See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225947374

Dimensions of mind

Article in Phenomenology and the Cognitive Sciences · December 2010

DOI: 10.1007/s11097-010-9186-7

CITATIONS READS

59 1,686

1 author:

Richard Menary
Macquarie University
47 PUBLICATIONS 2,492 CITATIONS

SEE PROFILE

All content following this page was uploaded by Richard Menary on 15 June 2015.

The user has requested enhancement of the downloaded file.

Phenom Cogn Sci (2010) 9:561–578
DOI 10.1007/s11097-010-9186-7

Dimensions of mind

Richard Menary

Published online: 10 November 2010

# Springer Science+Business Media B.V. 2010

Abstract In their papers for this issue, Sterelny and Sutton provide a dimensional
analysis of some of the ways in which mental and cognitive activities take place in
the world. I add two further dimensions, a dimension of manipulation and of
transformation. I also discuss the explanatory dimensions that we might use to
explain these cases.

Keywords Cognition . Integration . Manipulation thesis . Cognitive transformation .

Body schema . Epistemic action . Cognitive practices . Cognitive niche

In this issue, John Sutton and Kim Sterelny provide a dimensional analysis of some
of the ways in which mental and cognitive activities take place in the world. For
example in discussing the ways we interact with artefacts and representations in the
local cognitive niche, we might move along a dimension from trusting a notebook
implicitly to distrust of an environmental resource, from relationships that are
enduring to those that are fleeting, from individual interactions to collective
interactions and information sources that range from the easily transferable to the
individually entrenched.1 I will add some dimensions of my own: firstly along an
axis of manipulation from direct bodily manipulations that alter the informational
and physical structure of the environment, to manipulations of public representations
in the local environment. There is also a dimension of cognitive transformation
ranging from the transformation of body schemas for tool use to the transformation
of representational and cognitive capacities.

1
Thanks to John Sutton for clarifying these dimensions by personal communication.
The research for this paper was conducted whilst on research leave awarded by the Faculty of Arts. It is
supported by the Australian Research Council Discovery grant: Embodied Virtues and Expertise. Thanks
to John Sutton for comments.
R. Menary (*)
Department of Philosophy, The University of Wollongong, New South Wales, Australia
e-mail: rmenary@uow.edu.au
562 R. Menary

How should we understand these variations in dimensions of mind? One way is

through looking at the variations in the world in terms of cases, as above. However,
there are also theoretical dimensions at work, in terms of how we explain the
variations in dimensional cases as described above. So, for example, one can take a
weakly embodied view that cognition is embodied only by the brain and not the
wider body. One can also take a weakly embedded view which is that brains are
embodied and embedded in an environment which is the location of inputs and
outputs. In which case one would need to explain, case-by-case, how the brain is
doing all the cognitive work and how the environment is simply the location of
inputs and outputs.
A moderate version of embodiment and embedding might take cognition to be at
least dependent upon the activity of the body and the embedding environment, such
that some of the relevant explanatory factors will be located beyond the bounds of
the skull in a way that is richer and more detailed than a simple input–output story.2
A strongly embodied and embedded position would view cognition as, at least
sometimes, involving more than neural activity but also bodily activity in the
embedding environment. Such a position would need to show how cognitive
processes involve bodily manipulations and other forms of interaction with the
environment.

Embodied mind

& Embodied mind weak: the brain is embodied (internalism/individualism)

& Embodied mind moderate: some of our mental and cognitive processes and
states depend upon our non-neural body
& Embodied mind strong: some of our mental and cognitive processes and states
involve processes of the body acting in and on the environment (enactivism)

Embedded mind

& Embedded mind weak: the inputs to and outputs from mental and cognitive
processes and states are found in the environment (internalism/individualism)
& Embedded mind moderate: mental and cognitive processes and states are
scaffolded by/depend upon the environment.
& Embedded mind strong: mental and cognitive processes and states are integrated
with states and processes found in the environment (extended mind)
Depending upon the case, we might move back and forth across these explanatory
dimensions. Some examples might call for a weakly embodied and embedded
explanation and others for a strongly embodied and embedded explanation.
However, there will of course be those who deny that we can ever successfully
provide a strongly embodied and embedded explanation of cognition (Adams and
Aizawa 2008; Rupert 2009). Then there are those who would be likely to claim that

2
Sterelny (this issue) holds a moderate position.
Dimensions of mind 563

cognitive explanations are mostly moderate and only rarely of the strong variety. For
Sterelny (2010), this turns out to be the case because of the very rare conditions
under which a strongly embodied or embedded explanation would become active.
We are likely to find that certain dimensions go together; for example, it is unlikely
that a position located on the strongly embedded dimension will not also be located
on the strongly embodied dimension. This is mainly because most proponents of
the extended mind take mental states and processes to be composed of
environmental as well as brain and bodily states and processes—hence, the oft-
stated slogan, the mind is extended or distributed across brain, body and
environment. Similarly, individualists who think that the mind is ‘constituted’ or
‘realized’ only by the brain will also, sometimes, hold to a weak embedding—we
sometimes lean on or use props and tools to supplement our neurally realised
cognitive capacities.
It should also be noted that some positions, by occupying several dimensions at
once, absorb examples of the more restricted dimensions. Thus, strongly embedded
views can quite happily allow that there are examples of moderate and even weak
embedding. The question that holders of such views must address, and which is
raised by Sterelny 2010, is: how often is the strongly embedded dimension
occupied? My answer is: often, but only if we correctly understand what is meant by
strong embedding.
I understand strong embedding in terms of integration (Menary 2006, 2007a, b,
2009, 2010). One way to understand integration is as follows: cognition is the
coordination of bodily processes of the organism with salient features of the
environment, often created or maintained by the organism. A coordinated process
allows the organism to perform cognitive tasks that it otherwise would be unable to;
or allows it to perform tasks in a way that is distinctively different and is an
improvement upon the way that the organism performs those tasks via neural
processes alone. Integrationists think that some cognitive processes turn out to be
coordinated. It is important to note that integrationists are not committed to the view
that artefacts and tools are themselves cognitive or mental, nor that a simple causal
interaction between two states X and Y makes X part of Y (see Adams and Aizawa
2008).
The key to understanding coordinated processes is the manipulation thesis
(Rowlands 1999; Menary 2007a). We create, maintain and manipulate cognitive
niches (see Sterelny 2010) and we can give a causal or co-ordination dynamics style
explanation of these manipulations, but that is just to establish the physicalist
credentials of the explanations. There is no magic at work here (see the muggle
constraint in Wheeler 2005). These explanations must be located within the normative,
social and semantic structure of the cognitive niche. Cognitive manipulations are
bodily engagements with the niche that are regulated by norms. The norms and social
systems and structures of the environment regulate the coordination in two ways:
firstly by transformation of our basic cognitive capacities and secondly by
restructuring the informational and physical structure of the environment.
In the next section, I provide examples of the dimension of manipulation (and of
the structuring of the informational and physical environment); in the second, of
transformation. During the discussion, I locate the examples on the theoretical/
explanatory dimensions of strength of embodiment and embedding.
564 R. Menary

The manipulation thesis

A strongly embedded formulation of the manipulation thesis is due to Mark

Rowlands:
“[C]ognitive processes are not located exclusively in the skin of cognising
organisms because such processes are, in part, made up of physical or bodily
manipulation of structures in the environments of such organisms.” (Rowlands
1999, 23)
This early formulation of the manipulation thesis is insufficient on its own. The
manipulation thesis concerns our embodied engagements with the world, but it is not
simply a causal relation, bodily manipulations are also normative—they are
embodied practices developed through enculturation, learning and training (in
ontogeny). I will outline four different classes3 of bodily manipulation of the
environment and look at examples of three4 of them.
1. Biological coupling—organism–environment relations.
2. Epistemic actions—directly manipulating the environment to make the
completion of a cognitive task simpler.
3. Self-correcting actions—the use of language and props to direct and structure
practical actions in completing tasks.
4. Cognitive practices—manipulations of external representational and notational
systems regulated by cognitive norms.
Examples of biological coupling run from non-cognitive cases such as phonotaxis
in crickets (Webb 1994) and bee dances (Millikan 1993, 2004) up to sensorimotor
contingencies (O’Regan and Noë 2001) and animate vision (Ballard 1991).
Kirsh and Maglio (1994) have dubbed the second class of manipulations
“epistemic actions”. Epistemic actions are direct manipulations of the environment
that alter the informational state of the environment, resulting in a better state in a
problem solving/planning task. Epistemic actions are performed as part of the
cognitive processing of a problem and are to be contrasted with an internalist
account of problem solving where all of the processing takes place in the brain. I
shall be looking at epistemic actions in detail in the next sub-section.
An example of a self-correcting action is the role of spoken language in
structuring an activity. In these kinds of cases, we use speech as a corrective tool.
Self-correcting actions are to be distinguished from epistemic actions because they
do not involve a direct physical manipulation of the environment. It might be
thought that this is not really a case of manipulation, for we might ask what exactly
is being manipulated? The sense of ‘manipulation’ at work in these cases should not
be understood exclusively as a manual operation using the hands. It might involve a
saccade or a re-orientation of the body to better be able to see. Similarly we can
think of the creation of speech and gesture, or the integration of speech and action

3
The classes should be thought of as being broad with shifting and overlapping boundaries, rather than
too fixed.
4
I will focus on the final three classes of manipulation, because they better suit the argument of this paper.
For a detailed account of the first class, see Menary 2007a.
Dimensions of mind 565

as, broadly speaking, a manipulation of the environment. In the self-corrective cases,

we create linguistic structures which control and direct future action. We might
narrate the sequence of actions to ourselves in our heads, or out loud or we might
write them down. The linguistic sequencing might be exclusively for our own use, or
it might be designed for use by others. We might structure a sequence of actions
collectively as well as individually, by jointly creating linguistic control structures. I
shall look at a linguistic example of a self-correcting action.
The classic example of a cognitive practice is McClelland et al.’s (1986) example
of using pen and paper to complete a mathematical algorithm. Performing long
multiplication involves mastery over a notational system. Humans very often create
and manipulate external representational vehicles to complete a cognitive task. In
doing so, they are carrying out a cognitive practice which is governed by its own
norms—which I dub “cognitive norms”. The norms are cognitive because they are
aimed at completing cognitive tasks. I give an example of writing as a cognitive
practice.
The dimension of manipulation therefore runs from our direct biological
interactions with the environment, to actions that alter the informational structure
of the environment, through to the construction and manipulation of public systems
of representation, including natural languages and notational systems.
I will now look at examples of epistemic actions, self-corrective practices and
cognitive practices in more detail.

Epistemic action

Epistemic actions involve a tight coordination between body and world. They are
part of a coordinated process spanning environment, body and brain, which, when
taken together, provide the processing required for the completion of a cognitive
task. I will argue for this strongly embodied and embedded account of epistemic
actions against a weak internalist account, by reference to the work of Kirsch and
Maglio on epistemic actions.
Kirsh and Maglio (1994), henceforth K&M, are interested in a class of actions
which they call epistemic actions, which make mental computation (processing)
“easier, faster, or more reliable.” (1994, 513) These are physical actions which the
agent performs in order to alter their own computational state. They wish to shift the
emphasis from internal planning to performing actions which simplify the processing
required to achieve that goal. Instead of thinking of an agent internally processing
various plans and choosing an action which will most efficiently bring about the
desired goal, K&M want us to think of agents directly manipulating their
environment so as to reduce the need for such internal processing.5 Epistemic
actions alter the physical and informational structure of the environment, in doing so
they are taking the cognitive agent a step closer to a solution; they are an integral
part of the agent’s processing of the problem.
An example of such a task can be found in their studies of expert Tetris players. In
the game, falling geometric shapes, or “zoids”, have to be directed to available slots

5
The initial focus is on this simplifying process, but epistemic actions are moves in the problem solving
routines of epistemic agents.
566 R. Menary

in a continually emerging structure. A rotation button can be used to orient the

shapes relative to the slots—so that discrimination of fit can be judged. Classically,
we might think that the rotation of the zoids would first happen in the brain and
would then result in action in the world. However, K&M say that:
“the clearest reason to doubt that deciding where to place a zoid involves
mental rotation is that zoids can be physically rotated 90° in as few as 100 ms.
whereas we estimate that it takes in the neighbourhood of 800–1,200 ms to
mentally rotate a zoid 90°.” (Kirsh and Maglio 1994, 514)
Even allowing for an extra 200 ms for subjects to select the rotate button, the time-
saving benefits of physical over mental rotation are obvious; however, time is not the
only benefit. The cost on working memory and attention to produce and sustain
mental images of zoids would harm performance—physically rotating the zoid is
less demanding than mental rotation.
Classically, we might think of problem solving as involving a series of classical
computations (processes) leading to an output (an action). However, epistemic
actions are taking up some of the processing duties required to reach a solution to the
problem, they are not simply outputs. As Kirsh puts it: “the skills we have developed
for dealing with the external world go beyond those we have for dealing with the
internal world.” (Kirsh 1995, 64)
The actions that Tetris players perform are directly altering the informational
structure of the environment. They often perform actions that are not directly
tied to the pragmatic goal of dropping the zoids into the correctly shaped slots.
Expert players “regularly performed actions that helped them to recognise pieces,
verify the goodness of potential placements, and test plans (e.g. dropping a piece
from high up on the board).” (Kirsh 2009, 283) The actions are indirect; they often
take the zoid away from the optimal dropping point, by lining the piece up with a
wall for example.
Whether or not we consider epistemic actions to be part of our cognitive
processing depends, partly, upon how seriously we take K&M’s distinction between
epistemic and pragmatic actions. Pragmatic actions are aimed at achieving a non-
cognitive goal, like a full stomach. Epistemic actions are aimed at achieving a
cognitive goal, like solving a problem, by manipulating the environment. The
distinction is interestingly illustrated as follows:
“one significant consequence of recognising epistemic action as a category of
activity is that if we continue to view planning as state space search, we must
redefine the state space in which planning occurs. That is, instead of
interpreting the nodes of a state space graph to be physical states, we have to
interpret them as representing both physical and informational states.” (Kirsh
and Maglio 1994, 515)
Classically we might think of a plan and actions which implement the plan as
distinct; however, K&M’s point is that epistemic actions are part of strategy creation
and problem solving. Clearly, problem solving is cognitive. So if epistemic actions
are part of that problem-solving process, and not merely the result of problem
solving or the physical means by which plans are implemented, then we have a very
good reason for claiming that epistemic actions are part of the problem solving
Dimensions of mind 567

process.6 Accepting this depends upon taking seriously the distinction between
epistemic and pragmatic action and seeing a tighter integration between external
epistemic actions and internal processes.
K&M themselves say of their approach that: “its chief novelty lies in allowing
individual functional units inside the agent to be in closed-loop interaction with the
outside world.”7 (Kirsh and Maglio 1994, 542) Epistemic actions are coordinated
with processes such as working memory and attention and epistemic action plus
internal processing jointly govern pragmatic action. Therefore, the coordinated and
interactive process that spans physical manipulations and attention is the process that
is causally responsible for the pragmatic action. This is the reason why I prefer a
strongly embedded interpretation of cases of epistemic action.
K&M give us a way of thinking about the role of action such that we can unify
physical space and information processing space. A weakly embedded explanation
of these cases would distinguish between a realm of internal computations upon
inputted information and a realm of outputted physical behaviours, but this is
precisely the explanation that K&M are rejecting. The information-processing space
includes both processes in the head, and processes outside of the head. All of the
processing takes place within the same state space. This state space is the problem
solving state space. Consequently, epistemic actions (as bodily manipulations of
vehicles in the environment) are processes to be found in that state space.
However, the internalist thinks that there is an easy way of demonstrating how
this division works: the only restructuring of the environment which is relevant to
the problem solving/planning agent occurs inside the agent’s head. In other words,
“The structure of the environment which matters to cognition is the structure
the agent represents (or at least presupposes in the way it manipulates its
representations).” (Kirsh and Maglio 1994, 545)
If all the relevant structure is in the head, this obviates the need for epistemic action,
where the world is structured in such a way as to negate the need for all of the
structured representations being located in the head.
But K&M have established that there is such a class of actions as epistemic
actions and now the internalist is at a loss to account for them. If we were to agree
with the internalist that all the relevant structure, all the relevant states and processes,
are in the head, then we cannot explain the performance of Tetris players. If we
explain their performance by making reference to external manipulations as part of
their cognitive processing (the problem solving), then we can explain their
performance. The weak embedding strategy is blocked by the inability to explain
the Tetris player’s performance by internal cognitive resources alone. The internalist
is forced into this position because of the difference between epistemic actions and
pragmatic actions—epistemic actions are aimed at making a move in the problem
solving state space, pragmatic actions are not. Kirsh has recently made this very
clear: “The challenge epistemic actions pose to the classical approach… is that
without an analysis of the possible epistemic functions of an action it may be nearly

6
Remembering that ‘epistemic action’ is a term used to label a class of manipulations of the environment.
7
As opposed to simply being outputs of internal processes.
568 R. Menary

impossible to identify the primary function of an action and so label it correctly.”

(Kirsh 2009, 284)
K&M’s focus is on epistemic actions as interactions with the environment where
an internal representation is not required. Many complex cognitive behaviours arise
without the need for the processing of internal representations; because we can
always manipulate the environment instead. Cognitive integrationists avoid the
pitfall of the weak embedding strategy by locating epistemic actions in the problem
solving state space, not just as a clever strategy for off-loading complexity onto the
environment. Epistemic actions show that interaction with the environment replaces
the need for the processing of internal representations of the environment;
nonetheless, these classes of manipulations are still part of our cognitive economy.

Self-correcting actions

Verbal self-guidance or self-controlled action (Menary 2007a) has an important

cognitive function. This is clearly evident in the responses of children to the verbal
commands of adults in controlling and directing their actions (Morin 2005). They are
able to perform actions with the verbal and physical support of adult caregivers. They
are scaffolded by adults during their cognitive development. Eventually, they learn
how to deploy language without the scaffolding of caregiver directions. Instead they
auto-generate verbal commands, cues and interventions to control and direct activity.
The creation of linguistic structures that control and direct action, like epistemic
actions, can have a cognitive function. For example, Kendall and Hollon (1981) give
us four categories of the regulatory use of speech for problem solving: (1) the precise
definition of the problem; (2) the effective approach to the problem; (3) the focus on
the problem; (4) the progress evaluation that includes praise or strategy readjustment
(Morin 2005). Vygotsky has an account of the internalisation of egocentric speech
and the kind of cognitive functions such speech has. I shall use a core example from
his work to illustrate the kind of function that inner speech has as outlined by
Kendall and Hollon.
From a Vygotskian developmental point of view, higher cognition, for example
reasoning and memory, appears first on the ‘intermental’ plane, in other words, in
social interaction. Obvious examples would be language learning and joint adult–
child problem solving activities. Cognition, then, is primarily a social phenomenon.
However, Vygotsky did claim that higher cognition appears on the ‘intramental’
plane (individual), but only as it is shaped by and derived from intermental
cognition.8 It is crucial, then, to understand how intermental cognition works, for we
will be at a loss to understand cognition at the level of the individual. One typical
example of this phenomenon is the internalisation of speech.
Piaget labelled the speech that young children engage in when problem solving or
engaging in pretend play ‘egocentric speech.’ Vygotsky does not view this form of
speech as a manifestation of a child’s egocentricity, rather, Vygotsky argues (based

8
Of course, this does not preclude the possibility that there are inherited cognitive structures in the brain.
See for example Dehaene’s work below. Think also of the work of Meltzoff and Moore (1977) on early
infant imitation. The point here is that these kind of cognitive mechanisms are very basic and require
development by being scaffolded in the cognitive niche.
Dimensions of mind 569

on empirical studies of infants' speech) that this form of speech is merely the
internalisation of speech. It does not disappear with age, it becomes an internal
monologue.
The intermental development of cognition is understood in terms of ‘the zone of
proximal development.’ The zone of proximal development is the distance between the
actual level of development of an individual, what the individual can actually do, and the
potential level of development, which is what the individual can potentially do, with
guidance and collaboration from a tutor. It follows that the individual level of
development should not be the exclusive focus of interest. Intermental cognition as
mediated through language allows us to understand the intramental capabilities of an
individual.
In cases where the child must act in such a way as to bring about a goal, the
activity is accompanied by egocentric speech. As the child gains mastery of
egocentric and then inner speech, she gains access to self-controlled behaviour,
which helps her to complete cognitive tasks such as problem solving.
As an example of this development, Vygotsky cites an experiment (by a colleague
Levina) where a child’s speech arises spontaneously in a problem-solving situation.
The speech is continuous throughout the experiment as observed.
Levina’s experiments posed problems to 4- and 5-year olds, such as obtaining
candy/sweets from a cupboard. The candy was placed out of reach so that the child
could not reach it directly. Vygotsky describes the concurrent roles of speech and
action (including tool use) in the child in the following way:
“As the child got more and more involved in trying to obtain the candy,
‘egocentric’ speech began to manifest itself as part of her active striving. At
first this speech consisted of a description and analysis of the situation, but it
gradually took on the ‘planful’ character, reflecting possible paths to a solution
of the problem. Finally it was included as part of the solution.” (Vygotsky
1978, 25)
A four-and-a-half-year-old girl was asked to get candy from a cupboard with a stool
and a stick as tools. The experiment was described by Levina in the following way
(his descriptions are in parentheses, the girls speech is in quotation marks):
(Stands on a stool, quietly looking, feeling along a shelf with stick). “On the
stool.” (Glances at experimenter. Puts stick in other hand) “Is that really the
candy?” (Hesitates) “I can get it from that other stool, stand and get it.” (Gets
second stool) “No that doesn’t get it. I could use the stick.” (Takes stick,
knocks at the candy) “It will move now.” (knocks candy) “It moved, I couldn’t
get it with the stool, but the, but the stick worked.”

Vygotsky claims that activity is not just accompanied by speech in children, but
that speech plays a specific role in such activity. He claims that the experiments
show two important facts:
1. A child’s speech is as important as the role of action in attaining the goal.
Children not only speak about what they are doing; their speech and action are
part of one and the same complex psychological function, directed toward the
solution of the problem at hand.
570 R. Menary

2. The more complex the action demanded by the situation and the less direct its
solution, the greater the importance played by speech in the operation as a whole.
Sometimes speech becomes of such vital importance that, if not permitted to use it,
young children cannot accomplish the given task (Vygotsky 1978, 26).
We should note that the child’s speech gives a cognitive structuring to the activity.
The girl verbally structures a sequence of actions, searching for a solution and
controlling the actions that will reach the solution. Children gradually gain mastery over
language as part of their cognitive as well as communicative resources. Self-corrective
speech, whether private or public, is used to structure, direct and correct actions that lead
to the completion of cognitive tasks. Again, like epistemic actions, there can be moves
made in problem solving space. They are an example of one of the ways in which the
internalisation of public language transforms our cognitive capacities.

Cognitive practices

When we manipulate a public representation to complete a cognitive task, we are

performing a cognitive practice. We are able to manipulate public representational
vehicles because we acquire manipulative abilities that are governed by cognitive
norms. Public vehicles, such as written language and mathematical symbols, are
tokens of representational systems. Such systems have their own norms governing
how we are to manipulate token representational vehicles. These norms are cognitive
as opposed to moral or social norms; they are directly tied to the completion of
cognitive tasks.
There is a great variety of public representational systems, which mirrors the great
variety of tasks to which we put them. Examples of such tasks include: solving
problems, making inferences, planning, working out answers to questions and so on
(these are cognitive tasks). Public representational systems structure our cognitive
niches. The child’s cognitive development takes place in a representationally
structured niche and the child is trained and scaffolded by caregivers who have
already gained mastery over these systems.
The child learns a practice that is an established method of manipulating
representations to produce an end. For example, we write down the intermediate
stages in problem solving, which can function as part of the working memory space,
making information available for further manipulation. Plans are often written down
and then transformed, updated and shared. Lists and diaries allow us to retrieve
information that require long-term storage and are easily and conveniently
accessible. The representational properties of maps enable easy and shared
navigation. Public representational systems allow for the storage and manipulation
of detailed content, trusted sources of information, shared information and the
structuring of thought that internal representations alone cannot provide.
The implementation of a cognitive practice depends on cognitive norms that
guide that practice; these norms also structure, in part, our cognitive niches. So, for
example, there are:
1. Purposive norms. The activity is engaged in for a purpose, or end.
2. Corrective norms. These are norms for using representations to correct activity
in pursuit of an end.
Dimensions of mind 571

3. Manipulative norms. These are norms for manipulating inscriptions of a

representational system.
4. Interpretative norms. These are norms for interpreting inscriptions of a
representational system as having some wider significance, not just within the
representational system itself but also with regard to the wider world and
interests of others.
5. Creative norms. Norms that govern the creation of public representations.

We will need an account of how we learn cognitive practices. This will, in part,
involve the acquisition of capacities to manipulate representations and thereby
transform our cognitive abilities. However, acquiring these capacities should be
understood in the context of the cognitive practices required to complete cognitive tasks.
As such, a cognitive practice is essentially the embodying of norms in an activity.
Writing is such an activity: the crucial elements of writing, or typing, are the
ability to manipulate tools to create external vehicles: words, sentences and
paragraphs. The pen and paper or the CPU, keyboard and monitor are not
themselves cognitive, but the creation and manipulation of the external vehicles
and the coordination of internal and external vehicles are.
For example, my reading and rereading what I have written gives me new ideas
about what I should write next. Though it is true that tools such as keyboards and
pens enable me to write, it is the process of manipulating the written vehicles that is
the cognitive processing. The sentences can be rewritten, erased, moved to another
paragraph and so on. It is, moreover, precisely these kinds of manipulations that are
not easily, if ever, achieved in the head. Therefore, writing as an active and creative
process is enabled by tools such as pen and paper or word processors, but it is the
bodily manipulation of the external vehicles themselves that is where the cognitive
work gets done.
Once written, the vehicles are then available for further manipulations such as
restructuring, revising and redrafting. Manipulating written vehicles is a kind of
problem solving where a particular cognitive task must be completed: for example,
“how do I make this piece of writing clearer?” Completing these kinds of goals
without external media would be made more difficult by their absence. Without
them, behavioural competence will drop and the completion of the cognitive task be
made more difficult. Hence, cognitive integrationists are inclined to think that those
external manipulations play an important role in the processing of the task, one
different from the enabling role of tools.
What kinds of processes do external manipulations afford that purely internal
manipulations will not? If, for example, I tried to compose an essay in my head, the
likelihood of retaining much of the argument and structure would become very
limited. Making revisions and corrections would be almost impossible. Stable and
enduring external written sentences allow for manipulations, transformations,
reorderings, comparisons and deletions of text that are not available to neural
processes. Bodily manipulations of external vehicles are different from, but
complementary to, internal processes (Sutton 2010). The coordination of internal
and external processes in one extended dynamic process enables the completion of
complex cognitive tasks such as composition. I turn now to the dimension of
cognitive transformation.
572 R. Menary

Cognitive transformations

Cognitive transformations occur when the development of the cognitive capacities of

an individual are sculpted by the cognitive and social niche of that individual.
During the learning and training of a skill, such as a shot in tennis or cricket, we are
guided by norms. As I will show below, body schemas are governed by norms and
the norms have an organising function which results in new sensorimotor
programmes. During development, the individual also gains mastery over the
cognitive norms by which external representations are manipulated. Vygotsky
expresses this in the claim that children, "master the rules in accordance with which
external signs must be used." (Vygotsky 1981, 184–5) The child masters the creation
and deployment of signs (and later public systems of representation) by being
scaffolded by parents and care givers.
Vygotsky gives the following example of how such mastery of the creation and
deployment of signs may begin to develop in the child:
“At first the indicatory gesture is simply an unsuccessful grasping movement
directed at an object and designating a forthcoming action. The child tries to
grasp an object that is too far away. The child’s hands, reaching toward the
object, stop and hover in mid air…. Here we have a child’s movements that do
nothing more than objectively indicate an object.”
When the mother comes to the aid of the child and comprehends the movement as
an indicator, the situation changes in an important way. The indicatory gesture
becomes a gesture for others. In response to the child’s unsuccessful grasping
movement, a response emerges not on the part of the object, but on the part of
another human. Thus, other people introduce the primary sense into this
unsuccessful grasping movement. Eventually, the child grasps the indicatory import
of the gesture for themselves, thereby producing the gestures as communicative
signs. The functions of the movement itself have undergone an important change:
“from a movement directed toward an object it has become a movement
directed toward another human being. The grasping is converted into an
indication… this movement does not become a gesture for oneself except by
first being an indication, that is, functioning objectively as an indication and
gesture for others, being comprehended and understood by surrounding people
as an indicator. Thus the child is the last to become conscious of the gesture.”
(Vygotsky 1981, 160–1)
The child’s behaviour is shaped by the normative environment in which she acts.
Gesturing becomes indicating once it is interpreted as being so by another, thus the
child is brought into the practice of a social norm. It is in this way that the higher
cognitive capacities of the individual are sculpted, by interactions with the social/
representational environment.
In the next two subsections, I look at how the normative and social structure of
the environment, mediated by learning and training histories, has a direct
transformatory effect on the body. In particular, I look at how the development of
a body schema leads to its integration with the environment. Then I look at empirical
research which shows how our mathematical cognitive capacities are transformed by
Dimensions of mind 573

the internalisation of public symbolic representations of numerals and number

words.

The transformation of the body schema

Gallagher (2005) gives a historical overview of the development of the concepts of

body image and body schema, respectively. A body image “consists of a system of
perceptions, attitudes, and beliefs pertaining to one’s own body.” (2005, 24) A body
schema is “a system of sensory-motor capacities that function without awareness or
the necessity of perceptual monitoring.” (2005, 24)
“So the difference between body image and body schema is like the difference
between a perception (or conscious monitoring) of movement and the actual
accomplishment of one’s own movement…” (2005, 24)
I will concentrate on the body schema. Body schemas are subpersonal processes that
dynamically govern posture and movement and Gallagher claims that they do this in
a close to automatic way (Gallagher 2005, 26). Repertoires of body schemas
function together as motor programmes (Gallagher 2005). Some motor programmes
are learned such as riding a bike and writing and some are innate such as
swallowing.
They are only close to automatic, because they can be part of a goal-directed
activity, such as catching a ball in a game of cricket or baseball. There are higher
level goals and intentions involved in this action, but our consciousness is not
directed at the movements of our body, but at the ball. Hence, we are not aware of
the functions of the body schema in governing our posture and movement, in so far
as these are part of a goal-directed activity.
Consequently, the body schemas that control my bodily actions cannot be
understood in purely physiological terms.
“The physical environment, the size and shape of the ball, along with the
effects of all my previous practice (or lack thereof), and even the rules of the
game as they are habitually expressed in the practiced movements of my body,
may define how I jump to make a catch. Without a certain amount of
selectivity, built up by practice and the cultivation of habitual movements, the
body might move in any one of multiple ways, since the possibilities allowed
by physiology are much greater than the particular movements necessary to
catch the ball in the proper way. Thus the body schema is much more
selectively attuned to its environment than what physiology on its own will
specify.” (Gallagher 2005, 143)
The selective tuning of the body schema to its environment is developed through
habituation, practice and training. In the first instance one learns how to catch a ball
and then one trains to perfect the technique, one might do so by imitating the
movements of a teacher. The learned movements are practised until one is habituated
to make the right moves appropriate to the situation. Learning and training require us
to learn the right movements and to apply them in the right ways, to do so in
conformity with the rules and laws of the game. The cultural and social
574 R. Menary

sedimentation of our body schemas fundamentally shape the ways that we move,
intentionally act and come to communicate with others. There is a clear and
fundamental sense in which body schemas become integrated with the environment
and it is to this that I next turn.
One of the important things about schemas and motor programmes is that they
can exhibit a high degree of integration with the environment. In these cases, the
body schema incorporates parts of the environment that are not incorporated into the
body image9—such as the hammer in the carpenter’s hand (Gallagher 2005, 37).
The integration has a phenomenological aspect, where a part of the environment
can feel like an extension to the body; in these cases, the body schema goes beyond
the narrow boundary that is apparent from the body image. It also has a neural
aspect:
“This extension of the body schema into its surrounding environment is
reflected in its neural representations. Not only do bimodal premotor, parietal,
and putaminal neuronal areas that represent a given limb or body area also
respond to visual stimulation in the environmental space nearby, for some of
these neurons the visual receptive field remains ‘anchored’ to the body part
when it moves (Fogassi et al. 1996; Graziano and Gross 1998; Graziano et al.
1994).” (Gallagher 2005, 37)
A clear example of this is the work by Maravita and Iriki (2004) on training Japanese
Macaques in tool use. The Macaques were trained to use a rake to pull food placed
on a table towards them. Maravita and Iriki (2004) report that several changes to the
receptive field of bimodal neurons were observed in Macaques who were trained to
use the rake. These changes in the properties of bimodal neurons occurred only after
“active, intentional usage of the tool, not its mere grasping by the hand.” (Maravita
and Iriki 2004, 81) The vRFs of these bimodal neurons now extends to include the
hand and entire tool, the tool is now incorporated into the body schema.10
Body schemas do not just initiate behaviour they are fully integrated with the
environment. They are constrained by the environment because they often require its
perceptual navigation and the manipulation of environmental objects. Therefore, the
body shapes itself to meet the environment, to hold a glass in hand or grip a pencil
between fingers and thumb for writing. Body schemas are attuned to environmental
affordances for action (Gibson 1979), the glass affords drinking and the pencil
writing.
It is in the fluid manipulation of objects in the environment and in fluent skilled
activities that we are most likely to find the unconscious integration of the body
schema with the environment. Experienced drivers will understand the nature of this
integration which involves the seamless co-ordination of body and car, sometimes to
such an extent that one cannot recall the details of the journey when the destination
is reached (Gallagher 2005). Although the body image involves conscious
experience of our own bodies and that experience is of a bounded body the body
schema has no such boundary, it directs our primary embodied engagements with the
world and it is because of this that we feel ourselves to be both in and part of the

9
The body image ends at the periphery of the body.
10
See also Clark 2008 p.38, see Gallagher 2005 and Menary 2007a, b for earlier discussions.
Dimensions of mind 575

world. Furthermore, it is constitutive of our first cognitive engagements with the

world, our perceptual navigation, our imitation of others and our manipulation of the
environment.
The expert Tetris players in K&M’s study have developed motor programmes for
manipulating the buttons that transform the zoids on the screen. As experts, their
well-trained body schemas result in fluent and fluid actions, they do not need to
consciously rehearse the next action. Note though, that the manipulation of the
buttons is itself a normative practice, something that is learned and habituated. In
Vygotsky’s example of indication, the child develops a habit of gesturing and
reaching, which comes to have a communicative function—first for others, but then
also for the child itself. The body is also transformed to be able to write and draw
and create representational structures. It does so by incorporating the tools for
manipulating representations into the body schema, but also by mastery of the norms
which regulate those manipulations. In the next section, I look at the transformation
of our capacity to think mathematically by the internalisation of number systems.

The transformation of mathematical thinking

Stanislas Dehaene (a cognitive neuroscientist) and colleagues have produced stunning

results and a compelling model that shows that we think in an internalised conventional
representational system. Dehaene argues that we have a basic analogical and non-
linguistic capacity to recognise quantity and number located in the inferior parietal lobe
of the brain. However, that biological sense of number is not the whole of the story
because there is also an acquired ability to think in terms of discrete numbers in the left
frontal lobe. What we have here is the difference between a non-linguistic continuous
representation of quantities and an acquired, discrete and linguistic representation of
individual numbers from our numeral system, including individual words for numbers.
The two capacities work together but the internalisation of the public numeral system
allows us to perform the kind of digital mathematical operations that are required for
most arithmetic and mathematical operations (Nieder and Dehaene 2009, 197). Our
cognitive capacities for mathematical thinking are thereby transformed. The public
numeral system and its internalisation allow for new representational formats and new
operations—or manipulations of those representations. In Vygotsky’s sense, digital or
precise mathematical thinking occurs first on the intermental plane and only
secondarily on the intramental plane.
“The model that emerges suggests that we all possess an intuition about
numbers and a sense of quantities and of their additive nature. Upon this
central kernel of understanding are grafted the arbitrary cultural symbols of
words and numbers…. The arithmetic intuition that we inherit through
evolution is continuous and approximate. The learning of words and numbers
makes it digital and precise. Symbols give us access to sequential algorithms
for exact calculations.” (Dehaene 2007, 41)
Dehaene’s well-known experiments to establish the difference between the
inexact continuous and analogical part of our thinking and the exact discrete and
digital part were based around behavioural evidence and imaging data. In the first
576 R. Menary

series of experiments (Dehaene et al. 1999), Russian–English bilinguals were taught

a set of exact and approximate sums of two digit numbers in one of their languages
(1999, 970). Their tasks were split into giving exact answers to additions and giving
an approximate answer to the addition task. The interesting result was that “When
tested on trained exact addition problems, subjects performed faster in the teaching
language than in the untrained language, whether they were trained in Russian or
English.” (Dehaene et al. 1999, p. 971) This provided evidence that knowledge of
arithmetic was being stored in a linguistic format, and that there was a switching cost
between the trained and untrained languages. By contrast, there was equivalent
performance in the approximation task, and no switching cost between the trained
and untrained languages. Dehaene et al. conclude that this provides “evidence that
the knowledge acquired by exposure to approximate problems was stored in a
language-independent form.” (1999, p. 971)
The second set of studies used fMRI scans of the brain. The scans revealed
greater activation in the bilateral parietal lobes for the approximation tasks than for
the exact calculation (Dehaene et al. 1999, p. 971). The areas of the brain activated
during the approximation tasks are networks responsible for visuo-spatial represen-
tations. During exact calculation tasks, the scans revealed “a large and strictly left-
lateralized activation in the left inferior frontal lobe.” (Dehaene et al. 1999, p. 972)
These areas have been found to be activated in word-association tasks indicating that
for the exact calculations, there was a “network involved in the language-dependent
coding of exact addition facts as verbal associations.” (1999, p. 972)
The deeply transformative power of our learning histories in the cognitive niche is
one that reformats the representational capacities of the brain in terms of public
symbol systems. We internalise public symbol systems in the way Dehaene suggests,
but we also learn techniques for manipulating those symbol systems in public space.
In learning the manipulative techniques, the first transformation is one of sensory
motor abilities for creating and manipulating public symbols, we learn algorithms
like: ‘carry the two’ and this is an example of the application of the cognitive norms
(outlined above). This is something we learn to do on the page, in public space,
before we do it in our heads. We might argue that this is a case of scaffolding or
augmentation, but it looks clear that our capacities to think have been transformed,
but in this instance they are capacities to manipulate symbols in public space. This is
a way of showing that the transformation of our cognitive capacities has publicly
recognisable features. This ought not to be a surprise given that the cognitive niche
is socially constructed.
There are several explanatory options available at this point:

1. The only really cognitive elements are in the brain and although the brain is
transformed by public symbol systems nevertheless that’s where cognition takes
place,
2. The capacities to manipulate public symbols are mere scaffolds for our limited
online capacities.
3. Is quite different—that development in the cognitive niche results in an
integrated cognitive system (a position I have argued for elsewhere, Menary
2007a)—the integrated cognitive system is the system which has been through
the dual component process of transformation where one gains mastery over the
Dimensions of mind 577

symbol system in public space and that leads to a transformation of our limited
cognitive capacities such that one can complete cognitive tasks either by
manipulating symbols in public space or by manipulating symbols in neural
space or very possibly by a combination of both sets of resources. That is not to
say that notebooks, pieces of paper and so on are parts of the organism, it is
simply to say that certain cognitive capacities have both their inner and outer
manifestations and that seems to me to go beyond Sterelny’s scaffolding
position (2010) and not fall prey to the constitutive error which is so often
levelled at the extended mind (Adams and Aizawa 2008).

Conclusion

Mental capacities work across a multitude of dimensions. There is an increasing

abundance of evidence that some of those dimensions involve processes that are
composed of bodily manipulations of the environment. Against this evidence, one
can persist in denying the importance of the cognitive niche both for actual cases of
thinking and for our explanatory models of how we think. I have argued that we
should be willing to give a strongly embodied and embedded explanation of cases of
manipulation of the cognitive niche. The transformation of our cognitive capacities
by development in the niche also shows that when we think in our heads we are also
thinking with the representational resources of the cognitive niche. Furthermore,
when we manipulate those same representational forms publically, in the niche, we
are still thinking.

References

Adams, A., & Aizawa, K. (2008). Defending the bounds of cognition. Oxford: Basil Blackwell.
Ballard, D. H. (1991). Animate vision. Artificial Intelligence Journal, 48, 57–86.
Clark, A. (2008). Supersizing the mind: Embodiment, action and cognitive extension. Oxford: Oxford
University Press.
Dehaene, S. (2007). A few steps towards a science of mental life. Mind, Brain and Education, 1(no.1),
28–47.
Dehaene, S., Spelke, E., Pinel, P., Stanescu, R., & Tsivkin, S. (1999). Sources of mathematical thinking:
Behavioral and brain imaging evidence. Science, 284, 970–974.
Fogassi, L., Gallese, V., Fadiga, L., Luppino, G., Matelli, M., & Rizzolatti, G. (1996). Coding of
peripersonal space in inferior premotor cortex (area f4). Journal of Neurophysiology, 76, 141–157.
Gallagher, S. (2005). How the body shapes the mind. Oxford: OUP.
Gibson, J. J. (1979). The ecological approach to visual perception. Boston: Houghton Mifflin.
Graziano, M. S., & Gross, C. G. (1998). Spatial maps for the control of movement. Current Opinions in
Neurobiology, 8, 195–201.
Graziano, M. S., Yap, G. S., & Gross, C. G. (1994). Coding of visual space by premotor neurons. Science,
266, 1054–1057.
Kendall, P. C., & Hollon, S. D. (1981). Assessing self-referent speech: Methods in measurement of self-
statements. In P. C. Kendall & S. D. Hollon (Eds.), Assessment strategies for cognitive-behavioral
interventions. New York: Academic.
Kirsh, D. (1995). Complementary strategies: Why we use our hands when we think. In J. D. Moore & J. F.
Lehman (Eds.) Proceedings of the Seventheenth Annual Conference of the Cognitive Science Society.
212–217.
Kirsh, D. (2009). Problem solving and situated cognition. In P. Robbins & M. Aydede (Eds.), The
cambridge handbook of situated cognition. Cambridge University Press: Cambridge.
578 R. Menary

Kirsh, D., & Maglio, P. (1994). On distinguishing epistemic from pragmatic actions. Cognitive Science,
18, 513–549.
Maravita, A., & Iriki, A. (2004). Tools for the body schema. Trends in Cognitive Sciences, 8(2), 79–86.
McClelland, J., Rumelhart, D., et al. (Eds.). (1986). Parallel distributed processing: Explorations in the
microstructure of cognition. Vols. I and II. Cambridge: MIT Press/Bradford Books.
Meltzoff, A. N., & Moore, M. K. (1977). Imitation of facial and manual gestures by human neo-nates.
Science, 198, 7.
Menary, R. (2006). Attacking the bounds of cognition. Philosophical Psychology, 19(3), 329–344.
Menary, R. (2007a). Cognitive integration: Mind and cognition unbounded. Basingstoke: Palgrave
Macmillan.
Menary, R. (2007b). Writing as thinking. Language Sciences, 29(5), 621–632.
Menary, R. (2009). Intentionality, cognitive integration and the continuity thesis. Topoi, 28, 31–43.
Menary, R. (2010). Cognitive integration and the extended mind. In R. Menary (Ed.), The extended mind.
Cambridge: Bradford Books/MIT Press.
Millikan, R. (1993). White queen psychology and other essays for Alice. Cambridge: Bradford Books/MIT Press.
Millikan, R. (2004). The varieties of meaning: The Jean Nicod lectures. Cambridge: MIT Press.
Morin, A. (2005). Possible links between self-awareness and inner speech: Theoretical background, underlying
mechanisms, and empirical evidence. Journal of Consciousness Studies, 12(4–5), 115–134.
Nieder, A., & Dehaene, S. (2009). Representation of number in the brain. Annual Review of Neuroscience,
32, 185–208.
O’Regan, J. K., & Noë, A. (2001). A sensorimotor account of vision and visual consciousness.
Behavioural and Brain Sciences, 24, 939–1031.
Rowlands, M. (1999). The body in mind: Understanding cognitive processes. Cambridge: Cambridge
University Press.
Rupert, R. (2009). Cognitive systems and the extended mind. Oxford: Oxford University Press.
Sterelny, K. (2010). Minds: extended or scaffolded? Phenomenology and the Cognitive Sciences.
doi:10.1007/s11097-010-9174-y.
Sutton, J., Harris, C. B., Keil, P. G., & Barnier, A. J. (2010). The psychology of memory, extended
cognition, and socially distributed remembering. Phenomenology and the Cognitive Sciences.
doi:10.1007/s11097-010-9182-y.
Vygotsky, L. (1978). Mind in society. Cambridge: Harvard Press.
Vygotsky, L. (1981). The instrumental method in psychology. In J. Wertsch (Ed.), The concept of activity
in Soviet psychology. Armonk: Sharpe.
Webb, B. (1994). Robotic experiments in cricket phonotaxis. In D. Cliff, P. Husbands, J. Meyer, & S.
Wilson (Eds.). From animals to animals 3: Proceedings of the Third International Conference on the
Simulation of Adaptive Behaviour. Massachusetts: MIT Press. 45–54.
Wheeler, M. (2005). Reconstructing the cognitive world: The next step. Cambridge: MIT Press.

View publication stats