OPC on “Perception-Action Mutuality Obviates Mental Construction” by Martin Fultot et al.

Radical Enactivism and Ecological Psychology: friends or foes?

Karim Zahidi
Departement Wijsbegeerte, Universiteit Antwerpen, Belgium, [email protected]
Jan Van Eemeren
Departement Wijsbegeerte, Universiteit Antwerpen, Belgium, [email protected]

Upshot: We examine whether there are any irreducible contradictions between ecological psychology and
radical enactivism. We concentrate on two points of contention between the two approaches: the
(ir)relevance of neural structures in understanding perception and the use of semantically loaded concepts in
theorizing about perception.

Acknowledgements: The first author is supported through the FWO-NWO project G0C7315N
Getting real about words and numbers. An enactive approach to language and mathematics and
the FWO-project G048714N Offline cognition. The second author is supported through DOCPRO3
project Perceiving affordances in natural, social and moral environments. A study of the concept of
affordance, and its explanatory value of the BOF Research Fund of the University of Antwerpen.

1. In the target article, Fultot et al. criticize one form of enactivism, i.e. autopoetic enactivism, on
two accounts. Firstly, they find fault with autopoeitic enactivists’ insistence on the constructive
activity of the perceiving organisms. And, secondly they object to its focus on neural structures in
trying to understand perception. The aim of our commentary is, however not to engage directly with
this critique – with many points of which we agree, although perhaps for different reasons (see e.g.
Zahidi (2014)) - but rather to explore in how far this critique drives a wedge between ecological
psychology (EP hereafter) on the one hand and radical enactivism (REC hereafter) on the other
hand. In order to probe what separates these two approaches we will focus on two issues. First
there is the question of the explanandum phenomenon (i.e. what kind of phenomena are the target
explananda) and secondly the issue of meaning/information in ecological psychology.
2. Before we engage in this project, we’ll first give a very very brief summary of the basic tenets of
radical enactivism. Like EP, REC eschews all talk of mental (i.e. internal representations) in the
explanation of basic cognitive capacities such as perception. Like EP, REC rejects the dominant
representational- cum – computational paradigm. To explain cognitive capacities REC relies on
two theses, namely the Embodiment Thesis and the Developmental-Explanatory Thesis. The first
thesis essentially says that all forms of basic cognition are “concrete spatio-temporally extended
patterns of dynamic interaction between organisms and their environment” (Hutto & Myin 2013: 5).

1
Prime examples of such patterns of dynamic interaction between organism and environment in
perception are the so-called sensorimotor contingencies. Sensorimotor contingencies are the
lawful ways in which sensory stimulation changes as the organism moves or acts in the
environment. In perception the organism exploits these sensorimotor contingencies to find its way
around the environment. The second thesis concerns the explanation of these interactions, viz.
these interactions are “grounded in, shaped by, and explained by, and explained by nothing more,
or other, than the history of an organism’s previous interactions” (Hutto & Myin 2013: 8).
Importantly, the history to which the second thesis refers encompasses both the ontogenetic
development of an organism as well as its phylogenetic development. Hence, the relevant time-
scales for explaining cognitive abilities need not be restricted to the life-span of a single organism.
This short description of the basic tenets of REC shows that its approach and that of ecological
psychology share important emphases in their conception of perception and cognition. However,
despite this kinship there appear to be major differences as well.
3. The first such difference is connected with the importance of the neural machinery in explaining
perception. As already noted both REC and EP put great emphasis on the interaction between
organism and environment. However, from a REC point of view that interaction is always embodied
interaction. It is the embodied nature of the organism that grounds the kind of interactions that are
possible. Note however that the “body” referred to in the Embodiment Thesis is not our common-
sense notion of body but rather the body engaged in non-linear and far-reaching sensorimotor
interactions while engaging with certain salient features of the environment. In this interaction the
physical body ànd brain areas are in play. One way to grasp the importance of the neural
machinery in this interaction is by reflecting on the role of learning in the development of cognitive
capacities. For radical enactivists even the basic cognitive skills have an element of learning in
them. All interactions are made possible by previous interactions between the organism and its
environment (and recall that the history of interactions may extend beyond the organism’s life-
time). These interactions can and do shape the ways the organism will interact with the
environment in the future. How is this possible? The answer is that the interactions affect and
change the neural and non-neural body of the organism. In the neural machinery the past
experience is sedimented, and in order to explain current interactions this “sediment” is called
upon. Examples of explanations of perceptual phenomena in which an appeal to neural structure is
crucial concern for example the experience of subjects wearing right-left reversal goggles (see e.g.
Hurley 1998, Degenaar 2013). Note however that invoking neural facts in the explanation of
perceptual phenomena does not imply any kind of neurocentrism. Radical enactive explanations
do not claim that neural facts in and by themselves are sufficient to explain perceptual phenomena.
Nor does the reliance on neural facts about organisms necessarily lead to the acceptance of an
“inner realm” of mental representations. Indeed, the neural structures called upon on in radical
enactive explanations do not play any representational role whatsoever. There is thus no danger
that by invoking neural structure in enactivist explanation, representations slip back in by the back

2
door.
4. Radical enactivism’s acceptance of the relevance of neural structures in explaining perceptual
phenomena seems to put it squarely in opposition to EP, since the target article repeatedly
stresses that neural structure is irrelevant for the understanding of perception. However, some care
needs to be taken here. For the opposition between radical enactivism and ecological psychology
to be real and not merely apparent it has to be the case that REC invokes neural structure in its
explanation of perceptual phenomena for which ecological psychology claims neural irrelevancy.
To put the same point slightly different: if radical enactivism is mainly interested in a set of
perceptual phenomena X and ecological psychology is mainly interested in another non-
overlapping set of perceptual phenomena Y and if ecological psychology is silent on X; then radical
enactivism’s claim that neural structure is important for understanding “perception” (i.e. perceptual
phenomena from set X) need not be in contradiction with ecological psychology’s claim that neural
structure is irrelevant for understanding “perception” (i.e. perceptual phenomena from set Y).
Deciding whether there is any opposition between EP and REC depends on an answer to the
question whether they share the same target explananda. Answering that question is less than
straight forward however. Perception so Fultot et al. assure us is a capacity not only to be found in
all living things (humans, animals, plants, bacteria) but also in non-living matter. If the net of
perception is so widely cast and is constituted by certain types of behavior by which organisms
adapt to conditions in their environment, and if it is its aim to find in all these different kinds of
behavior a kind of common structure as the essence of perception thus conceived, then it is
unsurprising that perception does not depend on neural structure. One may wonder whether trying
to find a communal structure to such a diverse set of behaviors is going to result in something
genuinely explanatory. Whatever the answer to that question is, it should be clear that there
remains a substantial residue of perceptual phenomena which remains unexplained or
unthematized in ecological psychology. In particular, all perceptual phenomena where the current
interaction is shaped not only by how that interaction unfolds in the here and now but also by
previous engagements (i.e. where some form of “learning” is involved) seem to be out of bounds.
Fultot et al. do acknowledge that the history of the interactions do play a role, but that theme is
never fully addressed and remains unclear. Furthermore, it sits uneasy how they develop their
ideas further. Indeed, if the common core of perceptual behaviors is the explanandum of ecological
psychology then, since e.g. plants do not exhibit learning behavior, all perceptual behavior that
relies on learning falls out of explanatory purview of ecological psychology (at least as presented
by Fultot et al.). Furthermore, the question of perceptual experience or perceptual consciousness
seems to be by-passed completely. Arguably, these are the phenomena for which REC invokes
neural structure to explain them. So either, ecological psychology is not concerned with these
perceptual phenomena in which case ecological psychology and REC are potentially
complementary. Or, ecological psychology takes these phenomena to be within its non-neural
explanatory purview, in which case it is in contradiction with REC.

3
5. We now turn to what is potentially a second point of contention between REC and EP. This
issue revolves around the issue of meaning and information in ecological psychology. For REC
semantic notions such as meaning and information (semantically conceived) have no application to
non-linguistic creatures: meaning and semantic information only become relevant for creatures
who partake in socio-cultural practices such as linguistic practices. REC’s rejection of semantic
information (outside the domain of the socio-cultural) is based on good naturalistic reasons: up
until now nobody has been able to solve the Hard Problem of Content, i.e. to give a naturalistically
sound theory of semantic content (Hutto & Myin 2013: chapter 4). However, the language of EP is
infused with the notions of information and meaning, hence this seems to point to an irreducible
tension between the REC approach and the EP approach. However, whether the frequent use of
terms such as meaning” and “information” in EP drives a wedge between REC and EP depends on
the meaning these terms assume in EP. Starting with the term “meaning” in EP, in paragraph §9
Fultot et al. define the term meaning as the behavioral implications of objects”. Note that under this
definition the term meaning does not have any semantic connotations and hence is completely
acceptable to REC. Furthermore, note that the notion of meaning is one that is used by a third-
person to describe the interaction between organism and environment. There is, at this point no
reason to assume that the meaning of objects is in some way perceived by the organism. But EP
makes the further assumption that the meaning is actually perceived through the perception of
affordances. The way affordances are perceived is through the invariants in the optical array. The
latter (as well as the ambient light at a convergence point) is said to contain (as in “content”)
information about objects and their lay-out in the environment. Although information as conceived
by EP differs radically from the information-theoretic notion of information - Shannon-information –
in that it is not a commodity that can be stored, sent or processed, apart from this negative
qualification of information it is difficult to see what exactly is meant. The use of the word “about”
seems to indicate that information is a semantic notion and thus that information has semantic
content. It then also follows that their notion of meaning is intended to have semantic import. If so
then they are in fact operating with two different notions of meaning –one semantic and one non-
semantic, behavioral notion. That Fultot et al. take this semantic road is far from clear, however,
we suspect they do (cf. their complaint in § 51 that REC “doesn’t even tolerate content”). However,
given the Hard Problem of Content, the semantic road is a dead-end. On a positive side, one can
avoid going down that route without forsaking the explanatory strategies of EP. One way to do this
is to interpret EP information as co-variance of worldly lay-out in relation to a convergence point
and optical structure at that certain point. Meaning as a behavioral concept can than be
maintained. Rather than talking about the “perception of meaning” we should refer to the fact that
the organism has, through a history of interaction and reinforcement learning, learned to respond
to the structure of the optical array in appropriate ways. In other words, while the meaning of the
objects is something objective (which we can, from a third-person perspective, describe), that
meaning is not given in perception to the organism, it is something the organism enacts.

4
6. We believe that taking this road might save some forms of EP from some troublesome worries.
For example, in order to avoid falling back on some “impoverished stimulus”-type of psychology,
some ecological psychologists like Michael Turvey have swayed in the opposite direction,
postulating ultra-information-richness of the media (light, air, ...) around us. (cf. A. Chemero 2009:
106-107). Here, the structure of the optical array determines completely the (behavioral) meaning
of the object: there is no ambiguity possible. However, Rob Withagen and Anthony Chemero
(2009) have challenged this view on evolutionary grounds. If we think of the information in the
optical array not in a semantic way but as a certain correlation, which can be very strong or only
moderate, between that structure and the meaning of the object then we can move away from the
idea that the array-structure determines (or as ecological psychologists say “specifies”) meaning
(i.e. the correlation between the two is one). This is not to say that the stimulus is impoverished
and needs to be worked on to be enriched. It merely allows for information available through the
optic array to not fully specify affordances. That doesn't imply a need for it to be supplemented
through some type of mental construction. The organism can find out the affordances of the object
by interacting with it. When the available (optic) invariant co-varies only moderately with the
affordances, so to say, the history of interaction, c.q. learning history, explains the organism's
improved adjustment to its situation. No semantic information or enrichment is implied, because
the sedimentation of the historical interactions, as we called it, is characterized as a bodily change,
and not as an informational process.
7. To conclude, let us try to answer our question in the title with respect to the two point discussed
in this paper. Firstly, with respect to the question of neural structures, we suspect that, despite the
pronouncements made in the target article we do believe that neural structure is important in the
explanation of at least some perceptual phenomena. That Fultot et al. deny that, so we have
argued, is due to the fact that they are not primarily concerned with these aspects. If this
interpretation is correct, then EP and REC are complementary. With respect to the use of semantic
notions, we have argued that these can be banned from EP without loss of explanatory power. If
our interpretation is correct, EP and REC are not incompatible.

References
Chemero, A. (2009) Radical Embodied Cognitive Science. Cambridge (MA), MIT Press.
Degenaar, J. (2013) Through the inverting glass: first-person observations on spatial vision and
imagery. Phenomenology and the Cognitive Sciences 13(2): 373 – 393.
Gibson, J. J. (2015) The Ecological Approach to Visual Perception: Classic Edition. New York,
Psychology Press.
Hurley, S. (1998) Neuropsychology against the Input-Output picture. In: Hurley, S. Consciousness
in action. Harvard University Press, Cambridge(MA): 338 – 400.
Hutto, D. & Myin, E. (2013). Radicalizing enactivism. Basic Minds without Content. MIT Press,
Cambridge (MA).

5
Withagen, R. & Chemero, A. (2009) Naturalizing Perception: Developing the Gibsonian Approach
to Perception along Evolutionary Lines. Theory & Psychology 19(3): 363 – 389.
Zahidi, K. (2014) Non-representational cognitive science and realism. Phenomenology and the
Cognitive Sciences 13(3): 461 – 475.

Sciences