Content uploaded by Luca Tummolini
Author content
All content in this area was uploaded by Luca Tummolini on Jan 07, 2021
Content may be subject to copyright.
The Challenge of Abstract Concepts
Anna M. Borghi
University of Bologna and Italian National Research Council,
Rome, Italy
Ferdinand Binkofski
Aachen University
Cristiano Castelfranchi
Italian National Research Council, Rome, Italy
Felice Cimatti
University of Calabria
Claudia Scorolli
University of Bologna
Luca Tummolini
Italian National Research Council, Rome, Italy
Abstract concepts (“freedom”) differ from concrete ones (“cat”), as they do not have a bounded,
identifiable, and clearly perceivable referent. The way in which abstract concepts are represented has
recently become a topic of intense debate, especially because of the spread of the embodied approach to
cognition. Within this framework concepts derive their meaning from the same perception, motor, and
emotional systems that are involved in online interaction with the world. Most of the evidence in favor
of this view, however, has been gathered with regard to concrete concepts. Given the relevance of
abstract concepts for higher-order cognition, we argue that being able to explain how they are represented
is a crucial challenge that any theory of cognition needs to address. The aim of this article is to offer a
critical review of the latest theories on abstract concepts, focusing on embodied ones. Starting with
theories that question the distinction between abstract and concrete concepts, we review theories claiming
that abstract concepts are grounded in metaphors, in situations and introspection, and in emotion. We then
introduce multiple representation theories, according to which abstract concepts evoke both sensorimotor
and linguistic information. We argue that the most promising approach is given by multiple represen-
tation views that combine an embodied perspective with the recognition of the importance of linguistic
and social experience. We conclude by discussing whether or not a single theoretical framework might
be able to explain all different varieties of abstract concepts.
Keywords: abstract concepts, abstract words, embodied cognition, categorization, language acquisition
What do we talk about when we talk about “love”? There are
many reasons why this question is difficult to answer. One impor-
tant reason, though probably not one that makes for interesting
literature, is that the word “love”, in common with concepts
referring to commendable virtues such as “honesty” and “justice,”
cannot be easily pinned down to concrete and easily identifiable
referents. Indeed concrete concepts, such as “table” and “cat,”
typically have single, bounded, identifiable referents that can be
perceived with our senses—we can, for example, see and move a
table, and we can see and caress a cat and hear it meowing. In
contrast, abstract concepts such as “fantasy,” “freedom,” and “jus-
tice,” lack bounded and clearly perceivable referents, even if they
might evoke situations, scenes, introspection and emotional expe-
riences. This makes it harder to understand what we are talking
about when we talk about love, fantasy, freedom, and justice than
when we talk about cats. Furthermore, abstract concepts are more
detached by sensorial experience than concrete ones: for example,
a model based on five features related to sensorimotor experience
(sound, color, visual motion, shape, and manipulation), was able to
successfully predict brain patterns of concrete concepts, but not of
abstract ones (Fernandino et al., 2015).
Concrete and abstract words do not represent a dychotomy, even
if they differ. All concepts are highly dependent on context and are
variable, with. However, abstract concepts are less stable over time
and are more shaped by current life experiences, situations, and
culture compared with concrete concepts (Barsalou, 1987). More-
over, abstract concepts are also by far the most variable: people
agree more with one another when asked to define, produce
associations, or generate characteristics for “chair” than for “truth”
This article was published Online First January 16, 2017.
Anna M. Borghi, Department of Psychology, University of Bologna, and
Institute of Cognitive Sciences and Technologies, Italian National Re-
search Council, Rome, Italy; Ferdinand Binkofski, Uniklinik, Aachen
University; Cristiano Castelfranchi, Institute of Cognitive Sciences and
Technologies, Italian National Research Council; Felice Cimatti, Depart-
ment of Humanities, University of Calabria; Claudia Scorolli, Department
of Psychology, University of Bologna; Luca Tummolini, Institute of Cog-
nitive Sciences and Technologies, Italian National Research Council.
We thank Gillian Dunn Galvin and Stuart Fellows for the English
revision. We would also like to thank Guy Dove and Art Glenberg for
comments on an earlier draft.
Correspondence concerning this article should be addressed to Anna M.
Borghi, who is now at the Department of Clinical and Dynamical Psychol-
ogy, University of Rome La Sapienza, via degli Apuli, 1, 00185 Roma,
Italy. E-mail: anna.borghi@gmail.com
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Psychological Bulletin © 2017 American Psychological Association
2017, Vol. 143, No. 3, 263–292 0033-2909/17/$12.00 http://dx.doi.org/10.1037/bul0000089
263
or “love.” Furthermore, most concrete concepts can be inscribed
into the two broad categories of natural objects and artefacts (or
living and nonliving entities, such as “animals” and “furniture”).
Abstract concepts come instead in a great variety, as the difference
between “number,” “opinion,” and “philosophy” suggests.
The way in which abstract concepts are represented has been
discussed in the literature for over 30 years. However, it is now
becoming a topic of intense debate in psychological science,
because of the emergence of the embodied and grounded (from
now on referred to as embodied) approaches to cognition.
Embodied approaches consider human concepts to be influ-
enced by the kind of body that organisms possess to be “grounded”
in perception, action and emotion systems (Barsalou, 1999,2012;
Borghi, 2005;Gallese & Lakoff, 2005). This means, for example,
that, according to the embodied view, the concept of “cat” and the
word “cat” acquire meaning by the internal simulation or reenact-
ment of the perceptual, motor and emotional experiences linked to
seeing a cat, caressing it, or hearing it meowing (Barsalou, 2008;
Gallese, 2008). This simulation consists of the reactivation of the
neural patterns active when we experience a cat; this helps us to
interact with novel cats as we encounter them, forming predictions
about what they will do next.
In the last 10 –15 years researchers have collected a considerable
amount of evidence in support of the embodied view. Even if the
embodied approach has been recently under attack (e.g., Gold-
inger, Papesh, Barnhart, Hansen, & Hout, 2016;Mahon, 2015; but
see Barsalou, 2016, for a defense), many studies have demon-
strated that concrete concepts (e.g., “bottle,” “cat”) do activate
perceptual properties, actions and emotions. It is much more
difficult to provide compelling demonstrations that abstract con-
cepts such as “freedom” and “justice” are embodied, as they do not
have a clearly identifiable referent. Many proponents of embodied
theories have acknowledged this difficulty. Unfortunately, this
weakness has led others to suggest that embodied views cannot, in
the end, explain the formation, use and representation of abstract
concepts (e.g., Dove, 2009).
In this article we argue that abstract concepts pose a challenge
that can no longer be postponed. The challenge arises in demon-
strating that not only concrete but also abstract concepts can be
explained by adopting an extended embodied perspective: even if
abstract concepts lack a bounded, identifiable and perceivable
referent, they are grounded. We will contend that, beyond senso-
rimotor and emotional experience, language and sociality play a
crucial role in grounding them.
Because distributional theories of meaning can easily explain
how abstract concepts are represented (Andrews, Frank, &
Vigliocco, 2014), it could be argued that abstract concepts consti-
tute a problem only for the embodied approach. According to
distributional or statistical approaches (e.g., Lund & Burgess,
1996, HAL; Landauer & Dumais, 1997, LSA), meaning is com-
puted statistically. Meaning is given by the co-occurrence of words
in large masses (corpora) and it derives from the relationship
between associated words rather than between words and their
referents (for a recent overview, see Andrews et al., 2014; see also
work by Max Louwerse: e.g., Louwerse & Jeuniaux, 2010). Thus,
the meaning of both concrete and abstract words is derived from
their statistical distribution in large corpora. Because all meaning
is captured by associations between words, distributional theories
do not need to posit any difference between concrete and abstract
words.
However, the explanation offered by distributional theories of
both concrete and abstract concepts representation is incomplete,
since they are unable to solve the “symbol grounding problem”
(Harnad, 1990): to be understood and avoid circularity, symbols
ultimately need to be grounded in their world referents. (Even if
very powerful, notice however that Harnad’s argument is not
universally accepted.) As soon as one recognizes that abstract
concepts need to be grounded but do not have a bounded and
perceivable referent, the challenge arises. In the review, we will
argue that recent literature converges in showing that, to explain
grounding of abstract concepts, the embodied perspective must be
extended to take into account the important role of linguistic and
social experience, and that the study of conceptual acquisition can
widely contribute to account for abstract concepts representation.
More generally, the challenge to explain abstract concepts rep-
resentation is crucial for a simple but pivotal reason: the ability to
use abstract thought represents one of the most sophisticated
abilities of our species. We are not arguing that other animals lack
the capacity to represent at least some abstract concepts: for
example, nonhuman animals are able to group stimuli together
according to whether they differ or are equal (e.g., Flemming,
2012). Our point is that humans possess such a capacity at a
qualitative and quantitative different level than nonhuman animals
(Lupyan & Bergen, 2016). Any theory of higher-order cognition
would be incomplete if it was unable to explain how we make use
of abstract concepts.
In summary, providing an explanation of abstract concepts
representation is crucial to further the understanding of higher-
order human cognition. More important, such a challenge concerns
not only a subset of theories, that is, the embodied ones, since
nonembodied theories also fail to provide a thorough explanation
of concepts that addresses the symbol grounding problem.
In what follows we will present the current debate on abstract
concepts representation (for recent reviews, see Borghi & Binkof-
ski, 2014;Pecher, Boot, & van Dantzig, 2011;Reilly, Peelle,
Garcia, & Crutch, 2016). We will focus, in particular, on the most
recent embodied theories on abstract concepts, critically evaluating
recent evidence supporting them, and discussing their strengths
and their limitations. We will concentrate on embodied theories
also because the majority of the behavioral studies on abstract
concepts of the last years derive from the embodied cognition
approach, or address embodied approaches. While this review
focuses on abstract concepts, we will also discuss theories and
evidence on the representation of abstract words. We do not intend
to equate concepts and words: language influences categorization,
it augments it, rendering word meaning more compact and discrete
than concepts (Lupyan, 2012;Mirolli & Parisi, 2011). Where
possible, we will distinguish between concepts and word meanings
and focus on concepts; in most of the cases, however, it is impos-
sible, because tasks on conceptual representation in human adults
usually involve the use of words. In such cases we will not
distinguish between “abstract concepts” and “abstract word mean-
ings.”
Because the literature on abstract concepts is heterogeneous, this
review does not aim to be exhaustive. Specifically, we will not
treat neuropsychological and brain imaging evidence in detail,
unless the results of a study are crucial to test or support a specific
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
264 BORGHI ET AL.
proposal under review (for a recent review on insights from
neuropsychology on abstract concepts, see Hoffman, 2015). More-
over, even if numbers and emotions represent important domains
of abstract conceptualization, we will not consider theories that
focus only on numerical cognition or only on the conceptualization
of emotions, as these theories are beyond the scope of this work.
We will refer to emotion and number concepts only when dealing
with studies on wide samples of abstract concepts that include,
among others, emotional and numerical concepts. As to the theo-
ries we have selected, while we will only briefly overview classical
theories on abstract concepts, we have decided to describe the
embodied theories proposed from 2004 to 2015–2016, illustrating
if possible both supporting and contradictory evidence. For very
influential theories, as the Conceptual Metaphor theory, we will
not review all the available evidence, as this would require a
dedicated review. Instead, we have chosen to refer to some very
influential studies, as well as to reviews of studies, when they are
available.
In the review we will show that an important novelty of the
literature of the last years is the emergence of approaches based on
multiple strategies. In particular, it is becoming evident that to
explain abstract concepts a standard embodied approach must be
integrated with a perspective that emphasizes the importance of
language for concepts. In the course of the review we will thus
often refer to language and linguistic information. Linguistic in-
formation means different things in different theories. First, it can
mean information that is conveyed solely through language with-
out direct reference to the world (e.g., by pointing, gesture, or use
of deictic terms). We will refer to this meaning as “linguistically
conveyed information.” Note that linguistically conveyed informa-
tion might be embodied in its representation, that is, make use of
sensorimotor and emotional systems. Second, linguistic informa-
tion could refer to the literal words themselves and their represen-
tations. For example, a word can be a sequence of sounds or a
sequence of visual forms. We will refer to this concept as “lin-
guistic form information.” Note that linguistic form information
may very well be embodied. That is, the sounds may be repre-
sented in the auditory system and the visual forms in the visual
system. Some theories, however, propose that linguistic form
information is amodal, and so we will also use the terms “amodal
linguistic form information” and “modal linguistic form informa-
tion.” Third, linguistic information might refer to notions other
than words, such as syntactic relations. We will refer to this idea
as “syntactic linguistic information.”
The emphasis on language has long been one of the strong
points of distributional theories. In our view, the most promising
approaches take inspiration from distributional views in highlight-
ing the role of language for abstract concepts. At the same time,
they differ from distributional views as they do not focus simply
on word associations, but consider language as a holistic experi-
ence, in its bodily and social aspects. From our birth—and even
earlier (Moon, Lagercrantz, & Kuhl, 2013;Vouloumanos &
Werker, 2007)—we are immersed in a linguistic environment.
Language has been defined as a mode of participation in the world
(see Irwin, 2015, for developing this aspect in relation to the
philosophy of Merleau-Ponty). Language, which is primarily a
social experience, can become internalized and support our
thought processes: speaking to ourselves helps us to better mem-
orize and plan our actions (Alderson-Day & Fernyhough, 2015;
Clark, 1998;Vygotsky, 1986). Recent proposals have shown that
language is not only a communication system, but a control system
that programs human mind manipulating sensorimotor experiences
(Lupyan & Bergen, 2016). Language directed at others and inner
speech provide an important means of building predictions: for
example, listening to a word can help our visual system to process
noisy inputs (Lupyan & Clark, 2015). Language also involves a
bodily experience: we produce sounds, we listen to them, and it is
possible that when we think of abstract words we internally re-
produce their sound, reenacting the experience of their acquisition,
and explaining to ourselves their meaning, formulating predictions
against which sensory experiences can be assessed. More impor-
tant, it is highly plausible that the contribution of language to
thought processes is particularly crucial for abstract concepts and
for the corresponding words, given their lack of bounded and
perceivable referents.
Our aim is to convince the reader that the debate on abstract
concepts is not internal to embodied theories, but rather that it also
involves distributional theories of meaning, and that the most
exciting directions of research are emerging by bridging the strong
points of these two approaches, that is, the importance of ground-
ing abstract concepts in perception, action, and emotion systems
and the crucial role of language in this process. However, we will
also contend that integrating embodied and distributional theories
is not sufficient to account for abstract concepts for at least three
reasons. The first is that these approaches do not emphasize the
role of language as a tool to improve thinking (e.g., Clark, 1998;
Dove, 2015;Lupyan & Bergen, 2016;Vygotsky, 1986). The
second, more crucial reason, it that, focusing on the combinatorial
aspects of language, these approaches overlook the crucial role of
the social dimension for language. The third is that they do not
highlight the importance of the modality of acquisition of lan-
guage, and in particular the pivotal role of social competences for
word learning. In our view only new views and research directions
that highlight these three aspects—that is, that conceive language
as a prediction and thought instrument, that emphasize the social
dimension for concepts and language, and that take profound
inspiration and insights from studies on language acquisition—will
be able to meet the challenge to explain abstract concepts.
Classical Theories of Abstract Concepts
A Tale of Two Classical Theories: CAT Versus DCT
Concrete words have an advantage over abstract words: they are
processed faster and more accurately in a variety of tasks, such as
lexical decision (i.e., deciding whether a word exists or not), and
have a better performance in tasks as naming and recall. During the
1980’s and 1990’s, two classical theories on abstract concepts
proposed different explanations of this “concreteness effect.”
These are the Context Availability Theory and the Dual Coding
Theory.
According to the Context Availability Theory (CAT; e.g.,
Schwanenflugel, Harnishfeger, & Stowe, 1988;Schwanenflugel,
Akin, & Luh, 1992), concrete concepts are strongly associated to
a reduced number of contexts, while abstract concepts are weakly
associated to a much wider range and number of contexts
(Schwanenflugel et al., 1988). Therefore, the CAT explains the
better recall of concrete than abstract concepts arguing that pro-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
265
THE CHALLENGE OF ABSTRACT CONCEPTS
cessing abstract words is slower because activating their context
requires more effort.
An alternative theory, which has become far more influential, is
the Dual Coding Theory (DCT; e.g., Paivio, 1986;Paivio, Yuille,
& Madigan, 1968). This theory explains the concreteness effect
with the higher imageability of concrete compared to abstract
concepts. According to DCT, all concepts are represented through
the verbal system, but only concrete concepts have a direct con-
nection with images. For instance, the abstract concept of “reli-
gion” evokes images only through the mediation of concrete
concepts such as “church.”
The DCT posits a negative relationship between abstractness
and imageability: the more abstract a concept is, the less imageable
it is. However, this intuitive relationship is problematic because, as
was recently highlighted (Kousta, Vigliocco, Vinson, Andrews, &
Del Campo, 2011), the two dimensions are correlated but not
equivalent. An additional area of contention, shared by both CAT
and DCT, is the weak evidence-base for the “concreteness effect.”
While some research did not find this effect (Barca, Burani, &
Arduino, 2002), controlling for word valence some studies found
an opposite “abstractness effect,” that is, a processing advantage of
abstract over concrete concepts, (Kousta et al., 2011). Other stud-
ies replicated the abstractness effect in response times but found a
reverse concreteness effect in event related potentials (ERPs; Bar-
ber, Otten, Kousta, & Vigliocco, 2013): abstract words were
processed faster than concrete ones, but concrete ones elicited
larger N400 and N700 ERPs, typically associated with processing
meaningful stimuli and with imagery.
Beside the concreteness effect, results of recent studies are also
inconsistent with both CAT and DCT. Connell and Lynott (2012),
for instance, asked participants to rate concepts not only in terms
of concreteness and imageability, but also in terms of auditory,
gustatory, haptic, olfactory, and visual strength (see also Moffat,
Siakaluk, Sidhu, & Pexman, 2015). In contrast to CAT, they found
that concepts characterized by higher perceptual strength, which
CAT would categorize as concrete, evoke a higher number of
contexts compared to concepts with low perceptual strength. This
study also strongly contradicts the DCT. The results showed that
imageability is not strongly correlated with perceptual experience,
because it seems to reflect more the ease with which an image is
generated, rather than whether a concept is abstract or concrete. In
addition, imageability ratings appeared to be visually biased, while
perceptual strength was able to predict the results of lexical deci-
sion and naming tasks better than contextual availability and
imageability.
Furthermore, there is conflicting evidence regarding the neural
underpinnings of concrete and abstract words, which has led to
much debate in the literature. For example, some brain imaging
research supports the DCT, showing that processing concrete
words led to more bilateral activation than processing abstract
words (e.g., Binder, Westbury, McKiernan, Possing, & Medler,
2005;Sabsevitz, Medler, Seidenberg, & Binder, 2005), which are
more left-lateralized (e.g., Binder et al., 2005;Fiebach & Fried-
erici, 2004), while other studies contradict this research as they
found evidence for an opposite neural pattern (Kiehl et al., 1999;
Pexman, Hargreaves, Edwards, Henry, & Goodyear, 2007).
It is important to note that inconsistencies in this research area
(e.g., concreteness vs. abstractness effect, left- vs. right lateraliza-
tion in the brain; for a review see Borghi & Binkofski, 2014,
chapter 5) could be due to methodological differences between
studies, which have often used different methods to select stimuli.
For example, some studies selected abstract words on the basis of
the imageability ratings, others based their selection on abstract-
ness/concreteness ratings, while other studies selected abstract
words on the basis of intuition of researchers. It is possible,
therefore, that comparability of findings has been reduced by these
inconsistences in the selection of stimuli. It is clear from the
evidence mentioned above that future research needs to develop
validated criteria for the selection of abstract concepts, which
should then be implemented universally. In the final section of the
review we will briefly illustrate promising recent directions trying
to investigate different subtypes of abstract concepts, and at the
same time to propose more general criteria to use for stimuli
selection.
Embodied and Grounded Theories of
Abstract Concepts
As suggested above, the issue of how abstract concepts are
represented has been focus of intense debate over the last decade,
mainly because of the increasing diffusion of embodied theories of
cognition (Barsalou, 2008;Borghi & Caruana, 2015;Wilson,
2002). Our review will now focus on contributions adopting this
perspective and addressing the problem of abstract concepts rep-
resentations. We will cover the main current proposals and eval-
uate the supporting evidence, while highlighting their strengths
and limitations (see Table 1).
Within embodied theory, scholars are still divided into two
schools of thought: one group focuses on the similarities between
concrete and abstract concepts, while the other group emphasizes
their differences. The latter either argues for a sharp distinction
between concrete and abstract concepts, or views them on a
continuum that spans from less abstract to very abstract concepts
(Wiemer-Hastings, Krug, & Xu, 2001;Wiemer-Hastings & Xu,
2005).
One influential view that posits that abstract and concrete con-
cepts are profoundly different comes from neuropsychology. We
will describe it before starting to illustrate the embodied theories,
because it cannot be considered an example of an embodied
approach. According to this view, concrete and abstract concepts
are characterized by qualitatively and structurally different infor-
mation (Crutch & Warrington, 2005,2007,2010): while concrete
concepts mainly rely on categorical similarity relations (e.g.,
“theft-burglary”), abstract concepts primarily rely on semantic
associations (e.g., “theft-punishment”). This distinction is derived
from single-case studies on double dissociations: patients with
semantic refractory access dysphasia showed greater interference
for abstract words organized by associative relations, while the
opposite pattern was found with concrete words (see for example
Crutch, Ridha, & Warrington, 2006). This distinction has also been
found in healthy participants (see, e.g., Crutch, Connell, & War-
rington, 2009). Duñabeitia, Avilés, Afonso, Scheepers, and Car-
reiras (2009) provided evidence that healthy individuals tend to
fixate more and earlier on pictures associated with abstract words
than with concrete words (e.g., “nose-smell” compared with
“baby-crib”). Hence associative relations are seen to be more
important for abstract concepts than for concrete ones. However,
the evidence obtained both with clinical and with general popula-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
266 BORGHI ET AL.
Table 1
The Main Embodied Theories of Abstract Concepts
Theories
Difference between abstract and
concrete concepts
Level of
embodiment
Multiple
representation Role of acquisition Kind of evidence Mechanism or content?
Motor theory No: Abstract ⫽concrete concepts Strong No: Sensorimotor Unspecified Behavioral (ACE effect,
approach-avoidance
effect)
Mechanism, but limited for content
reasons (e.g., to transfer
sentences, or to valenced
stimuli)
Situation and
introspective view
Yes: Abstract concepts activate
more social aspects of
situations and introspective
properties
Weak Not specified Unspecified Behavioral (Feature
generation)
Mechanism, but possible content
limitations (e.g., introspective
properties mainly activated for
mental state concepts)
Affective Embodiment
Account (AEA)
Yes: Abstract concepts activate
more emotions
Weak Yes: Emotional and
sensorimotor;
also linguistic
(but not fully
discussed)
Emotions as bootstrapping
mechanism
Behavioral (Lexical
decision), fMRI,
ERPs, patients
Possible mechanism, but limited
for content reasons (e.g.,
emotional properties more
activated for emotional
concepts)
Conceptual Metaphor
View (CMT)
Yes Strong No: Sensorimotor Nonplausible
developmental trajectory
Mainly behavioral,
linguistics, and
psychology
Mechanism, but limited because of
content reasons
Language And
Situated Simulation
(LASS)
Yes Weak Yes: Sensorimotor
and linguistic
(language as
shortcut to
access meaning)
Unspecified Behavioral (feature
generation), fMRI
Mainly content. Mechanism related
to the task, not to the kind of
concepts
Representational
pluralism: Dove
Yes: Abstract concepts activate
more linguistic information
Hybrid Yes: Sensorimotor
and linguistic.
Both modal and
amodal codes.
disembodied
linguistic system.
Unspecified Not direct (but indirect
support from evidence
supporting Paivio’s
dual coding model)
Mechanism
Grounding and sign
tracking: Prinz
Yes Weak Yes: Multiple
strategies. Focus
on sensorimotor,
emotional,
linguistic
Unspecified Not direct (but indirect
support from evidence
on all the other
theories)
Both content and mechanism,
depending on the adopted
strategy
Words As social Tools
(WAT)
Yes: Abstract concepts activate
more linguistic, (emotional)
and social information than
concrete ones
Weak Yes: Sensorimotor,
emotional,
linguistic and
social
information
Very relevant: Acquisition
constrains representation
Behavioral (e.g., sorting,
categorization, feature
generation), fMRI,
TMS, studies on sign
languages
Mechanism:, not linked to the
content but to the abstractness
level (the more abstract concepts
are, the more they activate
linguistic and social
information). It is not
incompatible with content effect
Note. As to level of embodiment, we consider the following categories: Strong embodied views are those that assume that only sensorimotor areas are engaged during conceptual processing; weak
embodied views are those that assume that both sensorimotor and linguistic areas are engaged during conceptual processing; hybrid views assume that both embodied and nonembodied representations
are activated.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
267
THE CHALLENGE OF ABSTRACT CONCEPTS
tions is controversial. Failure to replicate the results of Crutch and
Warrington (2005) with aphasic patients has challenged the hy-
pothesis that abstract and concrete words differ in terms of the
conceptual relations they evoke (e.g., Hamilton & Coslett, 2008;
Hamilton & Martin, 2010). Furthermore, some behavioral results
contradicted the findings of Duñabeitia et al. (2009).Geng and
Schnur (2015) asked Chinese–English bilinguals to match a Chi-
nese auditory-presented word with one among several other visu-
ally presented English words. This study found that related words
are processed faster than unrelated words, but categorical relations
(e.g., “idea-attitude”) always led to a better performance compared
with associative relations (e.g., “math-examination”). In a recent
study with a semantic priming paradigm both within and across
languages, a priming effect was observed for concrete words for
both semantically similar and associative relations. For abstract
words, instead, it was present only when prime and target words
were associated (Ferré, Guasch, García-Chico, & Sánchez-Casas,
2015).
From these results, it could be concluded either that abstract and
concrete concepts appear to be represented similarly both in terms
of categorical and associative relations (see also Marques &
Nunes, 2012;Zhang, Han, & Bi, 2013, for a similar conclusion), or
that associative relations are more relevant to characterize abstract
concepts, but that for the representation of concrete ones both
categorical and associative relations are relevant (Ferré et al.,
2015).
Although evidence for the existence of a “marked distinction” is
conflicting, there is a compelling reason for these inconsistent
results. Crutch and Jackson (2011) recently suggested that the
relationship between concreteness and categorical and semantic
associations is not binary but rather graded. Therefore, the range of
results in the literature may actually provide evidence that these
associations are on a continuum. Drawing on this conclusion, in
what follows we will concentrate on embodied cognition studies
adopting the idea of a continuum spanning from concrete to
abstract items. The idea of a continuum is that most widely
accepted by most researchers. The idea of a continuum presents a
further important insight: that no real dichotomy exists between
concrete and abstract concepts, because even concepts that are
generally qualified as concrete have an abstract component and
vice versa. Consider the concept of “cent” or “euro”: It is mostly
a concrete one since its referent has specific perceptual character-
istics, as a given size, color, weight, but at the same time it has an
exchange value that cannot be easily pinned down to the concrete
aspects of its referent (see Guan, Meng, Yao, & Glenberg, 2013).
Strong Embodiment: Concrete and Abstract Concepts
Do Not Differ
According to some embodied theories, concrete and abstract
concepts do not differ substantially because they are both
grounded in the same systems engaged during perception, action
and emotion (e.g., Chen & Bargh, 1999;Connell & Lynott, 2012;
Glenberg, Sato, & Cattaneo, 2008a;Glenberg et al., 2008b). The
advocates of this perspective seek to demonstrate that the re-
enactment of perceptual states, actions, and emotions is not solely
limited to concrete concepts, but is also possible for abstract
concepts. Embodied theories that support this perspective are
based on evidence on the Action-sentence Compatibility Effect
(ACE; Glenberg & Kaschak, 2002), on the approach-avoidance
effect (Chen & Bargh, 1999), and on force dynamics (Talmy,
1988).
The ACE shows that, when a sentence implies action in one
direction (e.g., the sentence “give the cards to somebody” implies
action away from the body), the participants are slower in judging
whether the sentence makes sense or not if they are required to
respond performing a movement in the opposite direction. This
effect is equally present for both concrete and abstract transfer
sentences (e.g., “give the cards” and “give the responsibility”).
This behavioural evidence has been taken to suggest that the same
action schema underlies representation of both concrete and ab-
stract information (Glenberg et al., 2008a,2008b).
Neuro-cognitive studies using Transcranial Magnetic Stimula-
tion (TMS) complement these behavioral findings, as they have
observed higher Motor Evoked Potentials for transfer sentences
than for nontransfer ones. The absence of a difference between
sentences related to transfer of objects (concrete) and of informa-
tion (abstract) indicates that in both cases the motor system is
activated (Glenberg et al., 2008a). ACE effects also apply to
sentences involving time shifts (Sell & Kaschak, 2011): responses
away from the body were faster with sentences involving the
future, while responses toward the body were faster with sentences
involving past events.
Further evidence supporting the view that concrete and abstract
concepts do not differ substantially is provided by the approach-
avoidance effect. This effect states that positive words (e.g.,
“gold,” “sunshine”) evoke an attraction movement, while negative
words (e.g., “garbage,” “virus”) evoke a rejection movement,
regardless of whether they are abstract or concrete (Chen & Bargh,
1999).
Studies of cognitive linguistics on force dynamics also provide
converging evidence with these results. According to Talmy
(1988), events, be they physical, psychological, or social, can be
viewed as oppositions between conflicting forces, for example,
between an agonist and an antagonist force (e.g., “The ball kept on
rolling along the green”; “John can’t go out of the house”; “She’s
got to go to the park”). Force dynamics states that abstract and
concrete events rely on the same force mechanisms. In line with
this view, research using a sentence sensibility task has shown that
primes given by two shapes interacting following the same force
dynamics pattern, evoked faster responses than two shapes that did
not follow the same pattern (Madden & Pecher, 2010, reported in
Pecher et al., 2011). As predicted by Talmy (1988), the results
were the same with concrete and abstract sentences.
Strengths and limitations. To sum up, the evidence dis-
cussed so far is based on sound and compelling results, obtained
both through behavioral and TMS studies. However, these results
can be used to claim either that abstract concepts are grounded in
similar systems as concrete ones, or, more strongly, that abstract
and concrete concepts do not differ at all. Both of these interpre-
tations will now be discussed.
Abstract concepts are grounded in the same processes of
concrete concepts. Previous evidence indicates that abstract con-
cepts activate perception and action systems as concrete concepts.
Even if these findings are broadly consistent with an embodied
approach to abstract concepts, it is difficult to foresee whether this
strategy can be extended to account for all abstract concepts. For
example, ACE has been found with concrete and abstract transfer
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
268 BORGHI ET AL.
sentences, with sentences referring to past and future events (Sell
& Kaschak, 2011), and also with sentences referring to increasing
or decreasing quantities (e.g., Guan et al., 2013). Yet, it is difficult
to think that it can apply to all domains in which abstract concepts
exist: for example, how can it account for concepts as “philoso-
phy,” “thought,” or “freedom”? Similarly, current research sug-
gests that Approach Avoidance-Effect can be found with a variety
of words (e.g., Förster & Strack, 1996;Freina, Baroni, Borghi, &
Nicoletti, 2009;Eder & Hommel, 2013;Neumann & Strack, 2000;
Seibt, Neumann, Nussinson, & Strack, 2008;van Dantzig, Pecher,
& Zwaan, 2008; for a recent meta-analysis on approach avoidance
and affective stimuli in response times tasks, see Phaf, Mohr,
Rotteveel, & Wicherts, 2014). Both lines of research are, however,
limited to words or stimuli characterized by a positive or negative
valence, or to novel words designed in such a way that the same
kinematics used during deglution or expectoration (inward vs.
outward movement) are reproduced to pronounce them (Topolin-
ski, Maschmann, Pecher, & Winkielman, 2014). Moreover, even if
the theory of force dynamics requires further experimental support,
it is difficult to imagine that evidence based on force dynamics can
be extended beyond the domain of events: for example, how can it
explain concepts as “linguistics” or numerical concepts as “four”?
Finally, similar results do not necessarily imply similar processes.
An example is given by a recent study by Yao et al. (2013). The
authors found a size effect in word recognition: big words (e.g.,
“jungle”) were processed faster than small ones (e.g., “needle”),
independently of concreteness level of words. However, the ex-
planation of the effect might differ for concrete and abstract
concepts. While the explanation is more intuitive for concrete
concepts, it is possible that big abstract concepts (e.g., “disaster”)
include a wider range of introspective, social, and situational
associations compared to smaller concepts. Indeed, regression
analyses revealed that subjective ratings of emotional arousal were
the only significant predictor of size, and that arousal had a more
crucial effect in the recognition of abstract than of concrete words.
This is an example of a study in which apparently results do not
differ between abstract and concrete concepts, but the processes
they imply differ substantially.
Concrete and abstract concepts do not differ. This second
claim is not, in our view, justifiable by the available evidence.
First, it is statistically unsound to infer that two concepts (e.g.,
abstract and concrete) can be equated with one another, simply
from the absence of differences in performance. Second, people
tend to evaluate and use concrete and abstract concepts different-
ly—for example, they rate “justice” as more abstract than “bot-
tle”—and this difference should be explained. Third, an increasing
number of behavioral, neuropsychological, and neuroimaging
studies have demonstrated that these two kinds of concepts are
differently processed and recalled. To make an example, in a
recent study Binder et al. (2016) designed a componential model of
concepts based on functional divisions in the human brain. To this
aim they used a large set of attributes related to different experi-
ences and found that 57 of the 65 attributes distinguished abstract
from concrete categories. More specifically, abstract concepts re-
ceived higher ratings than concrete ones on attributes related to
temporal and causal experiences, social experiences, and emo-
tional experiences, while they received lower ratings in sensorial
experiences. For all these reasons, we suggest that a systematic
theory of abstract concepts should not only assume that differences
between concrete and abstract concepts exist, but also be able to
explain why these differences occur.
Recently there has been an interesting variation to the view that
abstract and concrete concepts are similar. Guan et al. (2013) used
EEG and found ACE effects with both concrete and abstract
concepts (for a further ERPs study see D’Angiulli, Griffiths, &
Marmolejo-Ramos, 2015). The authors explained their results by
arguing that both concrete and abstract words activate the motor
system. However, while concrete concepts are grounded in spe-
cific sensorimotor simulations, as they reenact previous experi-
ences with the category members, abstract concepts are mainly
grounded in the process of prediction. Examples introduced by the
authors provide weight to this argument: the concept of “banana”
activates the simulation of eating it, which may involve seeing its
shape and color, peeling it, and tasting it (vision, action, and taste).
Even if it is mainly concrete, the concept of banana has an abstract
component as well, consisting of the prediction that eating it will
reduce hunger. On the opposite site of the coin, an abstract concept
such as “democracy” involves some concrete components (e.g., the
simulation of voting), but most of its meaning derives from ab-
stract components linked to predictions of what follows from a
process being democratic (e.g., concern for human rights, higher
tolerance). More important, both kinds of components are
grounded upon and refer to different characteristics and mecha-
nisms of the motor system.
Compared with other views, this proposal has the advantage of
suggesting a mechanism underlying abstract concepts that is sup-
ported by a functional architecture of brain organization: the role
of forward models within the motor system, whose function is to
anticipate the sensory consequences of actions (Wolpert, Doya, &
Kawato, 2003;Glenberg & Gallese, 2012). The contribution of this
theory thus consists in specifying the fine-grained mechanisms that
underlie the involvement of the motor system in conceptual
grounding. Even if the theoretical framework is clearly outlined,
the extent to which predictions derived from this view differ from
those stemming from the view according to which abstract con-
cepts, similarly to concrete ones, are represented simply in terms
of their content, and are grounded in the motor system is not fully
clear. Furthermore, even if, as argued by Guan et al. (2013), many
abstract names are “names for processes” (e.g., democracy relies,
among other processes, on the processes of counting and voting),
this view is likely to be insufficiently general to be applied to all
varieties of abstract concepts.
Concrete and abstract concepts are different. Taken to-
gether the evidence discussed so far does not support the view that
there are no differences between concrete and abstract concepts.
Therefore, in what follows we will concentrate on the most recent
theories that provide evidence of differences between abstract and
concrete concepts, and discuss and evaluate the mechanisms and
cognitive components that might be responsible for such differ-
ences.
Conceptual Metaphor Theory (CMT). Since the 1980’s, Con-
ceptual Metaphor Theory (CMT), which is probably the most
influential embodied theory on abstract concepts, has proposed
that abstract and concrete concepts are indeed different (Lakoff &
Johnson, 1980,1999;Lakoff & Núñez, 2000). The CMT is pri-
marily based on observations of language use. This theory posits
that when talking about abstract concepts people tend to use
metaphors derived from concrete domains: for example, we say
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
269
THE CHALLENGE OF ABSTRACT CONCEPTS
that “life is a journey.” According to the CMT, metaphors concern
not only the way in which we use language, but also the way in
which we think of the world. Even if, for the most part, the
supporting evidence is derived from linguistics, the CMT has
benefited from empirical evidence obtained without linguistic
stimuli, mainly deriving from behavioral studies (for reviews, see
Gibbs, 1994,2006; for examples of very recent work e.g., Casa-
santo & Bottini, 2014;Jamrozik, McQuire, Cardillo, & Chatterjee,
2016;Sato, Schafer, & Bergen, 2015;Zhao, He, & Zhang, 2016).
Researchers favoring the CMT argue that abstract concepts are
understood by placing them in concrete knowledge domains, and
that this mapping guarantees their grounding. Important abstract
concepts are understood through multiple conceptual metaphors:
for example, the abstract concept of “communication” can be
understood as sending ideas from one container (head) to another,
as well as feeding someone with thoughts (Lakoff, 2014). Recent
behavioral evidence supporting this view shows that, for example,
the abstract notion of “category” is understood with reference to
the concrete concept of “container” (Boot & Pecher, 2011), the
concept of “similarity” relies on that of “closeness” (Boot &
Pecher, 2010), the notions of “good” and “bad” are associated to
the right and left space (for a recent study showing that this effect
is modulated by mental imagery, see de la Fuente, Casasanto,
Martínez-Cascales, & Santiago, 2016), and “power” is compre-
hended by referring to the vertical dimension (Lakens, Semin, &
Foroni, 2011;Zanolie et al., 2012), similarly to “morality” (Wang,
Lu, & Lu, 2016). Furthermore, we use abstract language to char-
acterize phenomena that we consider to be more distant from us:
for example, recent data from Twitter indicate that more abstract
language is used when referring to distant cities, to distant past or
future time points, and to people who are socially distant from us
(Snefjella & Kuperman, 2015).
Possibly, the concept which has been most extensively studied
is the concept of “time”: the underlying rationale is that, because
time is an abstract concept, it can be understood by mapping it on
the more concrete concept of “space” (e.g., Boroditsky & Ram-
scar, 2002;Casasanto & Boroditsky, 2008; see Flusberg, Thibo-
deau, Sternberg, & Glick, 2010, for a connectionist model). Imag-
ine presenting participants with an ambiguous question: “Next
Wednesday’s meeting has been moved forward two days. What
day is the meeting now that it has been rescheduled?” Answering
“Monday” implies adopting an ego-moving perspective, while
answering “Friday” a time-moving perspective: in the first case
forward is in the direction of motion of the observer, in the second
case forward is in the direction of motion of time. Boroditsky and
Ramscar (2002) have shown that people, who are at the beginning
of a train journey, of a lunch line, or people who had just flown in,
tend to respond using an ego-moving perspective, whereas when
they are at the end of a trip, or of a line, they use a time moving
perspective. This result indicates that space and time are strictly
interwoven. The perspective adopted varies depending on the
spoken language. Asking the same ambiguous question to Man-
darin Chinese and English speakers Lai and Boroditsky (2013)
have found that while English speakers tend to adopt an ego
moving perspective and Mandarin Chinese monolinguals a time
moving one, bilinguals shift between the two perspectives.
The proposition that we use the concrete concept of space to
reason and understand the abstract concept of time seems to be
confirmed by the observation that in language use the relationship
between space and time is asymmetrical: we tend to speak about
time in terms of space more often than we speak about space in
terms of time. Casasanto and Boroditsky (2008) have demon-
strated the asymmetrical relationship between space and time,
showing that distance affects duration, but duration does not in-
fluence distance.
That space and time are strictly related has been further sup-
ported by a variety of cross-cultural studies. Depending on the
culture, time metaphors can refer to quantity or to distance: for
example, in English people typically say “a long meeting,” in
Greek “a large meeting.” Consistent with the metaphors used,
English-speakers estimates of time are more influenced by length
than by quantity, while for Greek speakers the opposite is true
(Casasanto, 2008).
The way in which we think of the temporal flow, from the past
to the future, is influenced by culture as well as by the character-
istics of our own body (e.g., Bergen & Chan Lau, 2012). For
example, for English speakers time is conceptualized in terms of
the front-back dimension (the past is behind us, the future is ahead
of us), while the frequent use of vertical metaphors in Chinese has
led Chinese Mandarin speakers also to conceive time in terms of
the vertical dimension: the past is up, the future is down. This
difference influences responses to implicit tasks, such as priming
tasks. Boroditsky (2001) found that Chinese Mandarin speakers
were faster to confirm that March comes earlier than April after
being exposed to a prime where pairs of objects were vertically
displayed (e.g., a black and white worm-ball). In contrast, English
native-language speakers were faster when they were shown the
same two objects disposed horizontally, and bilinguals had an
intermediate performance. Though this study has been seen as very
influential by advocates of embodied cognition, a failure to repli-
cate these findings has also led to skepticism (see Chen, 2007).
A recent study challenged the assumption made by the CMT
that there is only an unidirectional influence from sensorimotor
experience to metaphors (Slepian & Ambady, 2014). If partici-
pants had been exposed to the metaphor that the past is heavy, they
tended to estimate old books as heavier, whereas if they had been
exposed to the metaphor that the present is heavy, they perceived
new books as heavier. Thus, novel metaphors can influence sen-
sorimotor processes, which suggests that the influence between
metaphors and sensorimotor states is bidirectional.
It should be noted that time is related not only to space, but also
to number, and that number is related to space, as evidenced by a
growing body of research (for a review, see Winter, Marghetis, &
Matlock, 2015; see also Jones, 2015). Lakoff and Nunez (2000)
have demonstrated that to speak about numbers, people tend to use
spatial metaphors: for example, they use the metaphor of numbers
as a point on a line, or metaphor of arithmetic as motion along a
path (Nunez & Marghetis, 2014). The close interaction between
space, time and number is also revealed by the similarities between
the Spatial Numerical Association of Response Codes, or SNARC
effect (Dehaene, Bossini, & Giraux, 1993), and the Spatial Tem-
poral Association of Response Codes, or STEARC effect (e.g.,
Ishihara, Keller, Rossetti, & Prinz, 2008;Santiago, Lupiáñez,
Pérez, & Funes, 2007;Torralbo, Santiago, & Lupiáñez, 2006).
The SNARC effect reveals that people in Western societies
associate small numbers with the left and large numbers with the
right, as well as subtraction to the left and addition to the right (see
also Anelli, Lugli, Baroni, Borghi, & Nicoletti, 2014;Lugli,
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
270 BORGHI ET AL.
Baroni, Anelli, Borghi, & Nicoletti, 2013). The STEARC effect
shows that people tend to associate the future to the right and the
past to the left. Recent evidence shows however that English-
speaking people systematically combine future-in-front and future-
to-right metaphors while gesturing, suggesting that the conceptual
representation of time involves the activation of multiple meta-
phors rather than the selection of one single metaphor (Walker &
Cooperrider, 2016). However, time does not always flow from left
to right because the writing direction influences people’s temporal
concepts. For example, Spanish people are faster in responding
with the right to words referring to the future and with the left with
words referring to the past, but the opposite is true for Hebrew
speakers (Flumini & Santiago, 2013;Ouellet, Santiago, Israeli, &
Gabay, 2010;Santiago et al., 2007; see also Fuhrmann &
Boroditsky, 2007). New evidence shows that the mental timeline
of Italians goes from left to right, following the writing direction;
importantly, this is true for both sighted and blind participants who
read with the hands (Bottini, Crepaldi, Casasanto, Crollen, &
Collignon, 2015).
The relationships between concepts of space, time and number
has been at the center of important recent debates (e.g., Lynott &
Coventry, 2014;Santiago & Lakens, 2015;Winter et al., 2015).
Two contrasting proposals highlighting the strict link between
space, time and numbers have been advanced, the Theory of
Magnitude (ATOM) and the CMT views.
The ATOM view is a general theory of magnitude that predicts
interactions between all domains involving magnitudes, while
CMT predicts interactions between space-number and space-time,
but not necessarily between number and time. Furthermore, CMT
posits that space representations are used asymmetrically to rep-
resent number and time, while no such asymmetry is assumed by
the ATOM view. These two views have always been considered to
be contrasting views, likely because of the different disciplines in
which they developed (ATOM in neuroscience, CMT in linguistics
and psychology) and to their differing focus. However, it is pos-
sible that they can be bridged: according to Winter et al. (2015),
ATOM is better at explaining low level associations that are
independent from language, while CMT is more precise in ac-
counting for higher level associations mediated by language. This
suggestion should be investigated further, as it would have inter-
esting implications for our understanding of abstract and concrete
concepts.
Strengths and limitations. In comparison to other theories,
the CMT has a number of advantages that have contributed to its
standing as the most influential embodied theory on abstract con-
cepts. Its most important strength is that it proposes a mechanism
underlying the formation and use of abstract concepts, rather than
arguing that a specific content characterizes them. However, the
scope of such a mechanism is limited to specific conceptual
domains. Using concrete-abstract mappings to understand the
meaning of abstract concepts is possible only if there are adequate
domains to be mapped in the first place. For example, one could
question the adequacy of the space-time mapping. The very ab-
stract idea of a general spatial container for all objects is neither
simple nor direct, and it is debatable whether it can be considered
as concrete.
Furthermore, it is unclear whether it is always possible to find
concrete domains corresponding to abstract ones. For example, it
is difficult to think of concrete domains suitable for mapping
abstract concepts such as “philosophy” or “linguistics” (for a
similar critique, see Dove, 2009;Goldman & De Vignemont,
2009). Furthermore, it should be better clarified whether the map-
ping always succeeds, and how it occurs. Consider metaphors such
as “a sea of stars”: one crucial aspect of the sea is that it is liquid,
but this aspect is not relevant for the mapped domain. Similarly,
which aspects of the notion of journey would be used to concep-
tualize “life,” or why is the notion of “power” conceptualized only
in terms of hierarchical structure/verticality? In addition, meta-
phors help to detect similarities between concepts, but not their
differences: life after all is not really a journey. How do we
represent what of life is not a journey? And how about the fact that,
while time flies, space does not (see, e.g., Galton, 2011, for such
a critique)?
In summary, even if conceptual metaphors can play a role in the
representation of abstract concepts, it is not necessarily so: many
abstract conceptualizations do not seem to rely on concrete do-
mains; furthermore, metaphorical mapping is likely to contribute
to the understanding of abstract concepts, but it hardly exhausts
their meaning, since metaphors cannot substitute direct experience
(Barsalou & Wiemer-Hastings, 2005). For example, neuronal re-
gions dedicated to time processing, and not only to space, should
be activated during comprehension of time concepts (Kranjec &
Chatterjee, 2010). These questions remain unanswered, and cast
doubts on the generalizability of the CMT.
Another important limitation of this view (see Dove, 2009, for
such a critique) concerns conceptual development: children start to
use metaphors rather late, and their comprehension of metaphors
remains quite poor until 8 –10 years of age (Winner, Rosenstiel, &
Gardner, 1976), while around 10% of the words used by 2 years-
olds are abstract ones. How can this developmental trajectory be
reconciled with the idea that we use metaphors to understand
abstract concepts (Murphy, 1996,1997)?
Finally, while an increasing body of linguistic and behavioral
evidence supports the CMT, neural evidence is still lacking. For
example, it is unclear how the CMT could explain the results of
brain imaging studies showing that differences exist between met-
aphorical and abstract concepts (e.g., Aziz-Zadeh, Wilson, Rizzo-
latti, & Iacoboni, 2006;Boulenger, Hauk, & Pulvermüller, 2009;
Boulenger, Shtyrov, & Pulvermüller, 2012;Desai, Binder, Conant,
Mano, & Seidenberg 2011;Rüschemeyer, Brass, & Friederici,
2007). Furthermore, recent evidence argues against a strong met-
aphoric account. Bardolph and Coulson (2014) recently recorded
electrocardiogram (EEG) as participants moved marbles either
upward or downward while reading words with literal or meta-
phorical spatial associations (e.g., “ascend,” “descend,” vs. “in-
spire,” “defeat”). They found early (200 –300 ms after word onset)
positive Event Related Potentials (ERPs) for literal words together
with congruent movements, while metaphorical words influenced
ERPs only 500 ms after word onset. Such a result suggests that
participants were sensitive to the association between abstract
concepts and vertical space in line with the metaphor “Good is up,”
but that such an integration does not occur in a rapid and automatic
manner.
In conclusion, the influence of the CMT is clear. However,
while there is evidence supporting this theory, there are a number
of gaps that suggest the need for a more systematic approach; in
particular, it is difficult to foresee that metaphors can be used in all
domains where abstract concepts exist, and it does not seem
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
271
THE CHALLENGE OF ABSTRACT CONCEPTS
plausible that the meaning of abstract words can be entirely ex-
plained by metaphors.
Introspective and situational properties. According to this
view abstract concepts, in comparison to concrete concepts, rely
more on introspection and on social and institutional aspects of
situations (Barsalou & Wiemer-Hastings, 2005). This view is
supported mainly by behavioral evidence, based, for example, on
tasks in which participants are asked to generate the characteristics
of a given concept.
Barsalou and Wiemer-Hastings (2005), for example, asked par-
ticipants to produce characteristics of three abstract (e.g., “truth”),
concrete (e.g., “bird”) and intermediate concepts (e.g., “farm”).
Concepts were either presented in isolation or preceded by a short
scenario. For example, for the concept “truth,” the situation de-
scribed a boy who told his mother that he had not broken a vase,
and his mother believing him. This study found that both concrete
and abstract concepts evoke properties related to situations. How-
ever, with concrete concepts the participants tended to focus on
entities within situations (e.g., “car-wheels”), while abstract con-
cepts focused “on the social, event, and introspective aspects of
situations (e.g., people, communication, beliefs, and complex re-
lations,” p. 152; e.g., “true-difficult to discuss after Postmodern-
ism,” “justice-beaurocratic”). Wiemer-Hastings and Xu (2005)
provided further evidence of this with a larger sample of concepts.
These results highlight that context availability per se is not unique
to abstract concepts, because concrete concepts also evoke situa-
tions (see Moffat et al., 2015, for convergent evidence). However,
even if abstract concepts do not evoke a higher number of contexts
compared with concrete concepts, it is possible that the contexts
they activate are more complex, because they involve more events
and actions (Borghi & Binkofski, 2014;Borghi, Flumini, Cimatti,
Marocco, & Scorolli, 2011;Connell & Lynott, 2012). This would
be in line with two unique aspects of abstract concepts. One unique
aspect of abstract concepts is their “relational” nature (Barsalou,
2003;Gentner, 1981;Gentner & Boroditsky, 2001;Markman &
Stilwell, 2001), that is, the fact that they are often “characterized
by their links to external concepts rather than by intrinsic prop-
erties, unlike most concrete concepts”(Wiemer-Hastings & Xu,
2005). The second unique characteristic is that their members are
more diverse and heterogeneous than those of concrete concepts,
thus the context can play an important role in connecting them,
acting as a sort of glue. A similar mechanism operates with
members of superordinate level concepts (e.g., “animal,” “vehi-
cle”), which is the presence of a common and broad context, where
many exemplars can coexist (e.g., “land” for animals). The pres-
ence of such a context facilitates the recognition of superordinate
category members more than of the members of basic concepts
(e.g., “car,” “dog”; Borghi, Caramelli, & Setti, 2005;Heit &
Barsalou, 1996;Murphy & Wisniewski, 1989).
However, there are limitations to these conclusions, and evi-
dence obtained in a recent dissertation (King, 2013) casts doubt on
the relational character of all abstract concepts, and also highlights
that context impacts processing in only some kinds of abstract
concepts. Participants were presented with short scenarios, and
were then asked to perform a lexical-decision task on abstract
words that were not mentioned during the scenario description.
The scenario facilitated processing of “relational” abstract con-
cepts (e.g., “ignore,” which describes an act, an actor, a patient
being ignored, but no internal feeling), but it did not influence
activation of mental states (e.g., “depressed”).
Roversi, Borghi, and Tummolini (2013) have provided evidence
supporting the introspective and situational view, but also limiting
its explanatory value to some kinds of abstract concepts. They
asked participants to generate properties of concrete and abstract
concepts of three different kinds: artifact, institutional, and social
concepts (e.g., “screwdriver-poetry”; “signature-ownership”; and
“party-friendship”). They found that situations play an important
role in characterizing abstract concepts. Their results also indicated
that the distinction between abstract and concrete concepts was
marked only in the case of social concepts, in which concrete
concepts elicited more specific spatial and temporal relations,
while abstract concepts evoked more general situational relations
(e.g., “friendship-infancy”) and free associations (e.g., “friendship-
pool”).
Zwaan (2016) has built upon Barsalou and Wiemer-Hasting’s
view on abstract concepts and situations. He has suggested that
abstract concepts play a double role during discourse processing.
They can assume an anaphoric function or a cataphoric function.
This duality depends on whether longer parts of text are read or
whether the abstract word is displayed early on in the discourse.
When enough contextual information is provided, an abstract term
such as “justice” activates a sensorimotor simulation. This simu-
lation is then linked with symbolic representations that work
anaphorically as “pointers to previously formed situational repre-
sentation.” Extending this proposition further, we can instead
consider the expression “Now justice is served.” This expression is
characterized by a high level of ambiguity, as context to which it
will be associated it is not yet clear. In such a case, the abstract
concept can work cataphorically to provide a focus to help textual
comprehension, in a similar way to pronouns. The sensorimotor
system, activated once the abstract word is encountered, would be
quickly disengaged to integrate further sensorimotor information.
The concept would thus work as “a placeholder in an active state
in working memory,” that can help to integrate sensorimotor
information acquired later (Zwaan, 2016). This interesting pro-
posal points out that concepts can differ depending on whether or
not they are presented in isolation, and when they are introduced
in a discourse.
However, even if the author presents this explanation as an
extension of the view according to which contextual information is
less focal for abstract than for concrete concepts (Barsalou &
Wiemer-Hastings, 2005), this perspective seems to be more in line
with the multiple representation theories. Zwaan (2016) appears
indeed to assume that an interplay between sensorimotor, emo-
tional, and linguistically conveyed information occurs, at different
moments in time, in line with such theories (see also Zwaan, 2014
and below).
Strengths and limitations. In common with many of the
theories evaluated within this review, the introspective and situa-
tional properties theory has strengths but also gaps in relation to its
applicability and explanatory power. The hypothesis that abstract
concepts activate introspective and social-situational components,
has the advantage of being rather general: it can be applied
intuitively to a wide range of concepts, from mental states to social
concepts and to purely abstract concepts. However, as we have
seen, the role played by context and by introspective properties
varies in function of the content of specific concepts. A further
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
272 BORGHI ET AL.
limitation that weakens the validity of this theory is that, so far,
evidence favoring this view has mainly been confined to property
generation tasks (Pecher et al., 2011), and, further, that supporting
evidence obtained with different paradigms is needed. More im-
portantly, in this theory, introspection is conceived as a form of
conscious self-awareness, which provides a severely biased image
of the relevant process. Indeed, it is possible that a more implicit
form of metacognition plays a more extensive role in grounding
abstract concepts (Frith, 2012), similar to the role it plays in
supporting social interaction (Bahrami et al., 2010). In other
words, the main limit of this view is that the mechanisms of this
internal grounding have not been fully elucidated. For example, it
could be hypothesized that, the more abstract a concept is, the
more it requires grounding on information of internal states and
processes (and possibly also some forms of internal speech), or
that the more abstract they are, the more they need a complex
context to put together the varied and sparse category members.
These are only two of the possible interpretations: therefore, it is
essential that further research is carried out to determine and
evaluate the underlying mechanisms.
Affective Embodiment Account (AEA). A third perspective
we will evaluate is the Affective Embodiment Account (AEA) and
we will show that it provides important evidence that abstract and
concrete concepts differ. The AEA is a new theory which hypoth-
esizes that abstract concepts evoke more emotions than concrete
ones (e.g., Kousta et al., 2011;Vigliocco, Kousta, Vinson, An-
drews, & Del Campo, 2013). The AEA view of abstract concepts,
outlined by Kousta et al. (2011; see also Kousta et al., 2009) and
by Vigliocco et al. (2013), proposes that abstract and concrete
concepts differ in terms of the experiences that characterize them.
While perceptual and motor experience is more crucial for con-
crete concepts, affective experience and emotional development is
more important for abstract concepts. Emotional development is
presented as the basis for the acquisition of an abstract vocabulary:
the AEA proposes that emotions can provide a bootstrapping
mechanism that favors the acquisition of abstract words (e.g.,
Vigliocco et al., 2013). AEA theorists argue that this is demon-
strated by the fact that emotional abstract words are learned earlier
than neutral abstract words. To illustrate and explain these con-
clusions, we will assess the behavioral and brain-imaging evidence
supporting this view. These studies provide evidence that, for
example, abstract words typically receive higher ratings for emo-
tional associations and that areas involved in emotion processing
are engaged during processing of abstract words (e.g., Vigliocco et
al., 2014).
Kousta et al. (2011) used a large sample of abstract and concrete
words, controlled for a variety of dimensions such as familiarity,
concreteness, abstractness, context availability (see CAT), image-
ability (see DCT), and age of acquisition. Words were also con-
trolled for mode of acquisition (Della Rosa, Catricala
`, Vigliocco,
& Cappa, 2010;Wauters, Tellings, Van Bon, & Van Haaften,
2003). Mode of acquisition determined whether the word was
acquired perceptually (e.g., by interacting with the object it refers
to; “bottle”), linguistically (e.g., by listening to explanations; “phi-
losophy”), or with a mixture of both modalities (e.g., being shown
a picture and explained the word meaning; “tundra”). The authors
found an advantage in lexical decision of abstract over concrete
words. That is to say, once imageability and context availability
are controlled for, the concreteness effect was substituted by an
abstractness effect. This same abstractness effect was found in
regression analyses of lexical decision response times for a large
sample of words (n⫽2,330).
The advantage of abstract over concrete words can be explained
by the difference in valence between concrete and abstract con-
cepts: Kousta et al. (2011), therefore, proposed that affective and
emotional information has a major weight in characterizing ab-
stract words. They concluded that, because context availability and
imageability were kept constant, classical theories such as the CAT
or the DCT cannot account for these data. Even if the DCT
recognizes the importance of emotions (Paivio, 2013), the pres-
ence of emotional connotations for nonemotional words would
derive from their link with imagistic representations. Thus, emo-
tional connotations should be more frequent with concrete than
with abstract concepts (Vigliocco et al., 2013). In addition, given
that modality of acquisition was kept constant, they argue that
differences in activation of linguistically conveyed information do
not exhaust the differences between concrete and abstract con-
cepts, and that emotions play a major role in abstract concept
representation. Valence however does not seem to exhaust the
meaning of abstract words: the advantage of abstract words in
accuracy was still maintained even when the effect of valence was
removed. This might be because of the role played by linguistically
conveyed information for abstract concepts. Furthermore, control-
ling for valence can lead to conflicting results: in a recent ERPs
study with stimuli controlled for imageability, context availability,
and valence, the abstractness effect was replicated in response
times but a reverse concreteness effect was obtained with ERPs
(Barber et al., 2013). Functional magnetic resonance imaging
(fMRI) results further support the AEA, but here too there is
controversial evidence on the role played by valence. Vigliocco et
al. (2014) performed an fRMI study with a lexical-decision task
and stated that, while concrete and abstract concepts activated the
visual processing system involving occipital, temporal, and sub-
cortical areas, only abstract concepts engaged the rostral anterior
cingulated cortex (rACC). rACC plays a regulatory role during
emotions processing. However, the conclusions drawn by
Vigliocco et al. (2014) have been questioned by Skipper and Olson
(2014) who confirmed that rACC responds to valence, but not to
abstract words: controlling for valence they found that the right
STS and right temporal pole were the only regions that remained
exclusively responsive to abstract concepts.
Further evidence partially favoring the AEA has been collected
with patients with neurodegenerative diseases. Catricala
`, Della
Rosa, Plebani, Vigliocco, & Cappa (2014) submitted patients with
Alzheimer’s disease (AD) and with primary progressive aphasia
(sv-PPA) to the same tasks (Della Rosa et al., 2014) using both
concrete and abstract words. Results showed that AD patients
differed from controls in all tasks with concrete and abstract items,
while the sv-PPA group differed from controls in all concrete
tasks, but only in the sentence completion task for abstract con-
cepts, that is, only in production and not in comprehension tasks.
The authors deepened their analysis by categorizing abstract con-
cepts into two types: social concepts and emotions. They demon-
strated that sv-PPA patients were impaired only in social concepts,
and AD patients in all abstract concepts with the exception of
emotion concepts.
Strengths and limitations. Overall, evidence favoring the
AEA is based on results obtained with different methods, from
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
273
THE CHALLENGE OF ABSTRACT CONCEPTS
behavioral studies to fMRI investigations. An obvious conclusion
is the importance of emotions for abstract concepts. The AEA
researchers have also identified a bootstrapping role played by
emotions to facilitate learning of abstract concepts. They add
weight to this conclusion by arguing that emotion development
precedes language acquisition. However, this could also be said for
other cognitive abilities, for example novelty detection (see
Crutch, Troche, Reilly, & Ridgway, 2013).
The AEA, however, has three limitations with relevant impli-
cations. First, the evidence supporting it is not fully consistent. For
example, ERP studies revealed a dissociation between an abstract-
ness effect in RTs and a concreteness effect in ERPs. A very recent
study in which facial muscle activity was measured during visual
word recognition found a valence effect in the m. corrugator
supercilii with concrete but not with abstract words (Kuenecke,
Sommer, Schacht, & Palazova, 2015). Another recent ERPs study
with a lexical-decision task with concrete and abstract verbs of
different valence has revealed that the emotion related EPN effect,
that is, an early posterior negativity effect associated with attention
to word meaning, emerged earlier for concrete than for abstract
verbs (Palazova, Sommer, & Schacht, 2013). These results clearly
contradict the AEA, which would predict the opposite result.
Second, the role of valence does not seem to account for all the
variance, and thus it cannot offer an exhaustive explanation of the
mechanisms underlying abstract concepts (see previous discussion
of the behavioral results by Kousta et al., 2011, and of the fRMI
results by Vigliocco et al., 2013). Third, it is possible that the
results obtained were biased by the presence, within the sample of
abstract concepts, of emotional concepts. Altarriba, Bauer, &
Benvenuto (1999) (see also Altarriba & Bauer, 2004;Setti &
Caramelli, 2005) have contended that emotions represent a subset
not to be included within abstract concepts, and have proposed that
a trichotomy between abstract-concrete-emotional words exists
rather than a simple abstract-concrete dichotomy.
Let us consider the issue from an embodied perspective. Emo-
tions are very basic functions, activated by rather primitive stimuli,
such as smell. Emotion concepts can be seen as either a peculiar or
a separate subset within abstract concepts, for which an embodied
explanation is more straightforward and intuitive than for other
abstract concepts, because they are typically associated with spe-
cific bodily expressions/states. It has been shown that the mecha-
nisms underlying representation of emotions and of sensorimotor
processes is the same, and are based on simulation and reenact-
ment of the situations experienced while interacting with the world
(Wilson-Mendenhall, Barrett, Simmons, & Barsalou, 2011). While
such mechanisms may play an important role in explaining purely
abstract concepts, they are clearly insufficient to fully explain
them.
In summary, all emotional concepts are to a certain extent
abstract, but it not clear whether all abstract concepts have affec-
tive/emotional connotations. Kousta, Vigliocco, Vinson, and An-
drews (2009) have suggested that valence is not a characteristic
limited to emotional words, but that it can be extended to many
concepts. The claim that emotional aspects characterize abstract
concepts is certainly plausible. However, it should be clarified
whether the effects found depend on the inclusion of emotional
concepts within the sample of abstract concepts, or whether they
depend on a mechanism underlying the representation of all ab-
stract concepts. Indeed, the contention that emotional valence
increases with abstractness needs to be corroborated by further
investigation.
The two last theories evaluated here share a common strength.
They do not characterize abstract concepts in terms of what they
miss, but, rather, in terms of their specific characteristics. Unfor-
tunately, they also share the same limitation. Data collected so far
convincingly demonstrates that abstract concepts evoke more in-
trospective and situational aspects, as well as more emotions.
However, this result can be linked to the specific content of the
concepts included in the database. For example, introspective
properties might be more frequently evoked by abstract mental
state concepts, such as “meditation” or “thought,” than by other
abstract concepts such as “situation” or “event.” It is certainly
important to distinguish concepts on the basis of their content, but
it is even more important to identify the mechanism underlying
abstract concept formation and use, as well as which mechanism
becomes more active as the abstractness level of a concept is
raised.
Multiple representation theories. Multiple representation
theories of abstract concepts represent a valuable and novel addi-
tion to current thinking and appear to be a promising avenue of
research. Researchers who endorse this perspective have claimed
that, even if abstract concepts are embodied in perception, action,
and emotion, they also rely more on reenactment of linguistic
experience. They also state that both sensorimotor and linguisti-
cally conveyed information is involved in the representation of
concrete and abstract concepts, but that these pieces of information
are differently distributed depending on the type of concept: per-
ception and action information is more important for concrete
concepts, emotion, and linguistically conveyed information for
abstract ones. As we will see, while all multiple representation
views highlight the importance of language for abstract concepts
representation, they characterize the role of language differently:
language can be considered simply as a shortcut to access word
meaning (LASS theory), or it can be given a more crucial role, as
a means of improving thought processes (Dove’s representational
pluralism), and as a complex bodily and social experience (WAT
theory).
This review will focus in particular on four multiple represen-
tation theories that are currently influential, or have the potential to
be so: the LASS (Language and Situated Simulation) view (e.g.,
Barsalou, Santos, Simmons, & Wilson, 2008), the representational
pluralism view proposed by Dove (2009,2011,2014), the WAT
(Words As social Tools) view (Borghi, 2013;Borghi & Binkofski,
2014;Borghi & Cimatti, 2009), and the grounding and sign track-
ing proposal by Prinz (2002,2012).
LASS (Language and Situated Simulation). According to the
LASS view (Barsalou et al., 2008), multiple systems underlie our
conceptual knowledge. LASS focuses on the linguistic and the
simulation systems, which continuously interact. Consider the
word “cat”. When we are presented with the word “cat”, we first
activate the linguistic system, recognizing the word form and
producing associated words, such as “dog,” “Siamese,” and so
forth. We then start to ground the concept, for example visualizing
the cat and reenacting our interaction with it.
As this example shows, according to the LASS theory, the
linguistic system is engaged to categorize words during their
perception. During a linguistic task, such as lexical decision, the
activation of linguistic forms peaks earlier, in line with the encod-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
274 BORGHI ET AL.
ing specificity principle (Tulving & Thomson, 1973). While at this
level words are processed only in a shallow way, when the simu-
lation system is involved, processing is deeper, in line with the
idea that, compared with words, images access the conceptual
system more directly (Glaser, 1992; see also Paivio, 1986). After
word recognition the simulation system is activated, that is, the
brain simulates the sensorimotor and mental states active during
interactions with the referents of the word. Word associations (e.g.,
“dog-dog’s bed”) can thus provide a shortcut, allowing fast access
to conceptual information (Barsalou et al., 2008, and converging
evidence: e.g., Connell & Lynott, 2013;Louwerse & Connell,
2011;Pecher et al., 2011). Put simply, according to LASS, words
work as “pointers” to the object, entity, or situation they refer to.
Property verification and fMRI research (review in Barsalou et
al., 2008) has provided evidence for activation of the simulation
system occurring in parallel with that of the linguistic system, but
slightly more slowly. Simmons, Hamann, Harenski, Hu, and Bar-
salou (2008) conducted an fMRI study where participants per-
formed a silent property generation task. Then, in a subsequent
scanning section, they underwent two localizer tasks. In the first
localizer task, they were asked to produce word associations for
different kinds of concepts (e.g., common objects such as car,
mental states such as “guess,” abstract concepts such as “exten-
sion” etc.), and in the second task to imagine a situation containing
the concept. Early phases of word generation in the silent property
generation task activated the same brain areas involved during the
first localizer task, that is, areas typically involved during linguis-
tic tasks, such as the left inferior frontal gyrus (Broca’s area) and
the right cerebellum. Late phases of word generation, however,
instead activated the same brain areas involved during the second
localizing task, that is, brain areas that are typically active during
mental imagery, such as precuneus and right middle temporal
gyrus.
Although these findings are novel and very compelling, there
exist some caveats. Given that fMRI has a scarce temporal reso-
lution, further research on timing would complement and further
the predictions of the LASS view. Most important, even if the
reported results provide evidence that the linguistic system is
engaged earlier than the imagery system, this does not necessarily
imply that meaning is only activated later, during activation of the
simulation system, as the LASS theorists suggest. The activation
of Broca’s area (see Rizzolatti & Craighero, 2004, for a review)
could instead be evidence of the activation of the motor system to
prepare for a situated action.
In addition, the LASS theory is not specifically aimed at ex-
plaining abstract concepts. In principle, one could conclude from
its principles that, while concrete concepts activate the simulation
system, abstract concepts activate the linguistic one. However, this
conclusion is not proposed by the LASS, even if is consistent with
its principles. In contrast, LASS theorists claim, on the basis of
fMRI evidence (Wilson-Mendenhall, Simmons, Martin, & Barsa-
lou, 2013), that both concrete and abstract concepts activate a
mixture of simulation and linguistic form information, that are
differently distributed, depending on the task. They also argue that
different brain areas are activated depending on the conceptual
content.
Wilson-Mendenhall et al. (2013) asked participants to think
deeply about an abstract concept (e.g., “convince”). They were
then asked to verify whether the concept applied to the picture of
a scene (e.g., a politician speaking) presented to them. This study
showed that the brain areas related to the content of the concept
were active, and that concrete and abstract concepts activated
nonlinguistic semantic information. For example, the concept
“convince” activates brain areas related to mentalizing, and the
concept “arithmetic” areas related to numerical processing. How-
ever, the lack of activation of linguistically conveyed semantic
information can be biased by the fact that in this study words were
processed only in relation to a subsequently presented picture.
Processing a word in relation to an image differs from processing
it in the context of other words, and such a control condition was
not present.
To summarize, the LASS theory argues for “multiple represen-
tation” of both concrete and abstract concepts, stating that each
concept, be it concrete or abstract, activates different brain areas
depending on its content. It is the task, not the kind of concept,
which determines a higher engagement of linguistic versus senso-
rimotor areas. For example, lexical decision tasks, that require one
simply to decide whether a word exists or not, activate the lin-
guistic system more, while imagination tasks activate the simula-
tion system preferentially.
Strengths and limitations. The LASS is an important pro-
posal because it is, to our knowledge, the first fully embodied
theory that recognizes the importance for concepts of not only
perception, action, and emotions, but also of language. Compared
with the perspectives we have previously outlined in this review,
LASS has a further strong advantage that sets it apart: it identifies
a mechanism underlying conceptual representation, that is, the
parallel— but also slightly different in time—activation of the
linguistic and sensorimotor system. However, such a mechanism is
not specifically applied to the representation of different kinds of
concepts, but rather to the adopted task.
One final potential limitation of LASS is that it argues that
language is simply used as a shortcut to access meaning, and it is
implied that shallow linguistic tasks do not allow access to mean-
ing. This is contradicted by evidence suggesting that access to
meaning is very fast, as it can occur 150 ms after word onset (e.g.,
Hauk, Johnsrude, & Pulvermüller, 2004).
Representational pluralism. The second Multiple Represen-
tation theory presented here is Representational Pluralism. This
theory was proposed by Dove (2009,2011,2014), and unlike the
LASS view, is devoted specifically to the explanation of abstract
concepts. In its original formulation, this view argued that, while
concrete concepts could be explained through an embodied view,
abstract concepts could not (Dove, 2009). More recently, Dove
(2011,2014) has suggested that both abstract and concrete con-
cepts are embodied, although to different extents. This view can,
however, be considered as being hybrid in nature, as it combines
embodied and nonembodied aspects. Dove is namely committed to
representational pluralism, and proposes that to account for the
representation of abstract concepts we must leave room for some
amodal representations.
Dove’s theory relies on Paivio’s (1986) DCT, and can be con-
sidered as the attempt to return some important aspects of DCT to
the forefront by adopting an embodied framework. Dove clarifies
that his theory departs from Paivio’s view in that he takes an
embodied stance, proposing that perceptual symbols (Barsalou,
1999) rather than mental images are the basic units of both verbal
and nonverbal representations. Perceptual symbols are perceptual
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
275
THE CHALLENGE OF ABSTRACT CONCEPTS
experiences that involve all senses and can be combined in the
same way as symbols. These symbols can be multimodal, sche-
matic, and they are not necessarily conscious. Language plays a
major role in abstract concepts representation. According to Dove,
linguistic representations are only partially embodied. They are
embodied because they rely on sensorimotor simulations, but they
acquire meaning not because of this form of embodiment, but
because of their relationship with other words. In this sense, this
view is also close to distributional approaches, according to which
meaning is derived from associations between words (e.g., An-
drews et al., 2014).
The prevailing importance of language defines what abstract
concepts are, allowing them to be distinguished from concrete
concepts. Language plays an essential role, as it empowers our
cognitive capabilities by serving as a medium of thought thanks to
its combinatorial characteristics (Dove, 2011,2014). According to
Dove, when we acquire language, we acquire a new and powerful
representational system, which interacts with other embodied sys-
tems but does not overlap with them. This system consists of
simulations that can be selective and partial, and do not involve
forms of inner speech, since simulations can be completely un-
conscious. Language in his view is “is an internalized amodal
symbol system that is built on an embodied substrate.” According
to Dove, a fully embodied account of the syntactic, morphological,
and phonological structure of language is unattainable: even psy-
choliguistics accounts that posit significant interaction between
comprehension and production systems (such as Pickering and
Garrod’s theory; Pickering & Garrod, 2013) are not fully embod-
ied, because they generally rely on intermediate representations
that work as a bridge between production and comprehension
(Dove, 2013).
Although Dove does not directly produce evidence to support
his theory, he discusses recent evidence that demonstrates the
importance of imageability for conceptual processing. He reports
studies on double dissociations showing that, while damage of the
left hemisphere leads to a selective semantic impairment for high
imageable words, the opposite case is also observed. Furthermore,
he discusses ERPs studies showing a larger and more anterior
N400 for abstract compared with concrete words, which would
suggest that different systems are used for the two kinds of
concepts. His conclusions are further supported by fMRI studies
that have shown a higher engagement of superior regions of the
left temporal lobe and inferior regions of the left prefrontal cortex
for abstract compared to concrete words, and right hemisphere or
bilateral activation in the case of concrete words. Dove notes that
while the evidence he uses clearly shows that two different neural
systems are engaged (e.g., Adorni & Proverbio, 2012), it is com-
patible both with the view that linguistic representations are modal
and with those claiming that they are amodal, and further research
is needed to disentangle this complex issue.
Dove (2015) further contributed to the debate surrounding em-
bodied cognition by suggesting that abstract concepts pose three
different problems to embodied cognition, for which different
solutions can be available. The first problem he notes is general-
ization, which is the problem of how we are able to represent
information that goes beyond our experience. The second issue is
the problem of flexibility, which arises because it is possible that
embodied representations are activated differently, or at different
levels of depth, depending on the context and task. The third and
final challenge is the problem of disembodiment: the “embodied”
nature of abstract concepts needs to be demonstrated, and repre-
sents a challenge for embodied cognition. All three problems
pertain to both concrete and abstract concepts, but are particularly
urgent and marked for the latter.
However, these issues are not without solutions and each theory
can be suitable to handle and solve different problems. Consider
for example the emotion-based AEA: while it partially solves the
problem of disembodiment, showing that abstract concepts are also
grounded in bodily states, it seems unable to handle the problems
of flexibility (why should emotional systems be differently acti-
vated depending on the task and context?), of generalization (e.g.,
how to account for the formation of high level superordinate
concepts, such as “animal”?), and to explain why sensorimotor
experience is so crucial for some concepts and less so for others.
Strengths and limitations. The theory proposed by Dove has
many advantages, among which is the important recognition of the
role played by linguistic experience in conceptual processing.
However, in common with many theories presented in this review,
further research is required to provide conclusive evidence in favor
of this theory. This limitation is particularly marked for this theory
as it is not directly supported by empirical work, even if the author
illustrates and discusses convergent evidence. Furthermore, much
of the behavioral and neuroscientific supporting evidence he re-
ports relies on the important role of imageability. However, it has
been recently shown that imageability is correlated but cannot be
conflated with abstractness (e.g., Kousta et al., 2011; see above).
An additional limitation is that this theory focuses on concept
representation in adults, without considering how concepts are
acquired. Furthermore in its different formulations it remains am-
biguous as to the role played by amodal and perceptual symbols.
Another limitation of this theory, is that Dove initially claimed
that abstract concepts are represented through amodal symbols
(Dove, 2009). He then argued that even if abstract concepts acti-
vate sensorimotor simulations, this is not the way in which they
acquire their meaning (Dove, 2011), and claimed that language
learning would lead to the acquisition of a “new dis-embodied
semantic system.” In this respect, the relation between this view
and the fully embodied the WAT proposal, which will be discussed
later in this review, should be further explored. At a general level,
the two theories partly overlap, because both claim that concepts
can either be associated mainly with nonlinguistic experience of
the world or with experience of language. Furthermore, the con-
ception of language as an instrument that extends thought capa-
bilities is fully compatible with the WAT view, and it helps in
clarifying why language is helpful to represent abstract concepts.
However, the two theories differ in their level of embodiment and
in the space they leave for amodal representations. According to
Dove (2011), his accounts differs from the WAT view
because it holds that the acquisition of language creates a new
dis-embodied semantic system, one that has many of the properties
usually associated with the amodal symbol systems favored by tradi-
tional cognitive science. In other words, natural language on my view
is not merely another source of information about the world but is also
another way of ‘thinking about the world.’
While we fully appreciate the idea that language can contribute to
improve our thinking capabilities, we believe that proposing to
leave room for amodal representations, as Dove does in his pro-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
276 BORGHI ET AL.
posal, risks incuring the symbol grounding problem that we dis-
cussed in the Introduction. In our view, this is a problem shared by
all nonembodied, but also by all hybrid approaches to cognition.
The amodal and arbitrary nature of language has been recently
questioned by studies on iconicity in spoken and sign languages,
showing the similarity between some characteristics of words and
some features of their referents. It is the case of sound-shape
correspondences: for example, words with rounded vowels—such
as “bouba” or “maluma”—as typically used to refer to rounded
shapes, words with unrounded vowels—as “kikki,” or “ta-
kete”—to jagged shapes (Gentilucci & Corballis, 2006;Maurer,
Pathman, & Mondloch, 2006;Perniss & Vigliocco, 2014). In
addition, we believe that not only amodal symbols, but also
grounded symbols possess combinatorial properties and can ex-
hibit productivity (Barsalou, 1999). Grounded linguistic symbols
exhibit, in our view, all the properties allowing them to support our
thought processes. Finally, if with amodal symbols Dove refers to
symbols represented in brain areas where information is integrated,
we prefer to intend convergence areas in the brain (Damasio, 1989)
as areas where multimodal information converges. The debate is
obviously open, and new evidence is needed to provide definite
answers, but we think that the most promising direction of research
to explain abstract concepts is one that not only emphasizes the
importance of language, as Dove’s theory does, but that also
conceives language as a fully modal and embodied system, as
Dove’s theory does only in part.
WAT (Words as Social Tools). WAT (Borghi, 2013;Borghi
& Binkofski, 2014;Borghi & Cimatti, 2009) proposes that con-
cepts are couched in representations that derive from both percep-
tual/motor and linguistic experiences. In common with the LASS
theory, it is a fully embodied view, because it maintains that
sensorimotor simulations need to be activated to grasp meaning.
We, however, will consider this as a weak embodied view,
similarly to the LASS theory, because it argues that not only
sensorimotor, but also linguistic areas—specifically, areas related
to auditory processing, language production and phonology, and so
forth—and social areas, for example areas engaged during recog-
nition of known others, are involved in representing the meaning
of abstract concepts.
According to the WAT proposal both concrete and abstract
concepts are embodied. At the same time in the case of abstract
concepts grounding in sensorimotor systems is not sufficient. In
contrast to other theories emphasizing the role of language, the
WAT view does not consider only the importance of word asso-
ciations in explaining meaning. Consider the distinction between
kinds of linguistic information we proposed at the beginning of the
review: this proposal assumes that embodied forms of all kinds of
linguistic information, that is, of linguistically conveyed informa-
tion, linguistic form information and syntactic linguistic informa-
tion, concur in their representation. Furthermore, this theory fo-
cuses also on the experience of language use itself, emphasizing
the role of the social dimension in word acquisition. In this view,
both concrete and abstract words are seen as social tools that help
us to interact with others, and with the physical and social envi-
ronment (for a more radical recent theory directly inspired by
WAT that further elaborate similar themes see the Words as
Cultivators (WAC) view; Schilhab, 2015a,2015b).
The WAT proposal has four main tenets, and these will be
evaluated in light of recent evidence (for an overview of the
theory, see Borghi & Binkofski, 2014). The first principle is that
concrete and abstract concepts are characterized by different ac-
quisition modalities (e.g., Bergelson & Swingley, 2013;Borghi et
al., 2011;Granito, Scorolli, & Borghi, 2015). This implies that,
given that abstract concepts do not have single, concrete referents,
but rather have sparse and diverse referents, they will be acquired
both through sensorimotor experience and linguistic input. In
addition, the physical environment is less available as a scaffold to
support the acquisition of abstract concepts than is the case for
concrete ones: language itself plays a scaffolding role for abstract
concepts. Literature on Modality of Acquisition (MoA; Wauters et
al., 2003) shows that some words are acquired mainly through
sensorimotor experience (e.g., “bottle”), more abstract words are
mainly acquired through linguistic input (e.g., “philosophy”),
while still other words have an intermediate status (e.g., “tundra”).
While in early grades words are mainly acquired through percep-
tual modality, in Grade 6 texts the majority of words are learnt
through linguistic input. Modality of acquisition and age of acqui-
sition are correlated but not overlapping dimensions. However,
studies on age of acquisition indicate that abstract concepts are
acquired later than concrete ones, at a time when children have
already mastered many words. One influential hypothesis on con-
ceptual development maintains that, to acquire abstract words,
children need to master a consistent amount of words and of
linguistic knowledge, and that the acquisition of syntax and se-
mantics are strictly interwoven (Gleitman, Cassidy, Papafragou,
Nappa, & Trueswell, 2005).
The WAT view contends that, while for concrete words children
may simply (or mostly) base themselves on associative mecha-
nisms between words and their referents, learning abstract con-
cepts is likely to require not only linguistic knowledge, but also
sophisticated social abilities. More important, not only the linguis-
tic but also the social competence (Tomasello & Akhtar, 1995) are
particularly relevant for learning abstract words, because the se-
lection of their referent is more difficult: it is not by chance that
they were called “hard” words! (Gentner, 2006;Gillette, Gleitman,
Gleitman, & Lederer, 1999;Gleitman et al., 2005;Wauters et al.,
2003). Studies investigating the early comprehension of abstract
concepts (e.g., Bergelson & Swingley, 2013) reveal that abstract
concepts (e.g., “all gone”) emerge at around 10 months and be-
come more stable at around 14 months. Abstract concepts start to
be comprehended in correspondence with the emergence of some
important social abilities: the ability to follow the gaze of others;
and the capability to develop forms of joint attention allowing
children to determine what they and the adults “know together”
(Carpenter, Nagell, Tomasello, Butterworth, & Moore, 1998). Lit-
erature on testimony found that preschool children are willing to
receive clarifications from adults, and that at around 3 or 4 years
of age they are able and keen to monitor the accuracy of the
information they receive (Corriveau & Harris, 2009;Sabbagh &
Baldwin, 2001). Crucially, testimony is particularly relevant in
domains where the environmental inputs are insufficient (Harris &
Koenig, 2006), such as those of abstract concepts.
The second principle of the WAT view is that different acqui-
sition modalities influence the conceptual representation in the
brain. Given the diversity and sparse character of their referents,
abstract concepts should engage brain areas that are more distrib-
uted and less focused than those associated with concrete concepts
(see Rodríguez-Ferreiro, Gennari, Davies, & Cuetos, 2011 for
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
277
THE CHALLENGE OF ABSTRACT CONCEPTS
consistent fMRI evidence with abstract verbs). According to the
WAT view, while both concrete and abstract concepts activate
sensorimotor networks, the areas that are part of the language
processing system are more activated by abstract than by con-
crete concepts processing. To verify this hypothesis, Sakreida et
al. (2013) and Scorolli et al. (2012) performed an fMRI and a TMS
study based on a previous behavioral study (Scorolli et al., 2011)
using sentences in which both the noun and the verb could be
concrete or abstract (e.g., “to caress the dog/the idea,” “to think of
the dog/the idea”). The fMRI study demonstrated that abstract
concepts activated the core sensorimotor areas, that is, the left
lateral (precentral gyrus) and the medial (supplementary motor
area) premotor cortex. While purely concrete expressions (e.g., “to
caress the dog”) elicited activation within the left inferior frontal
gyrus (pars triangularis) and two foci within the left inferior
parietal cortex, purely abstract word pairs (e.g., “to think of the
idea”) engaged the anterior part of left middle temporal gyrus, that
is to say part of the language processing system. Consistently, in
the TMS study, phrases containing both concrete and abstract
verbs activated the hand-related motor system (MEPs, motor
evoked potentials), but the activation of this system was delayed
with phrases containing abstract verbs. This might be because of a
cascade flow of activation from the mouth- to the hand-related
motor areas. There is a limitation to the TMS study, as its inter-
pretation remains partly speculative, since no control study was
run with the mouth’s MEPs. Furthermore, results of the TMS study
converged with the ones of the behavioral study (Scorolli et al.,
2011) in showing that response times were faster for congruent
pairs (abstract verb-abstract noun, concrete verb-concrete noun)
than for noncongruent ones (Scorolli, 2014). This suggests that
abstract and concrete concepts are represented in partially differing
circuits, and that switching from one circuit to the other (e.g., from
the core sensorimotor to the linguistic system; see Sakreida et al.,
2013) implies a cost in terms of processing time.
The third principle of the WAT view is that different acquisition
modalities lead to different embodied counterparts. WAT predicts
that, because of their link with language, abstract concepts activate
the mouth-related motor system more, while concrete concepts,
more linked to manipulative actions, activate the hand-related
motor system more. This prediction has been supported by two
studies (Borghi et al., 2011;Granito et al., 2015) in which partic-
ipants were taught new categories and then novel labels and by a
study in which participants were required to decide whether ab-
stract and concrete definitions matched with concrete and abstract
concepts (Borghi & Zarcone, 2016).
Borghi et al. (2011) operationalized concrete categories as novel
three-dimensional (3D) objects presented on the computer screen,
which differed in color and shape, and abstract words as groups of
moving objects that interacted in novel ways. Participants were
asked to “manipulate” the concrete objects by moving them on the
computer screen, and to observe the interaction of the members of
abstract categories. A subsequent property verification task (“is
XX a property of CALONA?”) revealed that, while responses to
concrete words were faster with the hand, that is, when participants
had to press a key on the keyboard, abstract words were facilitated
with responses with the mouth, that is, when participants had to
say “yes” on the microphone. The advantage of the mouth was
more pronounced when abstract concepts were introduced using
not only names, but also explanations of their meaning. Further
evidence supporting this principle of the WAT view was provided
by Granito et al. (2015), who replicated the mouth advantage with
abstract concepts. In this study novel names and explanations of
the category meaning were introduced by a researcher to mimic the
social situation that typically characterizes conceptual acquisition.
Results revealed that, with abstract words, participants who had
undergone the linguistic training performed better than the other
participants when they were required to respond with the mouth
(microphone) in a categorical recognition task (“Do XX and YY
belong to the same category?”; Granito et al., 2015).
Further rating studies in which participants were asked how
much different effectors were involved with action with target
words/sentences confirmed the association between abstract con-
cepts and the mouth. Granito et al. (2015) and Borghi and Zarcone
(2016) found that participants associated concrete categories with
the hand, and abstract categories with the mouth. Ghio, Vaghi, and
Tettamanti (2013) found convergent results: they showed that
while abstract concepts related to mental states and emotions are
more associated with the mouth, number concepts are more asso-
ciated with the hand, probably because of the influence of finger
counting on numerical cognition (Fischer & Brugger, 2011;Ran-
zini et al., 2011; see also Moseley, Carota, Hauk, Mohr, & Pul-
vermüller, 2012, for fMRI evidence showing that abstract emo-
tional words, beyond limbic regions, engage areas of the precentral
cortex activated somatotopically by mouth and hand words, and
Moseley et al., 2015, for evidence on the reduced brain activation
with abstract emotion words in motor areas and cingulate cortex in
Autism Spectrum Condition). Consistent results with numbers
were found by Grade, Badets, and Pesenti (2016) who recently
asked participants to randomly generate numbers after observing
stimuli representing aperture or closure of fingers, mouth, and
hand. They found that number meaning is mainly grounded in
mechanisms linked to finger control; mouth influence was limited
to opening, and led to an overproduction of large numbers. Borghi
and Zarcone (2016) recently performed the first behavioral study
in which the association between abstract words and the mouth
was demonstrated with real words and with an implicit task in
which response times were recorded. Participants were asked to
decide whether abstract and concrete definitions presented as
prime matched with abstract and concrete target words using a
device to press within the teeth or pressing a keyboard with the
index finger. Response times with the mouth were overall slower,
likely because of the device used, but the advantage of the hand
over the mouth responses was more pronounced with concrete than
with abstract target-words.
fMRI studies have provided further evidence potentially in line
with the WAT proposal, and have shown that abstract concepts
activate the left inferior frontal gyrus and the left middle temporal
gyrus (for meta-analyses see Binder, Desai, Graves, & Conant,
2009;Wang, Conder, Blitzer, & Shinkareva, 2010). The left infe-
rior frontal gyrus region is involved in language production and
phonological processing, in subvocalization (Fiebach, Ricker,
Friederici, & Jacobs, 2007), for example during processing of
pseudowords in lexical decision, or in working memory mainte-
nance (Petrides, 1994); a TMS study has shown that lexical deci-
sion with abstract words was less accurate after stimulation of this
area (Papagno, Fogliata, Catricala
`, & Miniussi, 2009). The activa-
tion of this region suggests that the linguistic areas related to the
vocal and motor aspects of word processing are activated (Lieber-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
278 BORGHI ET AL.
man, 2009). Recent fMRI evidence from Hoffman, Binney, and
Lambon Ralph (2015) confirmed that concrete concepts are more
associated with visual experience and abstract with acoustic expe-
rience, as revealed by the higher activation of dorsolateral tempo-
ral areas for abstract concepts and of ventromedial temporal areas
for concrete ones, while both kinds of concepts converge in acti-
vation of ventrolateral anterior temporal lobe (ATL). Further fMRI
evidence with an orthographic judgment task is in line with the
WAT view. It revealed that, compared with concrete concepts,
additional language areas were engaged by abstract concepts in
bilateral superior temporal and bilateral middle temporal region
(Kumar, 2015).To summarize, fMRI studies converge in showing
a higher engagement of the left inferior frontal gyrus, typically
engaged for language production and phonological processing, of
the superior temporal gyrus, involved in auditory processing but
also in social cognition (see Martin, 2016, for showing how
social/emotional concepts are grounded in a circuit connecting
anterior temporal and prefrontal regions, crucial for social func-
tioning), and of the middle temporal gyrus, which is engaged
during language comprehension while reading and during known
faces recognition. The behavioral results on the activation of the
mouth and the activation of the brain areas we have described are
compatible with the idea that to access the meaning of abstract
words we reenact previous acquisition experiences of the word,
and also internally rehearse words or re-explain their meaning to
ourselves. This can occur at least in part through internal language.
Further fMRI data has revealed that not only linguistic areas are
activated, but also areas related to experiential components, such
as the right hemisphere (superior frontal gryus, precuneus,
D’Esposito et al., 1997; anterior cingulated cortex, amygdala,
parieto-occipital junction, Perani et al., 1999), occipital gyrus
(Jessen et al., 2000). Overall, neural evidence reveals, in keeping
with the WAT view, that areas related to both sensorimotor and
linguistic experiences are activated by abstract concepts, that the
role of linguistic areas is more important for abstract than for
concrete concepts, and it suggests that these linguistic areas can
involve subvocalization.
The fourth principle of the WAT proposal is that, given the
crucial role of language in the formation of abstract concepts, these
are more likely to be influenced by differences between languages
than concrete concepts. For example, Malt, Sloman, Gennari, Shi,
and Wang (1999) have shown that the concrete concept of “con-
tainer” is highly variable across languages in terms of naming but
not in terms of knowing: when Chinese-, Spanish-, and English-
speaking participants were asked to perform a sorting task, the
crosslinguistic variability disappeared and they adopted the same
criteria in grouping containers. In contrast, much evidence has
revealed that the abstract concept of “time” varies across lan-
guages not only in terms of naming, but also in terms of the
conceptual content: for example, time evokes the vertical dimen-
sion for Mandarin Chinese, but the horizontal for North American
participants (see Borghi & Binkofski, 2014, for an analysis in
depth of this aspect).
Strengths and limitations. The WAT view has a number of
strengths. The first and most important is that it is based on a
mechanism that characterizes abstract concepts: the more abstract
the concepts are, the greater is the linguistically conveyed and
linguistic form information that should be necessary to compensate
the reduced perceptual experience and to keep together their dif-
ferent members. Being an embodied theory, the WAT view does
not suggest that this mechanism, that is, the activation of linguis-
tically conveyed and linguistic form information, is in contrast
with the fact that abstract concepts activate their content. Another
strength of the WAT view over other theories, is that it focuses not
only on conceptual representation but also on word acquisition. It
has the potential, therefore, to unify two separate streams of
research, the first being word acquisition in children and the
second representation of abstract concepts in adults.
Evidence on both acquisition and brain organization confirms
the importance of linguistically conveyed information and of lin-
guistic form information for abstract concepts. However, while the
claim that conceptual acquisition influences conceptual represen-
tation in the brain is very strong, it remains speculative, and should
be further developed and investigated. In addition, the concept of
modality of acquisition should be further analyzed to solve the
dilemma of whether what counts is only the initial acquisition or
also the successive times in which a given word is used, renego-
tiating its meaning. Furthermore, should such acquisition neces-
sarily occur through a social process or could it also occur through
written texts, such as when we read information on Wikipedia?
One final limitation of the WAT proposal is that the main
evidence collected so far derives from studies on word acquisition
in adults and studies on the neural underpinnings of abstract
concepts; further studies on infant and child acquisition should be
conducted, as should further cross-cultural studies, to test the
hypothesis that abstract concepts are more affected by linguistic
variability than concrete ones.
Grounding and Sign Tracking. The fourth Multiple Repre-
sentation theory discussed here is Grounding and Sign Tracking,
which is very close to the WAT view we have just discussed. The
philosopher Prinz (2002,2012) has recently proposed that we
understand words using a tracking strategy: to capture their mean-
ing, we need to anchor and link them to something nonverbal, their
referents. The same process holds for both concrete and abstract
categories, which, similarly to concrete ones, are correlated with
features that are perceivable and that can work as signs to track the
category. Prinz outlines different strategies that can be used to
understand abstract concepts.
One possible strategy is to ground them in concrete scenarios:
for example, “justice” can be simulated with a scenario reproduc-
ing inequality: one person gets two cookies, another person three.
Grounding concepts in situations might not, however, hold for all
abstract concepts. Besides sign tracking, we may adopt other
strategies that include metaphorical projection (see the CMT), the
reference to internal states and emotions, and labeling. Concepts
such as “meaningful activity” are comprehended introspecting
motivations and emotions, while other concepts, such as “truth”
and “identity,” can be understood through labeling, thanks to a
network of associate words. Linguistic associations alone, how-
ever, are not sufficient to wholly capture meaning: to fully under-
stand words we need to ground them. Crucially, Prinz has high-
lighted the importance of both mental imagery and verbal skills in
accessing the concept meaning. Overall, he contends that compre-
hending abstract concepts implies the capacity “to match mental
images with reality and sentences with testimony” (Prinz, 2012). A
novel aspect of his theory is that he underlines the importance of
testimony for getting the conceptual gist: for example, to learn the
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
279
THE CHALLENGE OF ABSTRACT CONCEPTS
concept of democracy we need to track definitions used by au-
thoritative members of our community.
Strengths and limitations. This theory proposed by Prinz
has, in our view, many advantages. This perspective not only
identifies some underlying mechanisms but highlights a range of
different possibilities that can underlie the formation of abstract
concepts. Therefore, it points out the flexibility of our conceptual
system. Another strength of this theory is that these mechanisms
are differently activated depending on the kind of concepts. This
points out an important aspect: the fact that abstract concepts are
very diverse and heterogeneous, and that different kinds of expla-
nations can be used to account for them. Furthermore, although
Prinz identifies different strategies, this is a fully embodied view,
which also extends to language. An important and novel aspect
that characterizes his theory is the role it ascribes to testimony: to
understand the meaning of abstract concepts we may need to refer
to authoritative members of our community.
Although it is evident that there are a number of merits to
Prinz’s work, the major weakness of this theory is that in its
present form it is still more of a proposal than a structured theory.
Furthermore, the proposal does not consider some important as-
pects, such as the link between concepts acquisition and concepts
representation. In addition, it highlights the role of words only in
terms of the past experiences they reenact, not as a possible means
to perform actions. However, the proposal could be extended to
include these aspects.
Multiple Representation Views: Further Evidence
As made clear in the course of the review, multiple representa-
tion views are emerging as the theoretical frameworks with the
potential to solve the challenge of abstract concepts. Specifically,
we believe that the most promising directions of research are
inspired by those multiple representation views that combine a
fully embodied approach with the recognition of the importance of
language intended not only in terms of word associations but in its
holistic aspects. We will now illustrate some recent studies that
further support multiple representation views, showing that— de-
pending on the task and on the level of processing it implies—
different kind of information is activated. Specifically, we will
focus on studies that highlight both the role of emotions and the
pivotal role played by language for abstract concepts.
Recent studies have investigated the role of emotions and
context availability for concrete and abstract concepts (Moffat et
al., 2015;Newcombe, Campbell, Siakaluk, & Pexman, 2012;
Siakaluk, Knol, & Pexman, 2014). Moffat et al. (2015) reported
evidence that, while confirming the important role of context
availability for conceptual processing in general, also revealed that
its relevance is not unique to abstract concepts. In a series exper-
iments Moffat et al. (2015) instructed participants to complete four
tasks; first to read aloud only abstract words, second to read aloud
only concrete words, third to name aloud all words that appeared
on the screen, and finally to name aloud all words, when concrete
and abstract words were separately presented and blocked for
emotions. The study concluded that, compared with other vari-
ables, context availability and emotions play an important role in
characterizing concepts. With abstract words both emotions and
context availability played a facilitatory role, such that higher
ratings on these two dimensions were associated with faster re-
sponses. Context availability played a facilitatory role for both
kinds of concepts: this suggests that it is not a unique dimension of
the first, but that it plays an important role for conceptual process-
ing overall. Emotions instead both facilitated processing of ab-
stract words and inhibited processing of concrete words: this
reveals that emotions play an important role for abstract concepts
(see also Newcombe et al., 2012). Even if this evidence is read by
the authors as consistent with the emotion-based view, it seems to
us that it is only partially consistent with it. Indeed, it shows that
emotions did not always play a role: when participants were
simply required to name words, which were presented together and
not blocked for emotions, only context availability and not emo-
tions influenced processing.
Further studies have revealed the importance of a variety of
semantic dimensions for representation of abstract concepts. Rec-
chia and Jones (2012) asked participants to generate 10 features for
concrete and abstract concepts, to allow another participant to
guess the target-word from their description. They found that
abstract concepts benefit from rich linguistic contexts, while con-
crete concepts benefit from rich physical contexts. Thill and
Twomey (2016) demonstrated that the variability of Age of Ac-
quisition data can be hardly explained with traditional factors, and
propose, in line with the WATs theory, that the degree of abstract-
ness of a concepts depends on how much it goes beyond simple
sensorimotor grounding, incorporating linguistic, social and also
interoceptive features. In a recent study, Danguecan and Buchanan
(2016) compared word recognition response times in four tasks
requiring increasing depth of semantic processes, from standard
lexical decision to sentence relatedness task. Manipulating ab-
stractness (abstract vs. concrete words) and semantic neighbor-
hood density (SND, high vs. low) they found that SND effects
were always present with abstract concepts, but not necessarily for
concrete ones. This evidence is clearly in line with the idea that
linguistically conveyed information is crucial for the representa-
tion of abstract concepts (for converging evidence see also Good-
hew, McGaw, & Kidd, 2014).
Further evidence has been provided by Zdrazilova and Pexman
(2013), who examined the effect of different semantic richness
variables (context availability, semantic neighborhood, number of
associates, sensory experience rating, valence, and arousal) on
lexical decision and semantic categorization tasks with abstract
concepts. They found faster response times (RTs) in lexical deci-
sion for abstract words that evoked more contextual information,
and faster RTs in semantic categorization for abstract words that
evoked both a richer sensory experience and a more positive
emotional experience. The results, which support multiple repre-
sentation views of abstract words, indicate that different dimen-
sions—perceptual, emotional, and linguistic— have a different and
dissociable influence on tasks.
In the same vein, a recent study by Farias, Garrido, and Semin
(2013) showed an interplay between semantic and perceptual
properties with abstract words referring to politics. Participants
were more likely to evaluate words associated with conservatism
as louder when presented to the right ear, and words associated
with socialism as louder when presented to the left ear, even if the
sounds did not differ in intensity. The association between left and
socialism-related words and right and conservativism-related
words was found also in a classification and a lexical-decision
task, controlling that the effects were not simply because of spatial
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
280 BORGHI ET AL.
compatibility (Farias, Garrido, & Semin, 2016). Further work on
political metaphors showed an influence of the left or right body
balance on their attributions to political parties (Dijkstra, Eerland,
Zijlmans, & Post, 2014) Overall, this evidence highlights the
multimodal character of abstract words; furthermore, it suggests
that “an opposition between symbolic representational and modal-
ity specific representations is misleading at best”(Farias et al.,
2013,p.5).
Strengths and limitations. The evidence assessed in this sec-
tion indicates that different experiential dimensions characterize
abstract concepts. Multiple representation views, which are based
on such an assumption, have for this reason multiple strengths.
One of the most important is that by taking into account both
linguistic and sensorimotor information, these proposals can pro-
vide a bridge between two approaches to meaning that were
traditionally considered to be in opposition: the embodied and the
distributional accounts.
As we have seen throughout this review, any embodied ap-
proach attempting to provide an explanation for abstract concepts,
is facing a major challenge. Distributional approaches do not share
this problem, since they derive the meaning of both concrete and
abstract words from their statistical distribution across languages
(Andrews et al., 2014).
However, as anticipated in the introduction, distributional ap-
proaches are unable to solve the “symbol grounding problem”
(Harnad, 1990) that asks: how can the meaning of a word be
explained only through reference to other words? And moreover,
how could we understand what “bottle” means if we could not see,
touch and experience bottles? Furthermore, distributional ap-
proaches do not take into account the emotional and social aspects
that accompany language acquisition and use (see the AEA and the
WAT theories). Finally, while distributional approaches are quite
successful in explaining experimental results, they cannot predict
results at the same level of detail and precision that characterizes
embodied account. Even authors who propose that language en-
codes perceptual, spatial, and temporal information recognize that
the language statistics cues can contribute to meaning comprehen-
sion, but are not the only factors. Recent data, for example, have
suggested that around 60% of the temporal relations such as cities
or the chronological sequence of days and historical figures can be
accessed through statistics (e.g., Louwerse, Raisig, Tillman, &
Hutchinson, 2015). This, in our opinion, demonstrates that ground-
ing of concepts is crucial to their full comprehension.
Multiple representation views have two further strengths. The
first is that they can utilize and explain neuropsychological evi-
dence showing double dissociations between the mastering of
abstract and concrete words. The second is that multiple represen-
tation views seem to us the only approach able to deal with the
different kinds and varieties of abstract concepts. We will discuss
these issues in the next two sections.
Multiple Representation Views and
Neuropsychological Evidence
Even if discussing this evidence in detail is beyond the scope of
the present work (see Borghi & Binkofski, 2014, chapter 5, for
details), it is important to consider the effect linked to two syn-
dromes, as well as their implications for the debate on concrete and
abstract concepts (see the review by Shallice & Cooper, 2013, but
also studies on Parkinson patients by Fernandino et al., 2013a,
2013b).
The first syndrome is deep dyslexia: patients make errors
while reading aloud, but the majority of errors are made with
words with concrete, or highly imageable, meaning (Shallice &
Warrington, 1975;Coltheart, Patterson, & Marshall, 1987). The
second syndrome is semantic dementia (Warrington, 1975) and
herpes encephalitis (e.g., Warrington & Shallice, 1984): the
results are rather inconsistent because according to some re-
ports patients showed a mild reverse concreteness effect, that is,
their performance was better with abstract than with concrete
concepts (e.g., Bonner et al., 2009). However, more recent
studies in which imageability and frequency were controlled
suggest that SD patients perform worse with abstract than with
concrete words (Giffard, Laisney, Desgranges, & Eustache,
2015;Hoffman & Lambon Ralph, 2011; see also recent evi-
dence on specific effects of mild amnesia while defining ab-
stract words, Kim, Kim, Baek, & Kim, 2015). The connectionist
model proposed by Plaut and Shallice (1993) provides a pos-
sible explanation for this impairment with abstract words, as it
assumes that a reduced number of features of abstract words
renders their representation weaker overall compared to con-
crete words.
It is worth spending a few words on the role played by
semantic dementia, caused by the bilateral degeneration of the
ATLs, in the debate between embodied and amodal theories of
semantics. Because the selective damage of a specific brain area
causes semantic dementia, it has been proposed that a single
neural area exists in the brain where all modalities are inte-
grated: the anterior temporal lobe would work as a hub inte-
grating white matter connections (spokes) and modality specific
associative cortices (hub-and-spoke model, Patterson, Nestor,
& Rogers, 2007). Crucially for the present work, ATLs would
play a pivotal role in representing abstract concepts (Pobric,
Ralph, & Jefferies, 2009). Evidence on semantic dementia has
been interpreted as favoring hybrid models, according to which
modal and amodal systems concur in representing abstract
concepts (e.g., Dove, 2015). We do not, however, believe that
these data lead to an unequivocal conclusion. First, the evidence
provided has been recently discussed (e.g., Gainotti, 2012;
Simmons & Martin, 2009); furthermore, the pattern of data is
complicated, first because of the limits of localization data, that
only allow validation of a specific theory in an indirect way,
and also because of the behavioral results: semantic dementia
does not impair all abstract concepts—for example, it spares
numerical concepts. Finally, proponents of embodied theories
tend to represent association areas (Damasio, 1989) where
information converge as multimodal (Simmons, Martin, & Bar-
salou, 2005), or “heteromodal” (Bonner, Peelle, Cook, & Gross-
man, 2013; for a recent critical review, see McCaffrey, 2015).
Hence, the debate remains open (see e.g., Binder, 2016), even
if explaining the impairment provoked by semantic dementia is
certainly a challenge for embodied theories. Reviewing evi-
dence on deep dyslexia and semantic dementia, Shallice and
Cooper (2013) contended that semantic representation of con-
crete and abstract words can be separable. They stated that the
higher complexity of abstract concepts requires, at a minimum,
mastery of high-level logical functions. For example, the con-
cept of “bicycle” can be accounted for by the conjunction of
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
281
THE CHALLENGE OF ABSTRACT CONCEPTS
features such as “is a vehicle,” “has a seat and wheels,” and so
forth, while to capture a concept such as “hope” we would need
a logical structure as “Hope (X) if and only if desire (X) AND
believe (possible [X]).” According to the authors, neither
feature-based theories (e.g., Plaut & Shallice, 1993) nor em-
bodied theories have sufficient computational power to be able
to account for how we compile the meaning of abstract words
such as “hope.”
How can researchers who endorse an embodied approach coun-
ter such objections? Different arguments can be advanced. First of
all, one could question the “separability” of the two systems. It
should be noted that not only behavioral and brain imaging studies
(see above) but also neuropsychological evidence suggests that
both concrete and abstract concepts are grounded in distributed
action-perception circuits. A recent study on a double dissociation
used a lexical-decision task to show that a patient with a lesion in
dorsolateral central sensorimotor systems was impaired in recog-
nizing tool words, while a patient with rather specific focal lesion
centered in the left supplementary motor area had a deficit with
abstract-emotional word processing (Dreyer et al., 2015). This
clearly indicates that sensorimotor systems are not peripheral for
abstract concepts processing, and suggests that the two systems for
concrete and abstract words are not completely separable but
overlap at least partially. Second, one should better consider the
potentialities of perceptual symbols and of linguistic symbols. One
response, that Shallice and Cooper (2013) do not seem to consider,
is to argue that perceptual symbols have the computational power
which characterizes the linguistic symbolic system, as suggested
by Larry Barsalou in his seminal 1999 article (Barsalou, 1999). A
second possibility is to claim, as was recently done by Dove
(2015), that the language system can offer the computational
abilities necessary for the separate system proposed by Shallice
and Cooper (2013). While Dove (2015) considers this linguistic
system to be amodal, we do not see the necessity to hypothesize a
transduction from linguistic experience to an amodal form of
representation.
Multiple Representational Views and Multiple Kinds
of Abstract Concepts
Many studies favoring multiple representation views have con-
tended that different systems might be activated, depending on the
task. For example, according to the LASS view the linguistic
system is more activated during lexical decision tasks, which make
small demands on semantic processing, while the simulation sys-
tem is more activated during tasks that imply a deeper semantic
comprehension, such as semantic decision tasks. Along the same
lines, Zdrazilova and Pexman (2013) demonstrated that different
measures of semantic richness are activated depending on the task.
However, it might be also suggested that different representation
systems are activated to a different extent, depending on the kind
of abstract concepts we consider. Abstract concepts can be very
different from each other: they include, for example, number
concepts, emotional concepts (e.g., “love”), mental state concepts
(e.g., “thought”), and purely abstract concepts (e.g., “truth”). Re-
cently, the focus of research has been on investigating fine-grained
distinctions between subkinds of abstract concepts. It is indeed
possible that, in a similar way to what happens with concepts
referring to sensorial experiences such as taste (Simmons et al.,
2005), or odor, and with perceptual properties, such as color and
shape (e.g., Martin, 2007;Martin, Haxby, Lalonde, Wiggs, &
Ungerleider, 1995), different distributed patterns of experiential
information characterize different kinds of concepts.
Wilson-Mendenhall et al. (2013) conducted a fMRI study that
showed that concepts such as “convince” are represented in areas
related to mentalizing and social cognition (e.g., medial prefrontal
cortex, superior temporal sulcus), while the concept “arithmetic”
engages brain regions underlying numerical cognition (e.g., bilat-
eral intraparietal sulcus). This study indicated that concepts are
represented in a distributed manner that depends on their content.
Further research should be devoted to identify eventual subsets of
abstract concepts and to determine a validated selection criterion
that should be used universally within research to allow for repli-
cation of findings.
A promising novel manner to investigate subsets of abstract
concepts, which is linked by different degrees of semantic relat-
edness, is to use multidimensional ratings to examine the contri-
bution of different dimensions in abstract concept representation.
(e.g., Crutch, Williams, Ridgway, & Borgenicht, 2012;Crutch et
al., 2013;Troche, Crutch, & Reilly, 2014). For example, Crutch et
al. (2013) asked participants to rate concepts by dimensions as
sensation, emotion, action, thought, social interaction, morality,
executive function, quantity, time, space, and polarity. They used
this approach starting from the hypothesis that further dimensions
beyond sensorimotor and emotional features characterize abstract
concepts; for example, the social dimension (see the WAT pro-
posal). On the basis of the ratings a high dimensional semantic
space was generated, where the semantic relatedness between
concepts was represented in terms of the distance between them.
The validity of this representation was tested and validated, show-
ing that the performance of a patient affected by global aphasia
was worse when she had to identify a target presented within word
pairs that were close together in the semantic space. It was also
tested using the eyetracker with a visual word paradigm (Prima-
tivo, Reilly, & Crutch, 2016).
This method of investigating subsets of abstract concepts was
also used by Ghio, Vaghi, and Tettamanti (2013) to evaluate
whether different abstract concepts (mental state-, emotion-, and
mathematics-related concepts) involved different parts of the body
(mouth, hand). They found that the three kinds of concepts varied
across different dimensions. For example, sentences referring to
emotions and mental states were similarly rated in concreteness,
context availability and familiarity. Emotion sentences were more
associated with mouth, hand, and leg movements than mental-state
and mathematics-related sentences, while math related sentences
involved more hand movements than the other concepts (see also
Ghio, Vaghi, Perani, & Tettamanti, 2016 for an fMRI study on the
same categories).
Further studies have made use of feature generation tasks to
identify differences between subtypes of abstract concepts. For
example, Setti and Caramelli (2005) used a feature production task
and found that concepts referring to nominal kinds (e.g., “error”),
states of the self (e.g., “worry”), cognitive processes (e.g., “mem-
ory”), and emotions (e.g., “fear”) were characterized by different
conceptual relations. In particular, emotion terms differed more
from the other three subtypes.
A further novel and promising way to identify subsets of ab-
stract concepts is to investigate how different kinds of abstract
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
282 BORGHI ET AL.
concepts are encoded into signs by sign languages. Borghi, Ca-
pirci, Gianfreda, and Volterra (2014) analyzed examples of ab-
stract concepts taken from LIS (Lingua dei Segni Italiana), the
visual-gestural language used within the Italian Deaf community
(for a similar attempt see Roush, 2011, who analyzed, starting
from the CMT, how concepts of politeness are represented in
ASL). The analysis suggests that a framework based on a single
dimension, that is, linguistic, sensorimotor or emotional, does not
fully capture the meaning of abstract concepts. Instead, multiple
sources of experiences converge, in different distributions, to
represent abstract concepts. Many signs in LIS provide support for
the CMT, as they make use of body parts in an iconic way to refer
to underlying metaphors, such as “knowing is seeing,” “the chest
contains feelings and emotions,” or “the head is a container”: for
example, in the sign TO LEARN all the extended digits rapidly
touch each other and move toward the signer’s forehead, as when
bringing something from the external space to the head. There are
a range of LIS signs that can be viewed as supporting evidence
for a number of theories discussed within this review. For
example, consistent with views that assign an important role to
linguistically conveyed information (WAT, Dove), purely ab-
stract signs such as TRUE and TRUTH (VERO and VERITA’
in Italian) or as LANGUAGE/LINGUA are conveyed in LIS
using a strategy known as “initialization,” that is, reporting the
initial letter of the correspondent word, as the letter V (VERO,
TRUTH) or L (LINGUA, LANGUAGE).
The results of this study on sign language led to two important
conclusions. First, these results show that, even if all examples are
consistent with an embodied view, none of the current theories or
proposals are of a sufficiently general nature to be able to account
for all examples of abstract concepts. Even the CMT, which could
be the most important and influential theory, because of the fact
that in sign languages meanings have to be expressed through
iconic gestures, does not exhaustively explain the meaning of all
concepts. More important, the results suggest that to express
purely abstract concepts LIS exploits linguistically conveyed in-
formation, regardless of whether it is derived from the same sign
language, from a foreign sign language as ASL or from the
corresponding spoken/written language (Italian, in this case).
Overall, the reviewed studies reveal that abstract concepts can
differ widely in terms of content, and cast some doubts on whether
a single theory might be able to explain all varieties of abstract
concepts. We do not intend to claim that the proposals we have
outlined have no merit. Instead, we suggest that all of them, even
if to a different degree, contribute to a better comprehension of
abstract concepts. However, to date no approach has proven able to
account for all abstract concepts. This opens the possibility that
multiple representation approaches are necessary to account for
specific abstract concepts features and content, in their varieties
and differences (Borghi et al., 2014;Crutch et al., 2013;Ghio et
al., 2013;Wilson-Mendenhall et al., 2013).
Among multiple representation approaches, in our view the most
promising ones are those that (a) take into account the important
differences in conceptual acquisition between concrete and abstract
concepts; (b) stress the importance not only of the linguistic but also
of the social dimension for abstract concepts. Such an approach would
imply a tighter interconnection between research on human sociality
and its development, developmental research on conceptual acquisi-
tion, and behavioral and neuroscientific research on abstract concepts.
Such an approach would be completely novel, and would go well
beyond the proposed integrations between distributional and embod-
ied approaches to abstract concepts that are becoming increasingly
popular (e.g., Andrews et al., 2014).
Conclusion
Being able to explain how we form and use abstract concepts is
pivotal to understand one of the most sophisticated abilities pos-
sessed by our species (Wang, Uhrig, Jarraya, & Dehaene, 2015).
We will first summarize the main characteristics that a successful
theory of abstract concepts should display, discussing them in light
of the proposals we have reviewed. We will then draw some
conclusions and point to some open issues.
Toward a Successful Theory of Abstract Concepts
a. Focus on mechanism rather than on content. We think that
an ideal theory should identify a mechanism rather than
simply a content which characterizes abstract concepts. The
most promising research directions, in our view, are offered
by the proposals according to which both sensorimotor and
linguistically conveyed information characterize abstract
concepts. Even if we think that these theories are those from
which more fruitful research directions can be derived,
much work needs to be done: a clear specification of the
mechanisms underlying the activation of linguistic, emo-
tional, and social information is needed.
b. Link between acquisition and conceptual representation.
Any thorough theory of abstract concepts should make
precise predictions on conceptual acquisition, and on how
the kind of acquisition influences conceptual representation.
Both the WAT proposal and (partially) the AEA focus on
how concepts are acquired, and on how this influences their
representation. WAT focuses on the different modality of
acquisition, AEA on the bootstrapping mechanism played
by emotions. However, much needs to be done to further
elucidate the relationship between acquisition and brain
representation.
c. Role of “introspective” information for grounding abstract
concepts (introspective and situational theory). Some form
of metacognition is a feature that a systematic theory of
abstract concepts should take into account. Interoception
(perception of one’s internal states and processes) is a form
of experience that might play an important role, especially
for abstract concepts. The introspective and situational the-
ory appears to focus on such a mechanism, even though the
kind of conscious self-awareness it implies can be too
limited if not complemented by an implicit form of meta-
cognition.
d. Explanation of cross-linguistic variability. It is important
for a thorough theory of abstract concepts to provide clear
predictions and to be able to account for the variability of
the different kinds of abstract concepts across languages.
While discussing the CMT and (partially) the WAT pro-
posals, we have seen that these views formulate specific
predictions on cross-linguistic differences in abstract con-
cepts representation.
e. Generabilizability. A thorough theory should be able to
account for all abstract concepts, in their differences and
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
283
THE CHALLENGE OF ABSTRACT CONCEPTS
varieties, and differences in their degree of abstractness.
Theories based on mechanisms are more promising than
content based theories, since they can be extended beyond
specific domains. However, as anticipated, it is an open
question whether to date theories are able to explain abstract
concepts of different domains.
Where the Field Is Going
Our analysis of the literature can help to comprehend where the
field is, and where it is going. We will summarize below our main
conclusions, and will then discuss them more in depth:
a. Views according to which abstract concepts representation
can be explained exclusively on the basis of their grounding
in sensorimotor systems or in linguistic systems are not
empirically supported;
b. Multiple representations views represent a viable alterna-
tive, because they highlight the importance of different
kinds of experiences: linguistic, emotional, social, and sen-
sorimotor. The major novelty in the field over the last few
years has been represented by these views;
c. Among multiple representation views, we believe that the
evidence so far has shown that an embodied view that takes
into account the importance of language experience can be
capable of explaining abstract concepts representations. In
our perspective, while it is crucial to recognize the pivotal
role language plays for abstract concepts, there is no need to
posit that language makes use of amodal representations. It
is important, instead, to underline the role of language for
abstract concepts acquisition, focusing on language in its
multifolded aspects, as a bodily, social, and emotional ex-
perience, as well as to underline the role of language as a
medium of thought, able to extend our cognitive abilities.
This idea of language as a bodily and social experience is
fully compatible with results showing a higher activation of
the mouth with abstract concepts, and with brain imaging
studies that indicate a higher activation of social cognition
areas, and of linguistic areas involved in language produc-
tion, in auditory processing, and in semantic processing.
However, in the field opinions diverge as to the modal or
amodal character of the language medium, and more re-
search and theories are needed to disentangle this complex
issue. Further research is also needed to explore more in
depth the role played by emotional (see the AEA theory)
and social and linguistic (see the WATs theory) informa-
tion;
d. While the most interesting perspectives in our view con-
verge in showing that, compared to concrete concepts,
abstract concepts are more (but not exclusively) character-
ized by social, emotional, and linguistically conveyed in-
formation than by sensorimotor information, the most re-
cent studies are also starting to identify subtypes of abstract
concepts, in terms of the different distribution of the expe-
riential information they evoke. This is a very fruitful
research direction, useful to prepare a solid ground for the
development of more compelling multiple representation
views.
While we hope we have been able to convey the message
summarized in the above four points, we believe that some issues
deserve more discussion. Our analysis of the literature comes
together to form one central conclusion, that states that explana-
tions based on a single strategy are insufficient and that we are
now compelled to recognize that different experiences (emotional,
social, linguistic, and sensorimotor) play a role in abstract concepts
representation. This claim can seem obvious, but we believe it is
not. Until some years ago theories on abstract concepts were
divided: distributional theories focused solely on the importance of
linguistically conveyed information, embodied theories high-
lighted primarily the role of sensorimotor information, and views
derived from Paivio’s DCT approach proposed that sensorimotor
information was important for concrete concepts, linguistically
conveyed information for abstract ones. More crucially, the pivotal
role played by emotional and social information for abstract con-
cepts has been underlined only very recently, respectively, by the
AEA and by the WAT theories. In addition, the research direction
that reconciles embodied and distributional theories of meaning,
ascribing importance to both sensorimotor and linguistically con-
veyed information, provides a novel way to account for all con-
cepts; the insights of distributional theories are, however particu-
larly, useful to explain abstract concepts, for which linguistically
conveyed information has a major relevance.
Still, one could argue that it is self-evident that concepts rely on
different sources of information and different kinds of experiences,
and that this also holds for concrete concepts. We hope to have
shown that, even if concepts are arranged along a continuum, the
distribution of these experiences differs substantially: while con-
crete concepts are grounded primarily in perception and action
system, abstract ones reenact more— but not exclusively—social,
emotional, and linguistic experiences. It appears that these differ-
ent strategies may have different activations, depending not only
on the kind of task, but also on the kind of abstract concept.
Further research is essential to deepen our understanding and
should focus on investigating and defining the fine-grained dis-
tinctions among abstract concepts, similar to those present for
distinguishing between concrete ones.
While the activation of a network of differently distributed brain
areas— depending on the conceptual content—is in line with an
embodied view, it remains to be clarified whether there is some-
thing distinctive about abstract concepts, which could explain why
we perceive, rate and use abstract concepts in a different manner
than with concrete concepts.
An increasing amount of evidence, obtained with different tech-
niques and methodologies, indicates that both linguistic and non-
linguistic experiences contribute to the representation of both
abstract and concrete concepts. This consideration is likely to be at
the root of the recent success of hybrid approaches, such as the
revitalization of the classic vision of Paivio (1986).
While it is largely established that multiple forms of represen-
tation are needed to account for the complexity of abstract con-
cepts, not all researchers in this area share the view that all systems
are modal and the modality of systems remains an issue for further
research (see Dove, 2009,2015;Tomasino & Rumiati, 2013; see
for a very recent proposal Reilly et al., 2016). We believe instead
that there is no need to call for amodal representations and that a
fully embodied view has the potential to meet the challenge of
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
284 BORGHI ET AL.
explaining abstract concepts. In addition, explaining abstract con-
cepts would not require the adoption of a unique strategy.
We have seen the advantages of multiple representation views
when combined with an embodied approach, and we have seen that
studies on activation of different dimensions are flourishing. We
have argued within this review that multiple representation theo-
ries can account for the variability of information activated across
both tasks and concepts. However, it is important to note that there
are weaknesses in this area and both “linguistic information” and
“linguistic experience” must be better defined. Language can be
seen simply as a shortcut to access the simulation system (Barsalou
et al., 2008), as a way to access meaning through the associations
between words (distributional views: e.g., Landauer & Dumais,
1997;Lund & Burgess, 1996), or as a way to provide us with
further computational abilities (Clark, 1998;Dove, 2014). How-
ever, language can be all of this, but also much more: words are
social tools to act in the world, and language is an important
embodied social experience (Borghi & Binkofski, 2014;Borghi &
Cimatti, 2009), which is likely to possess emotional and affective
implications (Vigliocco et al., 2014). Language can help us to
formulate predictions, and contributes to control our behavior
(Lupyan & Bergen, 2016;Lupyan & Clark, 2015) and to improve
our thought (Vygotsky, 1986). Language can provide a glue to
keep together different category members (Borghi & Binkofski,
2014), as well as a means to introspectively reason on them
(Barsalou & Wiemer-Hastings, 2005).
Such a perspective could encourage collaboration between a
range of specialties within psychology bridging, for example,
research from both developmental studies on language acquisition
and studies on conceptual representation in adults, which have
been the main focus of this review. Curiously, so far studies on
abstract concept acquisition and representation represent two sep-
arate lines of research (see Granito et al., 2015, for developing this
issue). One of the main challenges is the attempt to reunify them
under a unitary approach.
More importantly, the influence of language on concepts can be
seen as a powerful mechanism, useful for the comprehension of
abstract words of all kinds. In our view the activation of all sorts
of linguistic information, which is, of linguistically conveyed
information, linguistic form information and syntactic linguistic
information, becomes more important as the level of abstractness
increases. Furthermore, from what we have discussed above it
should be obvious that the activation of linguistic information, is
not in contrast with the activation of conceptual content. Indeed,
the embodied approach posits that abstract concepts evoke and
reenact previous experiences, and the linguistic and social experi-
ences are crucial ones.
It is plausible that the less sensorimotor and affective informa-
tion is activated, the more language is needed, because it can play
a scaffolding role to allow the acquisition of abstract words. More
important, this instrument is not simply a means to access mean-
ing, but it brings with it rich individual and social experiences.
References
Adorni, R., & Proverbio, A. M. (2012). The neural manifestation of the
word concreteness effect: An electrical neuroimaging study. Neuropsy-
chologia, 50, 880 – 891. http://dx.doi.org/10.1016/j.neuropsychologia
.2012.01.028
Alderson-Day, B., & Fernyhough, C. (2015). Inner speech: Development,
cognitive functions, phenomenology, and neurobiology. Psychological
Bulletin, 141, 931–965. http://dx.doi.org/10.1037/bul0000021
Altarriba, J., & Bauer, L. M. (2004). The distinctiveness of emotion
concepts: A comparison between emotion, abstract, and concrete words.
The American Journal of Psychology, 117, 389 – 410. http://dx.doi.org/
10.2307/4149007
Altarriba, J., Bauer, L. M., & Benvenuto, C. (1999). Concreteness, context
availability, and imageability ratings and word associations for abstract,
concrete, and emotion words. Behavior Research Methods, Instruments,
& Computers, 31, 578 – 602. http://dx.doi.org/10.3758/BF03200738
Andrews, M., Frank, S., & Vigliocco, G. (2014). Reconciling embodied
and distributional accounts of meaning in language. Topics in Cognitive
Science, 6, 359 –370. http://dx.doi.org/10.1111/tops.12096
Anelli, F., Lugli, L., Baroni, G., Borghi, A. M., & Nicoletti, R. (2014).
Walking boosts your performance in making additions and subtractions.
Frontiers in Psychology, 5, 1459. http://dx.doi.org/10.3389/fpsyg.2014
.01459
Aziz-Zadeh, L., Wilson, S. M., Rizzolatti, G., & Iacoboni, M. (2006).
Congruent embodied representations for visually presented actions and
linguistic phrases describing actions. Current Biology, 16, 1818 –1823.
http://dx.doi.org/10.1016/j.cub.2006.07.060
Bahrami, B., Olsen, K., Latham, P. E., Roepstorff, A., Rees, G., & Frith,
C. D. (2010). Optimally interacting minds. Science, 329, 1081–1085.
http://dx.doi.org/10.1126/science.1185718
Barber, H. A., Otten, L. J., Kousta, S. T., & Vigliocco, G. (2013).
Concreteness in word processing: ERP and behavioral effects in a lexical
decision task. Brain and Language, 125, 47–53. http://dx.doi.org/10
.1016/j.bandl.2013.01.005
Barca, L., Burani, C., & Arduino, L. S. (2002). Word naming times and
psycholinguistic norms for Italian nouns. Behavior Research, Methods,
Instruments, & Computers, 34, 424 – 434. http://dx.doi.org/10.3758/
BF03195471
Bardolph, M., & Coulson, S. (2014). How vertical hand movements impact
brain activity elicited by literally and metaphorically related words: An
ERP study of embodied metaphor. Frontiers in Human Neuroscience, 8,
1031. http://dx.doi.org/10.3389/fnhum.2014.01031
Barsalou, L. W. (1987). The instability of graded structure: Implications
for the nature of concepts. In U. Neisser (Ed.), Concepts and conceptual
development: Ecological and intellectual factors in categorization (pp.
101–140). Cambridge, United Kingdom: Cambridge University Press.
Barsalou, L. W. (1999). Perceptions of perceptual symbols. Behavioral and
Brain Sciences, 22, 637– 660. http://dx.doi.org/10.1017/S0140
525X99532147
Barsalou, L. W. (2003). Abstraction in perceptual symbol systems. Phil-
osophical Transactions of the Royal Society of London, Series B: Bio-
logical Sciences, 358, 1177–1187. http://dx.doi.org/10.1098/rstb.2003
.1319
Barsalou, L. W. (2008). Grounded cognition. Annual Review of Psychol-
ogy, 59, 617– 645. http://dx.doi.org/10.1146/annurev.psych.59.103006
.093639
Barsalou, L. W. (2012). The human conceptual system. In M. Spivey, K.
McRae, & M. Joanisse (Eds.), The Cambridge handbook of psycholin-
guistics (pp. 239 –258). New York, NY: Cambridge University Press.
http://dx.doi.org/10.1017/CBO9781139029377.017
Barsalou, L. W. (2016). On staying grounded and avoiding quixotic dead
ends. Psychonomic Bulletin & Review, 23, 1122–1142. http://dx.doi.org/
10.3758/s13423-016-1028-3
Barsalou, L. W., Santos, A., Simmons, K. W., & Wilson, C. D. (2008).
Language and simulations in conceptual processing. In M. De Vega,
A. M. Glenberg, & A. C. Graesser (Eds.), Symbols, embodiment and
meaning (pp. 245–284). Oxford, United kingdom: Oxford University
Press. http://dx.doi.org/10.1093/acprof:oso/9780199217274.003.0013
Barsalou, L. W., & Wiemer-Hastings, K. (2005). Situating abstract con-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
285
THE CHALLENGE OF ABSTRACT CONCEPTS
cepts. In D. Pecher & R. Zwaan (Eds.), Grounding cognition: The role
of perception and action in memory, language, and thought (pp. 129 –
163). New York, NY: Cambridge University Press. http://dx.doi.org/10
.1017/CBO9780511499968.007
Bergelson, E., & Swingley, D. (2013). The acquisition of abstract words by
young infants. Cognition, 127, 391–397. http://dx.doi.org/10.1016/j
.cognition.2013.02.011
Bergen, B. K., & Chan Lau, T. T. (2012). Writing direction affects how
people map space onto time. Frontiers in Psychology, 3, 109. http://dx
.doi.org/10.3389/fpsyg.2012.00109
Binder, J. R. (2016). In defense of abstract conceptual representations.
Psychonomic Bulletin & Review, 23, 1096 –1108. http://dx.doi.org/10
.3758/s13423-015-0909-1
Binder, J. R., Conant, L. L., Humphries, C. J., Fernandino, L., Simons,
S. B., Aguilar, M., & Desai, R. H. (2016). Toward a brain-based
componential semantic representation. Cognitive Neuropsychology,
33(3– 4), 130 –174. http://dx.doi.org/10.1080/02643294.2016.1147426
Binder, J. R., Desai, R. H., Graves, W. W., & Conant, L. L. (2009). Where
is the semantic system? A critical review and meta-analysis of 120
functional neuroimaging studies. Cerebral Cortex, 19, 2767–2796.
http://dx.doi.org/10.1093/cercor/bhp055
Binder, J. R., Westbury, C. F., McKiernan, K. A., Possing, E. T., & Medler,
D. A. (2005). Distinct brain systems for processing concrete and abstract
concepts. Journal of Cognitive Neuroscience, 17, 905–917. http://dx.doi
.org/10.1162/0898929054021102
Bonner, M. F., Peelle, J. E., Cook, P. A., & Grossman, M. (2013).
Heteromodal conceptual processing in the angular gyrus. Neuroimage,
71, 175–186.
Bonner, M. F., Vesely, L., Price, C., Anderson, C., Richmond, L., Farag,
C.,...Grossman, M. (2009). Reversal of the concreteness effect in
semantic dementia. Cognitive Neuropsychology, 26, 568 –579. http://dx
.doi.org/10.1080/02643290903512305
Boot, I., & Pecher, D. (2010). Similarity is closeness: Metaphorical map-
ping in a conceptual task. The Quarterly Journal of Experimental Psy-
chology: Human Experimental Psychology, 63, 942–954. http://dx.doi
.org/10.1080/17470210903134351
Boot, I., & Pecher, D. (2011). Representation of categories. Experimental
Psychology, 58, 162–170. http://dx.doi.org/10.1027/1618-3169/a000082
Borghi, A. M. (2005). Object concepts and action. In D. Pecher & R. A.
Zwaan (Eds.), Grounding cognition: The role of perception and action
in memory, language, and thinking (pp. 8 –34). Cambridge, United
kingdom: Cambridge University Press. http://dx.doi.org/10.1017/
CBO9780511499968.002
Borghi, A. M. (2013). Embodied cognition and word acquisition: The
challenge of abstract words. In C. Müller, A. Cienki, E. Fricke, S. H.
Ladewig, D. McNeill, & J. Bressem (Eds.) Body-language-communica-
tion: An international handbook on multimodality in human interaction.
Handbooks of linguistics and communication science (HSK) 38/2. Ber-
lin, Germany: De Gruyter Mouton.
Borghi, A. M., & Binkofski, F. (2014). Words as social tools: An embodied
view on abstract concepts. Berlin, Germany: Springer. http://dx.doi.org/
10.1007/978-1-4614-9539-0
Borghi, A. M., Capirci, O., Gianfreda, G., & Volterra, V. (2014). The body
and the fading away of abstract concepts and words: A sign language
analysis. Frontiers in Psychology, 5, 811. http://dx.doi.org/10.3389/
fpsyg.2014.00811
Borghi, A. M., Caramelli, N., & Setti, A. (2005). Conceptual information
on objects’ locations. Brain and Language, 93, 140 –151. http://dx.doi
.org/10.1016/j.bandl.2004.09.004
Borghi, A. M., & Caruana, F. (2015). Embodiment theory. In J. D. Wright
(Ed.), International encyclopedia of the social & behavioral sciences
(Vol 7, 2nd ed., pp. 420 –426). Oxford, United kingdom: Elsevier.
ISBN: 978008097086. http://dx.doi.org/10.1016/B978-0-08-097086-8
.56025-5
Borghi, A. M., & Cimatti, F. (2009). Words as tools and the problem of
abstract words meanings. In N. Taatgen & H. van Rijn (Eds.). Proceed-
ings of the 31st Annual Conference of the Cognitive Science Society (pp.
2304 –2309). Amsterdam, the Netherlands: Cognitive Science Society.
Borghi, A. M., Flumini, A., Cimatti, F., Marocco, D., & Scorolli, C.
(2011). Manipulating objects and telling words: A study on concrete and
abstract words acquisition. Frontiers in Psychology, 2, 15. http://dx.doi
.org/10.3389/fpsyg.2011.00015
Borghi, A. M., & Zarcone, E. (2016). Grounding abstractness: Abstract
concepts and the activation of the mouth. Frontiers in Psychology, 7,
1498. http://dx.doi.org/10.3389/fpsyg.2016.01498
Boroditsky, L. (2001). Does language shape thought? Mandarin and Eng-
lish speakers’ conceptions of time. Cognitive Psychology, 43, 1–22.
http://dx.doi.org/10.1006/cogp.2001.0748
Boroditsky, L., & Ramscar, M. (2002). The roles of body and mind in
abstract thought. Psychological Science, 13, 185–189. http://dx.doi.org/
10.1111/1467-9280.00434
Bottini, R., Crepaldi, D., Casasanto, D., Crollen, V., & Collignon, O.
(2015). Space and time in the sighted and blind. Cognition, 141, 67–72.
Boulenger, V., Hauk, O., & Pulvermüller, F. (2009). Grasping ideas with
the motor system: Semantic somatotopy in idiom comprehension. Ce-
rebral Cortex, 19, 1905–1914. http://dx.doi.org/10.1093/cercor/bhn217
Boulenger, V., Shtyrov, Y., & Pulvermüller, F. (2012). When do you grasp
the idea? MEG evidence for instantaneous idiom understanding. Neu-
roImage, 59, 3502–3513. http://dx.doi.org/10.1016/j.neuroimage.2011
.11.011
Carpenter, M., Nagell, K., Tomasello, M., Butterworth, G., & Moore, C.
(1998). Social cognition, joint attention, and communicative competence
from 9 to 15 months of age. Monographs of the Society for Research in
Child Development, 63, i⫹iii⫹v-vi⫹1–174. http://dx.doi.org/10.2307/
1166214
Casasanto, D. (2008). Who’s afraid of the big bad Whorf? Crosslinguistic
differences in temporal language and thought. Language Learning, 58,
63–79. http://dx.doi.org/10.1111/j.1467-9922.2008.00462.x
Casasanto, D., & Boroditsky, L. (2008). Time in the mind: Using space to
think about time. Cognition, 106, 579 –593. http://dx.doi.org/10.1016/j
.cognition.2007.03.004
Casasanto, D., & Bottini, R. (2014). Mirror reading can reverse the flow of
time. Journal of Experimental Psychology: General, 143, 473– 479.
http://dx.doi.org/10.1037/a0033297
Catricala
`, E., Della Rosa, P. A., Plebani, V., Vigliocco, G., & Cappa, S. F.
(2014). Abstract and concrete categories? Evidences from neurodegen-
erative diseases. Neuropsychologia, 64, 271–281. http://dx.doi.org/10
.1016/j.neuropsychologia.2014.09.041
Chen, J. Y. (2007). Do Chinese and English speakers think about time
differently? Failure of replicating Boroditsky (2001). Cognition, 104,
427– 436. http://dx.doi.org/10.1016/j.cognition.2006.09.012
Chen, M., & Bargh, J. A. (1999). Consequences of automatic evaluation:
Immediate behavioral predispositions to approach or avoid the stimulus.
Personality and Social Psychology Bulletin, 25, 215–224. http://dx.doi
.org/10.1177/0146167299025002007
Clark, A. (1998). Magic words: How language augments human compu-
tation. Language and thought: Interdisciplinary themes, 162–183.
Coltheart, M., Patterson, K., & Marshall, J. C. (1987). Deep dyslexia since
1980. In M. Coltheart, K. Patterson, & J. C. Marshall (Eds.), Deep
dyslexia (pp. 407– 451). London, United Kingdom: Routledge.
Connell, L., & Lynott, D. (2012). Strength of perceptual experience pre-
dicts word processing performance better than concreteness or image-
ability. Cognition, 125, 452– 465. http://dx.doi.org/10.1016/j.cognition
.2012.07.010
Connell, L., & Lynott, D. (2013). Flexible and fast: Linguistic shortcut
affects both shallow and deep conceptual processing. Psychonomic
Bulletin & Review, 20, 542–550. http://dx.doi.org/10.3758/s13423-012-
0368-x
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
286 BORGHI ET AL.
Corriveau, K., & Harris, P. L. (2009). Choosing your informant: Weighing
familiarity and recent accuracy. Developmental Science, 12, 426 – 437.
http://dx.doi.org/10.1111/j.1467-7687.2008.00792.x
Crutch, S. J., Connell, S., & Warrington, E. K. (2009). The different
representational frameworks underpinning abstract and concrete knowl-
edge: Evidence from odd-one-out judgements. The Quarterly Journal of
Experimental Psychology: Human Experimental Psychology, 62, 1377–
1390. http://dx.doi.org/10.1080/17470210802483834
Crutch, S. J., & Jackson, E. C. (2011). Contrasting graded effects of
semantic similarity and association across the concreteness spectrum.
The Quarterly Journal of Experimental Psychology: Human Exper-
imental Psychology, 64, 1388 –1408. http://dx.doi.org/10.1080/
17470218.2010.543285
Crutch, S. J., Ridha, B. H., & Warrington, E. K. (2006). The different
frameworks underlying abstract and concrete knowledge: Evidence from
a bilingual patient with a semantic refractory access dysphasia. Neuro-
case, 12, 151–163. http://dx.doi.org/10.1080/13554790600598832
Crutch, S. J., Troche, J., Reilly, J., & Ridgway, G. R. (2013). Abstract
conceptual feature ratings: The role of emotion, magnitude, and other
cognitive domains in the organization of abstract conceptual knowledge.
Frontiers in Human Neuroscience, 7, 186. http://dx.doi.org/10.3389/
fnhum.2013.00186
Crutch, S. J., & Warrington, E. K. (2005). Abstract and concrete concepts
have structurally different representational frameworks. Brain: A Jour-
nal of Neurology, 128, 615– 627. http://dx.doi.org/10.1093/brain/
awh349
Crutch, S. J., & Warrington, E. K. (2007). Semantic priming in deep-
phonological dyslexia: Contrasting effects of association and similarity
upon abstract and concrete word reading. Cognitive Neuropsychology,
24, 583– 602. http://dx.doi.org/10.1080/02643290701577351
Crutch, S. J., & Warrington, E. K. (2010). The differential dependence of
abstract and concrete words upon associative and similarity-based in-
formation: Complementary semantic interference and facilitation ef-
fects. Cognitive Neuropsychology, 27, 46 –71. http://dx.doi.org/10.1080/
02643294.2010.491359
Crutch, S. J., Williams, P., Ridgway, G. R., & Borgenicht, L. (2012). The
role of polarity in antonym and synonym conceptual knowledge: Evi-
dence from stroke aphasia and multidimensional ratings of abstract
words. Neuropsychologia, 50, 2636 –2644. http://dx.doi.org/10.1016/j
.neuropsychologia.2012.07.015
Damasio, A. R. (1989). The brain binds entities and events by multire-
gional activation from convergence zones. Neural Computation, 1, 123–
132.
D’Angiulli, A., Griffiths, G., & Marmolejo-Ramos, F. (2015). Neural
correlates of visualizations of concrete and abstract words in preschool
children: A developmental embodied approach. Frontiers in Psychology,
6, 856. http://dx.doi.org/10.3389/fpsyg.2015.00856
Danguecan, A. N., & Buchanan, L. (2016). Semantic neighborhood effects
for abstract versus concrete words. Frontiers in Psychology, 7, 1034.
Dehaene, S., Bossini, S., & Giraux, P. (1993). The mental representation of
parity and number magnitude. Journal of Experimental Psychology:
General, 122, 371–396. http://dx.doi.org/10.1037/0096-3445.122.3.371
de la Fuente, J., Casasanto, D., Martínez-Cascales, J. I., & Santiago, J.
(2016). Motor imagery shapes abstract concepts. Cognitive Science.
Advance online publication. http://dx.doi.org/10.1111/cogs.12406
Della Rosa, P. A., Catricala
`, E., De Battisti, S., Vinson, D., Vigliocco, G.,
& Cappa, S. F. (2014). How to assess abstract conceptual knowledge:
Construction, standardization and validation of a new battery of seman-
tic memory tests. Functional Neurology, 29, 47–55.
Della Rosa, P. A., Catricala
`, E., Vigliocco, G., & Cappa, S. F. (2010).
Beyond the abstract-concrete dichotomy: Mode of acquisition, concrete-
ness, imageability, familiarity, age of acquisition, context availability,
and abstractness norms for a set of 417 Italian words. Behavior Research
Methods, 42, 1042–1048. http://dx.doi.org/10.3758/BRM.42.4.1042
Desai, R. H., Binder, J. R., Conant, L. L., Mano, Q. R., & Seidenberg,
M. S. (2011). The neural career of sensory-motor metaphors. Journal of
Cognitive Neuroscience, 23, 2376 –2386. http://dx.doi.org/10.1162/jocn
.2010.21596
D’Esposito, M., Detre, J. A., Aguirre, G. K., Stallcup, M., Alsop, D. C.,
Tippet, L. J., & Farah, M. J. (1997). A functional MRI study of mental
image generation. Neuropsychologia, 35, 725–730. http://dx.doi.org/10
.1016/s0028-3932(96)00121-2
Dijkstra, K., Eerland, A., Zijlmans, J., & Post, L. S. (2014). Embodied
cognition, abstract concepts, and the benefits of new technology for
implicit body manipulation. Frontiers in Psychology, 5, 757. http://dx
.doi.org/10.3389/fpsyg.2014.00757
Dove, G. (2009). Beyond perceptual symbols: A call for representational
pluralism. Cognition, 110, 412– 431. http://dx.doi.org/10.1016/j
.cognition.2008.11.016
Dove, G. (2011). On the need for embodied and disembodied cognition.
Frontiers in Psychology, 1, 242. http://dx.doi.org/10.3389/fpsyg.2010
.00242
Dove, G. (2013). Intermediate representations exclude embodiment. Be-
havioral and Brain Sciences, 36, 353–354. http://dx.doi.org/10.1017/
S0140525X12002555
Dove, G. (2014). Thinking in words: Language as an embodied medium of
thought. Topics in Cognitive Science, 6, 371–389. http://dx.doi.org/10
.1111/tops.12102
Dove, G. (2015). Three symbol ungrounding problems: Abstract concepts
and the future of embodied cognition. Psychonomic Bulletin & Review,
23, 1–13.
Dreyer, F. R., Frey, D., Arana, S., von Saldern, S., Picht, T., Vajkoczy, P.,
& Pulvermüller, F. (2015). Is the motor system necessary for processing
action and abstract emotion words? Evidence from focal brain lesions.
Frontiers in Psychology, 6, 1661. http://dx.doi.org/10.3389/fpsyg.2015
.01661
Duñabeitia, J. A., Avilés, A., Afonso, O., Scheepers, C., & Carreiras, M.
(2009). Qualitative differences in the representation of abstract versus
concrete words: Evidence from the visual-world paradigm. Cognition,
110, 284 –292. http://dx.doi.org/10.1016/j.cognition.2008.11.012
Eder, A. B., & Hommel, B. (2013). Anticipatory control of approach and
avoidance: An ideomotor approach. Emotion Review, 5, 275–279. http://
dx.doi.org/10.1177/1754073913477505
Farias, A. R., Garrido, M. V., & Semin, G. R. (2013). Converging modal-
ities ground abstract categories: The case of politics. PLoS ONE, 8,
e60971. http://dx.doi.org/10.1371/journal.pone.0060971
Farias, A. R., Garrido, M. V., & Semin, G. R. (2016). Embodiment of
abstract categories in space...grounding or mere compatibility effects?
The case of politics. Acta Psychologica, 166, 49 –53. http://dx.doi.org/
10.1016/j.actpsy.2016.03.002
Fernandino, L., Conant, L. L., Binder, J. R., Blindauer, K., Hiner, B.,
Spangler, K., & Desai, R. H. (2013a). Parkinson’s disease disrupts both
automatic and controlled processing of action verbs. Brain and Lan-
guage, 127, 65–74. http://dx.doi.org/10.1016/j.bandl.2012.07.008
Fernandino, L., Conant, L. L., Binder, J. R., Blindauer, K., Hiner, B.,
Spangler, K., & Desai, R. H. (2013b). Where is the action? Action
sentence processing in Parkinson’s disease. Neuropsychologia, 51,
1510 –1517. http://dx.doi.org/10.1016/j.neuropsychologia.2013.04.008
Fernandino, L., Humphries, C. J., Seidenberg, M. S., Gross, W. L., Conant,
L. L., & Binder, J. R. (2015). Predicting brain activation patterns
associated with individual lexical concepts based on five sensory-motor
attributes. Neuropsychologia, 76, 17–26. http://dx.doi.org/10.1016/j
.neuropsychologia.2015.04.009
Ferré, P., Guasch, M., García-Chico, T., & Sánchez-Casas, R. (2015). Are
there qualitative differences in the representation of abstract and con-
crete words? Within-language and cross-language evidence from the
semantic priming paradigm. The Quarterly Journal of Experimental
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
287
THE CHALLENGE OF ABSTRACT CONCEPTS
Psychology: Human Experimental Psychology, 68, 2402–2418. http://
dx.doi.org/10.1080/17470218.2015.1016980
Fiebach, C. J., Friederici, A. D. (2004). Processing concrete words: fMRI
evidence against a specific right-hemisphere involvement. Neuropsycho-
logia, 42, 62–70. http://dx.doi.org/10.1016/S0028-3932(03)00145-3
Fiebach, C. J., Ricker, B., Friederici, A. D., & Jacobs, A. M. (2007).
Inhibition and facilitation in visual word recognition: Prefrontal contri-
bution to the orthographic neighborhood size effect. NeuroImage, 36,
901–911. http://dx.doi.org/10.1016/j.neuroimage.2007.04.004
Fischer, M. H., & Brugger, P. (2011). When digits help digits: Spatial–
numerical associations point to finger counting as prime example of
embodied cognition. Frontiers in Psychology, 2, 260. http://dx.doi.org/
10.3389/fpsyg.2011.00260
Flemming, T. M. (2012). Abstract concept learning in animals. In E. Seel
& M. Norbert (Eds.), Encyclopedia of the sciences of learning (pp.
28 –30). New York, NY: Springer.
Flumini, A., & Santiago, J. (2013). Time (also) flies from left to right if it
is needed! In M. Knauff, M. Pauen, N. Sebanz, & I. Wachmuz (Eds.),
Proceedings of the 36th Annual Conference of the Cognitive Science
Society (pp. 2315–2320). Austin, TX: Cognitive Science Society.
Flusberg, S. J., Thibodeau, P. H., Sternberg, D. A., & Glick, J. J. (2010).
A connectionist approach to embodied conceptual metaphor. Frontiers
in Psychology, 1, 197. http://dx.doi.org/10.3389/fpsyg.2010.00197
Förster, J., & Strack, F. (1996). Influence of overt head movements on
memory for valenced words: A case of conceptual-motor compatibility.
Journal of Personality and Social Psychology, 71, 421– 430. http://dx
.doi.org/10.1037/0022-3514.71.3.421
Freina, L., Baroni, G., Borghi, A. M., & Nicoletti, R. (2009). Emotive
concept nouns and motor responses: Attraction or repulsion? Memory &
Cognition, 37, 493– 499. http://dx.doi.org/10.3758/MC.37.4.493
Frith, C. D. (2012). The role of metacognition in human social interactions.
Philosophical Transactions of the Royal Society of London, Series B:
Biological Sciences, 367, 2213–2223. http://dx.doi.org/10.1098/rstb
.2012.0123
Fuhrman, O., & Boroditsky, L. (2007). Mental time-lines follow writing
direction: Comparing English and Hebrew speakers. In D. S. McNamara
& J. G. Trafton (Eds.), Proceedings of the 29th annual conference of the
Cognitive Science Society (pp. 1001–1007). Austin, TX: Cognitive Sci-
ence Society.
Gainotti, G. (2012). The format of conceptual representations disrupted in
semantic dementia: A position paper. Cortex, 48, 521–529. http://dx.doi
.org/10.1016/j.cortex.2011.06.019
Gallese, V. (2008). Mirror neurons and the social nature of language: The
neural exploitation hypothesis. Social Neuroscience, 3, 317–333. http://
dx.doi.org/10.1080/17470910701563608
Gallese, V., & Lakoff, G. (2005). The Brain’s concepts: The role of the
Sensory-motor system in conceptual knowledge. Cognitive Neuropsy-
chology, 22, 455– 479. http://dx.doi.org/10.1080/02643290442000310
Galton, A. (2011). Time flies but space does not: Limits to the spatialisa-
tion of time. Journal of Pragmatics, 43, 695–703. http://dx.doi.org/10
.1016/j.pragma.2010.07.002
Geng, J., & Schnur, T. T. (2015). The representation of concrete and
abstract concepts: Categorical versus associative relationships. Journal
of Experimental Psychology: Learning, Memory, and Cognition, 41,
22– 41. http://dx.doi.org/10.1037/a0037430
Gentilucci, M., & Corballis, M. C. (2006). From manual gesture to speech:
A gradual transition. Neuroscience and Biobehavioral Reviews, 30,
949 –960. http://dx.doi.org/10.1016/j.neubiorev.2006.02.004
Gentner, D. (1981). Some interesting differences between verbs and nouns.
Cognition and Brain Theory, 4, 161–178.
Gentner, D. (2006). Why verbs are hard to learn. In K. Hirsch-Pasek & R.
Golinkoff (Eds.), Action meets words: How children learn verbs (pp.
544 –564). New York, NY: Oxford University Press. http://dx.doi.org/
10.1093/acprof:oso/9780195170009.003.0022
Gentner, D., & Boroditsky, L. (2001). Individuation, relativity and early
word learning. In M. Bowerman & S. Levinson (Eds.), Language ac-
quisition and conceptual development (pp. 215–256). Cambridge,
United Kingdom: Cambridge University Press. http://dx.doi.org/10
.1017/CBO9780511620669.010
Ghio, M., Vaghi, M. M. S., Perani, D., & Tettamanti, M. (2016). Decoding
the neural representation of fine-grained conceptual categories. Neuro-
Image, 132, 93–103. http://dx.doi.org/10.1016/j.neuroimage.2016.02
.009
Ghio, M., Vaghi, M. M. S., & Tettamanti, M. (2013). Fine-grained seman-
tic categorization across the abstract and concrete domains. PLoS ONE
8, e67090. http://dx.doi.org/10.1371/journal.pone.0067090
Gibbs, R. W. J. (1994). The poetics of mind: Figurative thought, language,
and understanding. New York, NY: Cambridge University Press.
Gibbs, R. W. J. (2006). Embodiment and cognitive science. New York,
NY: Cambridge University Press.
Giffard, B., Laisney, M., Desgranges, B., & Eustache, F. (2015). An
exploration of the semantic network in Alzheimer’s disease: Influence of
emotion and concreteness of concepts. Cortex, 69, 201–211. http://dx
.doi.org/10.1016/j.cortex.2015.05.020
Gillette, J., Gleitman, H., Gleitman, L., & Lederer, A. (1999). Human
simulations of vocabulary learning. Cognition, 73, 135–176. http://dx
.doi.org/10.1016/S0010-0277(99)00036-0
Glaser, W. R. (1992). Picture naming. Cognition, 42(1-3), 61–105.
Gleitman, L. R., Cassidy, K., Papafragou, A., Nappa, R., & Trueswell, J. T.
(2005). Hard words. Language Learning and Development, 1, 23– 64.
http://dx.doi.org/10.1207/s15473341lld0101_4
Glenberg, A. M., & Gallese, V. (2012). Action-based language: A theory
of language acquisition, comprehension, and production. Cortex, 48,
905–922. http://dx.doi.org/10.1016/j.cortex.2011.04.010
Glenberg, A. M., & Kaschak, M. P. (2002). Grounding language in action.
Psychonomic Bulletin & Review, 9, 558 –565. http://dx.doi.org/10.3758/
BF03196313
Glenberg, A. M., Sato, M., & Cattaneo, L. (2008a). Use-induced motor
plasticity affects the processing of abstract and concrete language. Cur-
rent Biology, 18, R290 –R291. http://dx.doi.org/10.1016/j.cub.2008.02
.036
Glenberg, A. M., Sato, M., Cattaneo, L., Riggio, L., Palumbo, D., &
Buccino, G. (2008b). Processing abstract language modulates motor
system activity. The Quarterly Journal of Experimental Psychology, 61,
905–919. http://dx.doi.org/10.1080/17470210701625550
Goldinger, S. D., Papesh, M. H., Barnhart, A. S., Hansen, W. A., & Hout,
M. C. (2016). The poverty of embodied cognition. Psychonomic Bulletin
& Review, 23, 959 –978. http://dx.doi.org/10.3758/s13423-015-0860-1
Goldman, A., & de Vignemont, F. (2009). Is social cognition embodied?
Trends in Cognitive Sciences, 13, 154 –159. http://dx.doi.org/10.1016/j
.tics.2009.01.007
Goodhew, S. C., McGaw, B., & Kidd, E. (2014). Why is the sunny side
always up? Explaining the spatial mapping of concepts by language use.
Psychonomic Bulletin & Review, 21, 1287–1293. http://dx.doi.org/10
.3758/s13423-014-0593-6
Grade, S., Badets, A., & Pesenti, M. (2016). Influence of finger and mouth
action observation on random number generation: An instance of em-
bodied cognition for abstract concepts. Psychological Research, 1–11.
http://dx.doi.org/10.1007/s00426-016-0760-7
Granito, C., Scorolli, C., & Borghi, A. M. (2015). Naming a Lego world.
The role of language in the acquisition of abstract concepts. PLoS ONE,
10, e0114615. http://dx.doi.org/10.1371/journal.pone.0114615
Guan, C. Q., Meng, W., Yao, R., & Glenberg, A. M. (2013). The motor
system contributes to comprehension of abstract language. PLoS ONE,
8, e75183. http://dx.doi.org/10.1371/journal.pone.0075183
Hamilton, A. C., & Coslett, H. B. (2008). Refractory access disorders and
the organization of concrete and abstract semantics: Do they differ?
Neurocase, 14, 131–140. http://dx.doi.org/10.1080/13554790802032218
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
288 BORGHI ET AL.
Hamilton, A. C., & Martin, R. C. (2010). Inferring semantic organization
from refractory access dysphasia: Further replication in the domains of
geography and proper nouns but not concrete and abstract concepts.
Cognitive Neuropsychology, 27, 614 – 635. http://dx.doi.org/10.1080/
02643294.2011.609541
Harnad, S. (1990). The symbol grounding problem. Physica D. Nonlinear
Phenomena, 42, 335–346. http://dx.doi.org/10.1016/0167-2789(90)
90087-6
Harris, P. L., & Koenig, M. A. (2006). Trust in testimony: How children
learn about science and religion. Child Development, 77, 505–524.
http://dx.doi.org/10.1111/j.1467-8624.2006.00886.x
Hauk, O., Johnsrude, I., & Pulvermüller, F. (2004). Somatotopic represen-
tation of action words in human motor and premotor cortex. Neuron, 41,
301–307. http://dx.doi.org/10.1016/S0896-6273(03)00838-9
Heit, E., & Barsalou, L. W. (1996). The instantiation principle in
natural categories. Memory, 4, 413– 452. http://dx.doi.org/10.1080/
096582196388915
Hoffman, P. (2015). The meaning of ‘life’ and other abstract words:
Insights from neuropsychology. Journal of Neuropsychology, 10, 317–
343. http://dx.doi.org/10.1111/jnp.12065
Hoffman, P., Binney, R. J., & Lambon Ralph, M. A. (2015). Differing
contributions of inferior prefrontal and anterior temporal cortex to con-
crete and abstract conceptual knowledge. Cortex, 63, 250 –266. http://
dx.doi.org/10.1016/j.cortex.2014.09.001
Hoffman, P., & Lambon Ralph, M. A. (2011). Reverse concreteness effects
are not a typical feature of semantic dementia: Evidence for the hub-
and-spoke model of conceptual representation. Cerebral Cortex, 21,
2103–2112. http://dx.doi.org/10.1093/cercor/bhq288
Irwin, B. A. (2015). An enactivist account of abstract words: Lessons from
Merleau-Ponty. Phenomenology and the Cognitive Sciences, 1–21.
http://dx.doi.org/10.1007/s11097-015-9434-y
Ishihara, M., Keller, P. E., Rossetti, Y., & Prinz, W. (2008). Horizontal
spatial representations of time: Evidence for the STEARC effect. Cortex,
44, 454 – 461. http://dx.doi.org/10.1016/j.cortex.2007.08.010
Jamrozik, A., McQuire, M., Cardillo, E. R., & Chatterjee, A. (2016).
Metaphor: Bridging embodiment to abstraction. Psychonomic Bulletin &
Review, 23, 1080 –1089. http://dx.doi.org/10.3758/s13423-015-0861-0
Jessen, F., Heun, R., Erb, M., Granath, D. O., Klose, U., Papassotiropoulos,
A., & Grodd, W. (2000). The concreteness effect: Evidence for dual
coding and context availability. Brain and Language, 74, 103–112.
http://dx.doi.org/10.1006/brln.2000.2340
Jones, M. (2015). Number concepts for the concept empiricist. Philosoph-
ical Psychology, 29, 334 –348.
Kiehl, K. A., Liddle, P. F., Smith, A. M., Mendrek, A., Forster, B. B., &
Hare, R. D. (1999). Neural pathways involved in the processing of
concrete and abstract words. Human Brain Mapping, 7, 225–233. http://
dx.doi.org/10.1002/(SICI)1097-0193(1999)7:4⬍225::AID-HBM1⬎3.0
.CO;2-P
Kim, S. R., Kim, S., Baek, M. J., & Kim, H. (2015). Abstract word
definition in patients with amnestic mild cognitive impairment. Behav-
ioural Neurology, 2015, 580246. http://dx.doi.org/10.1155/2015/580246
King, L. (2013). The importance of situational information for abstract
concepts (Doctoral Dissertation). University of Western Ontario, Lon-
don, Ontario, Canada.
Kousta, S., Vigliocco, G., Vinson, D. P., & Andrews, M. (2009). Happiness
is an abstract word. The role of affect in abstract knowledge represen-
tation. In N. Taatgen & H. van Rijn (Eds.), Proceedings of the 31st
Annual Conference of the Cognitive Science Society. Amsterdam, the
Netherlands: Cognitive Science Society.
Kousta, S. T., Vigliocco, G., Vinson, D. P., Andrews, M., & Del Campo,
E. (2011). The representation of abstract words: Why emotion matters.
Journal of Experimental Psychology: General, 140, 14 –34. http://dx.doi
.org/10.1037/a0021446
Kranjec, A., & Chatterjee, A. (2010). Are temporal concepts embodied? A
challenge for cognitive neuroscience. Frontiers in Psychology, 1, 240.
http://dx.doi.org/10.3389/fpsyg.2010.00240
Kuenecke, J., Sommer, W., Schacht, A., & Palazova, M. (2015). Embodied
simulation of emotional valence: Facial muscle responses to abstract and
concrete words. Psychophysiology, 52, 1590 –1598. http://dx.doi.org/10
.1111/psyp.12555
Kumar, U. (2015). Neural dychotomy of word concreteness: A view from
functional neuroimaging. Cognitive Processing, 17, 39.
Lai, V. T., & Boroditsky, L. (2013). The immediate and chronic influence
of spatio-temporal metaphors on the mental representations of time in
English, Mandarin, and Mandarin-English speakers. Frontiers in Psy-
chology, 4, 142.
Lakens, D., Semin, G. R., & Foroni, F. (2011). Why your highness needs
the people. Social Psychology, 42, 205–213.
Lakoff, G. (2014). Mapping the brain’s metaphor circuitry: Metaphorical
thought in everyday reason. Frontiers in Human Neuroscience, 8, 958.
http://dx.doi.org/10.3389/fnhum.2014.00958
Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago, IL:
University of Chicago Press.
Lakoff, G., & Johnson, M. (1999). Philosophy in the flesh: The embodied
mind and its challenge to western thought. New York, NY: Basic books.
Lakoff, G., & Núñez, R. E. (2000). Where mathematics comes from: How
the embodied mind brings mathematics into being. New York, NY:
Basic books.
Landauer, T. K., & Dumais, S. T. (1997). A solution to Plato’s problem:
The latent semantic analysis theory of the acquisition, induction, and
representation of knowledge. Psychological Review, 104, 211–240.
http://dx.doi.org/10.1037/0033-295X.104.2.211
Lieberman, P. (2009). Human language and our reptilian brain: The
subcortical bases of speech, syntax, and thought. Cambridge, MA:
Harvard University Press.
Louwerse, M., & Connell, L. (2011). A taste of words: Linguistic context
and perceptual simulation predict the modality of words. Cognitive
Science, 35, 381–398. http://dx.doi.org/10.1111/j.1551-6709.2010
.01157.x
Louwerse, M. M., & Jeuniaux, P. (2010). The linguistic and embodied
nature of conceptual processing. Cognition, 114, 96 –104. http://dx.doi
.org/10.1016/j.cognition.2009.09.002
Louwerse, M. M., Raisig, S., Tillman, R., & Hutchinson, S. (2016). Time
after time in words: Chronology through language statistics. In Noelle,
D. C., Dale, R., Warlaumont, A. S., Yoshimi, J., Matlock, T., Jennings,
C. D., & Maglio, P. P. (Eds.) (2015). Proceedings of the 37th Annual
Meeting of the Cognitive Science Society. Austin, TX: Cognitive Sci-
ence Society. Advance online publication.
Lugli, L., Baroni, G., Anelli, F., Borghi, A. M., & Nicoletti, R. (2013).
Counting is easier while experiencing a congruent motion. PloS ONE,
8(5), e64500.
Lund, K., & Burgess, C. (1996). Producing high-dimensional semantic
spaces from lexical co-occurrence. Behavior Research Methods, In-
struments, & Computers, 28, 203–208. http://dx.doi.org/10.3758/
BF03204766
Lupyan, G. (2012). Linguistically modulated perception and cognition:
The label-feedback hypothesis. Frontiers in Cognition, 3, 54. http://dx
.doi.org/10.3389/fpsyg.2012.00054
Lupyan, G., & Bergen, B. (2016). How language programs the mind.
Topics in Cognitive Science, 8, 408 – 424. http://dx.doi.org/10.1111/tops
.12155
Lupyan, G., & Clark, A. (2015). Words and the world predictive coding
and the language-perception-cognition interface. Current Directions in
Psychological Science, 24, 279 –284. http://dx.doi.org/10.1177/
0963721415570732
Lynott, D., & Coventry, K. (2014). On the ups and downs of emotion:
Testing between conceptual-metaphor and polarity accounts of emo-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
289
THE CHALLENGE OF ABSTRACT CONCEPTS
tional valence-spatial location interactions. Psychonomic Bulletin &
Review, 21, 218 –226. http://dx.doi.org/10.3758/s13423-013-0481-5
Mahon, B. Z. (2015). What is embodied about cognition? Language,
Cognition and Neuroscience, 30, 420 – 429. http://dx.doi.org/10.1080/
23273798.2014.987791
Malt, B. C., Sloman, S. A., Gennari, S., Shi, M., & Wang, Y. (1999).
Knowing versus naming: Similarity and the linguistic categorization of
artifacts. Journal of Memory and Language, 40, 230 –262. http://dx.doi
.org/10.1006/jmla.1998.2593
Markman, A. B., & Stilwell, C. H. (2001). Role-governed categories.
Journal of Experimental & Theoretical Artificial Intelligence, 13, 329 –
358. http://dx.doi.org/10.1080/09528130110100252
Marques, J. F., & Nunes, L. D. (2012). The contributions of language and
experience to the representation of abstract and concrete words: Differ-
ent weights but similar organizations. Memory & Cognition, 40, 1266 –
1275. http://dx.doi.org/10.3758/s13421-012-0220-6
Martin, A. (2007). The representation of object concepts in the brain.
Annual Review of Psychology, 58, 25– 45.
Martin, A. (2016). GRAPES—Grounding representations in action, per-
ception, and emotion systems: How object properties and categories are
represented in the human brain. Psychonomic Bulletin & Review, 23,
979 –990.
Martin, A., Haxby, J. V., Lalonde, F. M., Wiggs, C. L., & Ungerleider,
L. G. (1995). Discrete cortical regions associated with knowledge of
color and knowledge of action. Science, 270, 102–105. http://dx.doi.org/
10.1126/science.270.5233.102
Maurer, D., Pathman, T., & Mondloch, C. J. (2006). The shape of boubas:
Sound-shape correspondences in toddlers and adults. Developmental
Science, 9, 316 –322. http://dx.doi.org/10.1111/j.1467-7687.2006
.00495.x
McCaffrey, J. (2015). Reconceiving conceptual vehicles: Lessons from
semantic dementia. Philosophical Psychology, 28, 337–354. http://dx
.doi.org/10.1080/09515089.2013.827561
Mirolli, M., & Parisi, D. (2011). Towards a Vygotskyan cognitive robotics:
The role of language as a cognitive tool. New Ideas in Psychology, 29,
298 –311. http://dx.doi.org/10.1016/j.newideapsych.2009.07.001
Moffat, M., Siakaluk, P. D., Sidhu, D. M., & Pexman, P. M. (2015).
Situated conceptualization and semantic processing: Effects of emo-
tional experience and context availability in semantic categorization and
naming tasks. Psychonomic Bulletin & Review, 22, 408 – 419. http://dx
.doi.org/10.3758/s13423-014-0696-0
Moon, C., Lagercrantz, H., & Kuhl, P. K. (2013). Language experienced in
utero affects vowel perception after birth: A two-country study. Acta
Paediatrica, 102, 156 –160. http://dx.doi.org/10.1111/apa.12098
Moseley, R., Carota, F., Hauk, O., Mohr, B., & Pulvermüller, F. (2012). A
role for the motor system in binding abstract emotional meaning. Cere-
bral Cortex, 22, 1634 –1647. http://dx.doi.org/10.1093/cercor/bhr238
Moseley, R. L., Shtyrov, Y., Mohr, B., Lombardo, M. V., Baron-Cohen, S.,
& Pulvermüller, F. (2015). Lost for emotion words: What motor and
limbic brain activity reveals about autism and semantic theory. Neuro-
Image, 104, 413– 422. http://dx.doi.org/10.1016/j.neuroimage.2014.09
.046
Murphy, G. L. (1996). On metaphoric representation. Cognition, 60, 173–
204. http://dx.doi.org/10.1016/0010-0277(96)00711-1
Murphy, G. L. (1997). Reasons to doubt the present evidence for meta-
phoric representation. Cognition, 62, 99 –108. http://dx.doi.org/10.1016/
S0010-0277(96)00725-1
Murphy, G. L., & Wisniewski, E. J. (1989). Categorizing objects in
isolation and in scenes: What a superordinate is good for. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 15, 572–
586. http://dx.doi.org/10.1037/0278-7393.15.4.572
Neumann, R., & Strack, F. (2000). Approach and avoidance: The influence
of proprioceptive and exteroceptive cues on encoding of affective infor-
mation. Journal of Personality and Social Psychology, 79, 39 – 48.
http://dx.doi.org/10.1037/0022-3514.79.1.39
Newcombe, P. I., Campbell, C., Siakaluk, P. D., & Pexman, P. M. (2012).
Effects of emotional and sensorimotor knowledge in semantic process-
ing of concrete and abstract nouns. Frontiers in Human Neuroscience, 6,
275. http://dx.doi.org/10.3389/fnhum.2012.00275
Núñez, R., & Marghetis, T. (2014). Cognitive linguistics and the concept
(s) of number. In R. Cohen-Kadosh & K. Dowker (Eds.), Oxford
handbook of numerical cognition (pp. 1–20). New York, NY: Oxford
University Press Oxford.
Ouellet, M., Santiago, J., Israeli, Z., & Gabay, S. (2010). Is the future the
right time? Experimental Psychology, 57, 308 –314. http://dx.doi.org/10
.1027/1618-3169/a000036
Paivio, A. (1986). Mental representations: A dual coding approach. New
York, NY: Oxford University Press.
Paivio, A. (2013). Dual coding theory, word abstractness, and emotion: A
critical review of Kousta et al. (2011). Journal of Experimental Psy-
chology: General, 142, 282–287, 2011. http://dx.doi.org/10.1037/
a0027004
Paivio, A., Yuille, J. C., & Madigan, S. A. (1968). Concreteness, imagery,
and meaningfulness values for 925 nouns. Journal of Experimental
Psychology, 76, 1–25. http://dx.doi.org/10.1037/h0025327
Palazova, M., Sommer, W., & Schacht, A. (2013). Interplay of emotional
valence and concreteness in word processing: An event-related potential
study with verbs. Brain and Language, 125, 264 –271. http://dx.doi.org/
10.1016/j.bandl.2013.02.008
Papagno, C., Fogliata, A., Catricala
`, E., & Miniussi, C. (2009). The lexical
processing of abstract and concrete nouns. Brain Research, 1263, 78 –
86. http://dx.doi.org/10.1016/j.brainres.2009.01.037
Patterson, K., Nestor, P. J., & Rogers, T. T. (2007). Where do you know
what you know? The representation of semantic knowledge in the
human brain. Nature Reviews Neuroscience, 8, 976 –987.
Pecher, D., Boot, I., & van Dantzig, S. (2011). Abstract concepts: Sensory
motor grounding, metaphors, and beyond. In B. Ross (Ed.), The psy-
chology of learning and motivation (Vol. 54, pp. 217–248). Burlington,
VT: Academic Press. http://dx.doi.org/10.1016/B978-0-12-385527-5
.00007-3
Perani, D., Cappa, S. F., Schnur, T., Tettamanti, M., Collina, S., Rosa,
M. M., & Fazio, F. (1999). The neural correlates of verb and noun
processing. A PET study. Brain: A Journal of Neurology, 122, 2337–
2344. http://dx.doi.org/10.1093/brain/122.12.2337
Perniss, P., & Vigliocco, G. (2014). The bridge of iconicity: From a world
of experience to the experience of language. Philosophical Transactions
of the Royal Society of London, Series B: Biological Sciences, 369,
20130300. http://dx.doi.org/10.1098/rstb.2013.0300
Petrides, M. (1994). Frontal lobes and working memory: Evidence from
investigations of the effects of cortical excisions in nonhuman primates.
In F. Boller & J. Grafman (Eds.), Handbook of neuropsychology (Vol. 9,
pp. 59 –82). Amsterdam, the Netherlands: Elsevier.
Pexman, P. M., Hargreaves, I. S., Edwards, J. D., Henry, L. C., &
Goodyear, B. G. (2007). Neural correlates of concreteness in semantic
categorization. Journal of Cognitive Neuroscience, 19, 1407–1419.
http://dx.doi.org/10.1162/jocn.2007.19.8.1407
Phaf, R. H., Mohr, S. E., Rotteveel, M., & Wicherts, J. M. (2014).
Approach, avoidance, and affect: A meta-analysis of approach-
avoidance tendencies in manual reaction time tasks. Frontiers in Psy-
chology, 5, 378. http://dx.doi.org/10.3389/fpsyg.2014.00378
Pickering, M. J., & Garrod, S. (2013). Forward models and their implica-
tions for production, comprehension, and dialogue. Behavioral and
Brain Sciences, 36, 377–392. http://dx.doi.org/10.1017/S0140525X
12003238
Plaut, D. C., & Shallice, T. (1993). Deep dyslexia: A case study of
connectionist neuropsychology. Cognitive Neuropsychology, 10, 377–
500.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
290 BORGHI ET AL.
Pobric, G., Ralph, M. A. L., & Jefferies, E. (2009). The role of the anterior
temporal lobes in the comprehension of concrete and abstract words:
RTMS evidence. Cortex, 45, 1104 –1110.
Primativo, S., Reilly, J., & Crutch, S. J. (2016). Abstract conceptual feature
ratings predict gaze within written word arrays: Evidence from a visual
world paradigm. Cognitive Science. Advance online publication. http://
dx.doi.org/10.1111/cogs.12348
Prinz, J. J. (2002). Furnishing the mind: Concepts and their perceptual
basis. Cambridge, MA: MIT Press.
Prinz, J. J. (2012). Beyond human nature. How culture and experience
shape our lives. London, United Kingdom: Penguin.
Ranzini, M., Lugli, L., Anelli, F., Carbone, R., Nicoletti, R., & Borghi,
A. M. (2011). Graspable objects shape number processing. Frontiers in
Human Neuroscience, 5, 147. http://dx.doi.org/10.3389/fnhum.2011
.00147
Recchia, G., & Jones, M. N. (2012). The semantic richness of abstract
concepts. Frontiers in Human Neuroscience, 6, 315. http://dx.doi.org/
10.3389/fnhum.2012.00315
Reilly, J., Peelle, J. E., Garcia, A., & Crutch, S. J. (2016). Linking somatic
and symbolic representation in semantic memory: The dynamic multi-
level reactivation framework. Psychonomic Bulletin & Review, 23,
1002–1014. http://dx.doi.org/10.3758/s13423-015-0824-5
Rizzolatti, G., & Craighero, L. (2004). The mirror-neuron system. Annual
Review of Neuroscience, 27, 169 –192.
Rodríguez-Ferreiro, J., Gennari, S. P., Davies, R., & Cuetos, F. (2011).
Neural correlates of abstract verb processing. Journal of Cognitive
Neuroscience, 23, 106 –118.
Roush, D. R. (2011). Language between bodies: A cognitive approach to
understanding linguistic politeness in American Sign Language. Sign
Language Studies, 11, 329 –374. http://dx.doi.org/10.1353/sls.2011.0000
Roversi, C., Borghi, A. M., & Tummolini, L. (2013). A marriage is an
artefact and not a walk that we take together: An experimental study on
the categorization of artefacts. Review of Philosophy and Psychology, 4,
527–542. http://dx.doi.org/10.1007/s13164-013-0150-7
Rüschemeyer, S. A., Brass, M., & Friederici, A. D. (2007). Comprehending
prehending: Neural correlates of processing verbs with motor stems.
Journal of Cognitive Neuroscience, 19, 855– 865. http://dx.doi.org/10
.1162/jocn.2007.19.5.855
Sabbagh, M. A., & Baldwin, D. A. (2001). Learning words from knowl-
edgeable versus ignorant speakers: Links between preschoolers’ theory
of mind and semantic development. Child Development, 72, 1054 –1070.
http://dx.doi.org/10.1111/1467-8624.00334
Sabsevitz, D. S., Medler, D. A., Seidenberg, M., & Binder, J. R. (2005).
Modulation of the semantic system by word imageability. NeuroImage,
27, 188 –200. http://dx.doi.org/10.1016/j.neuroimage.2005.04.012
Sakreida, K., Scorolli, C., Menz, M. M., Heim, S., Borghi, A. M., &
Binkofski, F. (2013). Are abstract action words embodied? An fMRI
investigation at the interface between language and motor cognition.
Frontiers in Human Neuroscience, 7, 125. http://dx.doi.org/10.3389/
fnhum.2013.00125
Santiago, J., & Lakens, D. (2015). Can conceptual congruency effects
between number, time, and space be accounted for by polarity corre-
spondence? Acta Psychologica, 156, 179 –191. http://dx.doi.org/10
.1016/j.actpsy.2014.09.016
Santiago, J., Lupiáñez, J., Pérez, E., & Funes, M. J. (2007). Time (also)
flies from left to right. Psychonomic Bulletin & Review, 14, 512–516.
http://dx.doi.org/10.3758/BF03194099
Sato, M., Schafer, A. J., & Bergen, B. K. (2015). Metaphor priming in
sentence production: Concrete pictures affect abstract language produc-
tion. Acta Psychologica, 156, 136 –142. http://dx.doi.org/10.1016/j
.actpsy.2014.09.010
Schilhab, T. (2015a). Double talk—Both biological and social constraints
on language. Biologically Inspired Cognitive Architectures, 13, 1– 8.
http://dx.doi.org/10.1016/j.bica.2015.06.002
Schilhab, T. S. (2015b). Words as cultivators of others minds. Frontiers in
Psychology, 6, 1690. http://dx.doi.org/10.3389/fpsyg.2015.01690
Schwanenflugel, P. J., Akin, C., & Luh, W. M. (1992). Context availability
and the recall of abstract and concrete words. Memory & Cognition, 20,
96 –104. http://dx.doi.org/10.3758/BF03208259
Schwanenflugel, P. J., Harnishfeger, K. K., & Stowe, R. W. (1988).
Context availability and lexical decisions for abstract and concrete
words. Journal of Memory and Language, 27, 499 –520. http://dx.doi
.org/10.1016/0749-596X(88)90022-8
Scorolli, C. (2014). Embodiment and language. In L. Shapiro (Ed.), The
Routledge handbook of embodied cognition (pp. 127–138). London,
United Kingdom: Routledge.
Scorolli, C., Binkofski, F., Buccino, G., Nicoletti, R., Riggio, L., & Borghi,
A. M. (2011). Abstract and concrete sentences, embodiment, and lan-
guages. Frontiers in Psychology, 2, 227. http://dx.doi.org/10.3389/fpsyg
.2011.00227
Scorolli, C., Jacquet, P. O., Binkofski, F., Nicoletti, R., Tessari, A., &
Borghi, A. M. (2012). Abstract and concrete phrases processing differ-
entially modulates cortico-spinal excitability. Brain Research, 1488,
60 –71. http://dx.doi.org/10.1016/j.brainres.2012.10.004
Seibt, B., Neumann, R., Nussinson, R., & Strack, F. (2008). Movement
direction or change in distance? Self-and object-related approach–
avoidance motions. Journal of Experimental Social Psychology, 44,
713–720. http://dx.doi.org/10.1016/j.jesp.2007.04.013
Sell, A. J., & Kaschak, M. P. (2011). Processing time shifts affects the
execution of motor responses. Brain and Language, 117, 39 – 44. http://
dx.doi.org/10.1016/j.bandl.2010.07.003
Setti, A., & Caramelli, N. (2005). Different domains in abstract concepts.
In B. Bara, B. Barsalou, & M. Bucciarelli (Eds.), Proceedings of the
XXVII Annual Conference of the Cognitive Science. Mahwah, NJ: Erl-
baum.
Shallice, T., & Cooper, R. P. (2013). Is there a semantic system for abstract
words? Frontiers in Human Neuroscience, 7, 175. http://dx.doi.org/10
.3389/fnhum.2013.00175
Shallice, T., & Warrington, E. K. (1975). Word recognition in a phonemic
dyslexic patient. The Quarterly Journal of Experimental Psychology, 27,
187–199. http://dx.doi.org/10.1080/14640747508400479
Siakaluk, P. D., Knol, N., & Pexman, P. M. (2014). Effects of emotional
experience for abstract words in the Stroop task. Cognitive Science, 38,
1698 –1717. http://dx.doi.org/10.1111/cogs.12137
Simmons, W. K., Hamann, S. B., Harenski, C. L., Hu, X. P., & Barsalou,
L. W. (2008). fMRI evidence for word association and situated simula-
tion in conceptual processing. Journal of Physiology, Paris, 102, 106 –
119. http://dx.doi.org/10.1016/j.jphysparis.2008.03.014
Simmons, W. K., & Martin, A. (2009). The anterior temporal lobes and the
functional architecture of semantic memory. Journal of the International
Neuropsychological Society, 15, 645– 649. http://dx.doi.org/10.1017/
S1355617709990348
Simmons, W. K., Martin, A., & Barsalou, L. W. (2005). Pictures of
appetizing foods activate gustatory cortices for taste and reward. Cere-
bral Cortex, 15, 1602–1608. http://dx.doi.org/10.1093/cercor/bhi038
Skipper, L. M., & Olson, I. R. (2014). Semantic memory: Distinct neural
representations for abstractness and valence. Brain and Language, 130,
1–10. http://dx.doi.org/10.1016/j.bandl.2014.01.001
Slepian, M. L., & Ambady, N. (2014). Simulating sensorimotor metaphors:
Novel metaphors influence sensory judgments. Cognition, 130, 309 –
314. http://dx.doi.org/10.1016/j.cognition.2013.11.006
Snefjella, B., & Kuperman, V. (2015). Concreteness and psychological
distance in natural language use. Psychological Science, 26, 1449 –1460.
http://dx.doi.org/10.1177/0956797615591771
Talmy, L. (1988). Force dynamics in language and cognition. Cognitive
Science, 12, 49 –100.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
291
THE CHALLENGE OF ABSTRACT CONCEPTS
Thill, S., & Twomey, K. E. (2016). What’s on the inside counts: A
grounded account of concept acquisition and development. Frontiers in
Psychology, 7, 402.
Tomasello, M., & Akhtar, N. (1995). Two-year-olds use pragmatic cues to
differentiate reference to objects and actions. Cognitive Development,
10, 201–224. http://dx.doi.org/10.1016/0885-2014(95)90009-8
Tomasino, B., & Rumiati, R. I. (2013). Introducing the special topic “The
when and why of sensorimotor processes in conceptual knowledge and
abstract concepts.” Frontiers in Human Neuroscience, 7, 498. http://dx
.doi.org/10.3389/fnhum.2013.00498
Topolinski, S., Maschmann, I. T., Pecher, D., & Winkielman, P. (2014).
Oral approach-avoidance: Affective consequences of muscular articula-
tion dynamics. Journal of Personality and Social Psychology, 106,
885– 896. http://dx.doi.org/10.1037/a0036477
Torralbo, A., Santiago, J., & Lupiáñez, J. (2006). Flexible conceptual
projection of time onto spatial frames of reference. Cognitive Science,
30, 745–757. http://dx.doi.org/10.1207/s15516709cog0000_67
Troche, J., Crutch, S., & Reilly, J. (2014). Clustering, hierarchical orga-
nization, and the topography of abstract and concrete nouns. Frontiers in
Psychology, 5, 360. http://dx.doi.org/10.3389/fpsyg.2014.00360
Tulving, E., & Thomson, D. M. (1973). Encoding specificity and retrieval
processes in episodic memory. Psychological Review, 80, 352–373.
van Dantzig, S., Pecher, D., & Zwaan, R. A. (2008). Approach and
avoidance as action effects. The Quarterly Journal of Experimental
Psychology: Human Experimental Psychology, 61, 1298 –1306. http://
dx.doi.org/10.1080/17470210802027987
Vigliocco, G., Kousta, S. T., Della Rosa, P. A., Vinson, D. P., Tettamanti,
M., Devlin, J. T., & Cappa, S. F. (2014). The neural representation of
abstract words: The role of emotion. Cerebral Cortex, 24, 1767–1777.
http://dx.doi.org/10.1093/cercor/bht025
Vigliocco, G., Kousta, S., Vinson, D., Andrews, M., & Del Campo, E.
(2013). The representation of abstract words: What matters? Reply to
Paivio’s (2013) comment on Kousta et al. (2011). Journal of Experi-
mental Psychology: General, 142, 288 –291. http://dx.doi.org/10.1037/
a0028749
Vouloumanos, A., & Werker, J. F. (2007). Listening to language at birth:
Evidence for a bias for speech in neonates. Developmental Science, 10,
159 –164. http://dx.doi.org/10.1111/j.1467-7687.2007.00549.x
Vygotsky, L. (1986). Thought and language. Boston, MA: MIT Press.
Walker, E., & Cooperrider, K. (2016). The continuity of metaphor: Evi-
dence from temporal gestures. Cognitive Science, 40, 481– 495.
Wang, H. L., Lu, Y. Q., & Lu, Z. Y. (2016). Moral-up first, immoral-down
last: The time course of moral metaphors on a vertical dimension.
NeuroReport, 27, 247–256. http://dx.doi.org/10.1097/WNR.0000
000000000528
Wang, J., Conder, J. A., Blitzer, D. N., & Shinkareva, S. V. (2010). Neural
representation of abstract and concrete concepts: A meta-analysis of
neuroimaging studies. Human Brain Mapping, 31, 1459 –1468. http://
dx.doi.org/10.1002/hbm.20950
Wang, L., Uhrig, L., Jarraya, B., & Dehaene, S. (2015). Representation of
numerical and sequential patterns in macaque and human brains. Current
Biology, 25, 1966 –1974. http://dx.doi.org/10.1016/j.cub.2015.06.035
Warrington, E. K. (1975). The selective impairment of semantic memory.
The Quarterly Journal of Experimental Psychology, 27, 635– 657.
Warrington, E. K., & Shallice, T. (1984). Category specific semantic
impairments. Brain: A Journal of Neurology, 107, 829 – 853. http://dx
.doi.org/10.1093/brain/107.3.829
Wauters, L. N., Tellings, A. E., Van Bon, W. H., & Van Haaften, A. W.
(2003). Mode of acquisition of word meanings: The viability of a
theoretical construct. Applied Psycholinguistics, 24, 385– 406. http://dx
.doi.org/10.1017/S0142716403000201
Wiemer-Hastings, K., Krug, J., & Xu, X. (2001). Imagery, context avail-
ability, contextual constraints and abstractness. Proceedings of the 23rd
Annual Meeting of the Cognitive Science Society (pp. 1106 –1111).
Hillsdale, NJ: Erlbaum.
Wiemer-Hastings, K., & Xu, X. (2005). Content differences for abstract
and concrete concepts. Cognitive Science, 29, 719 –736. http://dx.doi
.org/10.1207/s15516709cog0000_33
Wilcox, S., Rossini, P., & Antinoro Pizzuto, E. (2010). Grammaticalization
in sign languages. In D. Brentari (Ed.), Sign languages (pp. 332–354).
Cambridge, United Kingdom: Cambridge University Press. http://dx.doi
.org/10.1017/CBO9780511712203.016
Wilson, M. (2002). Six views of embodied cognition. Psychonomic Bul-
letin & Review, 9, 625– 636. http://dx.doi.org/10.3758/BF03196322
Wilson-Mendenhall, C. D., Barrett, L. F., Simmons, W. K., & Barsalou, L. W.
(2011). Grounding emotion in situated conceptualization. Neuropsycholo-
gia, 49, 1105–1127. http://dx.doi.org/10.1016/j.neuropsychologia.2010.12
.032
Wilson-Mendenhall, C. D., Simmons, W. K., Martin, A., & Barsalou,
L. W. (2013). Contextual processing of abstract concepts reveals neural
representations of nonlinguistic semantic content. Journal of Cognitive
Neuroscience, 25, 920 –935. http://dx.doi.org/10.1162/jocn_a_00361
Winner, E., Rosenstiel, A. K., & Gardner, H. (1976). The development of
metaphoric understanding. Developmental Psychology, 12, 289 –297.
http://dx.doi.org/10.1037/0012-1649.12.4.289
Winter, B., Marghetis, T., & Matlock, T. (2015). Of magnitudes and meta-
phors: Explaining cognitive interactions between space, time, and number.
Cortex, 64, 209 –224. http://dx.doi.org/10.1016/j.cortex.2014.10.015
Wolpert, D. M., Doya, K., & Kawato, M. (2003). A unifying computational
framework for motor control and social interaction. Philosophical
Transactions of the Royal Society of London Series B, Biological Sci-
ences, 358, 593– 602. http://dx.doi.org/10.1098/rstb.2002.1238
Yao, B., Vasiljevic, M., Weick, M., Sereno, M. E., O’Donnell, P. J., &
Sereno, S. C. (2013). Semantic size of abstract concepts: It gets emo-
tional when you can’t see it. PLoS ONE, 8, e75000. http://dx.doi.org/10
.1371/journal.pone.0075000
Zanolie, K., Dantzig, S., Boot, I., Wijnen, J., Schubert, T. W., Giessner,
S. R., & Pecher, D. (2012). Mighty metaphors: Behavioral and ERP
evidence that power shifts attention on a vertical dimension. Brain
and Cognition, 78, 50 –58. http://dx.doi.org/10.1016/j.bandc.2011.10
.006
Zdrazilova, L., & Pexman, P. M. (2013). Grasping the invisible: Semantic
processing of abstract words. Psychonomic Bulletin & Review, 20,
1312–1318. http://dx.doi.org/10.3758/s13423-013-0452-x
Zhang, X., Han, Z., & Bi, Y. (2013). Are abstract and concrete concepts
organized differently? Evidence from the blocked translation paradigm.
Applied Psycholinguistics, 34, 1059 –1092. http://dx.doi.org/10.1017/
S0142716412000124
Zhao, X., He, X., & Zhang, W. (2016). A heavy heart: The association
between weight and emotional words. Frontiers in Psychology, 7, 920.
http://dx.doi.org/10.3389/fpsyg.2016.00920
Zwaan, R. A. (2014). Embodiment and language comprehension: Refram-
ing the discussion. Trends in Cognitive Sciences, 18, 229 –234. http://
dx.doi.org/10.1016/j.tics.2014.02.008
Zwaan, R. A. (2016). Situation models, mental simulations, and abstract
concepts in discourse comprehension. Psychonomic Bulletin & Review,
23, 1028 –1034.
Received July 7, 2016
Revision received October 18, 2016
Accepted October 26, 2016 䡲
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
292 BORGHI ET AL.
A preview of this full-text is provided by American Psychological Association.
Content available from Psychological Bulletin
This content is subject to copyright. Terms and conditions apply.
