Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

The Evolution of Cognition: A 4E Perspective

Oxford Handbooks Online


The Evolution of Cognition: A 4E Perspective  
Louise Barrett
The Oxford Handbook of 4E Cognition
Edited by Albert Newen, Leon De Bruin, and Shaun Gallagher

Print Publication Date: Sep 2018 Subject: Psychology, Cognitive Psychology


Online Publication Date: Oct 2018 DOI: 10.1093/oxfordhb/9780198735410.013.38

Abstract and Keywords

4E cognition offers a viable alternative to the viewpoint of classic cognitivism, providing a


coherent, biologically grounded theory of cognitive evolution, which recognizes that
bodies evolved before brains, and argues that whatever cognition might be, it must be
grounded in the ability of organisms to coordinate and control action in a dynamic
environment. In this chapter, I first consider some ideas relating to the minimal criteria
for cognition as a means to introduce this more “biogenic” approach. This notion of
minimal cognition is then linked to a discussion of the origins of cognition in sensorimotor
coordination, as articulated in the “skin-brain thesis,” which argues that early nervous
systems enabled coordinated contractions across myoepithelia. Finally, I consider how
one particular school of 4E thought, radical enactivism, offers a view of evolutionary
continuity that seems well equipped to resist the anthropocentric impulses of traditional
cognitivist views.

Keywords: cognitive evolution, skin-brain thesis, minimal cognition, sensorimotor coordination, myoepithelia,
Umwelt, biogenic, radical, enactivism, evolutionary continuity, nervous systems

Page 1 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

What is Cognition Anyway?


How and why cognition evolved depends on what one thinks cognition is. The classic
definition by Neisser (1967) identified cognition as the processes by which sensory inputs
are transformed, manipulated, augmented, and used to give rise to motor outputs, with
the implicit assumption that these processes took place solely in the brain. There is a
distinctly anthropocentric tinge to this definition, grounded as it is in the cognitive
revolution, which aimed to model (or even recreate) human intelligence via the use of
computers. Consequently, the processes usually considered to be cognitive include
concept formation, reasoning, and problem-solving abilities, theory of mind, natural
language, memory, prospective planning, and the ability to represent objects in their
absence. This view of cognition often results in what Lyons (2006) terms an
“anthropogenic” approach to its evolution, in which we “assume, to a greater or lesser
extent, that human psychological attributes are the hallmarks of cognition and ask what
sort of biological or evolutionary story might account for them” (Lyons 2006, p. 12).

Accordingly, one of the tasks of comparative psychology is to identify which, if any,


cognitive traits are shared by other species. Basically, we engage in a sorting process that
distinguishes creatures capable of flexible “human-like” thought and behavior from those
capable only of fixed, “instinctive” noncognitive responses to environmental stimuli. Via
this sorting process, one can identify the conditions under which cognition was selected,
and how it contributes to organisms’ fitness. These views fit with Darwin’s dictum that
organisms should differ only by degree, rather than in kind, such that we should expect to
see the precursors of our own abilities in related species. Even in the context of a more
“ecological” approach to cognition, where the demands of the animal’s (p. 720) ecological
niche are argued to drive the evolution of particular cognitive capacities, cognitive
processes are conceptualized in the same representational, information-processing terms
that characterize the anthropogenic approach (Shettleworth 1999; Penn et al. 2008).

So far, so good. The problem is that, as Rowlands (2003, p. 167) has pointed out, there is
very little consensus on what entitles something to be called a cognitive process; often,
he says, we simply point toward the set of abilities we consider to be cognitive, like
reasoning and language, and leave it at that, without specifying exactly why such
processes count and others don’t. Deciding what counts as cognitive therefore becomes
rather fraught at times, with a continual battle over whether a particular behavior is the
result of a cognitive process or “merely” the outcome of associative learning (e.g.,
Buckner 2011; Byrne and Bates 2006). This state of affairs is made even more confusing
by the fact that the leading textbook on comparative cognition undercuts this distinction
by defining cognition as “any process by which animals acquire, process, store, and act
on information from the environment” (Shettleworth 1999), which would seem to render
debates over whether a particular result supports an associative or cognitive account
redundant, if not incoherent.

Page 2 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

Despite these various efforts, then, it seems that, as van Duijn et al. (2006) suggest, we
still lack a means of navigating between overly “mindless” accounts of behavior versus
“mindful” accounts that involve the use of mentalistic terms, where the latter, as Penn
(2011) points out, fail to identify any cognitive process at all, but simply fall back on folk
psychological descriptors (see also Ramsey 2007). In particular, we seem to lack the
appropriate vocabulary to discuss the abilities of simple creatures in ways that capture
the flexibility of their behavior, but that do not generate an inappropriately “mentalistic”
view of matters (van Duijn et al. 2006).

Determining when or why an organism can be deemed cognitive is, therefore, not clear-
cut, and tracing the evolutionary origins of cognition accordingly is more difficult.
Perhaps this is why accounts of cognitive evolution are sometimes couched in terms of
brain size (Roth and Dicke 2005; Dunbar and Shultz 2007). Brain size is, of course, more
easily and unambiguously measured than cognitive capacity, and has the advantage that
it can be obtained from species that are long dead and fossilized, as well as from extant
species. Although measuring brains avoids the difficulty of defining cognitive processes,
this approach rests on the assumption that brain capacity is an accurate proxy for
cognitive abilities. As van Duijn et al. (2006, p. 163) point out, however, it appears that
our best reason for linking brains and cognition is that it “seems so obviously true”: those
animals we characterize as obviously cognitive, like humans and apes, have very large
brains for their body size, making the connection seem entirely logical. Creatures like
bacteria, which lack nervous tissue cannot, by definition, engage in a cognitive process,
and this also seems reassuringly logical. Equating brains (and nervous systems) with
cognition thus provides another (simple) means by which different creatures can be
sorted into those that are cognitive and those that are not.

The problem with this solution is that bacteria have revealed themselves capable of a
remarkable flexibility in behavior, despite lacking the only structure, i.e., a nervous
(p. 721) system, allegedly capable of producing such flexibility (e.g., Shapiro 2007). The

brain = cognition equation is also undermined by creatures like salticid spiders, which
possess brains that are absolutely tiny, no bigger than a poppy seed, yet are capable of
remarkably flexible hunting tactics, including the ability to engage in deceptive mimicry,
create diversions to distract prey, and take long, complex detours in order to better
position themselves for prey capture (Tarsitano and Jackson 1994; Tarsitano 2006; Barrett
2011). If a certain degree of behavioral flexibility is possible without brains, or with only
very small ones, then the idea that more brain means more cognition seems rather shaky,
as is our ability to distinguish cognitive from noncognitive creatures using simple metrics.

The situation is made even worse by the fact that, among creatures in possession of fully
fledged brains, the manner in which cognitive capacity can be mapped onto brain
structure is a highly vexed issue—Healy and Rowe (2007, 2013), for example, identify a
number of problematic assumptions within the literature and dispute the idea that smart
behavior maps onto brain anatomy in the way some have argued. Even within the context
of this debate, however, the essential assumption that brains are the organs of cognition
remains intact: elsewhere, these same authors have argued for better controlled

Page 3 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

experiments that can eliminate the influence of “noncognitive” factors on behavioral


performance in order to achieve a “pure” estimate of brain-based cognitive capacity
(Rowe and Healy 2014).

The rise of 4E cognition (embodied, embedded, enactive, and extended) has raised
awareness of the inconsistencies and problems of the traditional cognitivist view, making
it clear that brains cannot be divorced from their bodily and environmental contexts
(Clark 1997; Gallagher 2005; Rosch, Thompson, and Varela 1992; Chemero 2009; Hutto
and Myin 2012; Barrett 2011). Whatever cognition might be, it is not best captured by a
purely brain-based computational perspective that places brains at a remove from the
rest of the body and the worlds they inhabit. This is true of humans, but even more so for
other creatures, given that the computational views on offer take some quite peculiarly
human abilities as a benchmark for cognitive processes in general; these might not apply
to other animals, or at least they might not operate in quite the same way. 4E cognition
thus offers an alternative to the anthropogenic approach, by providing a bottom-up or
“biogenic” perspective on cognitive evolution, where the principles of biological
organization, and their links to fitness, are taken to be the most productive means by
which we can understand what cognition is, what it does, and how and why it evolved
(Lyon 2006). A 4E biogenic approach thus seems worth considering, given that—although
psychologists readily accept that cognition occurs in a biological context—the inherent
anthropocentrism of the representational, computational view of cognition is less widely
recognized, even among authors who attempt a more ecological approach. Hence, many
authors clearly consider representational and computational theories of cognition to be
“species neutral,” when in fact these are highly human-oriented and hence much less
“biogenic” than they imagine. In what follows, I first consider some ideas relating to the
minimal criteria for cognition, before going on to discuss how cognitive evolution can be
conceived within a 4E framework. I conclude (p. 722) with a short consideration of why
more radical forms of embodied cognitive science may represent the way forward.

Cognition as Coordination
If we move away from thinking of cognition as brain-bound computational processes, and
move toward a more functional definition, the line between cognitive and noncognitive
creatures becomes increasingly blurred. Lyon (2006), for example, defines cognition as
the “capacity to infer relations between external circumstance and internal need to
facilitate agency” whose function is to “enable successful action and interaction in a
niche.” Anderson (2003) offers an even simpler definition of cognition as “situated
activity.” It is not immediately apparent that, under these definitions, cognition depends
on the possession of a particular kind of brain or nervous system organization, or that
there are particular kinds of representational or computational processes that qualify as
cognitive while others don’t.

Page 4 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

Brooks (1991, 1999), for example, famously rejected the idea that representation and
reason were necessary for intelligent behavior, and sought instead to study the adaptive
behavior of whole agents operating in an uncontrolled real-world environment (see also
Beer 1996; Pfeifer and Scheier 2001; Pfeifer and Bongard 2006). Brooks’s work
demonstrated that perception and action could often be so closely intertwined that there
was no need for “cognition”: the discrete action-producing systems of his robots resulted
in adaptive, robust behavior in the complete absence of a central processing unit that
could integrate inputs, plan, and produce coordinated action. His robots did not need a
complete internal representation of the world in order to produce coherent behavior.
Instead, they were “situated”: sensitive to relevant information in their surroundings,
with tightly coupled perception-action mechanisms that were, in turn, directly coupled to
an environment that they dynamically and frequently sampled. As Chemero (2013)
suggests, Brooks’s work was thus a logical extension of Gibson’s (1979) anti-
representational ecological theory of perception (where psychological phenomena are
seen as relations between an organism’s physical constitution and its environment) along
with Barwise and Perry’s (1983) notion of situation semantics (the idea that the meaning
of thoughts and utterances has nothing to do with mental representations, but with the
relationship between thinkers/speakers and their environment). For Brooks, like his
predecessors, intelligence was a relational property: it came about via possessing a
particular physical constitution operating within a particular environment.

This last point is key, as it helps illustrate how the rationale for Brooks’s research
program was genuinely evolutionary. As he pointed out, most of the four billion years of
life on earth has been spent refining the perception and action mechanisms that permit
organisms to cope with a dynamic, ever-changing environment (Brooks 1999). These
seemingly more “primitive” nonrepresentational forms of intelligence therefore must
have been the more difficult forms to evolve and implement, whereas the highly (p. 723)
representational processes that we humans deem intelligent—language, mathematical
reasoning, symbolic logic—must have been relatively easy to set in place. Following this
reasoning, Brooks decided to focus on insect-level intelligence, arguing it was essential to
understand the 97 percent of behavior that was controlled by nonrepresentational
processes, in order to understand the remaining 3 percent that could be deemed
representational (where these values refer to humans). As such, Brooks’s work
represented an early attempt to rid artificial intelligence of its anthropocentric
orientation, grounding cognition in biological processes, and not human artifacts.

This proposal was met with skepticism by some, however, who considered it unlikely that
insect intelligence would scale up to human intelligence. As Kirsh (1991, p. 162) put it:
“Insect ethologists are not cognitive scientists.” This latter pronouncement seems
increasingly insecure in the face of scientific findings that make a strong case for insect
cognition (e.g., Leadbeater and Chittka 2007; Lihoreau et al. 2012). One could argue,
though, that the decision to pursue insect-level intelligence as the basis of intelligent

Page 5 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

“cognitive” behavior was somewhat arbitrary, given that insects themselves appear pretty
late in the day, evolutionarily speaking.

Indeed, some in the classic enactivist camp suggest that any attempt to distinguish
cognition from other kinds of processes is, essentially, meaningless, and consider all life
to be inherently cognitive (Rosch et al. 1992). More specifically, Thompson (2007) offers
us the position that “life and mind share a set of basic organizational properties, and the
organizational properties of mind are an enriched version of those fundamental to life.
Mind is life-like, and life is mind-like” (p. 128). The problem with this position is that, if
everything is cognitive, then nothing is (van Duijn et al. 2006): we are left no means of
coming to grips with why a sea urchin differs from a shark or a slow worm from a squirrel
monkey. If we want to understand differences between minimally cognitive systems and
more advanced systems, van Duijn et al. (2006, p. 159) argue, then “it seems important to
differentiate between the cognitive processes of the rabbit and those of the carrot”: we
need to distinguish the subset of living systems that can be classified as cognitive
systems, and make clear the basis on which we make such a classification.

As noted earlier, one means of doing so is to assume minimal cognition requires the
possession of a nervous system. Moreno and colleagues, for example, argue that the
sensorimotor behavior of creatures without a nervous system is simply an extension of
their metabolic processes whereas, on their account, minimal cognition requires “meta-
metabolic functions” that can transcend these processes, generating an autonomous
sensorimotor domain that can sustain an “independent domain of patterns” (Moreno et al.
1997; Moreno and Etxeberria 2005). As van Duijn et al. (2006) point out, however, several
authors have now questioned whether a nervous system is essential for cognitive (i.e.,
meta-metabolic) behaviors, arguing that the evolution of nervous systems represents the
augmentation of abilities that already exist in unicellular organisms, rather than nervous
systems marking the crossing of some “cognitive Rubicon” (Lyon 2006; Taylor 2004).

To make this point more forcefully, van Duijn et al. (2006) give the specific example of the
sensorimotor abilities of the bacterium, Escherichia coli, and its two-component (p. 724)
signal transduction (TCST) system. In brief, and to oversimplify a little, chemotaxis in E.
coli relies on two interacting signaling pathways, a fast one (of the order of milliseconds)
that mediates perception, and a slower one (of the order of seconds) that mediates
adaptation to the signals in the environment (and therefore constitutes a form of
memory). The perception pathway consists of surface receptors for chemical attractants
or repellents, the occupation of which triggers phosphorylation of other molecules inside
the bacterium, and influences the direction in which the cells’ flagella rotate (so
controlling whether the organism “runs” or “tumbles”). The adaptation process involves
the methylation of occupied receptors. As this process is slower than the perceptual
process, the number of methylated receptors registers the concentration of chemicals in
the environment a few seconds previously. The receptors are continuously methylated and
demethylated in response to the occupation of receptor sites by attractants and repellant
chemicals as the organism moves through the environment while, at the same time,
receptor occupation and levels of methylation influence chemical processes inside the cell

Page 6 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

which, in turn, influence how the organism moves. The interaction and feedback between
the two systems thus produces a dynamic form of molecular memory. The difference
between the current state of the receptors and the internal level of methylation means
that the organism becomes sensitive to the relative concentration of chemicals in the
environment, rather than their absolute value, and hence can register any large-scale
abrupt changes in chemical concentration but will adapt to a constant level of signal.
This, in turn, means that organisms can move toward high concentrations of attractants
and away from repellents, but will remain in an environment with an optimum
concentration of attractants.

Van Duijn et al. (2006, p. 163) use this example to undercut the intuitively appealing idea
that minimally cognitive behavior requires a nervous system. Chemotaxis is more than
just an aspect of the cell’s metabolism because it involves physical changes of the
organism in relation to the environment that are distinct from metabolic processes.
Metabolic processes benefit from the changes in the environment the organism produces,
but the changes produced (moving toward a food source or away from a noxious stimulus)
are not themselves part of metabolism. Thus, they fulfill Moreno and colleagues’
requirement of autonomy and “meta-metabolic functions.” Given this, van Duijn et al.
(2006) suggest that sensorimotor coordination offers a conceptually clearer and more
effective starting point for minimal cognition.

Grounding cognition in sensorimotor coordination emphasizes that organisms are always


fully situated (or embedded) in their environments, and it naturally incorporates physical
embodiment. For example, E. coli uses a temporal system (memory) to act adaptively in
the environment because it is too small to employ any kind of spatial mechanism (e.g., it
cannot position its sensors/receptors in such a way that maintaining equal stimulation at
each sensor or receptor would drive the organism toward a target). Its rod-like shape
improves its ability to engage in chemotactic behavior, and it possesses flagella that are
activated by particular kinds of proteins, as well as the TCST system itself, which
provides an internal connection between sensors and effectors. This approach therefore
not only specifies what organisms are able to do, but also how they (p. 725) are able to do
it, something that any adequate theory of cognition must provide. This fits with a biogenic
perspective, as it identifies the need to optimize the conditions needed to maintain
metabolic processes within acceptable limits as the fundamental problem to be solved.
This, van Duijn et al. (2006) suggest, may lie at the origin of many basic forms of
cognition. This being so, it is crucial to understand the variety of solutions that exist if we
are to understand the nature of cognition, and how more complex forms of cognition
arise. In effect, then, van Duijn et al. (2006) reiterate Brooks’s (1999) earlier views, but
provide a more principled argument for where we should begin our exploration of the
origins and evolution of cognition, rather than taking insect-level cognition as a
convenient, but arbitrary, starting point.

Page 7 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

The Skin-Brain Thesis


This view has further implications for how to think about nervous system evolution.
Keijzer et al. (2013), for example, suggest that nervous systems may have evolved not to
enable more intelligent behavior (after all, bacteria, it seems, are pretty smart), but to
enable muscular behavior—what they refer to as the “skin-brain thesis” (SBT). Among
mobile organisms, the evolution of multicellular forms gave rise to selection for
movement via muscular contraction, rather than by the use of cilia or flagellae (the
slender hair-like or whip-like protuberances that single-cell organisms use to propel
themselves) in order to coordinate and control the movement of the body in adaptive
ways (Keijzer et al. 2013). (Other multicellular organisms, like plants, followed a different
evolutionary trajectory, and there is fascinating work on the nature of plant learning and
intelligence, see, e.g., Gagliano et al. 2014.) A simple analogy helps make clear why this
should be: building a pyramid from Lego is much easier than building it from massive
stone blocks—it is obvious that the greater size of the latter entails a completely different
means of moving things around. Thus, the rise of muscle-based motility is seen as a key
transition in animal evolution, with muscular contractions central to understanding the
evolution of nervous systems themselves: “while cilia—literally—only allow your skin to
crawl, internal sheets of contractile cells make it possible to use the body itself to
accomplish motility” (Keijzer et al. 2013, p. 12). According to this hypothesis, then, the
earliest nervous systems were coordination systems—diffuse nerve nets that spread out
over a large portion of the animal’s body (as seen in species like hydra and jellyfish)—and
not input-output systems (like the classic notion of a reflex arc, where a sensor is linked
to an effector by neural connections).

Keijzer et al. (2013) put forward a two-phase evolutionary process. The first (hypothetical)
phase involved the evolution of excitable myoepithelia (epithelia with contractile
properties that can conduct electrical activity across their surface) that was capable of
chemical transmission between cells. This myoepithelia is thought to have comprised an
unbroken sheet of protonervous and muscle tissue, with each cell signaling to its
immediate neighbors. Keizer et al. (2013) refer to these muscle surfaces as “Pantin
surfaces” (p. 726) (in honor of Carl Pantin, who first had the idea that the need for
mobility drove nervous system evolution; Pantin 1956), each of which has a specific size,
shape, and extension for each species (and indeed each individual, given that these will
not be identical to each other). They refer to the self-organized waves of coordinated
contraction that the surface produces as “patterning.” This aspect of the hypothesis is
therefore of particular relevance to an embodied perspective, because such patterning is
argued to be particular to the Pantin surface that produces it, i.e., it inherently reflects
the structure of an animal’s body. Patterning cannot, therefore, be interpreted as some
form of generic motor output function (e.g., feeding, mating), but should be viewed more
like the action of an organ: as a single coordinated system of movement that is

Page 8 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

characteristic of a particular species and individual. As such, there is no controller that


stands in contrast to the element being controlled.

The same goes for sensory capacities (Keijzer 2015). In order for patterning over the
Pantin surface to remain functional, the organism must be sensitive to disturbances in
such patterning (e.g., due to growth, short-term changes in activity, temperature,
chemical concentrations), as this will allow it to compensate and restore functionality.
The animal must therefore be sensitive to the ongoing dynamical changes in contraction
and extension across its Pantin surface, so that it can influence the processes that
produce and maintain these same contractions. This sensitivity to its own bodily dynamics
naturally incorporates sensitivity to any external disturbances that hinder or enhance
bodily contractions and extensions. Thus, the animal can sense the environment via the
skin-brain despite the absence of any dedicated external receptors. In this view, the
“animal body itself becomes a sensing device through its use of contractile tissues and
the environmental feedback this generates, both within and external to the body.” (Keijzer
2015, p. 14).

In the second phase of evolution, Keijzer et al. (2013) argue that a comparatively small
number of more specialized signaling cells evolved, with elongated axodendritic
processes (i.e., neurons) that enabled them to signal to non-adjacent cells. This would
give rise to a diffusely connected nerve net spread over the entire body. The advantage of
such nerve nets is that they would add to the self-organizing properties of existing
epithelial and muscular tissues. Specifically, Keijzer et al. (2013) argue that the addition
of long-range neural processes could thus enable wave-like patterning across larger
Pantin surfaces, releasing a constraint on size, as well as offering greater flexibility: in
contrast to sheet-like myoepithelia, nerve nets can take many forms, so removing
constraints on body form. Rather than providing specific, more precise connections from
sensors to effectors, as an input-output view of nervous system evolution would argue,
the function of nerve nets was to control, modify, and extend self-organized patterning
across a Pantin surface. Much like the memory system of E. coli, then, diffuse nerve nets
offered organisms an extra loop of control over behavior.

With diffuse neural nets in place, Keijzer et al. (2013) suggest that, evolutionarily, the
contractile apparatus could be cut loose from its placement on the animal’s skin and
repositioned as internal muscle. The general differentiation of muscle from skin would
then give rise to more complex nervous systems, and the kinds of animal bodies that we
(p. 727) are familiar with today, which fit well with the input-output conduction view of

nervous systems (see also Jekely et al. 2015). Thus, there is nothing in the SBT that
suggests we should abandon the conduction view. Rather the SBT offers a different, more
embodied standpoint from which to consider possible hypotheses about the origins of
nervous systems, and how we can think about cognitive processes. The SBT does,
however, present a major contrast to the one-size-fits-all view of classical cognitivism,
which “casts the operation of basic nervous systems in terms of abstract computational

Page 9 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

functions.” Instead, “the SBT draws it as a producer of complex but concrete dynamical
patterns of activation” (Keijzer et al. 2013, p. 15).

Behavioral Flexibility, the Elaboration of


Nervous Systems, and the Umwelt
I have explained the SBT in some detail because it offers a view of nervous system
evolution that fits perfectly with an argument I have offered elsewhere, which considers
cognitive evolution to be a process by which organisms became better equipped to track
and deal with unpredictable contingencies in real time in changeable environments
(Barrett 2011). In brief, what has taken place over the course of evolutionary history is
the progressive elaboration of animals’ central nervous systems, with larger, long-lived
species tending to possess more complex brains and more elaborate nervous systems
than small, short-lived species. Longevity is a crucial factor because long-lived creatures
are likely to encounter a wide variety of conditions across their lifetimes, and higher
levels of behavioral flexibility enable animals to cope with environmental change more
effectively (although, of course, this depends crucially on the rate of change: climate
change and habitat loss are exposing the limits of species’ flexibility).

It is important to emphasize the nervous system as a whole, and not just the brain,
because species with larger brains also tend to have more elaborate sensory apparatus.
Although we often speak of humans having a poorer sense of smell, or less sensitive
hearing, than other creatures, our senses of smell and hearing are actually pretty good,
and we also have good vision, along with a sense of touch, taste, and balance. There is
often the assumption that more brain tissue allows animals to engage in more complex
“thinking” processes, but the control of complex sensorimotor systems also requires a lot
of neural tissue. So, in order to employ our receptors effectively, they are incorporated
into perceptual systems (e.g., we have two eyes in our head, that in turn is positioned on
a moveable neck, which is attached to a body that can move in a variety of ways) that
allow us to actively explore and sample the environment, and not passively receive it
(Gibson 1979). Part of the reason why human brains are so big, then, is to help control
and coordinate our sensory and motor apparatus. This, in turn, is partly why humans
show such high levels of behavioral flexibility: our greater sensitivity to many different
(p. 728) aspects of our environments expands our Umwelt—the term used by Uexküll

(2014) to capture the notion that organisms are sensitive only to those aspects of the
environment that hold significance for their survival and reproduction. This also helps us
recognize why other animals, with different kinds of sensorimotor systems, which have
developed to differing degrees, vary in their flexibility and problem-solving capacities.

An embodied perspective on comparative cognition therefore includes the idea that other
features of an animal’s anatomy can contribute to successfully navigating the world,
including how morphology itself can contribute to successful problem-solving. For

Page 10 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

example, New Caledonian crows are large-brained birds, and have exceptionally good
tool-using abilities compared to other members of the crow family, despite the fact that
these are similarly large-brained. This is partly because they have much greater
convergence (binocular vision) and straighter beaks (Troscianko et al. 2012). It is much
easier for New Caledonian crows to visually guide the stick-like tools they use to obtain
food than it is for other crow species, and this helps explain why they perform so much
better on tool-using problems than other corvids in the laboratory, but show comparable
performance on non-tool-using tasks.

Once one is attuned to the idea that behavioral flexibility is a property of bodies as well as
brains, and that the exploitation of environmental structure can serve a similar function,
examples appear all over the place—from the plant hoppers that have a gearing
mechanism in their legs that synchronize their jumping movements more precisely than
neural control can achieve (Burrows and Sutton 2013), to the seahorses that have square
rather than cylindrical tails (Porter et al. 2015), which improve their ability to maintain a
controlled grasp onto coral reefs, to the way that mudskippers (fish with the ability to
feed on land) form a “hydrodynamic tongue” to help capture food: the fish emerge onto
land and eject a mouthful of water, using a pattern of motion that shows a clear
resemblance to the way newts protrude their sticky tongues to capture prey (Michel et al.
2015). From an evolutionary comparative perspective, a 4E approach allows us to
appreciate the capacities of other species on their own terms: it brings the organism’s
body and environment into clear focus, in a way that allows the anthropocentric biases of
the cognitivist position to be circumvented, and (hopefully) overcome.

Going Radical and Getting Our Continuity


Right
Originally, the argument presented here was couched in terms of “conservative”
embodied cognitive science (CEC) (Hutto and Myin 2012), which brings mind, body, and
world together by reconfiguring representations as controllers of action (“action-oriented
representations”) rather than the abstract, action-neutral “mirrors” of the world favored
by the classical cognitivist position (e.g., Clark 1997, 2008). More recently, (p. 729) I have
become persuaded by radical embodied and enactivist positions (REC), which embrace an
anti-representational stance (Chemero 2009; Hutto and Myin 2012).

One reason is because, as Chemero (2013, p. 147) notes, CEC maintains a commitment to
a computational theory of mind, “embracing some of the ideas of Gibson, Barwise and
Perry and Brooks, but backpedaling on the strongest claims these authors made.” Given
this, REC should not be seen as the radicalization of Clark-style embodied cognitive
science (as is often assumed); rather, Clark-style CEC should be seen as a watered-down
version of REC. Hence, in Chemero’s (2013) view, it is the computational elements of CEC
that actually requires justification, and not the radical anti-representational stance, which

Page 11 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

is inherent to a functionalist embodied approach. Given the anthropocentric origins of the


computational theory of mind (Barrett 2015a), ridding embodied theories of a
computational stance allows us to begin thinking of cognition in a truly biogenic, bottom-
up fashion.

Rejecting the notion of contentful representation for “basic minds” (i.e., those that are
non- or prelinguistic) thus seems to offer the same horizon-expanding promise as the SBT.
REC presents a new way of thinking about and conceiving of cognition, of understanding
what minds are and how they become more elaborate. A representational stance, in
contrast, seems to lead inexorably to an intellectualist view of other animals (and indeed
our own species; Barrett 2011), in ways that may hinder rather than help understand
cognitive evolution.

As I have argued elsewhere, the search for the precursors of our own cognitive and
linguistic abilities in other animals reflects the same “explanatory need” that plagues
studies of human cognition (Hutto 2013a): contentful representations are deemed to be
present because a need for them has been established by a logical argument that simply
insists this must be the case (Barrett 2015b). Specifically, the logic of evolutionary
continuity is used to insist that, as human contentful representational knowledge could
not have sprung from nowhere, then its precursors must be found in other species.

For example, Dorothy Cheney and Robert Seyfarth, pioneers and leaders in the field of
comparative cognition, have argued that baboon social knowledge “reveal[s] a
hierarchical, rule-governed structure with many properties similar to those found in
human language” (Seyfarth et al. 2005, p. 264), as well as suggesting that vervet alarm
calls are “best described as a proposition” and that “non-human primates certainly act as
if they are capable of thinking (as it were) in sentences” (Cheney and Seyfarth 2005, pp.
150–1). But it is only the assumption that nonhuman minds must contain content-bearing
representations that makes it necessary to characterize monkeys as possessing this kind
of knowledge, when many aspects of their behavior actually speak against this view.
Cheney and Seyfarth (2005) freely admit that their monkeys seem “unaware of their
effects on their audiences’ knowledge and beliefs” (p. 139), and do not “know that” others
have beliefs and knowledge, nor do they show any evidence of “knowing that” they
themselves possess knowledge and beliefs (Cheney and Seyfarth 1992; Cheney and
Seyfarth 2008). Indeed, it is this mismatch that requires certain concepts either to be
modified or watered down (e.g., a proposition is simply “a single utterance or thought
that simultaneously incorporates a subject and a predicate,” with no conditions of
(p. 730) satisfaction or truth-preserving component; Cheney and Seyfarth 2005, pp. 150–

1). As a result, it is unclear the extent to which these concepts can contribute to an
understanding of either monkey or human cognitive evolution.

A commitment to both evolutionary continuity and the computational theory of mind


forces comparative psychologists to ground human knowledge and language in more
basic forms of knowledge of our closest relatives—to accept that at the basis of our
knowledge is yet more knowledge. But this premise is shaky. As Wittgenstein (regarded

Page 12 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

by some as the first enactivist; Moyal-Sharrock 2013) argues, our most basic beliefs, our
“hinge certainties,” are non-epistemic and non-propositional: it is action that lies at the
base of our knowledge (Wittgenstein 1974, p. 204). If this is correct, then it makes no
sense to look for the precursors of human knowledge and linguistic thought in the
(putative) thought processes of the living primates (or other organisms). For RECers,
there are no precursors of this nature, and whatever we share in common with other
species is to be found by seeking similarity in the nonrepresentational ways they act and
control behavior, as these will form the common ground between humans and other living
beings. In other words, we have to get our continuity the right way around, and the
dominant computational view of mind seems to get it wrong.

REC is therefore attractive because it is explicitly pluralist in its approach. It does not
argue against all forms of representational thought—only the notion that basic minds
must possess content (Chemero 2009; Hutto and Myin 2012; Hutto 2013b). Linguistic,
socioculturally scaffolded minds like ours clearly deal in rules and representations (else
you would not be able to read and comprehend the words on this page), and it is quite
clear that some of the interesting issues of human psychology are best understood using
a representational framework. It is a mistake, however, to assume that, given this state of
affairs pertains in the human species, that contentful thought must therefore characterize
cognition across the animal kingdom.

Interestingly enough, this pluralist stance is seen as problematic precisely because it


seems to deny evolutionary continuity. Menary (2015), for example, suggests that, if
representational content is a feature only of cognitive systems scaffolded by language and
sociocultural practices, REC is left with the problem of bridging the gap between basic
cognitive processes and enculturated ones. That is, if we are to explain human abilities in
natural terms we must ensure a seamless psychological continuity across all forms of life,
and content-bearing representations provide the necessary link. In response to this, Hutto
and Satne (2017) ask: does evolutionary continuity logically require psychological
continuity? After all, as already discussed, this assumption raises its own problems. The
insistence that there be no discontinuity between humans and other species also seems to
deny the fact that humans occupy a unique sociocultural cognitive niche. If, as REC
argues, content arises with the mastery of sociocultural practices and artifacts in this
particular kind of niche—that is, if representations exist publicly externally first and are
then “internalized,” as Vygotsky (1997) argued—then the evolution and maintenance of
these public sociocultural practices can, in fact, explain both the initial emergence and
continued maintenance of content-involving minds (Hutto and Satne, 2017). It may be
that public practices like the high-fidelity transmission of (p. 731) cultural behaviors,
along with teaching (i.e., an elaboration of other skills also found in nonhuman species),
are crucial to generating human minds (Laland, 2017). Here, then, we come full circle,
returning to Brooks’s (1999) argument that, as human cognitive abilities appeared in a
mere blink of evolution’s eye, they must have been relatively simple to implement. “Basic
minds” were the hard part. REC therefore seems to get its continuity just right: it allows

Page 13 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

these abilities to emerge late, and builds them on top of basic minds, rather than
extending content-dependent processes far back into the evolutionary past.

Acknowledgments
Thanks to Jessica Parker, Gert Stulp, and an anonymous reviewer for comments on an
earlier draft. This work was supported by the Canada Research Chairs and NSERC
Discovery Grant Programs.

References
Anderson, M. L. (2003). Embodied cognition: a field guide. Artificial Intelligence, 149, 91–
130.

Barrett, L. (2011). Beyond the brain: how body and environment shape animal and human
minds. Princeton, NJ: Princeton University Press.

Barrett, L. (2015a). A better kind of continuity. The Southern Journal of Philosophy, 53,
28–49.

Barrett, L. (2015b). Back to the rough ground and into the hurly-burly. In: D. Moyal-
Sharrock, V. Munz, and A. Coliva (eds.), Mind, Language and Action: Proceedings of the
36th Wittgenstein Symposium. Berlin: De Gruyter, pp. 299–316.

Barwise, J., and Perry, J. (1983). Situations and attitudes. Cambridge, MA: MIT Press.

Beer, R.D. (1996). Toward the evolution of dynamical neural networks for minimally
cognitive behavior. From Animals to Animats, 4, 421–9.

Brooks, R.A. (1991). Intelligence without representation. Artificial Intelligence, 47, 139–
59.

Brooks, R.A. (1999). Cambrian intelligence: the early history of the new AI. Cambridge,
MA: MIT Press.

Buckner, C. (2011). Two approaches to the distinction between cognition and “mere
association.” International Journal of Comparative Psychology, 24, 314–48.

Burrows, M., and Sutton, G. (2013). Interacting gears synchronize propulsive leg
movements in a jumping insect.” Science, 341, 1254–6.

Byrne, R.W. and Bates, L.A. (2006). Why are animals cognitive? Current Biology, 16,
R445–48.

Chemero, A. (2009). Radical embodied cognitive science. Cambridge, MA: MIT Press.

Page 14 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

Chemero, A. (2013). Radical embodied cognitive science. Review of General Psychology,


17, 145–50.

Cheney, D.L., and Seyfarth, R.M. (1992). How monkeys see the world: inside the mind of
another species. Chicago: University of Chicago Press.

Cheney, D.L. and Seyfarth, R.M. (2005). Constraints and preadaptations in the earliest
stages of language evolution. The Linguistic Review, 22, 135–59.

Cheney, D.L., and Seyfarth, R.M. (2008). Baboon metaphysics: the evolution of a social
mind. Chicago: University of Chicago Press.

Clark, A. (1997). Being there: putting brain, body, and world together again.
(p. 732)

Cambridge, MA: MIT Press.

Clark, A. (2008). Supersizing the mind: embodiment, action, and cognitive extension:
embodiment, action, and cognitive extension. Oxford: Oxford University Press.

Dunbar, R.I.M. and Shultz, S. (2007). Evolution in the social brain. Science, 317, 1344–7.

Gagliano, M., Renton, R., Depczynski, M., and Mancuso, S. (2014). Experience teaches
plants to learn faster and forget slower in environments where it matters. Oecologia, 175,
63–72.

Gallagher, S. (2005). How the body shapes the mind. Cambridge: Cambridge University
Press.

Gibson, J.J. (1979). The ecological approach to visual perception. Brighton: Psychology
Press.

Healy, S.D. and Rowe, C. (2007). A critique of comparative studies of brain size.
Proceedings of the Royal Society B: Biological Sciences, 274, 453–64.

Healy, S.D. and Rowe, C. (2013). Costs and benefits of evolving a larger brain: doubts
over the evidence that large brains lead to better cognition. Animal Behaviour, 86, e1–e3.

Hutto, D.D. (2013a). Enactivism, from a Wittgensteinian point of view. American


Philosophical Quarterly, 50, 281–302.

Hutto, D.D. (2013b). Psychology unified: from folk psychology to radical enactivism.
Review of General Psychology, 17, 174–8.

Hutto, D.D., and Myin, E. (2012). Radicalizing enactivism. Cambridge, MA: MIT Press.

Hutto, D.D., and Satne, G. (2017). Continuity skepticism in doubt: a radically enactive
take. In: C. Durt, T. Fuchs, and C. Tewes (eds.), Embodiment, enaction, and culture:
investigating the constitution of the shared world. Cambridge, MA: MIT Press, pp. 107–
28.

Page 15 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

Jekely, G., Keijzer, F., and Godfrey-Smith. P. (2015). An option space for early neural
evolution. Philosophical transactions of the Royal Society of London. Series B, Biological
sciences, 370(1684), 20150181.

Keijzer, F. (2015). Moving and sensing without input and output: early nervous systems
and the origins of the animal sensorimotor organization. Biology and Philosophy, 30, 311–
31.

Keijzer, F., van Duijn, M., and Lyon, P. (2013). What nervous systems do: early evolution,
input-output, and the skin brain thesis. Adaptive Behavior, 21, 67–85.

Kirsh, D. (1991). Today the earwig, tomorrow man? Artificial Intelligence, 47, 161–84.

Laland, K. (2017). On the evolution of culture and the origins of human mind and
intelligence. Princeton, NJ: Princeton University Press.

Leadbeater, E. and Chittka, L. (2007). Social learning in insects—from miniature brains to


consensus building. Current Biology, 17, R703–13.

Lihoreau, M., Latty, T., and Chittka, L. (2012). An exploration of the social brain
hypothesis in insects. Frontiers in Physiology, 3, 442.

Lyon, P. (2006). The biogenic approach to cognition. Cognitive Processing, 7, 11–29.

Menary, R. (2015). Mathematical cognition: a case of enculturation. Open MIND, 25, 1–


20.

Michel, K.B, Heiss, E. Aerts, P., and Van Wassenbergh, S. (2015). A fish that uses its
hydrodynamic tongue to feed on land. Proceedings of the Royal Society B, Biological
Sciences, 282. doi:10.1098/rspb.2015.0057

Moreno, A. and Etxeberria, A. (2005). Agency in natural and artificial systems. Artificial
Life, 11, 161–75.

Moreno, A., Umerez, J., and Ibañez, J. (1997). Cognition and life: the autonomy of
cognition. Brain and Cognition, 34, 107–29.

Moyal-Sharrock, D. (2013). Wittgenstein’s razor: the cutting edge of enactivism. American


Philosophical Quarterly, 50, 265–79.

Neisser, U. (1967). Cognitive psychology. Brighton: Psychology Press.

Pantin, C.F.A. (1956). The origin of the nervous system. Pubblicazioni Della
(p. 733)

Stazione Zoologica Di Napoli, 28, 171–81.

Penn, D.C. (2011). How folk psychology ruined comparative psychology: and how scrub
jays can save it. In: J. Fischer (ed.), Animal thinking: contemporary issues in comparative
cognition. Cambridge, MA: MIT Press, pp. 253–65.

Page 16 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

Penn, D.C., Holyoak, K.J., and Povinelli, D.J. (2008). Darwin’s mistake: explaining the
discontinuity between human and nonhuman minds. Behavioral and Brain Sciences, 31,
109–78.

Pfeifer, R. and Bongard, J. (2006). How the body shapes the way we think: a new view of
intelligence. Cambridge, MA: MIT Press.

Pfeifer, R. and Scheier, C. (2001). Understanding intelligence. Cambridge, MA: MIT Press.

Porter, M.M., Adriaens, D., Hatton, R.L., Meyers, M.A., and McKittrick, J. (2015). Why the
seahorse tail is square. Science, 349, 6243.

Ramsey, W.M. (2007). Representation reconsidered. Cambridge: Cambridge University


Press.

Rosch, E., Thompson, E., and Varela, F.J. (1992). The embodied mind: cognitive science
and human experience. Cambridge, MA: MIT Press.

Roth, G. and Dicke, U. (2005). Evolution of the brain and intelligence. Trends in Cognitive
Sciences, 9, 250–7.

Rowe, C. and Healy, S.D. (2014). Measuring variation in cognition. Behavioral Ecology,
25, 1287–92.

Rowlands, M. (2003). Externalism: putting mind and world back together again.
Montreal: McGill-Queen’s Press.

Seyfarth, R.M., Cheney, D.L., and Bergman, T.J. (2005). Primate social cognition and the
origins of language. Trends in Cognitive Sciences, 9, 264–6.

Shapiro, J.A. (2007). Bacteria are small but not stupid: cognition, natural genetic
engineering and socio-bacteriology. Studies in History and Philosophy of Science, Part C:
Studies in History and Philosophy of Biological and Biomedical Sciences, 38, 807–19.

Shettleworth, S.J. (1999). Cognition, evolution, and behavior. Oxford: Oxford University
Press.

Tarsitano, M. (2006). Route selection by a jumping spider (Portia labiata) during the
locomotory phase of a detour. Animal Behaviour, 72, 1437–42.

Tarsitano, M.S., and Jackson, R.R. (1994). Jumping spiders make predatory detours
requiring movement away from prey. Behaviour, 131, 65–73.

Taylor, B.L. (2004). An alternative strategy for adaptation in bacterial behavior. Journal of
Bacteriology, 186, 3671–3.

Thompson, E. (2007). Mind in life: biology, phenomenology, and the sciences of mind.
Cambridge, MA: Harvard University Press.

Page 17 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018


The Evolution of Cognition: A 4E Perspective

Troscianko, J., von Bayern, A.M.P., Chappell, J., Rutz, C., and Martin, G.R. (2012). Extreme
binocular vision and a straight bill facilitate tool use in New Caledonian crows. Nature
Communications, 3, 1110.

Uexküll, Von J. (2014). Umwelt Und Innenwelt Der Tiere. Berlin: Springer-Verlag.

van Duijn, M., Keijzer, F., and Franken, D. (2006). Principles of minimal cognition: casting
cognition as sensorimotor coordination. Adaptive Behavior, 14, 157–70.

Vygotsky, L.V. (1997). The collected works of L.S. Vygotsky: problems of the theory and
history of psychology (vol. 3). Berlin: Springer Science & Business Media.

Wittgenstein, L. (1974). On certainty (ed. G.E.M. Anscombe and G.H. Von Wright; trans.
D. Paul and G.E.M. Anscombe). Oxford: Basil Blackwell. (p. 734)

Louise Barrett

Louise Barrett Psychology Department, University of Lethbridge, Canada

Access is brought to you by

Page 18 of 18

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights
Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in
Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: Gothenburg University Library; date: 16 October 2018

You might also like