Chapter 1

4E Cognition

Historical Roots, Key Concepts, and Central Issues

Albert Newen, Leon de Bruin, and Shaun Gallagher

HISTORICAL ROOTS OF THE DEBATE

The debate about the role of the body in cognition has been ongoing since close
to the beginnings of philosophy. In Plato’s dialogue The Phaedo, for example,
Socrates considers the idea, which he attributes to Anaxagoras, that one could
explain his decision to remain in prison by a purely material or physical
explanation in terms of bodily mechanisms. Socrates himself rejects this idea—
surely, he thinks, there is something more to reason than just bodily processes.
Aristotle, however, was motivated by the idea that Anaxagoras was not entirely
wrong. While Aristotle did not accept the radical view of Anaxagoras, he
considered that the body (with special reference to the hands) may play some
role in what makes for human rationality. Such debates considering the role of
the body for the mind can be traced through medieval texts authored by
Neoplatonists, Aquinas, and others, and are given their modern formulations in
thinkers such as Spinoza, La Mettrie, Condillac, and many others. Pragmatists,
phenomenologists, and philosophers of mind wrestle with the same issues
throughout the twentieth century. The more proximate background for the
current debates about embodied cognition, however, is to be found in the
disagreements between behaviorists and cognitivists. Continuing tensions within
cognitivism, and the cognitive sciences more generally, brought on by
contrasting functionalist and neurobiological accounts that tended to ignore the
role of body and environment and focus on internalist explanations of brain
function, set the stage for the emergence of contemporary views on embodied
cognition.

In the 1990s, Varela, Thompson, and Rosch’s (1991) The Embodied Mind,
drawing on phenomenological and neurobiological resources, proposed an
enactivist account of cognition that emphasized the role of the dynamical
coupling of brain–body– environment. Around the same time, a paper by Flor
und Hutchins (1991) introduced distributed cognition as a “new branch of
cognitive science” for which the unit of analysis includes external structures,
collectives, and artifacts organized as a system to perform a task. Hutchins’s
(1995) Cognition in the Wild was a direct influence on Clark and Chalmers’s
(1998) now-classic philosophical essay, “The Extended Mind.” Throughout this
time period, additional work inspired by Gibson’s ecological approach to
psychology contributed to a growing realization that cognition was not limited to
processes in the head, but was embodied, embedded, extended, and enactive.

Although the concept of 4E cognition1 brings these different approaches

together under one heading and conceives of them as coherently opposed to the
internalist, brain-centered views of cognitivism, there are continuing
disagreements about a variety of issues within and among these embodied
approaches. Is cognition embodied, embedded, extended, or enactive? The issues
that continue to be debated concern the very nature of embodiment, the precise
way that brain, body, and environment are coupled or integrated in cognition,
and how much we can generalize from the observation of embodiment in one
type of cognitive performance to others. Furthermore, there are questions about
the role of representations and what it means to say that cognition is
“constituted” by bodily and environmental processes.

KEY CONCEPTS

How to Individuate Cognitive Processes

Before introducing the key concepts in the debate, we first need to consider
whether there are certain constraints that need to be taken into account in order
to answer the question of what cognition is and how we should individuate
cognitive processes.

If we take cognition as a natural kind (even if we do not know the underlying

mechanisms) this would limit the nature of our investigation to a search for the
relevant mechanism constituting it. But there is no consensus on this question:
while Buckner (2015) argues that cognitive processes are indeed natural kinds,
evidence about neural plasticity presents a strong challenge to this claim
(Hübener and Bonhoeffer 2014). An alternative strategy for answering the
question is to focus on typical examples (Newen 2015). This seems to be a
promising strategy, but it is not without problems. One complication is that a
selection of the typical examples is already biased by certain assumptions
concerning the nature of cognition. Thus, where traditional cognitive science
focused primarily on playing chess and mastering the “Tower of Hanoi,” i.e.,
tasks that are strongly rule-governed, proponents of 4E cognition appeal to
experiments that involve spatial navigation, face-based recognition of emotion,
and basic forms of social interaction. It is therefore paramount to get a clear
view on the assumptions about cognition that are made by proponents of both
positions.

4E Cognition and Traditional Cognitive Science

The foundation of traditional cognitive science used to be the representational

and computational model of cognition (RCC). According to this model,
cognition is a kind of information processing that consists in the syntactically
driven manipulation of representational mental structures. In particular,
cognitive processes were said to be (1) abstract, a-modal processes that mediate
between modality-specific sensory inputs (perception) and motor outputs
(action), and (2) computations over mental representations that are either
symbolic (e.g., concepts in a “language of thought”; Fodor 1975) or sub-
symbolic (e.g., activations in neural networks; Rumelhart et al. 1986). The RCC
also involves a specific view of where cognition was supposed to take place—
some kind of “contingent intracranialism” (Adams and Aizawa 2008). On this
view, cognitive processes are, as far as their ontology is concerned, realized by
brain processes only (at least in the case of humans and other animals), and as
far as their explanation is concerned, understandable and explainable by
focusing on brain processes only.

During the past couple of decades, these key elements of the RCC—the pivotal
role of computation and representation in all cognitive processing and the pivotal
role of a central processing unit in the brain as the sole relevant factor of
cognitive processing—have come under pressure (Gallagher 2005; Walter
2014). Proponents of 4E cognition have argued against the assumption that
cognition is an isolated and abstract, quasi-Cartesian affair in a central
processing unit in a brain. This idea is typically associated with functionalism,
which claims that cognitive phenomena are fully determined by their functional
role and therefore form an autonomous level of analysis. According to
proponents of 4E cognition, however, the cognitive phenomena that are studied
by modern cognitive science, such as spatial navigation, action, perception, and
understanding other’s emotions, are in some sense all dependent on the
morphological, biological, and physiological details of an agent’s body, an
appropriately structured natural, technological, or social environment, and the
agent’s active and embodied interaction with this environment. Even most of the
phenomena studied by traditional cognitive science—such as language
processing (e.g., Glenberg and Kaschak 2002), memory (Casasanto and Dijkstra
2010), visual-motor recalibration (Bhalla and Proffitt 1999) and perception-
based distance estimation (Witt and Proffitt 2008)—are not abstract, modality-
unspecific processes in a central processing area either, but essentially rely on
the system’s body and its dynamical and reciprocal real-time interaction with its
environment.

Thus, by maintaining that cognition involves extracranial bodily processes, 4E

approaches depart markedly from the RCC view that the brain is the sole basis of
cognitive processes. But what precisely does it mean to say that cognition
involves extracranial processes? First of all, the involvement of extracranial
processes can be understood in a strong and a weak way. According to the strong
reading, cognitive processes are partially constituted by extracranial processes,
i.e., they are essentially based on them. By contrast, according to the weak
reading, they are non-constitutionally related, i.e., only causally dependent upon
extracranial processes. Furthermore, cognitive processes can count as
extracranial in two ways. Extracranial processes can be bodily (involving a
brain–body unit) or they can be extrabodily (involving a brain–body–
environment unit).

Following this line of reasoning, we can distinguish between four different

claims about embodied cognition:

a.

A cognitive process is strongly embodied by bodily processes if it is partially

constituted by (essentially based on) processes in the body that are not in the
brain;

b.

A cognitive process is strongly embodied by extrabodily processes if it is

partially constituted by extrabodily processes;

c.

A cognitive process is weakly embodied by bodily processes if it is not partially

constituted by but only partially dependent upon extracranial processes (bodily
processes outside of the brain);

d.

A cognitive process is weakly embodied by extrabodily processes if it is not

partially constituted by but only partially dependent upon extrabodily processes.

The last version of the claim (d) is identical with the property of being
embedded, i.e., being causally dependent on extrabodily processes in the
environment of the bodily system. Furthermore, being extended is a property of
a cognitive process if it is at least partially constituted by extrabodily processes
(b), i.e., if it extends into essentially involved extrabodily components or tools
(Stephan et al. 2014; Walter 2014).

Many proponents of 4E cognition not only maintain that cognition involves

extracranial processes, but also that cognition is enacted in the sense that it
involves an active engagement in and with an agent’s environment (Varela,
Thompson, and Rosch 1991). We can distinguish between two versions of this
claim:

e.

A cognitive process is strongly enacted if it is partially constituted by the ability

or disposition to act;

f.

A cognitive process is weakly enacted if it is only partially dependent upon the

ability or disposition to act.

It should be emphasized that proponents of 4E cognition differ greatly in terms

of their commitments to these claims, and consequently in their interpretation of
what it means for cognition to be embodied, embedded, extended, and enactive.
One famous example of an enacted theory of cognition is Noë’s (2004) theory of
perception, according to which perception is not something passive that happens
to us or in us but something we do: according to him, having a 3D-perceptual
experience of an object includes having a specific disposition to act which he
spells it out in terms of implicit knowledge of sensorimotor contingencies. It is
part of the discussion whether this justifies a strong or only a weak enactment
claim (Engel et al. 2013).
Constitution Versus Causal Dependency

As we saw earlier, the distinction between constitution and causal dependency

plays an important role in the debate on embodied cognition. But what exactly
grounds this distinction? Consider the example of cognitive processes involved
in solving a simple math problem. It likely involves visual perception (if the
problem is presented on paper), memory, language or symbol processing, etc.
This means it would depend on a variety of elements and processes that include
neuronal processes in the visual cortex, in motor areas, in language areas, the
hippocampus, frontal areas, etc. In addition, as I read the problem I move my
eyes, and likely my head. I posture my body so that my eyes are a certain
distance from the text. I may gesture with my hands as I work out the solution.
All of these factors can be involved even if I am solving the problem “in my
head,” without pencil and paper or other instruments. If I am involved in a
competition to solve the problem, that stressful fact may have an effect on my
cognitive performance. Can proponents of embodied cognition claim that not
only the neuronal processes, but also eye movements, head movements, posture,
use of pencil and paper, and perhaps even the competitive situation are all parts
of the cognitive system that constitutes cognition in this case? When they make
such claims, critics have accused them of the so-called coupling/constitution
fallacy (Adams and Aizawa 2008; Rupert 2009), according to which the strong
coupling between neural and extraneural processes, including bodily movement
and use of pencil and paper, for example, does not suffice to make the non-
neural processes constituents, rather than just causal or enabling conditions of
the cognitive process. Quite generally, the question is whether, and if so, how,
we are able to decide (either empirically, pragmatically, or a priori) whether a
particular cognitive process is constituted by or merely dependent upon
extracranial or extrabodily processes.

One strategy in this debate is to question whether the concept of constitution

necessarily involves just non-causal, part-whole relations (e.g., Craver 2007), or
in some cases requires diachronic and dynamical relations that depend on
reciprocal causality (e.g., Kirchhoff 2014, 2015; Leuridan 2012). Another
strategy is to take relevant features as constitutive of a cognitive process (e.g., an
emotion or an episode of self-consciousness) if it is a characteristic feature of the
phenomenon and part of a minimal pattern of integrated features sufficient to
realize this phenomenon (e.g., Newen et al. 2015; Gallagher 2013). It may be
that most of the features of mental phenomena are neither necessary nor
sufficient but only characteristic. For example, a facial expression of fear is
partially constitutive of fear although there are realizations of fear that do not
involve the typical facial expression, e.g., in the case of a trained poker face
(Newen et al. 2015). Issues about the relation of constitutive, causal, or
background conditions are unresolved, and are still subject to ongoing debate in
the embodied cognition literature.

Mental Representations

Another important question in the debate on embodied cognition concerns what

role, if any, mental representations play in cognitive processing. The theoretical
landscape is such that 4E approaches can and in fact do have supporters from
both the computational/representational and the anti-computational/anti-
representational camp. Dynamicists like Chemero (2009), for instance, defend a
decidedly anti-computational/anti-representational version of embodied
cognition (see also Barrett 2011), while Wilson’s (1994) “wide
computationalism” and Clark’s (2008) “extended functionalism,” according to
which the mind is the joint product of intracranial processing, bodily input, and
environmental scaffolding, are unequivocally computational/representational. In
a similar vein, while some proponents of embodied cognition, for instance, in the
area of vision research, explicitly try to supersede traditional
computational/representational approaches (Gibson 1979; Noë 2004; Hutto and
Myin 2013), others merely try to enrich them by integrating environmental
resources (Ballard et al. 1997; Clark 2013). Thus, embodied approaches range
from the computation/representation friendly variety (Alsmith and de Vignemont
2012; Prinz 2009) to accounts that are explicitly anti-computational and/or anti-
representational (see Thelen et al. 2001; Brooks 1991; Pfeifer and Bongard
2006). This shows that the 4E approach as such does not presuppose a specific
view on representation and computation.

AN OVERVIEW OF THIS BOOK

Since the volume is organized in nine additional parts, we will provide a short
overview of the main questions that are treated in these parts.

Part 2: What is Cognition?

The second part of essays explores the concept of cognition specifically from the
perspectives offered by 4E approaches to the mind. From a standard viewpoint,
the debates around embodied approaches seem to turn the “what” question into
the “where” question, so that the answer to the question about the nature of
cognition is first of all about location: precisely where is cognition located? In
this regard the line that demarcates between inside and outside plays an
important role. From the perspective of the 4Es, however, the question of
location is less critical; indeed, the distinction between inside and outside is
downplayed, and the boundary line turns out to be a movable and permeable
border. Thus, on the extended mind paradigm, if you happen to be using a piece
of the environment to assist memory or to solve a problem, then in that case the
mind extends into the environment; on the enactivist view, if there is a
dynamical coupling to others or to tools in joint action, then there is no line that
cuts the organism off from these other social and environmental factors.
Cognition is affordance-based, where affordances are always relational (between
the cognizing subject or some form of life and the possibilities offered by some
entity or complex of entities), and where entity may be some physical part of the
environment, another person who can provide information or opportunity, a
social or cultural structure, or even something more abstract, such as a concept
that, with some manipulation, offers a solution to a problem. Such approaches
transform the question about cognition into questions about the nature of
affordances, about whether cognition is extended or extensive, about what
precisely we mean by coupling, about whether a dynamical systems approach
can do without representations, and so forth.