Content uploaded by Alkis Gounaris
Author content
All content in this area was uploaded by Alkis Gounaris on Jul 15, 2020
Content may be subject to copyright.
Presentation for the 94th Joint Session of the Mind Association and the
Aristotelian Society | University of Kent, 10 -12 July 2020 (Open Session)1
Why do we need a
Unified Theory of Embodied Cognition?
Dr. Alkis Gounaris
National and Kapodistrian University of Athens (NKUA)
Abstract: The range, diversity and the different metaphysical assumptions of
Embodied Cognition (EC) hypotheses that have been formulated in recent years, a)
do not allow for their pragmatic, problem-solving oriented adoption by Cognitive
Science and (b) favor their fragmentary use by various scientific and practical fields,
resulting in misunderstandings concerning their content and validity. Through a brief
overview, this presentation aims to highlight the reasons why research towards a
unified theory of embodied cognition can contribute to solving old and modern
problems regarding Human Cognition and Artificial Intelligence.
Sub Disciplinary Area: Philosophy of Cognitive Science, Philosophy of Artificial
Intelligence
Keywords: Embodied Cognition, Embodied Mind
In this presentation I will attempt to answer the question of why we need a Unified
Theory of Embodied Cognition (UTEC), or why we need to untangle the thread of all
the hypotheses made over the last twenty years, regarding the role of the body and
the world in the process of understanding and explaining mental phenomena.
1 Gounaris, A. (2020). Why do we need a Unified Theory of Embodied Cognition? Presentation for the
94th Joint Session of the Mind Association and the Aristotelian Society. University of Kent, Online
Open Session. Retrieved July 15th, 2020 from https://alkisgounaris.gr/en/research/why-do-we-need-
a-unified-theory-of-embodied-cognition/
The 94th Joint Session of the Mind Association and the Aristotelian Society |
University of Kent, 10 -12 July 2020 (Open Session)
COPYRIGHT © 2020 ALKIS GOUNARIS | alkisg@philosophy.uoa.gr | PAGE 2 FROM 6
I will begin by outlining how we ended up talking about such a wide range of EC
hypotheses, in what ways they differ and what their levels of study and comparison
are, and I will conclude with (1) how clarifying and delineating them can lead us to a
unified theory and (2) why such a theory contributes to the evolution of Cognitive
Science and beyond.
It has been commonly known, ever since the early days of Standard Cognitive
Science (SCS) in the 1970s, that there have been many and important objections to
the philosophical assertions or the “dogma” of the computational and
representational theory of mind (Gomila & Calvo, 2008:11) mainly to: a) assuming
that Human Cognition (HC) is a computational process of symbolic representations
and b) assuming the existence of internal representations of the world inside the
brain. In the dominant model of SCS, cognitive functions have a definite beginning,
with data input into the brain, and a definite end with output from it, essentially
limiting research to processes “inside” the brain (Shapiro, 2011:26-27). From this
solipsistic view, the body and the world are absent.
Soon, the hypothesis that the environment in which a cognitive being operates is
nothing more than data entered into the system to be processed, without taking into
account the world itself, began to create problems in HC and AI research programs
(Dreyfus, 1972, 1992, 2007). For example, SCS could not provide an answer to the
frame problem (Wheeler, 2005; Zambak, 2013), or, the referential nature of
representations limited machine learning (Abrahamsen & Bechtel 2012:28).
Accordingly, the formal symbolic structure of language was unable to provide an
adequate answer to the grounding problem (Vogt, 2006:176-209), and then
questions regarding the nature of data and information itself occurred, as well as
regarding the way meaning is derived (Shapiro, 2007:339).
Subsequently, applied research in AI faced paradoxes, such as that of Moravec,
concerning the way complex cognitive computing processes (like chess) require
comparatively little computational resources, while simple cognitive tasks (such as
descending a staircase) require enormous amounts.
At the same time, in recent years, numerous findings of applied research in
education have demonstrated the important role of the body and the environment in
achieving cognitive goals such as learning, memorizing, language comprehension
etc. (Skulmowski, 2018; Kosmas, 2019). More research has also demonstrated the
anthropomorphic nature of cognition and the role of mirror neurons in monitoring
physical movements and behaviors (Gazzola & Rizzolatti, 2007).
From the 1990s and on, all this has led the field of SCS research to move “in depth”
(see cognitive neuroscience) and “outwards” (see EC) (Abrahamsen & Bechtel,
2012:29), resulting in new research programs, according to which cognition is not a
process that happens inside an isolated “box” but is something more complex, which
happens because an organism lives, moves and acts within an environment.
Other programs moving in this direction and often combined with EC research were:
(a) neural networks, which adopted distributed and coded patterns of
representations, carrying information in their internal properties (Clark, 1993:19) and
The 94th Joint Session of the Mind Association and the Aristotelian Society |
University of Kent, 10 -12 July 2020 (Open Session)
COPYRIGHT © 2020 ALKIS GOUNARIS | alkisg@philosophy.uoa.gr | PAGE 3 FROM 6
(b) dynamic systems theory, which introduced time variability as a determinant factor
in the cognitive process of the system (Van Gelder,1998).
A common assumption of all these hypotheses is that cognition is not an exclusive
function of a computational mind “passive” to external stimuli, but, in fact, the bodily
functions and the interactions of the cognitive being with its environment, contribute
directly and decisively to mental processes (Leitan & Chaffey, 2014).
However, despite the above common assumption, EC hypotheses are extremely
heterogeneous and diverse (Wilson, 2002). This may partly be due to the fact that
many of the research projects that have featured these hypotheses over the years
originate, either independently or combined, from different scientific disciplines such
as Philosophy, Psychology, Linguistics, Neuroscience, Artificial Intelligence, Robotics
etc. and have possibly coupled EC with other metaphysical theories such as the
emergence of properties of complex systems (Kuniyoshi et al., 2007) and others.
These divergences stemmed mainly from philosophical and conceptual assumptions
and resulted in significant differences in both their technical terminology and the
meaning of the common terms they use.
Influenced by theoretical traditions such as Phenomenology, American Naturalism,
the Ecological Theory of Visual Perception (Gibson, 1979), Gestalt psychology,
Cognitive Development Theory and others, EC theories have attempted to exempt
themselves (more or less) from the standard metaphysical or epistemological
assumptions of SCS, such as the “Cartesian theater”, the autonomy and causality of
mind, the formal syntactic structure of mental content, the meaningful data
hypothesis and others.
At the same time they made new claims that could be summarized according to
Shapiro (ibid:4) as follows: a) the bodily properties of an organism delimit or constrain
the concepts that the organism can acquire, b) the interaction of an organism with its
environment replaces the need for a representational process, c) the body and the
world play a constitutive, rather than simply a causal role in the cognitive process.
Such a classification, although extremely useful in understanding the general content
of the EC spectrum claims, is not “competent” enough to lead in the direction of
unification.
The need for a unified EC theory has been formulated for about a decade (Barsalou,
2008; Gomila & Calvo, 2008), while efforts have been made to classify individual
theories based on their ontological characteristics (Gallagher, 2011). However,
research in this direction has not been very fruitful, mainly because of the
researchers' attempt to somehow adapt a unified theory in the framework of their
own assertions.
Most will agree with Shapiro's (2011:3) claim that a UTEC, or at least a unified field
of research, will make it possible to compare and evaluate two fields that will then be
clearly defined –UTEC and SCS– rather than comparing a series of heterogeneous
research programs, as found in EC today, with a compact field such as SCS. This
comparison would essentially reveal whether these two approaches: a) offer
The 94th Joint Session of the Mind Association and the Aristotelian Society |
University of Kent, 10 -12 July 2020 (Open Session)
COPYRIGHT © 2020 ALKIS GOUNARIS | alkisg@philosophy.uoa.gr | PAGE 4 FROM 6
competing explanations for the same phenomena, b) aim at a different explanatory
model overall, or c) develop different yet compatible views through which they
understand the cognitive process.
This is certainly a good and useful reason why we should step up and intensify
research in this direction, but in my opinion it is not the main reason. The main
reasons are that through a UTEC we will be able a) to preserve and b) to utilize in an
interdisciplinary manner the wealth of EC ideas such as the notion of “affordances”
(Gibson, 1977), the systemic theories of meaning and its independency from
representations, as well as the possibility to eliminate the traditional assumptions of
“folk psychology” and its inherited terminology that lead us to dead ends.
Clearly, the main beneficiary of such an endeavor will be the theory of Next Day
Cognitive Science (NDCS), but those who are involved in applied research such as
neuroscience as well as robotics and artificial intelligence research programs, will
also benefit greatly, since a UTEC can become a tool for shaping research
hypotheses and can make a substantial contribution to the progress of their research
objectives.
Incidentally, as various EC theories are selectively, and rather superficially, adopted
by fields of applied science such as clinical psychology, education, sports science,
modern marketing and advertising (Leitan & Chaffey, 2014) among others, there will
be a clarification of the distorted assertions that create a veneer of science, and
above all, the authority of the sciences of the mind, cognition and intelligence will be
safeguarded.
If we consider an NDCS oriented towards problem solving as a commonly accepted
demand, we should be willing to move on to a cash value - pragmatic approach of a
UTEC, so that the differences in the spectrum of EC hypotheses will be addressed
at a practical level. An important first step in such a venture may be to classify
existing EC hypotheses according to the area and the objectives of their study. For
example:
[a] at a level of compatibility or not with SCS, [b] at a level of scientific fields of
application, [c] according to characteristics mainly related to the philosophical origins,
[d] at a conceptual level, [e] at a terminology level, [f] at a methodology level, etc.
Finally, an additional (visionary) reason that a UTEC can prove to be extremely
useful is the possibility of reaping a significant payoff: The unification of EC theories
is likely to pave the way for the unification of the sciences of the mind, cognition and
intelligence and possibly explain the precise role that the body, the environment,
language and other systemic factors play in mental processes.
Such an understanding may help to conceptualize and formulate new “bridging laws”
between the various description levels among research fields, from the level of the
cognitive organism acting in the world and interacting with others, to the level of the
basic elements (i.e. neurons, molecules, etc) or, respectively for AI, from the level of
the agent coping in the world to the basic level of data.
The 94th Joint Session of the Mind Association and the Aristotelian Society |
University of Kent, 10 -12 July 2020 (Open Session)
COPYRIGHT © 2020 ALKIS GOUNARIS | alkisg@philosophy.uoa.gr | PAGE 5 FROM 6
Selected Bibliography:
Abrahamsen, Α. Bechtel. W. (2012). History and Core Themes. In: Frankish, K. -
Ramsey, W. The Cambridge Handbook of Cognitive Science. Cambridge University
Press
Barsalou, L.W. (2008). Grounded Cognition. Annual Review of Psychology. 59,
pp.617-645. http://dx.doi.org/10.1146/annurev.psych.59.103006.093639
Clark, A. (1993). Associative Engines: Connectionism, Concepts, and
Representational Change. Cambridge, MA: MIT Press.
Dreyfus, H. (1992). What Computers Still Can't Do: A Critique of Artificial Reason.
Cambridge, Massachusetts: MIT Press
Dreyfus, H. (2007). Why Heideggerian AI Failed and How Fixing It Would Require
Making it More Heideggerian. Philosophical Psychology. Routledge. Vol.20. No.2,
pp.247-268
Gallagher, S. (2011). Interpretations of embodied cognition. In W. Tschacher and C.
Bergomi (Eds.). The Implications of Embodiment: Cognition and Communication.
pp.59-70. United Kingdom: Imprint Academic
Gazzola, V. Rizzolatti, G., et, al. (2007). The anthropomorphic brain: The mirror
neuron system responds to human and robotic actions. DOI:
10.1016/j.neuroimage.2007.02.003
Gibson, J. J. (1977). The theory of affordances. In Shaw, R. and Bransford, J. (eds.).
Perceiving, acting, and knowing: Toward an ecological psychology. NJ: Erlbaum,
pp.67-82
Gibson, J. J. (1979). The ecological approach to visual perception. Boston, MA:
Houghton Mifflin
Gomila, T. & Calvo, P. (2008). Directions for an Embodied Cognitive Science:
Toward an Integrated Approach. In Calvo P. & Gomila A. (Eds.). Handbook of
Cognitive Science: An Embodied Approach. San Diego, Elsevier
Kosmas, P. (2019), Implementing embodied learning in the classroom: effects on
children’s memory and language skills. DOI: 10.1080/09523987.2018.1547948
Kuniyoshi, Y. et al. (2007). Emergence and development of embodied cognition: A
constructivist approach using robots. Progress in brain research. 164:425-45. DOI:
10.1016/S0079-6123(07)64023-0
Leitan, N. & Chaffey, L. (2014). Embodied cognition and its applications. Sensoria: A
Journal of Mind, Brain & Culture. DOI: 10.7790/sa.v10i1.384
Shapiro, L. (2007). The Embodied Cognition Research Programme. Philosophy
Compass. 2/2 (2007): pp.338-346, 10.1111/j.1747-9991.2007.00064.x
The 94th Joint Session of the Mind Association and the Aristotelian Society |
University of Kent, 10 -12 July 2020 (Open Session)
COPYRIGHT © 2020 ALKIS GOUNARIS | alkisg@philosophy.uoa.gr | PAGE 6 FROM 6
Shapiro, L. (2011). Embodied Cognition. New York: Routledge
Skulmowski, Α. (2018). Embodied learning: introducing a taxonomy based on bodily
engagement and task integration. DOI: 10.1186/s41235-018-0092-9
Van Gelder, T. et al. (1998). The dynamical hypothesis in cognitive science.
Behavioral and Brain Sciences. 21, pp.615-665
Vogt, P. (2006). Language evolution and robotics: Issues on symbol grounding and
language acquisition. In: Loula, A. Gudwin, R. Queiroz, J. (Eds.). Artificial Cognition
Systems. Hershey, PA.: Idea Group. pp. 176-209
Wilson, M. (2002). Six views of embodied cognition. Psychonomic Bulletin and
Review. 9: pp.625-636
Wheeler, Μ. (2005). Reconstructing the Cognitive World: The Next Step. Cambridge,
MA: A Bradford Book. The MIT Press
Zambak, A. (2013). The Frame Problem. Autonomy Approach versus Designer
Approach. In V.C. Müller (ed.): Philosophy and Theory of Artificial Intelligence.
Springer
