PreprintPDF Available

The Trouble with 'Embodied' Cognitive Science

Preprints and early-stage research may not have been peer reviewed yet.


This argumentative essay puts forward the case that even ‘radical’ and ‘embodied’ forms of cognitive science are seriously flawed because of a series of key assumptions that prevent them from being adequately either ‘radical’ or ‘embodied’. These assumptions concern the role of the subject, the framing of theory, the reliance on dichotomised approaches, the mis-naming of the whole discipline, the lack of awareness of dialectical features in the mind/brain, and the adoption of the dogmas of ‘ecological psychology’.
The Trouble with ‘Embodied’ Cognitive Science
Robert M. Ellis
This argumentative essay puts forward the case that even ‘radical’ and ‘embodied’ forms of
cognitive science are seriously flawed because of a series of key assumptions that prevent them
from being adequately either ‘radical’ or ‘embodied’. These assumptions concern the role of the
subject, the framing of theory, the reliance on dichotomised approaches, the mis-naming of the
whole discipline, the lack of awareness of dialectical features in the mind/brain, and the adoption of
the dogmas of ‘ecological psychology’.
This paper comes from the standpoint of a non-specialist, cross-disciplinary philosopher, offering a
critique of supposedly ‘embodied’ cognitive science from the outside, on the basis of identifiable
general assumptions. My work so far consists in the development of Middle Way Philosophy, which
brings a rigorously practical standpoint to the critique of absolute claims, both positive and negative,
so as to allow us to focus on provisional claims that can be modified in response to experience (Ellis
2015, 2019). Embodied meaning as developed by Lakoff and Johnson (e.g. 1999, 2007) provides
important support for this approach, by showing the dependence of the language of even absolute
claims on associative links made in the body and brain in the course of active experience.
This paper consists in an application of that Middle Way method to ‘embodied’ cognitive science (of
a kind that I would sharply distinguish from embodied meaning theory: meaning is not cognition).
Cognitive science in general is a label for the intersection between artificial intelligence,
neuroscience, psychology and philosophy of mind, where the nature of ‘cognition’ is investigated. It
is also the site of a clash between the majority of ‘representationalists’, who treat the mind like a
computer and think of it as ‘processing’ symbols that represent aspects of reality, and the minority
of ‘anti-representationalists’ or ‘eliminativists’, who want to explain the operations of the mind
without the use of ‘internal representations’. I will be particularly focusing on Chemero (2011) as a
representative of the ‘anti-representationalist’, ‘radical embodied’ approach to cognitive science. I
work on the assumption that if he represents the supposedly ‘radical’ and ‘embodied’ edge of
cognitive science, all these issues will apply even more to the assumptions of the vast majority of
other cognitive scientists, who are liable to be more conservative than him.
Chemero’s ‘radical embodied’ cognitive science seems to be subject to the Holy Roman Empire
problem (famously neither holy, nor Roman, nor an empire) his ideas may qualify as ‘cognitive
science’, but I shall argue that they are neither radical nor embodied. Instead they are pretty much
more of the same old mainstream appeal to absolutized abstractions, pretty much oblivious of the
subject, of the extent of uncertainty, of the belief v meaning distinction, or of the ways in which
reformers can merely reinforce the polarised framing of what they are trying to reform. How can any
view be ‘radical’ that doesn’t properly question its assumptions? How can it be ‘embodied’ when it
adopts an abstracted explanatory viewpoint without any consideration whatsoever for the
developing practical experience of the embodied individual who interprets that viewpoint?
I’ve identified six major issues in the assumptions made by Chemero, which are outlined in the
remainder of this paper.
1. Exclusion of the observing subject and its uncertainty
2. The ontological obsession
3. Anti-representationalism reinforces representationalist framing
4. An emphasis on cognition frames an incremental process in absolute and inverted terms
5. Neglect of the bilateral and dialectical features of experience
6. Ecological psychology is not
1. Exclusion of the observing subject and its uncertainty
The first and most basic problem is that even the approach to cognitive science that prides itself on
adequacy to the full complexity of conditions, drawing on dynamical systems theory, excludes the
most basic condition us. Of course, ‘embodied’ cognitive science tries to take into account the
psychology of perception and ‘cognition’, but only in a highly impersonal way that takes no account
of the dynamics of the subject. The most immediate effect of complexity in a system is the difficulty
for any subject understanding it when they are an intrinsic part of the problem. ‘Embodied’ cognitive
science uses complex analyses of systems, such as multi-dimensional variables attempting to model
every aspect of the system that is subject to change, and how all those changes relate to each other,
in terms of non-linear equations. However, this complexity is useless in practice if it ignores the
biggest variable in the system you whose judgement is constantly at stake in working out how
each of the mathematical variables relates to experience. Without more effort to take into account
the uncertainty attending the subject’s judgements, and make those in turn a core part of your
whole way of framing a situation, your mathematical models, however complex, may still be based
on flawed assumptions and thus of very little use to anyone.
The cause of this may be the way in which the attempt to exclude the subject is bred into the bone
of science, and forms such a basic part of the culture passed on through scientific education. This
does not mean that in various senses, cognitive scientists don’t investigate the process of individual
judgement. However, they do so by developing theories and attempting to verify those theories in
publicly accessible ways. That process is useful up to a point, but only as long as some attention is
paid to the ways that those theories and the judgements made about their justification are held by
those who advocate and investigate them. General theories of ‘cognition’ also need to be
understood in relation to the ‘cognitive’ development of those who hold the theories, and of how
that development may impact how they are understood. In short, complexity requires the
recognition of uncertainty, which requires provisionality (see Ellis 2015, 4.2.a & 2019 3.c).
Provisionality is not adequately supported only through ‘rigorous’ scientific method, but also
requires constant attention to the psychology of how beliefs are held by the subject. Contrary to
traditional scientistic assumptions, then, the appearance of more first-person ‘I’ in cognitive science
writing would thus probably be a sign of greater objectivity and universality, showing that the
subject as part of a system is being taken into account.
2. The ontological obsession
The second problem is again one that is widespread in academia, namely the relentless emphasis on
explanatory theory about what is generally the case as opposed to a processual emphasis on how we
come to understand things in a particular way. I call this ‘the ontological obsession’ (Ellis 2019, 4.b),
because it appears in all sorts of places, including even in Buddhism (where Emptiness, the general
description, often supplants the Middle Way, the processual method). In cognitive science, this takes
the form of philosophically-based arguments about what is generally the case: representationalism v
anti-representationalism, computationalism v anti-computationalism, behaviourism v ‘ecological
psychology’, and so on. Of course, theoretical formulations of what we take to be the case have their
place, but generally, no attention whatsoever is focused on either how we hold these beliefs or how
they operate in practice.
I’ve come to expect the ontological obsession to dominate in many areas of academia – Christian
theology, for instance, which is still dominated by the obsession with whether God ‘exists’, and
whether to ‘believe’ in him. However, I think I can be forgiven for being disappointed when the very
people who bill themselves as being at the forefront of human thought on creating an adequate
account of complexity and human understanding seem to be completely unaware that they are
stuck in it. The problem is that the theoretical battle lines on which the cognitive scientists line up
are actually irrelevant to any adequate theorisation of what they are discussing. ‘Representation’, for
instance, is not a problem because of whether we can know whether or not it ‘exists’, or whether it
forms the ‘correct’ basis for explaining the process of human judgement. Rather, the problem with
representationalism lies in how we hold beliefs about it. If you think representation provides a total
explanation, you are deluded so far the ‘anti-representationalist’ arguments hold good. If,
however, you think it plays no useful role in explanation, you are again absolutising your explanation
in a way that shows that you are holding your anti-representationalist view in an unhelpful way.
3. Anti-representationalism reinforces representationalist framing
That takes us to a further point, which again also applies much more widely, but is particularly ironic
in the context of cognitive science. The general issue here is that of unconscious dualism or
dichotomisation, whereby those who reject an inadequate but established view fail to challenge all
the framing that holds that makes that view damaging, but merely oppose it within the terms of that
framing. A clear example of this is theism v atheism, where those who deny that God exists continue
to reinforce the framing belief that God’s ‘existence’ is important or relevant. Atheism of this kind
merely reinforces the worst features of theism, since the dogmatism of that assumption, and the
way it can then support appeals to authority and a great many other biases, is the main problem
with theism, rather than the meaning of the God people encounter in their experience (Ellis 2015,
4.4.f). The same can be said of anti-representationalism in cognitive science, which, most ironically,
presents itself as a total explanation in a way that is highly reminiscent of the worst features of
Anti-representationalism presents itself as a total explanation because it holds to a claim about the
lack of representations in reality, regardless of their apparent role in experience. It lacks a wider
theory that takes their practical role into account. The status of anti-representationalist claims
themselves also needs attention. What is the status of anti-representationalist theory? Surely it is
itself implicitly a representation, unless it is accompanied by some explanation of its own
provisionality? But this requires attention to the way in which the theory is held and applied rather
than merely to the content of the theory and whether it is in itself correct or incorrect. It seems
impossible to deny that people (including anti-representationalists!) think they have representations
of the world in their minds much of the time, even though they are deluded for all the reasons that
anti-representationalists advance. What people then do with that belief in representations is more
important than whether or not it is ultimately true. The impression of representation in practice
causes people to absolutise by assuming a direct relationship between the language or symbols in
their minds and some form of actual or hypothetical reality (Ellis 2015,3.3). That practical effect is
the only reason that the argument about representationalism is important, and apparently the only
thing that cognitive scientists are not very interested in!
4. An emphasis on ‘cognition’ frames an incremental process in absolute and inverted terms
The whole framing of cognitive science as ‘cognitive’ creates another set of issues. ‘Cognition’ is a
concept deriving from that of knowledge, with a whole range of other responses to experience
(perception, meaning, emotion and belief) bracketed under ‘cognition’ on the assumption that
‘knowledge’ is their most basic feature. This itself involves representationalist assumptions, namely
that we should understand all the features of human response to their experience in terms of their
response to ‘reality’. This is also an aspect of the ontological obsession mentioned above. All the
evidence collected in psychology, linguistics and indeed cognitive science itself suggests that
‘knowledge’ is at best the tip of an iceberg of experience based in the human body, not the founding
principle on the basis of which that experience should be understood, and yet a thoroughly outdated
Platonic framework is still being enshrined in the very name of the discipline!
Human responses to their experience can best be described as following an incrementally rising
profile of definiteness, beginning with perception that is not necessarily meaningful, continuing with
meaning (where increasing associative links are made between experiences, including between
experiences and symbols), continuing further with increasing complexity in the formation of
cognitive models through the development of schemas and metaphor to structure our meaningful
symbolisations, then with the formation of (implicit or explicit) beliefs in terms of those cognitive
models. Beliefs enable us to act, on the assumption of the world around us being a particular way.
Long-established sceptical arguments make it clear that there is no certainty of any of those beliefs
being correct, and without representationalism there are no grounds to even think that the language
of our beliefs is even capable of matching reality (Ellis 2015 1.1). So ‘knowledge’ is merely a
hypothetical tip to this incremental pyramid. In this case, why should the concept of ‘knowledge’ be
extended into that of ‘cognition’ to encompass this entire scale, rather than the more basic concept
of ‘experience’? Given the still widespread false dichotomy between ‘cognition’ and ‘emotion’
(which cognitive scientists, such as Kiverstein and Miller (2015), are happy to question), why is it
‘cognitive science’ rather than ‘emotional science’ or better still ‘experiential science’?
To make ‘cognition’ descriptive of this whole pyramid of experience is absolutising, because
‘cognition’ can only either apply as a correspondence to reality or not, even though the phenomena
are incremental and ranged on degrees of probability. It is also inverted, because it makes the
supposed apex of the pyramid the top-down basis on which to understand the whole of that
pyramid, rather than building up our understanding of that pyramid in a bottom-up fashion from the
most basic experiences and responses that create it. Thus instead of ‘embodied cognition’, it would
be far more appropriate to talk of ‘embodied perception’, ‘embodied meaning’, and ‘embodied
epistemology’, focusing on elements of the pyramid and working from the bottom up, with due
regard for the practical role of the observing subject.
This objection to the name of the discipline is not based on any claim that embodied cognitive
scientists are not fully aware of all the points mentioned here they obviously are. It is a question of
the effects of naming the discipline in this way, and the assumptions of priority embedded in that
naming tradition. These need to be much more questioned and discussed.
5. Neglect of the bilateral and dialectical features of experience
Another major problem in cognitive science is the failure, so far, to take any account of the work of
Iain McGilchrist (2009), or see its implications for the base assumptions with which the discipline is
pursued. Iain McGilchrist develops our understanding of the impact of brain lateralisation, using a
wealth of prior evidence from neuroscience but pursuing it in a multi-disciplinary way. Whatever
specific technical objections may be raised regarding the uncertainty of the neuroscience of brain
lateralisation, McGilchrist’s work is much more far-reaching than their scope, because his
perspective on the brain also offers an accompanying basis for the reassessment of mental
experience as bilateral and dialectical. Regardless of their neural basis, ‘left hemisphere’ in
McGilchrist’s work corresponds to the patterning of self-sufficient, representational, goal-orientated
and dominant aspects of our mental experience, whilst ‘right hemisphere’ corresponds to the
receptive, sensory, corporeal, imaginative and metaphorical. Some commentators unfairly dismiss
McGilchrist because they take the import of his work to be a descriptive theory about neuroscience
on which his multi-disciplinary development depends. This is mistaken: instead the neuroscientific
evidence reinforces a view of mental experience as bilateral and dialectical that can also be justified
from several other directions.
There are at least three such directions. One is dynamical systems theory of the very kind that
‘embodied’ cognitive science claims to be using. McGilchrist’s account is that human experience is
bilateral and dialectical, implying that it cannot be encompassed in a single model, no matter how
complex, but rather that each left-hemisphere representation either attempts to maintain itself self-
sufficiently (forming a closed feedback loop), or changes according to the impact of new information
from the right hemisphere (forming an open feedback loop). A helpful application of dynamic
systems theory to our understanding of human experience is thus not simply one of constantly
reinforcing a particular complex explanation of how things are by adding further complexity. Rather
it is one in which the model itself becomes dynamic by constantly monitoring itself for provisionality.
A second reinforcement of McGilchrist’s approach comes from the practice of mindfulness, where
closed loops of self-reinforcing mental states can be challenged by increasingly open awareness
grounded in the body. There is no substitute for directly experiencing this in practice, to see the
attempt at self-sufficiency of what McGilchrist labels as the left hemisphere at work. A third
reinforcement comes from bias theory in cognitive psychology (e.g. Kahneman 2011), where biases
can be readily accounted for as ‘fast thinking’ shortcuts. These shortcuts are substituted for the
slower approach that is required to open oneself to awareness from beyond a currently dominant
self-sufficient set of implicit beliefs, again illustrating the dialectical interplay between ‘left
hemisphere’ and ‘right hemisphere’ function that could still stand even in the unlikely the event that
the neuroscience proved to be entirely wrong. I could go on to discuss a number of other
reinforcements to this model, including those of Popperian falsificationism, Jungian analytic
psychology and Ellen Langer’s research on ‘mindfulness’ in her specific sense, but there is no need to
multiply examples excessively.
To return to ‘radical’, ‘embodied’ cognitive science, however, so far as I can see, any synthetic
appreciation and application of this developing perspective is deafening in its absence. The base
theories of cognitive science constantly draw people back to a completely illusory basis of supposed
explanatory omniscience, rather than helping them adapt to the obvious absence of any such
omniscience in the dialectical interplay of their mental experience. The mode of all-encompassing
explanatory theory is that of the over-dominant left hemisphere. Welcome though the emphasis is
in much cognitive science of paying full attention to new empirical evidence, this will be of little help
if the framework applied for seeking and interpreting that evidence is one that assumes that the
mind/brain is unilateral.
A bilateral understanding of the mind/brain can also help to resolve the equally false dichotomy
between facts and values, description and prescription. Instead of trying to reduce ethics to further
description, and thus reinforcing relativism of values, cognitive science can provide the basis of an
account of justifiable, prescriptive ethics in terms of the openness of the self-sufficient left
hemisphere of the mind to the right. An adjustable set of beliefs is a set of beliefs that is more
adequate to conditions, and thus a better set of beliefs. Although the justification of our moral
judgements thus remains particular to each judgement in each situation rather than being a matter
of absolute rules, it is nevertheless no more ‘subjective’ than the alleged ‘facts’ (actually heavily
laden with values) that are still too often appealed to in science.
6. ‘Ecological psychology’ is not
Finally, it seems that the main influence on ‘radical embodied’ cognitive science in its approach to
mental processes and judgements is ‘ecological psychology’, an influential approach developed by
James Gibson (1979). This approach appears to have supplanted awareness amongst cognitive
scientists of many much more interesting and genuine developments in psychology since that date
(such as those mentioned above), and suffers from another Holy Roman Empire problem. It is not
‘ecological’, because it does not take all the elements of a complex dynamic system into account
(especially the subject!), and it is not ‘psychology’ in the sense of theory that helps us understand
human judgement or behaviour. Instead, it offers three dogmas, summarised by Chemero (2011,
p.98) as ‘perception is direct’, ‘perception is for action’, and ‘perception is of affordances’ (i.e.
opportunities to act). Rather than offering insights into how far perception may have these features,
‘ecological psychology’ tells us that they should be assumed true and makes them the basis for all
subsequent investigation, regardless of the massive crashing assumptions they involve and the
effects of making those assumptions. To be fair to Chemero, he does spend a lot of time critiquing
and refining these principles so as to improve on Gibson’s version of them (such as defining
‘affordances’), but he does not do so in a way that questions the basic assumptions.
The belief in ‘direct perception’ is a dogma because it cuts out all consideration of the experience or
judgement of the subject, insisting that perception is a direct response to what is perceived in the
environment. This appears to imply the metaphysical dogmas of realism and determinism (and
Chemero indeed makes the realism explicit). Although the ‘embodied’ relationship between an
organism and its activities in the environment is thus emphasised, this is at the expense of all
consideration of human awareness and its effects on judgement. In relation to McGilchrist’s work, it
seems to reduce all mental activity to the left hemisphere perspective namely that of a goal-
oriented response to the world. Even though it denies representations, it does so in a way that
neglects the absolute representations that must be maintained in the holder of the anti-
representational theory. It could only possibly be seen as an advance by completely ignoring those
dogmatic features, and narrowing the field of options so that it is seen as the only alternative to
Though I have little doubt that I will be accused of straw-manning (one always is when attempting to
comment on the affairs of specialists, as one then inevitably does not go into the detail and
complexity they would like), these six points seem to me to have a pretty clear basis in relatively
unambiguous key statements made by Chemero and others. However, it will be welcome news if it
turns out that some ‘embodied’ cognitive scientists do not either make some of those six key
assumptions, or make other assumptions that are even less radical and less embodied than
Chemeros. It needs to be noted that the merely theoretical recognition of the problem is not always
enough to shake it off, if one is not willing to follow through the implications of thinking differently.
A genuinely radical, genuinely embodied cognitive science must be possible. Such a cognitive science
would give as much attention to the dynamic development of the subject understanding the theory
is it does to complex mathematical modelling of the ‘objective’ phenomena beyond that subject. It
would focus primarily on increasingly adequate theorisations of adequate investigatory processes
and principles, rather than on general ‘ontological’ or naturalistic claims about systems themselves.
It would develop forms that genuinely challenge representationalism in its damaging sense, by
remaining agnostic about final claims regarding the role of representations, rather than being ‘anti-
representationalist’. It would change its name, or at least show constant awareness of the limiting
effects of using the name it has assumed. It would be much more broadly based in the alternative
discourses it draws on, and thus have a much more diverse and adequate set of models available to
it than at present. In the process it might also assume a much greater relevance for the public, and
be able to contribute much more fully to a practical discourse about the direction of human society.
Chemero, Anthony (2011) Radical Embodied Cognitive Science. MIT Press.
Ellis, Robert M. (2015) Middle Way Philosophy: Omnibus Edition. Lulu.
Ellis, Robert M. (2019) The Buddha’s Middle Way: Experiential Judgement in his Life and Teachings.
Gibson, James (1979) The Ecological Approach to Visual Perception. Houghton-Mifflin.
Johnson, Mark (2007) The Meaning of the Body. University of Chicago Press
Kahneman, Daniel (2011) Thinking Fast and Slow. Penguin.
Kiverstein, Julian & Miller, Mark (2015) ‘The Embodied Brain: Towards a radical embodied cognitive
neuroscience’. Frontiers in Human Neuroscience 9, p.237.
Lakoff, Georg & Johnson, Mark (1999) Philosophy in the Flesh. Basic Books.
McGilchrist, Iain (2009) The Master and his Emissary: The Divided Brain and the Making of the
Western World. Yale University Press.
Full-text available
In this paper I attempt to reflect upon several claims made by Robert M Ellis in his critique of contemporary 'embodied' or 'anti-representational' cognitive science. My main objective will be to clarify his approach and to provide other relevant aspects that will add complexity to his claims and extend his overall view. I will also reflect upon the broader context and scope of the Middle Way approach, understood as a practical means for conducting scientific and philosophical inquiry.
Full-text available
Middle Way Philosophy is not about compromise, but about the avoidance of dogma and the integration of conflicting assumptions. To rely genuinely on experience as our guide, we need to avoid the interpretation of experience through the lense of unnecessary metaphysical dogmas. Very often the dogmas we need to find our way between are ones conventionally assumed in modern thinking. Middle Way Philosophy questions alike the assumptions of scientific naturalism, religious revelation and political absolutism, trying to separate what addresses experience in these doctrines from what is merely assumed. Robert M. Ellis draws on balanced Pyrrhonian scepticism, experience of Buddhist practice, falsificationist philosophy, cognitive psychology, Jungian integration and archetype theory, embodied meaning, and the neuroscience of brain lateralisation to develop a synthetic philosophy that is not limited by the false divisions between arts, religion and science. Seeking a universal philosophy without dependence on metaphysics or tradition, he develops a new account of objectivity based on the adequacy of our judgement rather than on the traditional obsession with a ‘truth’ that is always speculative. This Omnibus Edition includes all four of the volumes of the 'Middle Way Philosophy' series that were previously published separately: 1. The Path of Objectivity (2012), 2. The Integration of Desire (2013), 3. The Integration of Meaning (2013) and 4. The Integration of Belief (2015).
The Middle Way was first taught explicitly by the Buddha. It is the first teaching offered by the Buddha in his first address, and the basis of his practical method in meditation, ethics, and wisdom. It is often mentioned in connection with Buddhist teachings, yet the full case for its importance has not yet been made. This book aims to make that case. The Middle Way can be understood from the Buddha's life and metaphors as well as his teachings, and it is these that are used to provide an accessible way into it. In the traditional story, he moved from Palace to Forest, finally realising that neither offered the whole story. The acceptance of rice-porridge marks the moment of recognition. The Middle Way can also be found in the Buddha's practice throughout the rest of his life: his teaching, his politics, even his death. Well-known similes such as the raft, the lute-strings, the arrow, and the blind people with the elephant also offer a profound source of the Middle Way. They are not just allegories of Buddhist teachings, but relate closely to universal judgement in human experience. This book emphasises a positive case, but also has a critical one. Although it has transmitted the Middle Way, the Buddhist tradition has also often ignored or distorted it. The Middle Way is experiential, authentic and creative, and thus threatening to the power of a tradition that has instead emphasised the Buddha's authority as a source of revelation. That authority is allegedly based on the Buddha's enlightenment, which is often interpreted as abstract, discontinuous and absolute. Too many other Buddhist teachings are routinely interpreted in these absolute terms, when they would be much more helpful and universal interpreted in the terms of the Middle Way. Although it engages fully with Buddhist material, this book shows the Middle Way to be a principle of experiential judgement based on awareness that goes far beyond Buddhism. Because it's about how humans judge things, it's universal, not a Buddhist monopoly. In its final section it offers ten alternative non-Buddhist sources for the Middle Way, many of them recent.
The Meaning of the Body
  • Mark Johnson
Johnson, Mark (2007) The Meaning of the Body. University of Chicago Press Kahneman, Daniel (2011) Thinking Fast and Slow. Penguin.