ArticlePDF Available

A Taxonomy of Cognitive Artifacts: Function, Information, and Categories



The goal of this paper is to develop a systematic taxonomy of cognitive artifacts, i.e., human-made, physical objects that functionally contribute to performing a cognitive task. First, I identify the target domain by conceptualizing the category of cognitive artifacts as a functional kind: a kind of artifact that is defined purely by its function. Next, on the basis of their informational properties, I develop a set of related subcategories in which cognitive artifacts with similar properties can be grouped. In this taxonomy, I distinguish between three taxa, those of family, genus, and species. The family includes all cognitive artifacts without further specifying their informational properties. Two genera are then distinguished: representational and non-representational (or ecological) cognitive artifacts. These genera are further divided into species. In case of representational artifacts, these species are iconic, indexical, or symbolic. In case of ecological artifacts, these species are spatial or structural. Within species, token artifacts are identified. The proposed taxonomy is an important first step towards a better understanding of the range and variety of cognitive artifacts and is a helpful point of departure, both for conceptualizing how different artifacts augment or impair cognitive performance and how they transform and are integrated into our cognitive system and practices.
A Taxonomy of Cognitive Artifacts: Function,
Information, and Categories
Richard Heersmink
#Springer Science+Business Media Dordrecht 2013
Abstract The goal of this paper is to develop a systematic taxonomy of cognitive
artifacts, i.e., human-made, physical objects that functionally contribute to
performing a cognitive task. First, I identify the target domain by conceptualizing
the category of cognitive artifacts as a functional kind: a kind of artifact that is
defined purely by its function. Next, on the basis of their informational properties, I
develop a set of related subcategories in which cognitive artifacts with similar
properties can be grouped. In this taxonomy, I distinguish between three taxa, those
of family, genus, and species. The family includes all cognitive artifacts without
further specifying their informational properties. Two genera are then distinguished:
representational and non-representational (or ecological) cognitive artifacts. These
genera are further divided into species. In case of representational artifacts, these
species are iconic, indexical, or symbolic. In case of ecological artifacts, these species
are spatial or structural. Within species, token artifacts are identified. The proposed
taxonomy is an important first step towards a better understanding of the range and
variety of cognitive artifacts and is a helpful point of departure, both for conceptu-
alizing how different artifacts augment or impair cognitive performance and how they
transform and are integrated into our cognitive system and practices.
1 Introduction
Situated cognition theorists argue that in order to properly understand a substantial
part of our cognitive capacities, we should take into account the artifacts we deploy to
help us perform our cognitive tasks. We use maps to navigate, notebooks to remem-
ber, rulers to measure, calculators to calculate, sketchpads to design, agendas to plan,
textbooks to learn, and so on. Without such artifacts we would not be the same
cognitive agents, as they allow us to perform cognitive tasks we would otherwise not
DOI 10.1007/s13164-013-0148-1
R. Heersmink (*)
Department of Cognitive Science, Macquarie University, Sydney, Australia
be able to perform. Cognitive artifacts are thus important to study, not only because
they make us more powerful and versatile thinkers, but also because they shape and
transform our cognitive system and cognitive practices, ontogenetically (Menary
2007; Kirchhoff 2011) and phylogenetically (Donald 1991).
Numerous philosophical and empirical studies conceptualize how we deploy
artifacts to perform cognitive tasks (Kirsh and Maglio 1994; Kirsh 1995; Hutchins
1995; Clark and Chalmers 1998; Clark 2008; Heersmink 2013). These studies focus
on well-chosen examples and demonstrate that particular artifacts play a crucial role
in performing particular cognitive tasks. But, although there is a great variety in
cognitive artifacts, there has been no substantial or detailed attempt to categorize
different types of such artifacts. Some brief attempts have been made, but these focus
only on a segment of representational artifacts (see e.g. Norman 1993; Nersessian et
al. 2003; Brey 2005) and neglect non-representational cognitive artifacts. Due to this
variety and neglect of non-representational artifacts in current categorizations, a
taxonomy providing a systematic understanding of the different categories of cogni-
tive artifacts and their different cognition-aiding properties would be of great help for
situated cognition theorists. Andy Clark has aptly put it as follows:
The single most important task, it seems to me, is to better understand the
range and variety of types of cognitive scaffolding, and the different ways in
which non-biological scaffoldings can augment or impair performance on a
task.... The Holy Grail here is a taxonomy of different types of external prop,
and a systematic understanding of how they help (and hinder) human perfor-
mance(Clark 2004,p.3233).
Clarks suggestion has two interrelated aspects. In order to acquire a systematic
understanding of how different types of cognitive scaffolds augment or impair perfor-
mance on different cognitive tasks, it is helpful to first have an understanding of the range
and variety of such scaffolds. In other words, a first possible step to obtain ClarksHoly
Grailconsists of a taxonomy of cognitive scaffolds, outlining their distinctive cognition-
aiding properties. In this paper, I focus on one (important) type of such scaffoldsnamely,
artifactsby developing an elaborate and detailed taxonomy in which cognitive artifacts
with similar informational properties can be grouped. To the best of my knowledge, such
a taxonomy does not yet exist in the literature. Although the functional role of token
artifacts in performing cognitive tasks has been widely discussed in the philosophical and
cognitive science literatures under the heading of situated cognition (Dourish 2001;Clark
2003,2008; Robbins and Aydede 2009), metaphysical issues concerning (the categories
of) cognitive artifacts have not received much attention by those authors and are in some
cases even deliberately ignored (Hutchins 1995).
Whilst other authors working in philosophy of technology have addressed meta-
physical issues concerning artifacts, they focus on technological artifacts in general
and have little, if anything, particular to say about cognitive artifacts (Preston 1998,
2009,2013; Houkes and Vermaas 2004,2010; Kroes 2012). So there is a gap in the
literature which this paper addresses by conceptualizing functional and informational
properties of cognitive artifacts. By doing so, this paper brings into contact concepts
and theories in (philosophy of) cognitive science and philosophy of technology and
strengthens the rather thin ties between those fields. The taxonomy proposed in this
paper thus builds intradisciplinary and interdisciplinary bridges.
R. Heersmink
This paper is structured as follows. It starts by identifying some of the components of
situated cognitive systems and argues for the usefulness of a taxonomy of cognitive
artifacts (section 2). Next, the target domain is identified and a method for taxonomizing
is proposed (section 3). A taxonomy is then developed in which cognitive artifacts with
similar informational properties can be grouped into categories (sections 4,5&6).
2 Components of Situated Cognitive Systems
Situated cognitive systems have components that interact, transform each other, and
are (in varying degrees) integrated into a larger cognitive system. A possible strategy
to better understand the larger system is by decomposing it into its components and
by conceptualizing the particular cognitive properties of those components. This
strategy is helpful because those properties largely determine how the situated system
performs its cognitive tasks and to what degree the components are integrated and
transform each other. But before those properties can be conceptualized, we first need
to identify the components. Some situated cognitive systems consist only of human
agents such as, e.g., transactive memory systems in dyads (Harris et al. 2011)or
larger groups (Theiner 2013); others consist of humans and artifacts; and yet others
consist of humans and non-artifactual objects. The focus in this paper is on situated
systems that consist of human agents and cognitive artifacts, while acknowledging
the existence of systems consisting of other components.
2.1 Agents, Techniques, Artifacts, and Naturefacts
It may seem obvious that artifacts are easily identifiable components of certain
situated systems. However, cognitive anthropologist Edwin Hutchins is critical about
identifying and developing a distinct category of cognitive artifacts. In his seminal
book, Cognition in the Wild, he writes:
We are all cognitive bricoleursopportunistic assemblers of functional systems
composed of internal and external structures. In developing this argument I
have been careful not to develop a class, such as cognitive artifacts, of designed
external tools for thinking. The problem with that view is that it makes it
difficult to see the role of internal artifacts, and difficult to see the power of
the sort of situated seeing that is present in the Micronesian navigators images
of the stars(Hutchins 1995, p. 172).
Hutchinsworries here are twofold: If we focus on cognitive artifacts as the most
important component of situated cognitive systems, it makes it difficult to see the
functional roles of (a) internal artifacts and structures, and (b) of external structures
that are not human-made. To make his case, Hutchins points out that Micronesian
navigators use the stars as perceptual anchors to navigate at sea. He argues that the
interaction between internal artifacts, or perceptual strategies, and the external stars
makes it possible for the Micronesians to navigate. Thus, in order to explain the
navigational capacities of the Micronesians, we have to take into account their
learned perceptual strategies and the stars, neither of which are proper artifacts. If
we only focus on situated systems consisting of embodied agents and cognitive
A Taxonomy of Cognitive Artifacts
artifacts, we might overlook interesting cases like these, which would reduce explan-
atory power. And, furthermore, if we only focus on the artifactual component in such
situated systems, we might overlook the functional role of what Hutchins calls
internal artifacts. I am sympathetic to Hutchinsworries, but they can be overcome
by acknowledging and emphasizing that we should be aware that artifacts are only
one possible component of particular situated cognitive systems that interact and are
integrated with other components. Developing a category of cognitive artifacts does,
in my view, not mean that other components of situated systems are ignored or
overlooked. Ultimately, we should study situated cognitive systems, rather than their
components, but this does not mean that we cannot develop categories and vocabu-
laries for their components, as this would equally reduce explanatory power.
Drawing on Hutchinsquote, I now distinguish some of these components. First, it
is clarifying to make a distinction between technology and technique. A technology
(or artifact) is usually defined as a physical object intentionally designed, made, and
used for a particular purpose,
whereas a technique (or skill) is a method or procedure
for doing something. Both technologies and techniques are intentionally developed
and used for some purpose and are in that sense artificial, i.e., human-made.
However, it is important to note, or so I claim, that they are not both artifactual.
Only technologies are artifactual in that they are designed and manufactured physical
objects and in this sense what Hutchins refers to as internal artifacts, such as
perceptual strategies, can best be seen as cognitive techniques, rather than as internal
artifacts. Moreover, given that these cognitive techniques are learned from other
navigators and are thus first external to the embodied agent, it is perhaps more
accurate to refer to them as internalized cognitive techniques, rather than as internal
cognitive techniques.
Second, Hutchins writes that Micronesian navigators use the stars in the same way
as manufactured navigational artifacts are used, loosely implying that they are a kind
of cognitive artifact. Whilst stars are neither artificial nor artifactual in the sense just
explained, they are nevertheless the perceptual object of a cognitive technique and
have an important functional role in navigation. Using natural objects or structures for
some purpose is not uncommon, for example, we might use a dead branch of a tree as
a walking stick, a stone as a hammer, or, indeed, the stars to navigate. When doing so,
the branch, stone, and stars are not intentionally made for those purposes and may be
seen to form a bridge between natural objects and artifacts. Risto Hilpinen (2011), in
his entry on artifacts in the Stanford Encyclopedia of Philosophy, refers to such
objects as naturefacts. So, following Hilpinens terminology, I suggest referring to
natural objects that are used for cognitive purposes as cognitive naturefacts. This does
not make them less important for performing certain cognitive tasks. My point is that
because stars are not intentionally designed, manufactured, or modified for some
purpose, they do not belong in a category of artifacts. Hutchinsexample is apt in that
it shows that humans as cognitive agents not only intentionally construct and modify
This definition is sufficiently broad as to include less prototypical cases of artifacts such as domesticated
animals (e.g., guide dogs) and genetically modified organisms (e.g., biofuel producing algae). Guide dogs
and biofuel producing algae are intentionally modified (or trained) by humans to perform a particular
function, i.e., guiding blind people or producing biofuel. Thus the material of which (cognitive) artifacts are
made can be biological or non-biological and in some cases (cognitive) artifacts may even be alive, e.g., in
the case of a guide dog.
R. Heersmink
their environment for cognitive purposes, but even exploit natural objects for cogni-
tive purposes. Thus, to conclude this subsection, a preliminary and high-level taxon-
omy of cognitive scaffolds consists of embodied agents (or social scaffolds),
cognitive techniques, cognitive artifacts, and cognitive naturefacts.
2.2 Why Focus on (Taxonomizing) Artifacts?
As Hutchins rightly pointed out, we are opportunistic assemblers of functional
systems that are composed of internal (or internalized) structures and external struc-
tures. From an agent-centered perspective, it does not matter whether these external
structures are artifactual or natural. What matters is that they functionally contribute
to performing a cognitive task. So, in one sense, artifacts and naturefacts are
continuous in that they can both function as external cognition-aiding structures.
There is, however, one relevant difference between artifacts and naturefacts that
justifies paying more attention to (taxonomizing) artifacts: our intentional control
over the informational content and functions of cognitive artifacts is considerably
larger and results in significantly more variety, as compared to the intentional control
over cognitive naturefacts and the resulting variety. We have intentional control over
the content and functions of cognitive naturefacts only insofar as we can choose
which natural objects or structures to use for some cognitive task. We cannot actually
modify naturefacts to help us perform a cognitive task, because when we do, they
enter the realm of the artifactual.
One could argue that we also do not have full intentional control over the content
and functions of all cognitive artifacts, as some are designed and made by others.
Maps, timetables, textbooks, manuals, encyclopaedias, and roads signs, for example,
are designed and made by agents outside the situated system and a user typically has
no control over the content of such artifacts. The informational content of other
cognitive artifacts, however, is designed and made by the user of the artifact.
Notebook and agenda entries, to-do lists, shopping-lists, PowerPoint slides, and an
architects sketch, for example, are typically made by the user of the artifact. So the
intentional control an agent has over the content and functions of a cognitive artifact
differs, depending on the kind of artifact (Heersmink 2012). But, in either case,
cognitive artifacts are intentionally designed and made to aid human cognition. The
intentional design and making of cognition-aiding artifacts (either by designers or
users) results in a broad range of cognitive artifacts exhibiting different kinds of
informational properties that are specifically geared towards realizing a broad range
of cognitive tasks, including navigating, calculating, remembering, measuring, plan-
ning, designing, etc. This kind of intentional control not only results in a much richer
variety of cognitive artifacts, as compared to cognitive naturefacts, but also results in
external artifactual structures that can be integrated much deeper into the onboard
cognitive system, because they are functionally and informationally malleable.
Consequently, the transformative impact of artifacts on our cognitive system and
practices, both ontogenetically and phylogenetically, seems much more substantial as
compared to naturefacts.
There are at least two other reasons why it is important to pay attention to
(taxonomizing) cognitive artifacts. First, it gives us a much deeper conceptual
understanding of a particular kind of artifact, namely, cognitive artifacts. This is
A Taxonomy of Cognitive Artifacts
important because it contributes to expanding and further developing a relatively
small and emerging subfield in the philosophy of technology, which is sometimes
referred to as philosophy of artifacts. As Randall Dipert (1993) has argued, an
adequate philosophical theory of artifacts is largely lacking in the history of analytical
Western thought (compare Houkes and Vermaas 2010; Kroes 2012; Preston 2013).
Given the ubiquitousness of artifacts and their fundamental role in our lifeworld,
culture, and cognition, an adequate philosophical theory of artifacts would be very
much welcome. Developing a taxonomy of cognitive artifacts contributes to such a
Second, the reason cognitive artifacts are important to better understand is because
they have different properties and affordances as compared to cognitive techniques.
Although cognitive techniques are also physical in that they supervene on neural and
sometimes bodily structures, the physical material of artifacts is very different and
allows operations that are very difficult to perform in the brain. The specific phys-
icality and operations external artifacts allow, gives them particularly distinct func-
tional and informational properties, which are important to study in their own right.
As Merlin Donald (1991) has pointed out, exograms (or external representational
systems) have properties that are different from engrams (internalized information in
biomemory). Engrams are internalized and realized in the medium and format of the
brain, whereas exograms are external and much less restrained in their format and
capacity. The storage capacity of exograms far exceeds the storage capacity of both
single entries and clusters of entries in biological memory. Exograms are flexible in
that they can be reformatted and easily transmitted across different media, whereas
engrams are less flexible. These differences certainly do not always apply, but when
they do apply, they are enabled by the particular physicality, malleability, and format
of external artifacts. Because exograms have such different functional and represen-
tational properties, as compared to engrams, they have the capacity to complement the
shortcomings and limitations of engrams (Sutton 2006,2010). This is arguably why
we have developed exograms in the first place and if we want to understand how they
complement the working of engrams and cognitive techniques, it is very helpful to
have a taxonomy that distinguishes between different types of cognitive artifacts,
conceptualizing their distinctive cognition-aiding properties.
3 Taxonomizing Cognitive Artifacts
3.1 Identifying the Target Domain
Having identified some important components of situated cognitive systems, I now
focus on one of those components, i.e., cognitive artifacts. To get some clarity on
what they precisely are, it is helpful to start with discussing some existing definitions:
–“Cognitive artifacts are artificial devices that maintain, display, or operate upon
information in order to serve a representational function and that affect human
cognitive performance(Norman 1991, p. 17, emphasis added).
–“Cognitive artifacts are physical objects made by humans for the purpose of
aiding, enhancing, or improving cognition(Hutchins 1999, p. 126).
R. Heersmink
–“Cognitive artifacts are a special class of artifacts that are distinguished by their
ability to represent, store, retrieve and manipulate information(Brey 2005,p.385,
emphasis added).
–“Cognitive artifacts are material media possessing the cognitive properties of
generating, manipulating, or propagating representations(Nersessian 2005,p
41, emphasis added).
These definitions have two elements in common: cognitive artifacts are defined as (a)
human-made, physical objects that (b) are deployed by human agents for the purpose of
functionally contributing to performing a cognitive task. A third and more specific
element which only Donald Norman, Philip Brey, and Nancy Nersessiansdefinitions
contain is that cognitive artifacts provide (and sometimes manipulate or process)
representational information. Hutchinsdefinition is neutral about the specific informa-
tional properties of cognitive artifacts. Of the above three elements, only the second is
truly distinctive for the category of cognitive artifacts. In other words, their most
distinctive property is their function, which is to contribute to performing a cognitive
task. Given the centrality offunction, I demarcate the boundaries of the target domain by
conceptualizing cognitive artifacts as a functional kind, i.e., a kind of artifact that is
defined purely by its function (Carrara and Vermaas 2009). In Hilary Kornblithswords:
At least for the most part, it seems that what makes two artifacts members of the same
kind is that they perform the same function(Kornblith 1980, p. 112). Thus, any artifact
that functions to aid it user(s) in performing a cognitive task is a member of the
functional kind of cognitive artifacts.
By contrast, artifacts with pragmatic functions such as hammers, coffee cups, chairs,
screwdrivers, flower pots, light switches, and trash bins are not used for cognitive
purposes, at least not in their normal use, because they have no functional role in
performing a cognitive task. Hammering a nail into a wall, using a screwdriver to open
a can of paint, or turning on a light switch, does not straightforwardly contribute to
performing a cognitive task. We interact with those artifacts to create a change in the
state or location of the artifact, not because it aids our cognition, but because that goal
state is desirable for some practical purpose. The purpose for which such artifacts are
deployed is therefore not cognitive.
Artifacts with pragmatic functions may of course influence human cognition,
because sometimes we have to think about how to interact with those artifacts.
However, being the object of perception and cognition is necessary but not sufficient
for functionally contributing to a cognitive task. Many things are the object of
perception and cognition, but do not in any obvious sense help us to perform or
complete a cognitive task. However, this is not to say that such artifacts can never
have cognitive functions. For example, when I leave a rented DVD on my desk to
remind myself to bring it back to the video store, the DVD (in virtue of its location)
arguably functions as a mnemonic aid. So during improvisation, we can attribute
cognitive functions to initially non-cognitive artifacts, thereby demonstrating that we
cognitively exploit not only our natural environment, but also our artifactual envi-
ronment quite opportunistically. This example shows that a cognitive artifact is
neither defined by intrinsic properties of the artifact nor by the intentions of the
designer, but by its function, which is established by the intentions of the user and by
how it is used. Thus, the target domain that is to be categorized are artifacts with
A Taxonomy of Cognitive Artifacts
cognitive functions, which may either be intentionally designed or improvised to
perform that function.
3.2 Methodology
Current categorizations focus on representational artifacts and thus neglect non-
representational artifacts. The taxonomy proposed in this paper does include non-
representational artifacts and is thus more inclusive and has more explanatory power.
Current categorizations are, moreover, anthropocentric, i.e., they start with human
cognition and then categorize artifacts in terms of the cognitive process to which they
contribute. So, for example, Norman (1993) distinguishes between two types of
representational artifacts, those that aid experiential cognition and those that aid
reflective cognition. Likewise, Brey (2005) develops a brief taxonomy of represen-
tational artifacts on the basis of the cognitive process to which they contribute (i.e.,
memory, interpretation, searching, and conceptual thought). In this paper I take an
artifact-centered approach in categorizing cognitive artifacts: I take as my point of
departure the specific properties of cognitive artifacts and then categorize them on the
basis of those properties, not on the properties or goals of the agents that design,
make, or use them.
The properties I focus on are informational properties, because a cognitive function
supervenes on informational properties. Cognitive artifacts are used in virtue of the
information they provide. A map, for example, is used to navigate because the infor-
mation it provides is helpful for navigating. In other words, cognitive functions are
established only when an artifact exhibits an information-structure that is used in
performing some cognitive task. If there are no information-structures, then there are
no cognitive functions. Exhibiting an information-structure is thus a necessary and
sufficient condition for being a cognitive artifact. I distinguish between two kinds of
information-structures: representational and non-representational information-
structures. Representational artifacts contain information-structures about the world
(i.e. representational information), whereas non-representational artifacts contain
information-structures as the world (i.e. ecological information). These distinct infor-
mational properties are essential for better understanding the range and variety of
cognitive artifacts and are therefore further conceptualized below.
The proposed taxonomy is hierarchical and starts at a very general and inclusive level
and gets increasingly more specific and less inclusive when one goes deeper into the
taxonomy. I distinguish between three levels or taxa, those of family, genus, and species
(see Fig. 1below). The family includes all cognitive artifacts without further specifying
their informational properties. The family is thus a functional kind defined purely by the
cognitive function of the artifact. On the second level in the taxonomy, i.e., the taxon of
genus, I distinguish between two genera: representational and ecological artifacts. On
the third level, these two genera are further divided into species. In case of representa-
tional artifacts, three species are identified, those that are iconic, indexical, and symbolic.
In case of ecological artifacts, two species are identified, those that are spatial and
structural. Within species, token artifacts are identified.
In the above taxonomy, family, genus, and species membership are monothetic,
which means that one property is the determining factor for category membership. In
case of family, having a cognitive function is the determining property for
R. Heersmink
membership. In case of genus and species, exhibiting a particular informational
property is the determining factor for membership. So this taxonomy combines
functional and informational properties for category membership and is in that sense
eclectic. Finally, upward deduction in the taxonomy is valid, but downward deduction
is not. So, when membership at the species level is known, one can deduce the
membership of the higher levels, but not vice versa.
4 Representational Cognitive Artifacts
Some artifacts exhibit representational properties. But what exactly are representa-
tional properties? To answer this question, we need to take a closer look at what a
representation is. A useful starting point is the work of American pragmatist Charles
Saunders Peirce, who has argued that a representational system involves a triadic
relation between an interpretant (an understanding of the signs object), sign (repre-
sentation or representational vehicle), and object (represented world). A defining
property of a representation is that it stands in for something else. In John
Haugelands words: That which stands in for something else in this way is a
representation; that which it stands in for is its content; and its standing in for that
content is representing it(Haugeland 1991, p. 62, original italics). A representation
thus has aboutness in that its representational content is about something external to
the representation.
4.1 Icon, Index, and Symbol
According to Peirce, a representational system is irreducibly triadic, but for analytical
purposes, we can decompose Peirces triadic relation into two dyadic relations; one
between the representation and its object, and one between the human agent and the
representation. Let us first have a look at the representation-object relation. Peirce
Cognitive artifacts
Car keys
Fig. 1 A diagram of a taxonomy of cognitive artifacts, outlining three taxa: family, genus, and species
A Taxonomy of Cognitive Artifacts
(1935a,b) distinguished between three possible grounds for representation, which are
iconic, indexical, and symbolic (see also Von Eckhardt 1995). An icon is relevantly
similar or isomorphic to what it represents. Straightforward examples are maps, radar
systems, and blueprints, which are iconic because there is a high structural isomor-
phism between the content of the map, radar system, or blueprint and their objects.
Isomorphism is a quite general concept and because of its generality it can be
pushed rather far, as to include representations such as graphs and other diagrams.
Consider a line graph representing the amount of CO
in the earths atmosphere
plotted against the time. The amount of CO
has been increasing over the last hundred
and 50 years or so and hence the graph will show a line going up. There is indeed
some kind of isomorphism between an increasing amount of CO
particles in the
atmosphere over a certain period of time and a line going up, but this isomorphism is
of a very different kind than, for example, the isomorphism between a map and its
object which is much stronger. Isomorphism between properties of the representation
and its object may thus be abstract.
Indices have a direct causal connection between the index and its object. There is,
for instance, a direct causal connection between the direction of the wind and the
direction of a weathervane. If the direction of the wind changes, then the direction of
the weathervane automatically changes, too. So the position of the weathervane is an
index for the direction of the wind. Note that it is also partly iconic, since there is an
isomorphism between the direction of the wind and the direction of the weathervane.
Other examples of indices are thermometers, compasses, scales, voltmeters, speed
dials, barometers, and many other measuring instruments. These are indices because
there is a direct causal connection between the representational state of the index and
its object. In these indices, the following things are causally connected: temperature
and the expansion of mercury in a thermometer, the location of the North Pole and the
direction of an arrow in a compass, the mass of a particular object and the reading on
a scale, the amount of volts in an electrical current and the reading on a voltmeter, the
speed of a car and the reading on a carsspeed dial, and atmospheric pressure and the
reading on a barometer.
The last class of representations that Peirce identified are symbols, which acquire
their meaning and content through shared use, agreement, and logical rules. Typical
cases are words and sentences in natural language, numerals, or symbols in mathe-
matical and scientific formula. The form or structure of symbols is quite often (though
certainly not always) arbitrary. There is, for example, nothing intrinsic in the structure
of the word buildingsthat makes it represent the category of buildings, there is
likewise nothing intrinsic in the structure of the numeral 4that makes it represent 4
and there is nothing intrinsic in the structure of the sign for wavelength λ
that makes it represent wavelength. There is no resemblance or causal connection
between symbols and their object. It is shared use, agreement, and logical rules that
establish the meaning of symbols, not their representational structure.
Sometimes other kinds of representations have symbolic properties as well.
Certain icons such as diagrams cannot function as icons merely by their isomorphic
relations to their object. There needs to be an agreement or convention that indicates
which elements of the icon are to be interpreted as being isomorphic to its object.
Although this is different in the Roman numeral system, see Zhang and Norman (1995).
R. Heersmink
Measurement outcomes of certain indices are also partly symbolic, e.g., temperature
may be measured in degrees Celsius or Fahrenheit, depending on ones geographical
location and convention. Thus, as Peirce recognises, any particular representational
vehicle may display a combination of iconic, indexical, and symbolic characteristics.
Consequently, a useful way of conceptualizing Peirces trichotomy is by seeing a
token representation as predominantly iconic, indexical, or symbolic (Atkin 2008).
The power of Peirces trichotomy is that it can account for all forms of represen-
tation as predominantly iconic, indexical, or symbolic. Donald Peterson, in his edited
book, Forms of Representation: An Interdisciplinary Theme for Cognitive Science,
correctly points out that we use a great variety of representational systems, including:
algebras, alphabets, animations, architectural drawings, choreographic nota-
tions, computer interfaces, computer programming languages, computer models
and simulations, diagrams, flow charts, graphs, ideograms, knitting patterns,
knowledge representation formalisms, logical formalisms, maps, mathematical
formalisms, mechanical models, musical notations, numeral systems, phonetic
scripts, punctuation systems, tables, and so on(Peterson 1996, p. 7).
Peirces trichotomy can be used to categorize these representational systems.
Algebras, alphabets, computer programming languagesknowledge representation,
logical, or mathematical formalismsmusical notation systems, numeral systems,
phonetic scripts, punctuation systems, and tables are predominantly symbolic, as
they mainly acquire their meaning through logical rules and convention. But some
of these systems may also display isomorphism with their objects. For example, in the
case of musical notations, the order of the symbols for the notes is isomorphic with
the order of the notes in the actual piece of music.
Animations, architectural drawings, choreographic notations, computer models
and simulations, diagrams, flow charts, graphs, ideograms, knitting patterns, maps,
and mechanical models are predominantly iconic, as they acquire their meaning
through exhibiting some kind of isomorphism with their objects. However, some of
these systems, for example, architectural drawings, computer models, and diagrams
may also contain symbolic representations such as words, sentences, and numerals.
And, more importantly, there are rules for interpreting certain elements within those
representations as isomorphic to their objects. For example, the legend on a map often
contains rules and guidelines for interpreting certain elements in the map as being
isomorphic to certain objects and structures in the world.
Finally, computer interfaces do not really belong in that list, as they are not a form
of representation, but more a medium in which a variety of representational systems
can be expressed and computed. So, to conclude this subsection, iconic, indexical,
and symbolic cognitive artifacts are species of the genus of representational cognitive
artifacts and can account for all forms of representation.
4.2 Interpretation and Consumption
Having looked at different species of representational cognitive artifacts and some of
their informational properties, let us now have a closer look at the dyadic relation
between agent and representational vehicle. To establish a genuine triadic relation-
ship, an agent has to interpret or consume the representation. For this to happen, an
A Taxonomy of Cognitive Artifacts
agent has to understand the dyadic relationship between the representation and what it
represents. An agent thus has to understand the representation qua representation.
This means that an interpreter has to realize that icons display isomorphic relations to
their objects, that indices have causal relations to their objects, and that symbols are
based on rules and conventions. So when I perceive a map of the London subway
system, I understand that I perceive a predominantly iconic representation. Due to the
properties of the vehicle, I understand that the London subway system is an existing
structure in the world and because of this understanding I am able to form beliefs
about certain properties of the London subway system (e.g. the order of the stations).
Thus representations are psychologically efficacious in that they cause beliefs about
the content of the vehicle. In this sense, they causally and informationally mediate
between agent and world.
Representations can be misinterpreted, which will hinder cognitive performance.
This occurs when the vehicle accurately captures its object, but for whatever reason,
the agent misinterprets the vehicle and attributes certain properties to its object that do
(in fact) not exist, resulting in false beliefs about aspects of its object. In such cases,
there is no genuine triadic relationship between agent, representation, and world, but
a mere dyadic relationship between agent and representation. It is of course possible
that an agent only misinterprets certain elements of the representation. For example,
when navigating the London subway system with the aid of a map (not an easy
cognitive task for a tourist), I may correctly interpret my current location on the map,
but misinterpret how many stations it is from my current location to Piccadilly Circus
and, consequently, form a false belief about its location. In this example, there is a
mixture of a true and false belief, but due to the true belief, a genuine triadic relation
is established.
5 Ecological Cognitive Artifacts
Having outlined the genus of representational cognitive artifacts, I now focus on the
genus of ecological cognitive artifacts. These are characterized not by exhibiting
information about the world, but as the world and thus exhibit non-representational
or ecological information. To give an example: When playing Tetris, one has to rotate
a geometrical shape called a zoidso that it fits into a specific socket in the lower
regions of a geometrical template. One can either choose to mentally rotate the zoid or
to rotate it by means of a button-push. Experienced players have learned to rotate the
zoid by means of a button-push, not only because it is significantly faster, but also
because it relieves the brain from performing mental rotation (Kirsh and Maglio
1994; see also Clark and Chalmers 1998). The artificial rotation of a zoid in playing
Tetris clearly has a functional role in performing a cognitive task. This functional
role, however, is not established by exhibiting representational properties. Zoids do
not stand in for something else, they have no representational content, and neither
causally nor informationally mediate between agent and world. They are mere
abstract geometrical shapes that are interacted with directly, without them
representing something outside the game. So, rather than triadic situated cognitive
systems (agent-representational artifact-world), ecological artifacts concern dyadic
situated cognitive systems (agent-artifact). The functional role of the artifactual
R. Heersmink
elements in such dyadic systems is thus established through non-representational or
ecological information. But what exactly is ecological information? How can it aid
their users in performing a cognitive task? And how can it be categorized?
The work of cognitive scientist David Kirsh (1995,2006,2009;Hollanetal.2000)is
a useful starting point in answering these questions. In his seminal paper, Kirsh (1995)
makes a tripartite distinction between spatial arrangements that simplify choice, percep-
tion, or internal computation. Although this a helpful and insightful categorization, it is
(like all other currently existing categorizations) anthropocentric, i.e., it starts with a
cognitive agent and then categorizes artifacts on the basis of the cognitive processes to
which they contribute. In the taxonomy I am developing in this paper, I take an artifact-
centered approach and taxonomize cognitive artifacts on the basis of their informational
properties. With this non-anthropocentric approach in mind, I recycle and reclassify
some of Kirshs examples and focus on their informational properties, rather than on the
cognitive processes to which they contribute. By doing so, I distinguish between two
species of ecological artifacts, those that obtain their cognitive function in virtue of
physical-spatial or virtual-spatial structures and those that obtain their cognitive function
in virtue of manipulable physical or virtual structures.
5.1 Spatial Ecological Cognitive Artifacts
Human agents quite frequently make use of space for cognitive purposes. Kirsh refers
to this as the intelligent use of space, which is so commonplace, he argues, that we
should not assume that such cognitive or informational structuring is not taking place
all the time(Kirsh 1995, p. 33). The intelligent use of space enables us to encode
important information into artifacts that are typically neither designed nor made for
cognitive purposes and thus mainly (though not solely) concerns improvised uses of
artifacts. Some straightforward examples include consistently leaving car keys on a
certain spot in your apartment so that you know where they are, putting an article you
have to read on top of the pile on your desk, leaving a book open and turned upside
down so that you know where you have stopped reading, tying a string around your
finger as a reminder, or leaving a rented DVD on your desk as a prompt to bring it
back to the video store. By putting artifacts in certain locations that are either
deliberately usual or deliberately unusual, we intentionally encode information into
the artifact and its location, thereby creating what may be referred to as spatial
ecological cognitive artifacts.
Consider another example: When doing the dishes, it is not always clear which
items have been washed and which ones have not. In order to keep track of the items
that have been washed, it is helpful to create spatial categories of washed itemsand
unwashed itemsby putting them in certain locations. Most kitchen sinks have
designated areas for items that have been recently washed. These areas are not just
practical, so that residual dishwater can drip away without spilling it on the kitchen
counter, but also have cognitive functions, as they simplify perception and reduce
memory load in a task.
Kirsh (1995) describes a more idiosyncratic case of someone who is dismantling a
bicycle and then puts certain parts on a sheet of newspaper placed on the floor. The
newspaper demarcates a spatial boundary within which certain items are placed, in that
way structuring the task space and making items easier to locate. The user may also
A Taxonomy of Cognitive Artifacts
place the dismantled items in such a way that, when reassembling the bicycle, the items
that need to be reassembled first are located closer to the user than the items that are
reassembled later in the process. The virtue of spatially decomposing the task is that
one need not consult a plan, except at the very highest level, to know what to do. Each
task context affords only certain possibilities for action(Kirsh 1995, p. 44). Kirsh here
is referring to a mental plan, but it could also be an external one, e.g., an assembly guide,
manual, or blueprint of some kind. There is no need for a manual (or an elaborate mental
plan) if all the parts are placed such that they correspond to the correct order of actions
for reassembling the bicycle. Spatial structuring of artifacts thus makes both complex
internal representations and external representations superfluous. It is much more
efficient to spatially structure the artifactual task environment such that it affords the
most efficient and environmentally-embedded plan.
The intelligent use of space, however, is not restricted to physical or actual space,
as it also frequently occurs in virtual space. In connection to Human-Computer
Interaction, Hollan, Hutchins, and Kirsh provide some suitable examples. Computer
users, they write;
Leave certain portals open to remind them of potentially useful information or
to keep changes nicely visualized; they shift objects in size to emphasize their
relative importance; and they move collections of things in and out of their
primary workspace when they want to keep certain information around but have
other concerns that are more pressing(Hollan et al. 2000, p. 190).
So the way we spatially organize the items and structures on our screen helps us in
performing certain computer-related tasks. We may, for example, prioritize certain
information by leaving certain portals open or by making them larger than other
portals. We may also organize the items on our desktop or in our navigation menus
such that they reflect their importance. Often used items typically inhabit a more
prominent position than items that are used less often (for example, by placing them
in preferred positions on ones desktop or menu) thereby making them easier to
locate. These everyday examples show that we organize artifactual elements (physical
or virtual) within space (actual or virtual) for cognitive purposes, thereby encoding
important information into physical-spatial structures or virtual-spatial structures.
5.2 Structural Ecological Cognitive Artifacts
Some artifacts obtain their cognitive function in virtue of their manipulable physical
or virtual structure. During or after the manipulation of such artifacts, new informa-
tion emerges that is important for performing a cognitive task. For example, when
rearranging letter tiles in Scrabble to prompt word recall, new information emerges
from their spatial configurations (Kirsh 2009). In this case, the non-representational
information of the tiles (i.e. individual letters), their spatial location in relation to the
other tiles, and the words and openings on the board are important for performing the
task. Novel and larger information-structures emerge when the tiles are rearranged,
which (ideally) prompts the recall of words with as many letters as possible and that
fit into existing letter structures on the board. Kirsh points out that one of the purposes
for rearranging letters tiles, is to generate and draw attention to often occurring two-
and three-letter combinations that figure in words such as, for instance, ES,TH,
R. Heersmink
IN,REA, et cetera. Note that such information (i.e. letter combinations) are often
not representational in nature, although there is a symbolic element to it, because
there are logical rules and conventions for creating words out of letters. So perhaps
letter tiles can be seen as sub-representational, as they are the building blocks of
proper representations. The point here is that the cognitive function of the artifacts
supervenes on their manipulable physical structures and by manipulating their phys-
ical structure one automatically manipulates the information they contain, too.
Virtual structures, too, can be manipulated such that new information emerges
from their spatial configurations. The functional role of the artificial rotation of a zoid
in playing Tetris is established due to a constantly manipulable virtual structure, i.e.,
because the zoids can be manipulated, a user delegates rotation to the computing
device, in that way enabling a user to decide quicker whether or not it fits into a
socket in the lower regions of the task space. Like with rearranging Scrabble tiles,
new task-relevant information emerges only in relation to some other structure. In
Scrabble, new information emerges when two or more tiles are positioned in a certain
way. In Tetris, new information emerges when the zoid is spatially orientated in
relation to a template in a certain way. The spatial orientation of the zoid in itself, i.e.,
without taking into account the structure of the template in which it has to fit, is not
sufficient for performing the task. Both zoid and template are important for generat-
ing the task-relevant information. Finally, the above artifacts are much more dynamic
than cases of spatial ecological cognitive artifacts. They are not about developing
static spatial categories in which artifacts are placed, but about dynamically and
constantly changing information in an ongoing task.
6 Representational and Ecological Artifacts
Ecological artifacts, I think, are instances of what roboticist Rodney Brooks (1999)
calls the world is its own best modeland what Clark (1989) calls the 007
principle. The point of these notions is: Why create an expensive internal represen-
tation of the world, when you can use the world itself as a model? In a similar way we
may ask: Why create an expensive external representation of the world, when you can
use the world itself to interact with and use it to facilitate your cognitive tasks? One
could, for example, consult a manual for how to reassemble a dismantled bicycle. But
one can also structure the items such that they facilitate the reassembling process, in
that way streamlining the task and making a cost-expensive external representation
superfluous. In Clarks words, evolved creatures will neither store nor process
information in costly ways when they can use the structure of the environment and
their operations upon it as a convenient stand-in for the information-processing
operations concerned(Clark 1989, p. 64).
For analytical purposes, I presented representational and ecological artifacts as
distinct genera. Up to this point, a reader may get the impression that a cognitive
artifact either exhibits representational or ecological information. This analytical
distinction was helpful in that it allowed me to emphasize and conceptualize their
distinct cognition-aiding informational properties. However, in some cases, cognitive
artifacts exhibit a combination of representational and ecological properties. To give
an example: In a bookcase in which the books are alphabetically organized,
A Taxonomy of Cognitive Artifacts
representational and ecological structures jointly facilitate a perceptual task. The
representational structures are the names and titles on the back of the books and the
ecological properties are the spatial order in which they have been placed. In this
example, the representational properties determine the spatial structure, which in turn,
supports the representational structure. So alphabetically organized books may be
seen as a predominantly representational cognitive artifactual structure. The genera of
representational and ecological artifacts are thus not mutually exclusive.
7 Conclusion
This paper has developed a taxonomy in which cognitive artifacts with similar
informational properties can be grouped into categories. An artifact-centered ap-
proach was taken: I took as my point of departure the specific informational proper-
ties of cognitive artifacts and then categorized them on the basis of those properties,
not on the properties or goals of the agents that design, make, or use them. In the
developed taxonomy, I distinguished between three taxa, those of family, genus, and
species. This taxonomy is an important first step towards a better understanding of the
variety and range of cognitive artifacts and their distinct cognition-aiding properties.
Acknowledgments I would like to thank my supervisors John Sutton and Richard Menary, Peter Woelert,
and two anonymous reviewers for helpful comments on an earlier version of this paper. I would also like to
thank Sadjad Soltanzadeh for inviting me to present this paper at a colloquium at the Centre for Applied
Ethics and Public Philosophy (CAPPE) at Charles Sturt University in Canberra as well as the audience,
particularly Steve Clarke.
Atkin, A. 2008. Icon, index, and symbol. In The Cambridge encyclopaedia of language sciences, ed. P.
Hogan, 367368. Cambridge University Press.
Brey, P. 2005. The epistemology and ontology of human-computer interaction. Minds & Machines 15(34):
Brooks, R. 1999. Cambrian intelligence: The early history of the new AI. MIT Press.
Carrara, M., and P. Vermaas. 2009. The fine-grained metaphysics of artifactual and biological functional
kinds. Synthese 169(1): 125143.
Clark, A. 1989. Microcognition: Philosophy, cognitive science and parallel distributed processing.MITPress.
Clark, A. 2003. Natural born cyborgs: Minds, technologies, and the future of human intelligence. Oxford
University Press.
Clark, A. 2004. Towards a science of the biotechnological mind. In Cognition and technology: Co-
existence, convergence, and co-evolution, ed. B. Gorayska and J. Mey, 2536. John Benjamins
Publishing Company.
Clark, A. 2008. Supersizing the mind: Embodiment, action, and cognitive extension. Oxford University
Clark, A., and D. Chalmers. 1998. The extended mind. Analysis 58: 1023.
Dipert, R. 1993. Artifacts, art works, and agency. Temple University Press.
Donald, M. 1991. Origins of the modern mind: Three stages in the evolution of culture and cognition.
Harvard University Press.
Dourish, P. 2001. Where the action is: The foundations of embodied interaction. MIT Press.
Harris, C.B., P.G. Keil, J. Sutton, A.J. Barnier, and D.J.F. McIlwain. 2011. We remember, we forget:
Collaborative remembering in older couples. Discourse Processes 48(4): 267303.
R. Heersmink
Haugeland, J. 1991. Representational genera. In Philosophy and connectionist theory, ed. W. Ramsey, S.
Stich, and D. Rumelhart, 6189. Lawrence Erlbaum Associates.
Heersmink, R. 2012. Mind and artifact: A multidimensional matrix for exploring cognition-artifact re-
lations. In Proceedings of the 5th AISB Symposium on Computing and Philosophy, ed. J.M. Bishop and
Y.J. Erden, 5461.
Heersmink, R. 2013. Embodied tools, cognitive tools, and brain-computer interfaces. Neuroethics 6(1):
Hilpinen, R. 2011. Artifact. Stanford Encyclopaedia of Philosophy. Retrieved 23 February 2012, from
Hollan, J., E. Hutchins, and D. Kirsh. 2000. Distributed cognition: Toward a new foundation for human-
computer interaction research. Transactions on Computer-Human Interaction 7(2): 174196.
Houkes, W., and P. Vermaas. 2004. Actions versus functions: A plea for an alternative metaphysics of
artifacts. The Monist 87(1): 5271.
Houkes, W., and P. Vermaas. 2010. Technical functions: On the use and design of artefacts. Springer.
Hutchins, E. 1995. Cognition in the wild. MIT Press.
Hutchins, E. 1999. Cognitive artifacts. In The MIT encyclopedia of the cognitive sciences, ed. R.A. Wilson
and F.C. Keil, 126128. MIT Press.
Kirchhoff, M.D. 2011. Extended cognition and fixed properties: Steps to a third-wave version of extended
cognition. Phenomenology and the Cognitive Sciences 11(2): 287308.
Kirsh, D. 1995. The intelligent use of space. Artificial Intelligence 72: 3168.
Kirsh, D. 2006. Distributed cognition: A methodological note. Pragmatics and Cognition 14(2): 249262.
Kirsh, D. 2009. Problem-solving and situated cognition. In The Cambridge handbook of situated cognition,
ed. P. Robbins and M. Aydede, 264306. Cambridge University Press.
Kirsh, D., and P. Maglio. 1994. On distinguishing epistemic from pragmatic action. Cognitive Science 18:
Kornblith, H. 1980. Referring to artifacts. Philosophical Review 89(1): 109114.
Kroes, P. 2012. Technical artefacts: Creations of mind and matter. Springer.
Menary, R. 2007. Cognitive integration: Mind and cognition unbounded. Palgrave McMillan.
Nersessian, N.J. 2005. Interpreting scientific and engineering practices: Integrating the cognitive, social,
and cultural dimensions. In New directions in scientific and technical thinking, ed. M. Gorman, R.
Tweney, D. Gooding, and A. Kincannon, 1756. Erlbaum.
Nersessian, N.J., E. Kurz-Milcke, W.C. Newstetter, and J. Davies. 2003. Research laboratories as evolving
distributed cognitive systems. In Proceedings of the 25th Annual Conference of the Cognitive Science
Society, ed. R. Alterman and D. Kirsh, 857862.
Norman, D. 1991. Cognitive artifacts. In Designing interaction: Psychology at the human-computer
interface, ed. J.M. Carroll, 1738. Cambridge University Press.
Norman, D. 1993. Things that make us smart: Defending human attributes in the age of the machine. Basic
Peirce, C.S. 1935a. The collected papers of Charles S. Peirce vol 2. Harvard University Press.
Peirce, C.S. 1935b. The collected papers of Charles S. Peirce vol 3. Harvard University Press.
Peterson, D. 1996. Introduction. In Forms of representation: An interdisciplinary theme for cognitive
science, ed. D. Peterson, 727. Intellect Books.
Preston, B. 1998. Why is a wing like a spoon? a pluralist theory of function. The Journal of Philosophy
95(5): 215254.
Preston, B. 2009. Philosophical theories of artifact function. In Philosophy of technology and engineering
sciences, ed. A. Meijers, 213234. Elsevier.
Preston, B. 2013. A philosophy of material culture:Action, function, and mind. Routledge.
Robbins, P., and M. Aydede (eds.). 2009. The Cambridge handbook of situated cognition. Cambridge
University Press.
Sutton, J. 2006. Distributed cognition: Domains and dimensions. Pragmatics and Cognition 14(2): 235247.
Sutton, J. 2010. Exograms and interdisciplinarity: History, the extended mind and the civilizing process. In
The extended mind, ed. R. Menary, 189225. MIT Press.
Theiner, G. 2013. Transactive memory systems: A mechanistic analysis of emergent group memory. Review
of Philosophy and Psychology 4(1): 6589.
Von Eckhardt, B. 1995. What is cognitive science? MIT Press.
Zhang, J., and D. Norman. 1995. A representational analysis of numeration systems. Cognition 57: 271295.
A Taxonomy of Cognitive Artifacts
... Cognitive Artefacts refer to "artificial devices designed to maintain, display, or operate upon information in order to serve a representational function and that affect human cognitive performance" (Norman, 1991). Using products inevitably involves users' cognition, but only those that functionally contribute to cognitive tasks, or those that "help people deal with information processing related tasks" legitimise as Cognitive Artefacts (Heersmink, 2013;Morana et al., 2010). For example, a memo becomes an Cognitive Artefact when it is used by people as a reminder since it helps offload the users' memory. ...
... Cognitive Artefact is one of the pioneering design theories that integrate theories of human cognition in to product interactions (Norman, 1991 (Heersmink, 2013) tributed Cognition which suggests that human cognition is a composite of personal aptitude, computational tools, and embedded socio-cultural organisations (Hutchins, 1995). The dominant reductionism approach considers user cognition isolated from the external world, leading to the inadequate understanding of how a product interacts with the mind (Norman, 1991). ...
... In addition, the concept of Cognitive Artefact was initially raised for "physical objects", and its explanatory power is compromised when moving to the design of digital interfaces (e.g. Calcagno and Collavizza, 2018;Heersmink, 2013). ...
The advance in digital technologies is placing increasingly high demands on users’ cognitive capabilities, resulting in ubiquitous cognitive challenges (e.g. memory overload) in everyday interaction with products. Such challenges can be attributed to the inadequate understanding of cognitive capability for informing design. The literature review further specifies that current research has largely ignored how user cognition is understood and implemented by designers. This study thus aims to advance the understanding of user cognition by adding the designers’ perspective. This research adopts a mixed method approach, which helps highlight and address the designerly ways of knowing as well as the implicit nature of cognition in design. The research framework consists of three phases. First, the language differences between design practise and cognitive research were confirmed by interviewing designers and analysing designers’ articles. Second, a review of design languages of cognition and ethnographic fieldwork were conducted to explore the designerly ways of understanding and implementing cognition. Finally, the findings were tested in a larger-scale survey. By explicating the related design languages, this research elucidates designers’ Micro, Meso, and Macro perspectives on user cognition, thereby expanding the current rational information-processing conception of cognition to a more context-based understand- ing. The multiple perspectives also add more embodied and situated considerations to the conventional focus on elementary information processing capacity, advancing the existing understanding of users’ cognitive capabilities. Hence, this study enhances the predominant way of conceptualising user cognition in design research and empirically supports the postulated situated nature of cognition in cognitive research. Furthermore, it recognises a specific application field within the broader landscape of cognitive research, suggesting that designing interactions has its own research agenda rather than relying on other established sub-areas.
... Many studies, both conceptual and empirical, in the situated and 4E cognition literature focus on our relation to particular artifacts (e.g., Brey, 2005;Clark & Chalmers 1998;Hutchins, 1995;Kirsh & Maglio 1994;Sutton, 2010). There are also taxonomies of particular classes of artifacts such as cognitive artifacts (Brey, 2005;Casati, 2017;Fasoli, 2018a;Heersmink, 2013a;Nickerson, 2005). However, a comprehensive general taxonomy of artifacts and the way they are co-opted by the human embodied mind is lacking. ...
... Cognitive artifacts are human-made, material objects or structures that functionally contribute to performing a cognitive task (Heersmink, 2013a(Heersmink, , 2016compare Vaccari, 2016). Examples of cognitive artifacts include maps, labels, abacuses, diaries, indexes, search engines, diagrams, models, computer systems, and leaving a rented DVD on your desk as a reminder to bring it back to the video store. ...
... By putting these artifacts in certain locations that are either deliberately usual or deliberately unusual, I intentionally encode information into the artifact and its location. Heersmink (2013a) has referred to such artifacts as "ecological cognitive artifacts." These aid us in performing cognitive tasks such as remembering, but not by having representational properties in the Peircean sense. ...
Full-text available
The primary goal of this essay is to provide a comprehensive overview and analysis of the various relations between material artifacts and the embodied mind. A secondary goal of this essay is to identify some of the trends in the design and use of artifacts. First, based on their functional properties , I identify four categories of artifacts co-opted by the embodied mind, namely (a) embodied artifacts, (b) perceptual artifacts, (c) cognitive artifacts, and (d) affective artifacts. These categories can overlap and so some artifacts are members of more than one category. I also identify some of the techniques (or skills) we use when interacting with artifacts. Identifying these categories of artifacts and techniques allows us to map the landscape of relations between embodied minds and the artifactual world. Second, having identified categories of artifacts and techniques, this essay then outlines some of the trends in the design and use of artifacts, focusing on neuroprosthetics, brain-computer interfaces, and personalization algorithms nudging their users toward particular epistemic paths of information consumption.
... When used in this context, the robots do not primarily focus on social interactions with the learners. They, instead, represent pedagogical instruments that are mainly used as cognitive artifacts, i.e., "human-made physical objects that functionally contribute to performing a cognitive task" (Heersmink, 2013). In fact, such "objects-to-think-with", represent a key element of Papert's constructionism (Papert, 1980), which has been considered the historical origin of ER. ...
... Along these lines and as an approach to taxonomize cognitive artifacts, Heersmink (2013) has introduced the notion of situated cognitive systems. In his work, Heersmink presented three possible configurations for such systems: they can be composed by only human agents, by human agents and non-artifactual objects, or by human agents and cognitive artifacts. ...
... However, many times, ER activities also involve other objects supporting the human agents in performing the task: apart from the robots, they also often involve programming or interaction interfaces and playgrounds for task accomplishment. According to the taxonomy of Heersmink (2013), each of these objects can be classified as a specific type of cognitive artifact. ...
The technological advances of the past years have impressively demonstrated that the digital age is no longer just a science fiction vision of the future - we are in the midst of it. The digital transformation is affecting many areas of our lives, including the educational system. For instance, recent years have shown an increased adoption of educational robotics activities in classrooms. Although educators and researchers have acknowledged the potential of educational robotics, further efforts are needed to effectively integrate them into formal education. As for any kind of educational tool, one key factor for the successful integration of educational robots in classrooms is the proper alignment of such tools with the classroom activities and vice versa. In practice, however, this may pose challenges. On the one hand, it still appears that only few educators have already developed the required know-how allowing them to leverage educational robotics for classroom activities. On the other hand, the issue of poor alignment may already arise in the design stages of educational robotics tools, due to the limited experience of developers with the pedagogy and learning theories related to such tools. This thesis, therefore, seeks to study how instructional alignment can be attained in the context of educational robotics. To this end, an alignment framework will be devised aimed at supporting developers in the design of educational robotics tools, as well as educators in the design and implementation of classroom activities involving educational robotics. In this context, we will introduce the notion of Educational Robotics Learning Systems (ERLS), a model that conceptualizes alignment for educational robotics activities. In different studies, we will illustrate how the devised alignment framework can be used to guide the development of educational robotics tools and activities. The PaPL and CreroBot projects will consider the alignment issue from the developers' perspective. The former is concerned with the development of a tangible programming interface based on accessible materials such as paper and cardboard. The latter will then expand these ideas to the development of a do-it-yourself educational robot. In two other projects, we will then illustrate how the devised framework can be used to guide the development of educational robotics activities. We will first discuss the Thymio lawnmower mission, an activity devised to foster the development of students' computational thinking competencies. Finally, we will present the Thymio Escape Game, an immersive learning activity that has been inspired by escape room experiences. The findings from these studies are intended to make a contribution to the field of educational robotics, specifically by addressing the global research question of this thesis: How can instructional alignment be attained for educational robotics activities?
... Many theorists posit that the mind acquires much of its power from its ability to create and interface with external objects, props, aids, scaffolding, and tools [1,2,4,7,[22][23][24][25][26][27][28][29][30]. Although other species act upon their environment in ways that may feed back to influence cognition, such as when ants leave pheromone trails that guide themselves and other ants [31][32][33], humans appear to flexibly and intentionally exploit the environment for cognitive ends in unparalleled fashion [34,35]. ...
... memory and memory strategies [36][37][38][39][40][41][42][43][44][45][46][47], our findings empirically chart how human children first begin to augment their cognitive processing with external support to achieve their goals. Children used the tokens and pens to create external representations of hidden stickers [26], thereby transforming otherwise inert objects into cognitive artifacts: tools to aid in cognitive processes [2,48,49]. By setting reminders in such fashion, people establish ''retrieval cues for communicating with themselves in the future'' [50]. ...
Metacognition plays an essential role in adults’ cognitive offloading decisions. Despite possessing basic metacognitive capacities, however, preschool-aged children often fail to offload effectively. Here, we introduced 3- to 5-year-olds to a novel search task in which they were unlikely to perform optimally across trials without setting external reminders about the location of a target. Children watched as an experimenter first hid a target in one of three identical opaque containers. The containers were then shuffled out of view before children had to guess where the target was hidden. In the test phase, children could perform perfectly by simply placing a marker in a transparent jar attached to the target container prior to shuffling, and then later selecting the marked container. Children of all ages used this strategy above chance levels if they had seen it demonstrated to them, but only the 4- and 5-year-olds independently devised this external strategy to improve their future performance. These results suggest that, when necessary for optimal performance, even 4- and 5-year-olds can use metacognitive knowledge about their own future uncertainty to deploy effective external solutions. This article is protected by copyright. All rights reserved
... I shall analyse the original debate and the early developments on mindshaping, considering my proposal a further advancement in the field, complementary to them. Afterwards ( § § 5-6), I argue that whenever we design an imaginal space, we create public representations of stories: we design an ecological niche (Odling-Smee, Laland, and Feldman 2003) with both narrative and descriptive features by means of the representational artefacts (Heersmink 2013;Norman 1991) we publicly manipulate in our environment. ...
... I stress here the importance of physical space because, just as a cultural niche interpenetrates an ecological niche(Laland, Matthews, and Feldman 2016), there is an interpenetration between a physical and an imaginal space: indeed, images in the niche are supported by representational artefacts(Heersmink 2013). ...
Full-text available
In this article I shall argue that the categories a subject employs to codify her perceptions are emergent elements of the social niche her community inhabits. Hence, I defend the claim that categories are primarily elements of the social ontology a certain subject experiences. I then claim that public representations (e.g. icons) shared in a social niche play a crucial regulative role for the members of that community: in fact, they offer a rule (a canon) to conceive a certain type or a certain category, e.g. ‘movement’, ‘time’ or ‘space’. In this sense, categories function as normative elements.
... As suggested by Risko and Gilbert (2016), cognitive offloading can be subdivided into actions that offload cognitive demands into the world and onto the body. The former can correspond to leaving car keys in a certain spot in the house to remember where they are (Heersmink, 2013;Kirsh, 1995) and the latter to tilting the head to perceive a rotated image (Clark & Chalmers, 1998;Risko, Medimorec, Chisholm, & Kingstone, 2014). The use of cognitive tools (e.g., diary, calculator, smartphone) represents instances of cognitive offloading, thereby modifying our way of interacting with the environment. ...
Cognitive tools (e.g., calculators) provide all users with the same potential. Yet when people use such cognitive tools, interindividual variations are observed. Previous findings have indicated that 2 main factors could explain these variations: intrinsic cognitive skills (i.e., the “non–tool use” cognitive skills associated with the task targeted) and metacognitive beliefs about one's performance with tool use. In this study we sought to reproduce these findings and to investigate in more detail the nature of the relationships (i.e., linear vs. exponential) between tool use performance and intrinsic cognitive skills. In Experiment 1, 200 participants completed 2 cognitive tasks (calculation and geography) in 2 conditions (non–tool use vs. tool use). In Experiment 2, 70 participants performed a geography task in 2 conditions (non–tool use vs. tool use) and estimated their performance in each condition before completing the task. Results indicated that intrinsic cognitive skills and, to a lesser extent, metacognitive beliefs improved tool use performance: The higher the intrinsic cognitive skills and the higher participants estimated their tool use performance, the higher this tool use performance was. The nature of the relationship between tool use performance and intrinsic cognitive skills appeared to be linear rather than exponential. These findings extend previous research showing a strong impact of intrinsic cognitive skills on the performance associated with the use of cognitive tools or external aids.
... There have been a number of analyses of cognitive tools. For example, those by: Clark (1997), Heersmink (2013), Humphreys (2004), Hutchins (1995), Norman (1991Norman ( , 1993, and Menary (2007aMenary ( , 2018b). Norman's influential analysis introduces the important concept of how cognitive tools change the nature of the task being performed and not just our ability to complete the task. ...
Full-text available
We propose an account of cognitive tools that takes into account the process of enculturation by which tools are integrated into our cognitive systems. Drawing on work in cultural evolution and developmental psychology, we argue that cognitive tools are complex entities consisting of physical objects, representational systems, and cognitive practices for the physical manipulation of the tool. We use an extensive case study of spatial navigation to demonstrate the core claims. The account we provide is contrasted with conceptions of cognitive tools that simplify cognition, in particular that they offload cognitive work, or that the tools themselves are temporary developmental scaffolds or props. Enculturation results in transformed cognitive systems, and we can now think and act in new ways with cognitive tools.
... To get a better grip on the intricacies of how this notion of transparency relates to AI systems, it will be useful to employ Richard Heersmink's (2013) distinction between procedural and representational transparency: a distinction that Heersmink introduces in his taxonomy of dimensions of integration of informational technologies into an agent's cognitive system. 24 Heersmink (2013) refers to representational transparency as the ease with which a user can interpret the representational system or arrangement of an informational device. This concerns being able to read some sort of abstract pattern or interpret a representational system. ...
Full-text available
AI systems play an increasingly important role in shaping and regulating the lives of millions of human beings across the world. Calls for greater transparency from such systems have been widespread. However, there is considerable ambiguity concerning what “transparency” actually means, and therefore, what greater transparency might entail. While, according to some debates, transparency requires seeing through the artefact or device, widespread calls for transparency imply seeing into different aspects of AI systems. These two notions are in apparent tension with each other, and they are present in two lively but largely disconnected debates. In this paper, we aim to further analyse what these calls for transparency entail, and in so doing, clarify the sorts of transparency that we should want from AI systems. We do so by offering a taxonomy that classifies different notions of transparency. After a careful exploration of the different varieties of transparency, we show how this taxonomy can help us to navigate various domains of human–technology interactions, and more usefully discuss the relationship between technological transparency and human agency. We conclude by arguing that all of these different notions of transparency should be taken into account when designing more ethically adequate AI systems.
... While there is a long and rich history of research on metacognition, much of it has focused on cognitive acts that rely largely on internal processes (e.g., recalling items studied from a list then estimating the number correct). However, individuals frequently integrate "external" aids into cognitive acts (Clark, 1997;Gray et al., 2006;Heersmink, 2013;Hutchins, 1995;Michaelian & Sutton, 2013;Risko & Gilbert, 2016;Sterelny, 2010;Sutton, 2010;Wilson, 2002;Wilson & Clark, 2008). For instance, individuals offload a tremendous amount of information retrieval onto the Internet with various consequences (Ferguson et al., 2016;Fisher et al., 2015;Risko & Gilbert, 2016;Sparrow et al., 2011;Ward, 2013). ...
Full-text available
Our metacognitive ability to monitor and evaluate our cognitive performance is central to efficient and adaptive behaviors. Research investigating this ability has focused largely on tasks that rely exclusively on internal processes (e.g., memory). However, our day-today cognitive activities often consist of mixes of internal and external processes. In the present investigation, we expand research on metacognition to this distributed domain. We examined participants' ability to accurately monitor their performance in a knowledge retrieval task when they were required to rely on only their internal knowledge and when required to rely on both internal knowledge and utilizing the Internet. One hundred and ninety-four participants completed an online study consisting of answering general knowledge questions. Individuals were also randomly assigned to provide accuracy judgments either prospectively or retrospectively. Results revealed metacognitive bias (i.e., overconfidence) increased when using the Internet and when making retrospective judgments. Metacognitive sensitivity was also worse when using the Internet, especially when individuals made prospective judgments about what their performance would be. Furthermore, metacognitive bias was positively related across the internal knowledge and Internet conditions. These results provide the beginnings of an understanding of metacognition and behavior in distributed cognitive contexts involving the Internet.
Since the middle of the last century, normative language has been much studied. In particular, the normative function performed by certain sentences and by certain speech acts has been investigated in depth. Still, the normative function performed by certain physical artifacts designed and built to regulate human behaviors has not yet been thoroughly investigated. We propose to call this specific type of artifacts with normative intent ‘deontic artifacts’. This article aims to investigate this normative phenomenon that is so widespread in our daily reality, but so often forgotten by scholars of norms and normativity.
Where are the borders of mind and where does the rest of the world begin? There are two standard answers possible: Some philosophers argue that these borders are defined by our scull and skin. Everything outside the body is also outside the mind. The others argue that the meanings of our words "simply are not in our heads" and insist that this meaning externalism applies also to the mind. The authors are suggesting a third position, i.e. quite another form of externalism. Their so called active externalism implies an active involvement of the background in controlling the cognitive processes.
This book proposes a theory of human cognitive evolution, drawing from paleontology, linguistics, anthropology, cognitive science, and especially neuropsychology. The properties of humankind's brain, culture, and cognition have coevolved in a tight iterative loop; the main event in human evolution has occurred at the cognitive level, however, mediating change at the anatomical and cultural levels. During the past two million years humans have passed through three major cognitive transitions, each of which has left the human mind with a new way of representing reality and a new form of culture. Modern humans consequently have three systems of memory representation that were not available to our closest primate relatives: mimetic skill, language, and external symbols. These three systems are supported by new types of ''hard'' storage devices, two of which (mimetic and linguistic) are biological, one technological. Full symbolic literacy consists of a complex of skills for interacting with the external memory system. The independence of these three uniquely human ways of representing knowledge is suggested in the way the mind breaks down after brain injury and confirmed by various other lines of evidence. Each of the three systems is based on an inventive capacity, and the products of those capacities - such as languages, symbols, gestures, social rituals, and images - continue to be invented and vetted in the social arena. Cognitive evolution is not yet complete: the externalization of memory has altered the actual memory architecture within which humans think. This is changing the role of biological memory and the way in which the human brain deploys its resources; it is also changing the form of modern culture.
Function theory is a growth industry in philosophy of biology, but artifact function has received comparatively little attention. Perhaps the most important factor contributing to this neglect is the lack-both historical and contemporary-of a philosophical literature dedicated to the study of artifacts. There is no philosophy of material culture in general, and thus no breeding ground for theories of artifact function. Artifact function has not emerged as a major topic of discussion in either of these areas, despite their interest in specific aspects of material culture. Important factor contributing to the neglect of artifact function is the widespread perception that it is unproblematic in comparison to biological function. The background intuition here is that because of the uncontroversial existence of intelligent makers in the artifact case, what the function of an artifact is and how that function is established can be easily traced to the purposes of the human agents who made and/or used it.
This book focuses on material culture as a subject of philosophical inquiry and promotes the philosophical study of material culture by articulating some of the central and difficult issues raised by this topic and providing innovative solutions to them, most notably an account of improvised action and a non-intentionalist account of function in material culture. Preston argues that material culture essentially involves activities of production and use; she therefore adopts an action-theoretic foundation for a philosophy of material culture. Part 1 illustrates this foundation through a critique, revision, and extension of existing philosophical theories of action. Part 2 investigates a salient feature of material culture itself its functionality. A basic account of function in material culture is constructed by revising and extending existing theories of biological function to fit the cultural case. Here the adjustments are for the most part necessitated by special features of function in material culture. These two parts of the project are held together by a trio of overarching themes: the relationship between individual and society, the problem of centralized control, and creativity.
Introduction: what the brain's-eye view tells the mind's-eye view. Part 1 The mind's-eye view: classical cognitivism situation and substance folk psychology, thought, and context biological constraints. Part 2 The brain's-eye view: parallel distributed processing informational holism the multiplicity of mind - a limited defence of classical cognitivism structured through reassembling the jigsaw. Epilogue: The parable of the high-level architect. Appendix: Beyond eliminativism.
L'A. replique a S. Schwartz, "Putnam on Artifacts" (ibid., 87, 1978) que le retour a Locke ne se justifie pas davantage pour la reference artificielle des noms pour celle des noms naturels. Il s'agit d'abord d'expliquer cette distinction entre les deux references. La distinction analytique/synthetique s'avere importante.