do no work toward an understanding of how and why populations evolve. So a conception of the environment that merely identifies the set of physical factors external to organisms is insufficient.

The second conception of the environment is the ecological environment, which utilizes organisms as ‘measuring instruments’ (Brandon and Antonovics 1996, 164) to determine the

131

external factors as they affect population growth. This conception effectively picks out features

of the environment that are relevant to a particular lineage such that identifying them sheds light

on that lineage’s evolutionary trajectory (Griffiths and Gray 2001). Brandon and Antonovics

note that a further step is needed in order to compare fitnesses of different genotypes. A third

conception of the environment as selective allows for comparison between genotypes in relation

to pressures from the environment. As the goal with this approach is to be able to measure

organism-environment coevolution by way of assessing organism-relative factors, specifying an environment as a narrower set of features can aid in understanding why some genotypes fare better than others. The selective environment, then, is the appropriate sense of the environment to consider when comparing individual genotypes, and thus individual differences that may over time be selected either for or against.

One way to further parse out these differing senses of environment is in a developmental context, albeit one with relevant evolutionary implications. Each sense of environment can be said to be constituted by a set of resources organisms can make use of, and some of which are necessary for the transgenerational stability of form (Griffiths and Gray 1994). Identifying the structure of resources available to organisms is a central goal of developmental systems theory

(Oyama 2000). By specifying individual domains with unique sets of resources, it is possible to identify developmental resources which may arise during the interaction of organisms as developmental processes and the environments in which they are situated. For example, persistent resources may be those specified by the notion of an external environment, such as temperature, gravity, and light. These resources play a role in the developmental process and may potentially be relevant to organismal fitness, but they are not identified with reference to a particular organism and exist independently and regardless of any organismal interaction. The

132

sets of resources specified in more fine-grained scales are organism-dependent and are thus

factors in the ecological sense of environment. Resources specified at an even finer-grained scale

can arguably fit into the sense of a selective environment. The availability of these resources

highlights interactions that take into consideration individual differences in behavior.

Organisms, as developmental processes themselves, make use of resources across each domain depending on a shared history of interaction and with regard to individual needs. What is relevant for the sake of providing an evolutionary explanation is identifying recurrent interactions between resources and organisms with the capacity to utilize those resources.

Variation arises when resources are utilized in a new manner, when new resources are introduced or existent resources are removed, the relationship with resources is altered, and so on. In this way, the developmental system is comprised of not solely an individual organism interacting with an environment over the course of its life cycle, but rather it extends over both the organism as a developmental process and the developmental resources with which it is coupled such that interactions with those resources constitute a species-typical life cycle. This picture places greater emphasis on the environment itself in understanding the life activity of the organism than traditional accounts of ontogeny do, as developmental resources (potential or actual), are integral to the specification of the system as a developmental system. Without reference to these features, the resources available for explaining the transgenerational stability of organismal form are impoverished. Specifying the environment and building that specification into an understanding of organisms as developmental processes embedded within a larger developmental system results in a richer account of ontogeny, with greater explanatory power across a developmental timescale, but also a behavioral and, on a population scale, an evolutionary one.

133

In ecological psychology, a research field in perceptual psychology that shares some

similarities with the enactive approach, comparable attempts have been made to parse out

different senses of the environment. Baggs and Chemero (2019) distinguish between the physical

world, a species habitat, and an individual organism’s umwelt. Here I think it is helpful to map

this distinction onto the three-way distinction between senses of the environment described by

Brandon and Antonovics. The physical world approximately corresponds to the notion of an

external environment– it is not specified in relation to any particular organism. The sense of the environment as a habitat is species-specific, and contains affordances as resources typical for that species. The third sense of environment, the umwelt, references Jakob von Uexküll’s concept of a particular organism’s lived environment; it is a behavior setting that is “shaped by the places where that individual dwells, and by the history of interactions that the individual participates in”

(Baggs and Chemero 2019, 16). An individual organism’s umwelt, then, references its unique abilities and experiences to determine which features of the world are especially salient to it given these properties.

The goal in introducing this three-way distinction between the physical world, a habitat, and an umwelt is to resolve tensions in ecological psychologist James Gibson’s original distinction between a perceiver-independent physical world and an affordance-containing yet ambiguous surrounding environment, which roughly corresponds to the notion of a habitat on

Baggs and Chemero’s three-way distinction. What the sense of an umwelt is meant to do, in this context, is specify exactly how individual differences result in different affordance spaces. A species-specific habitat contains environment features that are utilized in a species-typical fashion, thus referencing an idealized member of that species. The bee orchid (Ophrys apifera) that successful tricks male bees into thinking they are encountering female bees presumably

134

tricks all male bees, but the one male bee that does not fall for this trick does not act on the

affordance in a species-typical manner. The world, to this clever bee, appears differently – there

is no female bee to encounter, only an equally clever orchid plant. Thus, an umwelt, as an

individual-specific, third sense of environment, allows for individual variation, which as

Withagen and Chemero (2009) note, is essential for an understanding of evolution on a naturalized account of perception.

5.3.2 The developmental niche

The sense of the environment as an individual-specific environment, I want to suggest, can be built upon by further specifying how structural features of the individual-specific environmental milieu both shape and are shaped by coupled features of the individual organism. This task requires looking at the environment, understood as an organism’s cognitive domain, as an ontogenetic or developmental niche34. Individual variation in perceptual activity can be

investigated in relation to the developmental niche in which individual organisms live. Here I

suggest that specifying the cognitive domain in which organisms are situated as their own

developmental niches provides a framework for understanding the environment as an organism-

specific enacted world.

West and King (1987; see also Stotz 2014) suggest that the concept of an ontogenetic

niche can aid in identifying the set of developmental resources that an individual inherits in

addition to genes. The social, cultural, and ecological circumstances that an organism is born into

play a prominent role in its developmental trajectory. For example, West and King (1988; see

34 While West and King use the term ‘ontogenetic niche’, Griffiths and Stotz note that they use ‘developmental niche’ as a synonym in their work (Griffiths and Stotz 2018; see also Stotz 2008; Stotz 2010; Griffiths and Stotz 2013). I will use ‘developmental niche’ here to make it clear that I am drawing mainly from DST. 135

also Smith, King and West 2000) found that the presence and response of female cowbirds had a significant effect on male song development. Identifying this social influence as a parameter in the male cowbird’s ontogenetic niche guides understanding of what factors are relevant in the species-typical development of singing behavior. This influence on song learning and development can have transgenerational effects, making it the case that the multimodal (both visual and auditory) sensory feedback from social interactions can serve as an inherited resource.

Griffiths and Stotz (2018) describe a developmental niche as the “set of parameters that must be within certain bounds for an evolved life to occur (or, in more traditional terms, for the organism to develop normally” (Griffiths and Stotz 2018, 237). Importantly, they distinguish between a developmental niche and the selective niche described by niche construction theory, which they define as “the set of parameters that determine the relative fitness of competing types in the population” (ibid.; see also Stotz 2017). While the selective niche picks out elements that generate selective pressure on an organism, the developmental niche picks out elements that are relevant for the species-typical development of an organism.

The developmental niche is part of the larger developmental system; it identifies the environmental setting or context in which a developmental system constructs a life cycle. It is the set of parameters that “play a role in the modification and reproduction of the life cycle” (Stotz

2017, 2). The relevant parameters may be not just physical resources but also “social, ecological and epistemic” (ibid.) resources that aid in the reliable reconstruction of a life cycle (in other words, an individual organism). These resources are inherited in the reconstructing of a life cycle. The claim that extragenetic resources are inherited within the context of a developmental system is a key aspect to DST and differentiates it from traditional accounts of ontogeny.

136

An individual organism’s developmental niche is its own unique environmental setting,

morphed by its interaction with resources within the niche just as those resources impact it. It

enjoys a unique history that unfolds over time through the specific interactions that take place

within the larger developmental system. Whether or not species-typical phenotypes are exhibited is dependent upon specific kinds of interaction between an organism and the resources within its developmental niche. Changes in interaction potentially have a generative effect over time. Shifts in developmental niche are possible through variation in behavior.

Recall that in Oyama (2000), organisms are conceived of as integral parts of a larger developmental system which contains environmental resources that act on and are acted upon by the organism in that system. The developmental system is comprised of a complex web of interactions that impact how the organism develops and changes over its lifetime. In this context, an organism’s developmental niche can be thought of as the specific environmental setting that is comprised of inherited developmental resources part of a larger developmental system. Thinking of the environment as an individual organism’s developmental niche makes it clearer how organisms form certain relationships with certain environmental elements (including conspecifics) and how those relationships can change (and new ones created) over developmental, behavioral, and evolutionary time. This conceptualization leaves room for the creation of new coupling processes via individual innovation, potentially leading to new features of both the organism and its environment.

Developmental systems are dynamic and in constant flux, with new iterations (i.e., new generations) impacted by prior individual variations acted upon by selection over time. This view of the environment thus avoids potential issues with circularity that may arise if the improper environmental scale is considered. The concern here is that on a generalized account of the

137

environment, an organism-environment system is markedly circular – organisms pick up relevant

environmental stimuli, and environmental stimuli serve as a resource for those organisms. This

picture does not tell us why the coupling has arisen or why it persists. Bees perceive UV light,

and UV light is perceivable by bees. But this was not always the case; bees did not pop into

existence ready to utilize UV light as an action-guiding visual cue. In a similar vein, one hypothesis for the evolution of trichromatic color vision in some primates was the ability to pick out colored fruit (Allen 1984 [1879]). A generalization of this coupling does not tell us how organisms move, evolutionarily, through color space. It properly identifies a coupled system, but provides only a synchronic account of that phenomenon. If we are convinced that a dynamic approach to understanding perceptual, cognitive activity is a fruitful way forward, we must look at the environmental setting in which this activity occurs across multiple timescales – developmentally, behaviorally, and evolutionarily. I have attempted to illustrate that the biological resources for looking at the individual organism across these scales (both spatial and temporal) are plentiful, and thus a biological approach to cognition ought to make good use of them.

In accordance with DST’s concept of a developmental niche and Baggs and Chemero’s sense of the environment as an individual-specific umwelt, I suggest that one fruitful way of specifying the cognitive domain is to characterize it as an individual-specific developmental niche. On this view, conceptual tools from the enactive approach can be put to use alongside conceptual tools from DST to result in a cohesive framework for understanding cognition as a biological phenomenon.

Enactivists speak of the ‘enacted world’ of an organism as emerging from perceptual interactions contingent upon sensorimotor capacities. The enacted world is populated by

138

environmental features with potential valence to an organism depending on the internal needs of

the organism at a specific time. The enactive approach is thus a naturalistic one, as it draws on

the biological factors involved in cognition for explanatory purposes. But exactly what features

populate an organism’s enacted world is dependent upon not only a history of coupled

interactions between its species and the ecological environment, but also between an individual

and its developmental niche. Importantly, these interactions are dynamic, with some couplings

strongly conserved over time (such as eye structures and light stress [Nilsson 2009; Oakley and

Speiser 2015]) and others in flux during an individual’s life cycle (such as differentiation in

abilities enabled by learning). Thus, the cognitive domain of an organism shapes and is shaped

by that organism’s influences on various timescales, making the enacted world a dynamic one

emerging out of a complex web of interactions as a result of both individual experience and

innovation35 as well as species-typical behavior.

Importantly, the developmental niche need not be thought of as being populated solely by

biological or ecological factors. Social and cultural affordances play a large role in guiding

action for humans (Rietveld et al. 2013) and arguably for non-human animals as well (Avital and

Jablonka 2000). The resources an organism inherits in a niche include both physical resources

such as food and shelter but also the potential for social interaction with conspecifics and

behavioral traditions such as those seen in West and King’s cowbirds. Individuals inherit a

developmental niche, but continue to shape it over time via their own behaviors and with regard

to their own interests. Although beyond the scope of this chapter, the close investigation of the

35 One example of (often individual) innovation having a downstream effect is the notion of ‘cumulative culture’, or the “ever-increasing, additive complexity or efficiency of cultural performance over time” (Schofield et al. 2018). The sweet potato washing behavioral repertoire of Japanese macaques began with a single innovator, with the behavior quickly spreading to others in the group. It is noted that the washing has the effect of reducing parasitic infections, thus suggesting an adaptive aspect to the behavior as well. 139

social and cultural affordances in an individual’s developmental niche may reveal valuable

insights about individual variation and change.

By specifying the cognitive domain as an individual’s developmental niche comprised of

developmental resources, and as an integral part of the larger developmental system, it is

possible to gain a better understanding of why organisms engage with the sort of features that

they do, and how they are able to act on perceptual information in the way that they do.

Importantly, this account provides us with a way of looking at how novelty, such as the move

from dichromacy to trichromacy, may have been generated as a result of complex interactions

between organisms and features within their environment. But it also suggests the need for a

complementary psychological view that emphasizes the dynamic relationship between organisms

and their environments, across multiple spatial and temporal scales, and it is here that I think

biological enactivism has resources to contribute.

In this chapter, I investigate the enactive concept of sense-making, which captures how

organisms enact worlds with significance and meaning. I suggest that looking at sense-making from a developmental perspective entails thinking of it as the generation of coherence between organisms and environmental resources or elements. Talk of coherence between elements of the organism and elements of its developmental niche provides an avenue for understanding the dynamic coupling between organisms and their environments as developmental niches. This picture elucidates the mechanisms underlying the enactive constructing of a world in such a way that fits neatly with claims from DST.

140

Chapter 6: Conclusion

6.1 Summary

In this dissertation, I argue for a new variety of enactivism intended to bring the enactive

approach to cognition together with recent developments across a number of fields in biology, including developmental biology, evolutionary biology, and behavioral ecology. I suggest that this bridging results in a research framework that is well-suited for studying cognition in a

biological context. It is one that takes seriously the complexities and nuances in those subfields

of biology that are relevant to it, and directly addresses the project of how to weave together

research findings across cognate fields. This highly interdisciplinary, pluralistic approach has the

potential to shed new light on the biological basis of cognition while at the same time providing

more support for the robustness of the enactive approach as an alternative to traditional theories

in cognitive science.

I present this framework by working through conceptual claims in the enactive approach

and analyzing their application in the biological domain. Chapter 2 begins this project by

providing an overview of the enactive approach to cognition. I suggest that what is in places

referred to as ‘autopoietic enactivism’ is the variety of enactivism that is best suited for

integration with theoretical perspectives from alternative views in biology such as developmental

systems theory and extended interpretations of evolutionary theory. Indeed, as Thompson (2007)

points out, developmental systems theory and the enactive approach put forth in that work are

complementary to one another. The catalyst for this project is, in part, this claim, and Chapter 2

141

serves as a call for further work on how precisely these two perspectives fit together. I suggest that we can think of this project as a synthesis, and one that I refer to as ‘biological enactivism’ for the sake of differentiating it from extant varieties. Biological enactivism investigates the dynamic relationship between organisms and their environments, emphasizing their reciprocal nature, and focuses on how certain biological phenomena give rise to cognizing systems qua living systems. While general claims in this line of thinking parallel those found in writings in the enactive tradition, my goal is to explore new insights from biology through an enactive lens.

Chapter 3 introduces the Extended Evolutionary Synthesis (EES) as a novel research program intended to apply new theoretical, conceptual, and empirical insights to standard evolutionary theory. I offer two reasons why extended interpretations of evolutionary biology are particularly well suited both for the integration with the enactive approach to cognition and to the study of cognition in nature more generally. The first reason is that some of the theoretical commitments in the EES parallel those in views in cognitive science that reject the standard computationalist approach to cognition. Recall that the enactive approach is meant to serve as an alternative to computationalism; therefore lines of argument in the debate in evolutionary biology over the EES and views originating from the Modern Sythesis (MS) research program may have bearing on the similar debate in cognitive science. The second reason is that the central commitments in the EES, reciprocal causation and constructive development, are also central commitments in biological enactivism. An investigation of phenomena that fall under these considerations on a shared EES/enactive view has the potential to shed new light on some of the developmental and evolutionary mechanisms that are responsible for generating those phenomena. As many of these mechanisms, such as non-genetic inheritance, are arguably ignored or are deemed secondary by traditional evolutionary theory, a framework that has the

142

theoretical and conceptual tools to explain their relevance may put pressure on the traditional

view – in both biology and cognitive science.

Returning to the project specified in Chapter 2, Chapter 4 investigates the organization of

cognizing systems qua living systems. There, I argue that, fundamentally, organisms are self-

organizing, operationally closed, plastic systems embedded within a larger developmental

system. In virtue of this organizational makeup, they have the ability to actively assess and

modulate their coupling with their environment. This organizational makeup serves as the

precondition for cognition. In order to support these claims, I draw on resources from dynamical

systems theory, DST, evo-devo, and enactivism. Using these resources, I offer an account of

plasticity that describes how organisms undergo structural modification and how this capacity

serves as a precursor to cognitive behavior.

In Chapter 5, I analyze the enactive concept of sense-making, which on some enactive readings, simply is cognition. Sense-making captures an organism’s engagement with its enacted world as a world that is rich with significance and meaning in virtue of its organismal needs. I argue that a reformulation of sense-making as coherence between an organism and its enacted world fits more naturally with the view of development and evolution that I am putting forth in this work. With this formulation of sense-making in mind, I suggest that we can think of an organism’s cognitive domain – its enacted world – as its developmental niche. In so doing DST is brought even closer together with the enactive approach for a shared conceptualization of the environment in organism-environment systems.

By looking more closely at the mechanisms at work in relationships of structural coupling (and thus codetermination) of organism-environment systems, the enactive approach and complementary views in biology can be brought closer together for a shared space in which

143

to investigate these phenomena. Thinking of the organism as a developmental process and the locus of interaction between multifarious developmental resources highlights the reciprocal, dynamic nature of the relationship between an organism and its environment. In turn, it supports a view of the organism’s developmental niche as its enacted world, a world which develops as a result of these reciprocal interactions. This picture rejects the shared computationalist and adaptationist view of the organism as a passive receiver of environmental stimuli in favor of an enactive, developmental view of the organism as an active constructor of itself and its niche. As discussed in the beginning of the dissertation, the organism is both the subject and the object of evolution.

6.2 Future directions

My main hope is that this work inspires further collaboration between researchers working in enactivism and the non-mainstream research programs in biology that I have discussed here. As a project that is at its heart about the integration of several fields that I think yearn for integration, it has a necessarily broad scope. There are many further projects that deserve attention. In the following section, I will address a few of these projects, and some potential ways forward for pursuing them from a biological enactive perspective.

Ethology is the study of animal behavior; cognitive ethology is the study of animal cognition. Allen and Bekoff (1999) define cognitive ethology more precisely as “the comparative, evolutionary, and ecological study of animal thought processes, beliefs, rationality, information processing, and consciousness” (Allen and Bekoff 1999, ix). The field of cognitive ethology, defined in this way, is situated within the cognitivist, computationalist paradigm. To

144

study animal cognition, according to cognitive ethology, is to study the mental content of

animals.

A cognitivist approach to animal minds may not necessarily be subject to the same

criticisms and complaints waged against traditional cognitivism. Some animal cognition might

be more straightforward to study and might lend to success on an information processing

paradigm, such as in the cases of classic model organisms Drosophila and C. elegans. However,

this approach may end up generating significant issues as well. For example, on a view of

cognition as information processing, some organisms may end up looking like simplistic

“minirobots” when in fact some bodies of research suggest that they actually have quite complex

cognitive capacities (see Chapter 4 for discussion). If the rebuttal is to claim that they perhaps then have complex information processing capacities, the same argument against the information processing view for humans can then apply to non-human organisms. The problem is not

increasing the degree of complexity to the appropriate level, it is capturing what is explanatorily

relevant in understanding the cognitive agent. And as the enactive approach argues, information

processing is explanatorily insufficient.

With these concerns in mind, it is worth considering what an enactive approach to

cognitive ethology might look like. To date, these fields have yet to interact, though they have

much in terms of overlapping interests. Cognitive ethology aims to explain animal minds;

enactivism aims to explain the minds of living things. If enactivism has the theoretical and

conceptual tools to solve some internal issues within cognitive ethology, it is of worth to that

field. But importantly as well, if it can serve as a more appropriate theoretical framework for

studying animal minds than traditional cognitivism, then that works as evidence to show it can be

145

a more productive research paradigm than cognitivism. So the application of enactive thinking to the study of animal minds is a worthwhile endeavor.

Cognitive ethology themes can be found in Darwin’s writings on animal behavior and mentality (Darwin 1871, 1896). Darwin offered anecdotal accounts of complex animal behavior and suggested a mental continuity between humans and non-human animals. Indeed, in Chapter

2 of The Descent of Man Darwin explicitly states that his “object in this chapter is solely to show that there is no fundamental difference between man and the higher mammals in their mental faculties” (Darwin 1871, 35). Jamieson and Bekoff (1992) refer to Darwin’s work on animal mentality as ‘anecdotal cognitivism’, given his reliance upon field observation rather than controlled experiments for evidence of such mentality.

Lorenz and Tinbergen employed similar anecdotal approaches in studying animal behavior. Tinbergen’s development of a matrix of four levels on which behavior could be analyzed helped to formalize the research program. Each level corresponded to a distinct question that biologists were posing.

146

Figure 5. Bird song as an example of an integrative approach to Tinbergen’s ‘four questions’. (A) details bird- song learning mechanisms. (B) details current utility of song function. (C) details developmental mechanisms for singing capacity. (D) details the evolutionary history of songbird populations. (From Bateson and Laland 2013)

Referred to as Tinbergen’s ‘four questions’, (Tinbergen 1963; Bateson and Laland 2013), these questions were intended to guide sufficient explanations for particular behaviors and traits.

Lorenz and Tinbergen’s work helped to shape the field of ethology, and brought along valuable insights for the later development of the subfield of cognitive ethology. Yet there remains a question of how independent research that addresses one of the four questions can be brought together with research addressing the remaining questions. Figure 5 illustrates one example of how an integrative approach in ethology might look. Bird song can be studied in cognate but

147

distinct fields, with that research then being brought together to study the phenomenon of bird song more generally.

Bateson and Laland 2013 (see also Burghardt 2005) suggest that an additional question may be incorporated into the ‘four questions’ paradigm. They ask: Of what is an animal aware?

Here, I think the biological enactive approach has application. On this approach, we might answer this additional question by employing the tools of biological enactivism, namely sense- making. What an animal is aware of might be a matter of what it is attuned to via sense-making.

Recall that the enactive approach rejects a representationalist view of the mind, so answering this question on an enactive view would not involve making reference to mental representations. This approach may avoid some issues that have arisen in the field of cognitive ethology, which I will provide a brief overview of in what follows.

Jamieson and Bekoff (1992) note that criticisms of cognitive ethology have focused on whether or not it is possible to apply a cognitivist research framework to the study of animal minds, with some researchers arguing that “cognitive or mental concepts cannot be operationally defined, thus there are no researchable questions in cognitive ethology” (113). More recent criticisms have focused on philosophical concerns about whether or not we can even ascribe consciousness (or any sort of mentality) to non-human animals (Allen and Bekoff 1999, Chapter

8) as well as whether empirical work on animal intentionality produces any significant results, especially fieldwork (Allen and Bekoff 1999, Chapter 9). Controlling variables in a laboratory setting is challenging, but doing so in a field setting, critics stress, is nearly impossible (Heyes and Dickinson 1990; Heyes and Dickinson 1995).

Despite developments in both biology and comparative psychology that could potentially shed light on questions asked by cognitive ethologists, the field has seemed to not yet recover

148

from these criticisms. Here, I want to suggest that this is primarily due to its commitment to a

cognitivist research framework. For example, critics have demanded that if cognitive ethology is

to refer to mentalistic concepts such as consciousness, then it must be able to operationally

define those concepts. But needing to operationally define consciousness in order to get off the

ground as a scientific enterprise is something of a herculean task, especially for studying animal

minds. It also does not necessarily need to halt work in cognitive ethology – as a common

objection goes, the gene did not need to be defined in order for work in genetics to proceed. This

concern aside, the enactive approach offers a way forward for understanding cognition that does

not make reference to mental representations, thus avoiding these criticisms. Cognition can be studied in terms of perception-action dynamics, developed via structural coupling in organism- environment systems.

In the closing remarks of their book Species of Mind, Allen and Bekoff stress that, in moving the field of cognitive ethology forward, an animal’s sensorimotor capacities must be taken into consideration when studying their behavior. They state:

Although obvious in one sense, in designing studies it is important to try only to ask animals to do what is within their capabilities … it is important to know as much as possible about the sensory world of the animals being studied. Experiments should not be designed that ask animals to do things that they are unable to do because they are insensitive to the experimental stimuli or unmotivated by the stimuli. (Allen and Bekoff 1999, 179)

This suggestion falls squarely in line with the enactive approach. It emphasizes the importance of

studying cognition as it is in nature, as emerging from the coupling of organism and environment

that share a history of recurrent interactions. Thus, biological enactivism may have application in the field of cognitive ethology.

149

On a biological enactive view informed by developmental systems theory, an individual

organism is a life cycle, a process that occurs within a larger developmental system. One

promising route to pursue as an application of this framework is the task of identifying what

qualifies as an organism in organism-environment systems. For example, we might think of the

organism in the organism-environment (or developmental) system as the process of an individual

life cycle (see Griffiths and Stotz 2018; Griffiths and Gray 1994; Griffiths and Gray 2001). This

view may potentially shed light on what we might think of as qualifying as an individual

organism. It may also help in dealing with more complex cases of individuality. For example, we

might think of a single ant as qualifying as an individual, as it undergoes a unique life cycle.

However, the ant colony itself may also be viewed as a life cycle process, with distinct

developmental stages (see Gordon 2001). Recent discussions on process ontology in biology (see

Nicholson and Dupré 2018) are also resourceful for investigating life cycles as processes. A process view of the organism would fit quite naturally with the enactive approach put forth here.

My final comments are normative in nature. Although I have for the most part abstracted away from individual instances of cognitive abilities, one hope I have is that the reader comes away with a sense that cognition is a complex phenomenon, no matter how seemingly simplistic the organisms are which possess it. In Chapter 3 I made reference to a rather complex analysis of invertebrate behavior that suggests that even organisms as “robotic” as cockroaches may not be operating solely on the basis of simple innate behavioral mechanisms. My wish is for the picture of biological cognition that I am putting forward in this dissertation to put pressure on the commonplace view that most “lower” animals are mere automata. I also hope that it puts equal pressure on the view that there is a sharp distinction to be made between the cognitive capacities

150

of Homo sapiens and non-human animals. As suggested in Chapter 5, cognitive ability is bound up in the structural coupling between an organism and its environment; organisms are only as

“intelligent” as their structural interactions require. Recall, too, Godfrey-Smith’s (1996) claim that the function of cognition is to deal with environmental complexity. Along this line of thinking, we can imagine a coevolutionary relationship between complexity in environmental structure and complexity in cognitive capacity. Naturally, this picture requires looking more carefully at the intricacies of the enacted world of each organism.

The great complexity of the human environment, including but especially the constructed human environment, has coevolved with the great complexity of the human mind. Research on the coevolution of culture and genes (Richerson and Boyd 2004) may serve as backing for this argument. Cultural niche construction (Laland and O’Brien 2011) is an extension of niche construction theory that can shed light on the role of the constructed human environment in human evolution. There is certainly more to be said about the relationship between the rich, complex human environment and human cognition. The upshot here is that this is the appropriate place to look for understanding human cognition, rather than any innate ability for complex cognitive functioning. I have shied away from saying much if anything about human cognition, because my focus is on the sort of cognitive capacities that arise as a result of these dynamic interactions. On this picture, human cognition is “special” only in virtue of the complexity of our structural coupling. We may find that equally complex arrangements of structural coupling challenge what is special about the human mind. Further research on eusocial insects may surprise us in this regard (e.g. Feinerman and Korman 2017).

I say that these comments are normative in nature because I wish to emphasize how these perspectives challenge traditional thinking about cognition, and how this challenge to tradition is

151

important to progress in the fields of both biology and cognitive science. We have much to gain by revisiting old assumptions about cognition in nature. Kralj-Fišer and Schuett (2014) provides a good case in point. They argue that the lack of research on personality variation in invertebrates is a significant oversight by scientists; taking seriously the study of personality variation and relevant behavioral patterns in invertebrate species can not only shed light on the biology of those organisms but can also inform the study of personality more generally. By challenging assumptions about cognition, we can unveil new features of biological life that had previously either been ignored, glossed over, or deemed unimportant. Science must be in the business of mindful exploration rather than cynical rejection. Anthropogenic hubris must be placed aside, and cognition must be studied wherever it is found in nature.

152

Bibliography

Adams, Fred, and Ken Aizawa. 2001. “The Bounds of Cognition.” Philosophical Psychology 14

(2): 43-64.

Aguilar, Leonel, Stefano Bennati, and Dirk Helbing. 2019. “How Learning Can Change the

Course of Evolution.” PLoS ONE 14 (9): e0219502.

Alcock, John. 1972. “The Evolution of the Use of Tools by Feeding Animals.” Evolution 26:

464-73.

Allen, Colin and Marc Bekoff. 1999. Species of Mind: The Philosophy and Biology of Cognitive

Ethology. Cambridge, MA: MIT Press.

Allen, Grant. 1879. The Colour Sense. London: Turbner & Co.

Avital, Eytan and Eva Jablonka. 2000. Animal Traditions: Behavioural Inheritance in

Evolution. Cambridge: Cambridge University Press.

Badyaev, Alexander V. 2005. "Stress-Induced Variation in Evolution: From Behavioural Plasticity

to Genetic Assimilation." Proceedings of the Royal Society B: Biological Sciences 272

(1566): 877-86.

Baggs, Edward and Anthony Chemero. 2019. "The Third Sense of Environment." In Perception

as Information Detection: Reflections on Gibson’s Ecological Approach to Visual

153

Perception, edited by Jeffrey B. Wagman and Julia J. C. Blau. New York, NY: Taylor &

Francis.

Baldwin, James Mark. 1902. Development and Evolution. New York: Macmillan.

Bateson, Patrick and Kevin N. Laland. 2013. "Tinbergen's Four Questions: An Appreciation and

an Update." Trends in Ecology & Evolution 28 (12): 712-18.

Bayne, T., D. Brainard, R. W. Byrne, L. Chittka, N. Clayton, C. Heyes, J. Mather, et al. 2019.

"What is Cognition?" Current Biology 29 (13): R608-R615.

Bennett, Nigel C. and Chris G. Faulkes. 2000. African Mole-Rats: Ecology and Eusociality.

Cambridge: Cambridge University Press.

Bertalanffy, Ludwig von. 1950. "An Outline of General System Theory." The British Journal for

the Philosophy of Science 1 (2): 134-65.

Blix, Arnoldus Schytte. 2016. "Adaptations to Polar Life in Mammals and Birds." The Journal of

Experimental Biology 219 (8): 1093.

Bonduriansky, Russell, and Troy Day. 2009. “Nongenetic Inheritance and Its Evolutionary

Implications.” Annual Review of Ecology, Evolution, and Systematics 40: 103-25.

Brandon, Robert and Janis Antonovics. 1996. "The Coevolution of Organism and Environment."

In Concepts and Methods in Evolutionary Biology, edited by Robert Brandon, 161-78.

Cambridge: Cambridge University Press.

Burghardt, Gordon M. 2005. The Genesis of Animal Play: Testing the Limits. Cambridge,

MA: MIT Press. 154

Cameron, Tom C., Daniel O'Sullivan, Alan Reynolds, Stuart B. Piertney, and Tim G. Benton.

2013. “Eco-evolutionary dynamics in response to selection on life-history.” Ecology

Letters 16 (6): 754-63.

Capra, Fritjof. 1996. The Web of Life: A New Scientific Understanding of Living Systems. New

York: Anchor Books.

Carroll, Sean B. “Evo-Devo and an Expanding Evolutionary Synthesis: A Genetic Theory of

Morphological Evolution.” Cell 134 (1): 25-36.

Chemero, Anthony. 2009. Radical Embodied Cognitive Science. Cambridge: The MIT Press.

Chittka, Lars. 2017. “Bee Cognition.” Current Biology 27 (19): R1049-R1053.

Danchin, Étienne, Anne Charmantier, Frances A. Champagne, Alex Mesoudi, Benoit Pujol, and

Simon Blanchet. 2011. “Beyond DNA: Integrating inclusive inheritance into an extended

theory of evolution.” Nature Reviews Genetics 12 (7): 475-86.

Darwin, Charles. 1871. The Descent of Man and Selection in Relation to Sex. London: John

Murray.

Darwin, Charles. 1896. Insectivorous Plants. New York: D. Appleton.

Darwin, Charles. 1966. Darwin on humus and the earthworm: The formation of vegetable mould

through the action of worms with observations on their habits. London: Faber and Faber.

De Jong, Gerdien. 2005. “Evolution of Phenotypic Plasticity: Patterns of Plasticity and the

Emergence of Ecotypes.” New Phytologist 166 (1): 101-17.

155

Depew, David J. and Bruce H. Weber. 1998. "What does Natural Selection have to be Like in

Order to Work with Self-Organization?" Cybernetics & Human Knowing 5 (1).

Di Paolo, Ezequiel A. 2005. "Autopoiesis, Adaptivity, Teleology, Agency." Phenomenology and

the Cognitive Sciences 4 (4): 429-52.

Di Paolo, Ezequiel A., Thomas Buhrmann, and Xabier E. Barandiaran. 2017. Sensorimotor Life:

An Enactive Proposal. Oxford, UK: Oxford University Press.

Di Paolo, Ezequiel. 2008. "Extended Life." Topoi 28 (1): 9. doi:10.1007/s11245-008-9042-3.

Fábregas-Tejeda A, and F Vergara-Silva. 2018. "The emerging structure of the Extended

Evolutionary Synthesis: where does Evo-Devo fit in?" Theory in Biosciences 137 (2):

169-84.

Feinerman, Ofer and Amos Korman. 2017. "Individual Versus Collective Cognition in Social

Insects." The Journal of Experimental Biology 220 (1): 73.

Fisher, James and Hinde, Robert A., 1949. The opening of milk bottles by birds. British Birds

42: 347-57.

Foster, Geoffrey. G. and David T. Suzuki. 1970. "Temperature-Sensitive Mutations in

Drosophila Melanogaster, IV. A Mutation Affecting Eye Facet Arrangement in a

Polarized Manner." Proceedings of the National Academy of Sciences of the United

States of America 67 (2) (Oct): 738-45.

Gallagher, Shaun. 2017. Enactivist Interventions: Rethinking the Mind. Oxford, UK: Oxford

University Press.

156

Gibson, James Jerome. 1979. The Ecological Approach to Visual Perception. Boston, MA, US:

Houghton, Mifflin and Company.

Gilbert, Scott F., and David Epel. 2009. Ecological Developmental Biology: Integrating

Epigenetics, Medicine, and Evolution. Sunderland: Sinauer.

Gilbert, Scott F., John M. Opitz, and Rudolf A. Raff. “Resynthesizing Evolutionary and

Developmental Biology.” Developmental Biology 173 (2): 357-72.

Godfrey-Smith, Peter. 1996. Complexity and the Function of Mind in Nature. Cambridge, UK:

Cambridge University Press.

Godfrey-Smith, Peter. 2001. "On the Status and Explanatory Structure of Developmental

Systems Theory." In Cycles of Contingency: Developmental Systems and Evolution,

edited by Susan Oyama, Paul E. Griffiths and Russell D. Gray, 283-98. Cambridge, MA:

MIT Press.

Gordon, Deborah M. 2001. "The Development of Ant Colony Behavior." In Cycles of

Contingency: Developmental Systems and Evolution, edited by Susan Oyama, Paul

Griffiths and Russell Gray, 141-48. Cambridge, MA: MIT Press.

Gottlieb, Gilbert and John G. Vandenbergh. 1968. “Ontogeny of vocalization in duck and chick

embryos.” Journal of Experimental Zoology 168: 307-25.

Gottlieb, Gilbert. 1970. “Conceptions of prenatal behavior.” In Development and Evolution of

Behavior, edited by Lester R. Aronson, Ethel Tobach, Daniel S. Lehrman, and Jay S.

Rosenblatt, 111-37. San Fransisco: W.H. Freeman.

157

Gottlieb, Gilbert. 2001. “A developmental psychobiological systems view.” In Cycles of

Contingency: Developmental Systems and Evolution, edited by Susan Oyama, Paul

Griffiths and Russell Gray, 41-54. Cambridge, MA: MIT Press.

Gottlieb, Gilbert. 2002. Individual Development and Evolution: The Genesis of Novel

Behavior. New York: Psychology Press.

Gould, Stephen Jay. 1983. “The Hardening of the Modern Synthesis.” In Dimensions of

Darwinism, edited by Marjorie Greene, 71-93. Cambridge: Cambridge University Press.

Gould, Stephen Jay. 2002. The Structure of Evolutionary History. Cambridge, MA: Belknap

Press of Harvard University Press.

Griffiths, Paul and Karola Stotz. 2013. Genetics and Philosophy: An Introduction. Cambridge,

UK: Cambridge University Press.

Griffiths, Paul and Karola Stotz. 2018. "Developmental Systems Theory as a Process Theory."

In Everything Flows: Towards a Processual Philosophy of Biology, edited by Daniel

Nicholson and John Dupré, 282-310. Oxford: Oxford University Press.

Griffiths, Paul and Russell Gray. 2001. "Darwinism and Developmental Systems." In Cycles of

Contingency: Developmental Systems and Evolution, edited by Susan Oyama, Paul

Griffiths and Russell Gray, 195-218. Cambridge, MA: MIT Press.

Griffiths, Paul and Adam Hochman. 2015. “Developmental Systems Theory.” eLS

doi:10.1002/9780470015902.a0003452.pub2

158

Griffiths, Paul E. and Russell D. Gray. 1994. "Developmental Systems and Evolutionary

Explanation." The Journal of Philosophy 91 (6): 277-304.

Griffiths, Paul E. and Russell D. Gray. 2004. "The Developmental Systems Perspective:

Organism-Environment Systems as Units of Development and Evolution." In Phenotypic

Integration: Studying the Ecology and Evolution of Complex Phenotypes, edited by

Massimo Pigliucci and Katherine Preston, 409-431. Oxford: Oxford University Press.

Griffiths, Paul E. and Russell D. Gray. 2005. "Discussion: Three Ways to Misunderstand

Developmental Systems Theory." Biology and Philosophy 20 (2): 417-25.

Hart, Simon P., Martin M. Turcotte, and Jonathan M. Levine. 2019. “Effects of rapid evolution

on species coexistence.” Proceedings of the National Academy of Sciences of the United

States of America 116 (6): 2112-2117.

Head, Megan L., Michael D. Jennions, and Susanne RK Zajitschek. 2016. “Sexual selection:

incorporating non-genetic inheritance.” Current Opinion in Behavioral

Sciences 1 (12): 129-37.

Heras-Escribano, Manuel, and Paulo de Jesus. 2018. “Biosemiotics, the Extended Synthesis, and

Ecological Information: Making Sense of the Organism-Environment Relation at the

Cognitive Level.” Biosemiotics 11 (2): 245-62.

Heyes, Cecilia and Anthony Dickinson. 1990. "The Intentionality of Animal Action." Mind &

Language 5 (1): 87-103.

159

Heyes, Cecilia and Anthony Dickinson. 1995. "Folk Psychology Won't Go Away: Response to

Allen and Bekoff." Mind & Language 10 (4): 329-32.

Howard, Ensign Evans. 1953. "Comparative Ethology and the Systematics of Spider Wasps."

Systematic Zoology 2 (4): 155-72.

Hutto, Daniel D. 2017. "REC: Revolution Effected by Clarification." Topoi 36 (3): 377-91.

Hutto, Daniel D. and Erik Myin. 2012. Radicalizing Enactivism. Cambridge, MA: The MIT

Press.

Jablonka, Eva and Marion Lamb. 2005. Evolution in Four Dimensions: Genetic, Epigenetic,

Behavioral, and Symbolic Variation in the History of Life. Cambridge, MA: The MIT

Press.

Jablonka, Eva and Marion Lamb. 1995. Epigenetic Inheritance and Evolution : the Lamarckian

Dimension. Oxford, UK: Oxford University Press.

Jacob, F. 1977. "Evolution and Tinkering." Science 196 (4295): 1161.

Jamieson, Dale and Marc Bekoff. 1992. "On Aims and Methods of Cognitive Ethology." PSA:

Proceedings of the Biennial Meeting of the Philosophy of Science Association 1992 (2):

110-24.

Jonas, Hans. 1973. Organismus Und Freiheit: Ansätze Zu Einer Philosophischen Biologie.

Göttingen, Germany: Vandenhoeck & Ruprecht.

Jones, Clive G., Fred B. Samson, Fritz L. Knopf, John H. Lawton, and Moshe Shachak. 1994.

“Organisms as Ecosystem Engineers.” Oikos 69 (3): 373-86. 160

Kauffman, Stuart. A. 1993. The Origins of Order: Self-Organization and Selection in Evolution.

Oxford, UK: Oxford University Press.

Kralj-Fišer, Simona and Wiebke Schuett. 2014. "Studying Personality Variation in Invertebrates:

Why Bother?" Animal Behaviour 91 (2014): 41-52.

Kuhn, Thomas S. 1970. The Structure of Scientific Revolutions. Chicago, IL: University of

Chicago Press.

Laforsch, Christian, and Ralph Tollrian. 2004. “Inducible Defenses in Multipredator

Environments: Cyclomorphosis in Daphnia Cucullata.” Ecology 85 (8): 2302-11.

Laland, Kevin N. and Michael J. O’Brien. 2011. "Cultural Niche Construction: An Introduction."

Biological Theory 6 (3): 191-202.

Laland, Kevin N., John Odling-Smee, and Marcus W. Feldman. 2001. “Niche Construction,

Ecological Inheritance, and Cycles of Contingency in Evolution.” In Cycles of

Contingency: Developmental Systems and Evolution, edited by Susan Oyama, Paul

Griffiths and Russell Gray, 117-26. Cambridge: MIT Press.

Laland, Kevin N., John Odling-Smee, William Hoppitt, and Tobias Uller. 2013. “More on how

and why: Cause and effect in biology revisited.” Biology and Philosophy 28 (5): 719-45.

Laland, Kevin N., Tobias Uller, Marcus W. Feldman, Kim Sterelny, Gerd B. Müller, Armin

Moczek, Eva Jablonka, and John Odling-Smee. 2015. “The extended evolutionary

synthesis: Its structure, assumptions and predictions.” Proceedings of the Royal Society

B: Biological Sciences 282 (1813) https://doi.org/10.1098/rspb.2015.1019

161

Laland, Kevin N., Kim Sterelny, John Odling-Smee, William Hoppitt and Tobias Uller. 2011.

“Cause and Effect in Biology Revisited: Is Mayr’s Proximate-Ultimate Dichotomy Still

Useful?” Science 334 (6062): 1512-16.

Laland, Kevin N., John Odling-Smee, and Marcus. W. Feldman. 1999. “Evolutionary

consequences of niche construction and their implications for ecology.” Proceedings of

the National Academy of Sciences of the United States of America 96 (18): 10242-10247.

Laland, Kevin, Tobias Uller, Marc Feldman, Kim Sterelny, Gerd B. Müller, Armin Moczek, Eva

Jablonka, et al. 2014. “Does evolutionary theory need a rethink?” Nature 514 (7521):

161-64.

Lefebvre, Louis. 1995. "The Opening of Milk Bottles by Birds: Evidence for Accelerating

Learning Rates, but Against the Wave-of-Advance Model of Cultural

Transmission." Behavioural Processes 34 (1): 43-53.

Lehrman, Daniel S. 2001. “A critique of Konrad Lorenz's theory of instinctive behavior.”

In Cycles of Contingency: Developmental Systems and Evolution, edited by Susan

Oyama, Paul Griffiths and Russell Gray, 25-39. Cambridge: MIT Press.

Lehrman, Daniel S. 1970. “Semantic & conceptual issues in the nature-nurture

problem.” In Development & Evolution of Behaviour, edited by D. S. Lehrman, 17–

52. San Francisco: W. H. Freeman.

Levins, Richard, and Richard C. Lewontin. 1985. The Dialectical Biologist. Cambridge, MA:

Harvard University Press.

162

Lewens, Tim. 2019. “The Extended Evolutionary Synthesis: what is the debate about, and what

might success for the extenders look like?” Biological Journal of the Linnean Society 127

(4): 707-21.

Love, Alan C. 2010. “Rethinking the Structure of Evolutionary Theory for an Extended

Synthesis.” In Evolution, the Extended Synthesis, edited by Massimo Pigliucci & Gerd B.

Müller, 403-41. Cambridge, MA: The MIT Press.

Love, Alan C. 2017. “Evo-devo and the structure(s) of evolutionary theory: a different kind of

challenge.” In Challenging the Modern Synthesis, edited by Philippe Huneman & Denis

M. Walsh, 159-87. New York, NY: Oxford University Press.

Maiese, Michelle. 2015. Embodied Selves and Divided Minds. Oxford, UK: Oxford University

Press UK.

Marshall, Paul, and Timothy W Bredy. 2016. “Cognitive neuroepigenetics: the next evolution in

our understanding of the molecular mechanisms underlying learning and memory?” NPJ

Science of Learning 1 (16014): doi: 10.1038/npjscilearn.2016.14

Maturana, Humberto. R. and Francisco. J. Varela. 1980. Autopoiesis and Cognition: The

Realization of the Living. Netherlands: Springer Netherlands.

Maturana, Humberto R. and Francisco J. Varela. 1987. The Tree of Knowledge: The Biological

Roots of Human Understanding. Boston, MA: Shambhala Publications.

Mayr, Ernst. 1961. “Cause and effect in biology.” Science 134 (3489): 1501-06.

Merleau-Ponty, Maurice. 1963. The Structure of Behavior. Boston, MA: Beacon Press.

163

Mesoudi, Alex, Simon Blanchet, Anne Charmantier, Étienne Danchin, Laurel Fogarty, Eva

Jablonka, Kevin N. Laland, et al. 2013. “Is Non-genetic Inheritance Just a Proximate

Mechanism? A Corroboration of the Extended Evolutionary Synthesis.” Biological

Theory 7 (3): 189-95.

Müller, Gerd. 2007. “Evo-devo: extending the evolutionary synthesis.” Nature Reviews

Genetics 8, 943-49.

Müller, Gerd. 2017. “Why an extended evolutionary synthesis is necessary.” Interface Focus,

7 (5): 20170015. doi: 10.1098/rsfs.2017.0065

Nettle, Daniel, and Melissa Bateson. 2015. “Adaptive developmental plasticity: What is it, how

can we recognize it and when can it evolve?” Proceedings of the Royal Society B:

Biological Sciences 282 (1812) 20151005. doi: 10.1098/rspb.2015.1005

Nevo, Eviatar. 1979. "Adaptive Convergence and Divergence of Subterranean

Mammals." Annual Review of Ecology and Systematics 10 (1): 269-308.

Newen, Albert, Leon De Bruin, and Shaun Gallagher. 2018. The Oxford Handbook of 4E

Cognition. Oxford, UK: Oxford University Press.

Nicholson, Daniel & John Dupré. 2018. Everything Flows: Towards a Processual Philosophy of

Biology. Oxford, UK: Oxford University Press.

Nijhout, Frederik H. 2001. "The Ontogeny of Phenotypes." In Cycles of Contingency:

Development Systems and Evolution, edited by Susan Oyama, Paul Griffiths and Russell

Gray, 129-40. Cambridge, MA: The MIT Press.

164

Nikitina, Natalya. V., Bronwen Maughan-Brown, M. Justin O'Riain, and Susan H. Kidson. 2004.

"Postnatal Development of the Eye in the Naked Mole Rat (Heterocephalus Glaber)." The

Anatomical Record Part A, Discoveries in Molecular, Cellular, and Evolutionary

Biology 277 (2): 317-37.

Nilsson, Dan-Eric. 2009. "The Evolution of Eyes and Visually Guided Behaviour." Philosophical

Transactions of the Royal Society B: Biological Sciences 364 (1531): 2833-47.

Noble, Denis. 2015a. “Conrad Waddington and the origin of epigenetics.” Journal of

Experimental Biology. 218 (6): 816-18.

Noble, Denis. 2015b. “Evolution beyond neo-Darwinism: A new conceptual

framework.” Journal of Experimental Biology. 218 (1): 7-13.

Noë, Alva. 2012. Varieties of Presence. Cambridge, MA: Harvard University Press.

Oakley, Todd H. and Daniel I. Speiser. 2015. "How Complexity Originates: The Evolution of

Animal Eyes." Annual Review of Ecology, Evolution, and Systematics 46 (1): 237-60

Odling-Smee, John, and Kevin N. Laland. 2011. “Ecological Inheritance and Cultural

Inheritance: What Are They and How Do They Differ?” Biological Theory 6 (3): 220-

30.

Odling-Smee, John, Marcus W. Feldman, and Kevin N. Laland. 2003. Niche Construction: the

Neglected Process in Evolution. Princeton, MA: Princeton University Press.

O'Regan, J. Kevin and Alva Noë. 2001. "A Sensorimotor Account of Vision and Visual

Consciousness." Behavioral and Brain Sciences 24 (5): 939-73.

165

O'Regan, Kevin J. 2011. Why Red Doesn't Sound Like a Bell: Understanding the Feel of

Consciousness. New York, NY: Oxford University Press.

Oyama, Susan. 2000 [1985]. The Ontogeny of Information: Developmental Systems and

Evolution. Durham, NC: Duke University Press.

Oyama, Susan, Paul E. Griffiths, and Russell D. Gray. 2001. Cycles of Contingency:

Developmental Systems and Evolution. Cambridge, MA: The MIT Press.

Paolo, Ezequiel A. Di, Marieke Rohde, and Hanne De Jaegher, 2010. “Horizons for the Enactive

Mind: Values, Social Interaction, and Play.” In Enaction: Toward a New Paradigm for

Cognitive Science, edited by John Stewart, Olivier Gapenne and Ezequiel A. Di Paolo.

Cambridge, MA: The MIT Press.

Pigliucci, Massimo, and Gerd B. Müller. 2010. Evolution, the Extended Synthesis. Cambridge,

MA: The MIT Press.

Pigliucci, Massimo, Courtney J. Murren, and Carl D. Schlichting. 2006. “Phenotypic plasticity

and evolution by genetic assimilation.” Journal of Experimental Biology 209 (12): 2362-

Pigliucci, Massimo. 2003. "Phenotypic Integration: Studying the Ecology and Evolution of

Complex Phenotypes." Ecology Letters 6 (3): 265-72.

Pigliucci, Massimo. 2007. “Do We Need an Extended Evolutionary Synthesis?” Evolution 61

(12): 2743-9.

166

Pigliucci, Massimo. 2009. “An Extended Synthesis for Evolutionary Biology.” Annals of the

New York Academy of Sciences 1168 (1): 218-28.

Platnick, Norman, and Donn Rosen. 1987. “Popper and evolutionary novelties.” History

and Philosophy of the Life Sciences 9 (1): 5–16.

Plutynski, Anya. 2018. “The modern synthesis.” In Routledge Encyclopedia of

Philosophy. https://www.rep.routledge.com/

Potochnik, Angela. Forthcoming. “What Constitutes an Explanation in Biology?” In Philosophy

of Science for Biologists, edited by Kostas Kampourakis & Tobias Uller. Cambridge.

MA: Cambridge University Press.

Prigogine, Ilya. 1967. "Dissipative Structures in Chemical Systems." In Fast Reactions and

Primary Processes in Chemical Kinetics, edited by Stig Claesson. New York, NY:

Interscience.

Rago, Alfredo, Kostas Kouvaris, Tobias Uller, and Richard Watson. 2019. “How adaptive

plasticity evolves when selected against.” PLOS Computational Biology 15 (3):

e1006260. doi: 10.1371/journal.pcbi.1006260

Ranciaro, Alessia, Michael C. Campbell, Jibril B. Hirbo, Wen Ya Ko, Alain Froment, Paolo

Anagnostou, Maritha J. Kotze, et al. 2014. “Genetic origins of lactase persistence and the

spread of pastoralism in Africa.” American Journal of Human Genetics 94 (4): 496-510.

Reynolds, Vernon. 1970. "The "Man of the Woods"." In Field Studies in Natural History, 202-

210. New York: Van Nostrand.

167

Richerson, Peter J. and Robert Boyd. 2005. Not by Genes Alone: How Culture Transformed

Human Evolution. Chicago, IL: University of Chicago Press.

Rietveld, Erik, Sanneke de Haan, and Damiaan Denys. 2013. "Social Affordances in Context:

What is it that we are Bodily Responsive to?" Behavioral and Brain Sciences 36 (4): 436.

Robert, Jason. 2004. Embryology, epigenesis, and evolution: Taking development seriously.

Cambridge, UK: Cambridge University Press.

Schlichting, Carl D. and Massimo Pigliucci. 1998. Phenotypic Evolution: A Reaction Norm

Perspective. Sunderland, MA: Sinauer Associates Incorporated.

Schneirla, Theodore C. 1957. “The concept of development in comparative

psychology.” In The Concept of Development, edited by D. B. Harris, 78–

108. Minneapolis, MN: University of Minnesota Press.

Schofield, Daniel P., William C. McGrew, Akiko Takahashi, and Satoshi Hirata. 2018.

"Cumulative Culture in Nonhumans: Overlooked Findings from Japanese

Monkeys?" Primates 59 (2): 113-22.

Schwenk, Kurt and Gunter P. Wagner. 2004. "The Relativism of Constraints on Phenotypic

Evolution." In Phenotypic Integration: Studying the Ecology and Evolution of Complex

Phenotypes, edited by Massimo Pigliucci and Katherine Preston, 390-408. Oxford, UK:

Oxford University Press.

Sherman, Paul W., Jennifer U. M. Jarvis, and Richard D. Alexander. 2017. The Biology of the

Naked Mole-Rat. Princeton, MA: Princeton University Press.

168

Smith, V. Anne, Andrew P. King, and Meredith J. West. 2000. "A Role of Her Own: Female

Cowbirds, Molothrus Ater, Influence the Development and Outcome of Song

Learning." Animal Behaviour 60 (5): 599-609.

Sober, Elliott R. 1982. “The Modern Synthesis: Its Scope and Limits.” PSA: Proceedings of the

Biennial Meeting of the Philosophy of Science Association. Vol. 1982. 314-21.

Stadler, Bernhard and Anthony F. G. Dixon. 2005. "Ecology and Evolution of Aphid-Ant

Interactions." Annual Review of Ecology, Evolution, and Systematics 36 (1): 345-72.

Stotz, Karola and Paul Griffiths. 2018. “A Developmental Systems Account of Human Nature.”

In Why we Disagree about Human Nature, edited by Elizabeth Hannon and Tim Lewens,

58-75. Oxford, UK: Oxford University Press.

Stotz, Karola. 2008. “The Ingredients for a Postgenomic Synthesis of Nature and

Nurture.” Philosophical Psychology 21 (3): 359-81.

Stotz, Karola. 2010. “Human nature and cognitive-developmental niche construction.”

Phenomenology and the Cognitive Sciences 9 (4): 483-501.

Stotz, Karola. 2014. “Extended evolutionary psychology: The importance of transgenerational

developmental plasticity.” Frontiers in Psychology 5 (908): doi:

10.3389/fpsyg.2014.00908

Stotz, Karola. 2017. "Why Developmental Niche Construction is Not Selective Niche

Construction: And Why it Matters." Interface Focus 7 (5): 20160157 doi:

10.1098/rsfs.2016.0157

169

Sznajder, B., M. W. Sabelis, and M. Egas. 2012. “How Adaptive Learning Affects Evolution:

Reviewing Theory on the Baldwin Effect.” Evolutionary Biology 39 (3): 301-10.

Thompson, Evan and Mog Stapleton. 2009. "Making Sense of Sense-Making: Reflections on

Enactive and Extended Mind Theories." Topoi 28 (1): 23–30.

Thompson, Evan, Adrian Palacios, and Fransisco J. Varela. 1992. "Ways of Coloring:

Comparative Color Vision as a Case Study in Cognitive Science." Behavioral and Brain

Sciences 15 (1): 1-26.

Thompson, Evan. 2004. "Life and Mind: From Autopoiesis to Neurophenomenology. A Tribute

to Francisco Varela." Phenomenology and the Cognitive Sciences 3 (4): 381-98.

Thompson, Evan. 2007. Mind in Life: Biology, Phenomenology, and the Sciences of Mind.

Cambridge, MA: Harvard University Press.

Tinbergen, N. 1963. "On Aims and Methods of Ethology." Zeitschrift Für Tierpsychologie 20

(4): 410-33.

Toth, Miklos. 2015. “Mechanisms of non-genetic inheritance and psychiatric disorders.”

Neuropsychopharmacology 40 (1): 129-40.

Van Tienderen, Peter H. 1997. "Generalists, Specialists, and the Evolution of Phenotypic

Plasticity in Sympatric Populations of Distinct Species." Evolution 51 (5): 1372-80.

Varela, Fransisco J., Humberto Maturana, and Ricardo Uribe. 1974. "Autopoiesis: The

Organization of Living Systems, its Characterization and a Model." Currents in Modern

Biology 5 (4): 187-96.

170

Varela, Francisco J., Evan Thompson, and Eleanor Rosch. 1991. The Embodied Mind: Cognitive

Science and Human Experience. Cambridge, MA: The MIT Press.

Varela, Fransisco J. 1979. Principles of Biological Autonomy. New York, NY: Elsevier North

Holland.

Varela, Fransisco J. 1997. "Patterns of Life: Intertwining Identity and Cognition." Brain and

Cognition 34: 72-87.

Von Uexküll, Jakob. 2010. A Foray into the Worlds of Animals and Humans, with A Theory of

Meaning. Minneapolis, MN: University of Minnesota Press.

Waddington, Conrad. H. 2014. The Strategy of the Genes. Taylor & Francis.

Wallace, Bruce. 1986. “Can Embryologists Contribute to an Understanding of Evolutionary

Mechanisms?” In Integrating Scientific Disciplines, edited by William Bechtel, 149-63.

Dordrecht: Springer.

Ward, Dave, David Silverman, and Mario Villalobos. 2017. "Introduction: The Varieties of

Enactivism." Topoi 36 (3): 365-75.

Weber, Andreas and Fransisco J. Varela. 2002. "Life After Kant: Natural Purposes and the

Autopoietic Foundations of Biological Individuality." Phenomenology and the Cognitive

Sciences 1: 97-125.

Weber, Bruce H. and David J. Depew. 1996. "Natural Selection and Self-Organization:

Dynamical Models as Clues to a New Evolutionary Synthesis." Biology and

Philosophy 11: 33-65.

171

Weber, Bruce H. and David J. Depew. 2001. "Developmental Systems, Darwinian Evolution,

and the Unity of Science." In Cycles of Contingency: Developmental Systems and

Evolution, edited by Susan Oyama, Paul Griffiths and Russell Gray, 239-54. Cambridge,

MA: MIT Press.

West, Meredith. J. and Andrew P. King. 1987. "Settling Nature and Nurture into an Ontogenetic

Niche." Developmental Psychobiology 20 (5): 549-62.

West, Meredith J. and Andrew P. King. 1988. "Female Visual Displays Affect the Development

of Male Song in the Cowbird." Nature 334 (6179): 244-6.

West-Eberhard, Mary Jane. 2003. Developmental Plasticity and Evolution. New York, NY:

Oxford University Press.

Wheeler, William M. 1910. Ants: Their Structure, Development and Behavior. New York, NY:

Columbia University Press.

Williams, George C. 1992. Natural Selection. New York, NY: Oxford University Press.

Wilson, Edmund. 1894. "The Embryological Criterion of Homology." Biological Lectures of the

Marine Biology Laboratory of Wood's Hole (1894): 101-24.

Wimsatt, William C. 2001. "Generative Entrenchment and the Developmental Systems Approach

to Evolutionary Processes." In Cycles of Contingency: Developmental Systems and

Evolution, edited by Susan Oyama, Paul Griffiths and Russell Gray, 219-38. Cambridge,

MA: MIT Press.

172

Wimsatt, William. 1986. "Developmental Constraints, Generative Entrenchment and the Innate-

Acquired Distinction." In Integrating Scientific Disciplines, edited by William Bechtel,

185-208. Dordrecht: Martinus Nijhoff.

Withagen, Rob and Anthony Chemero. 2009. "Naturalizing Perception: Developing the

Gibsonian Approach to Perception Along Evolutionary Lines." Theory & Psychology 19

(3): 363-89.

173