Social semantics: the organization and grounding of abstract concepts

Abstract

Abstract concepts, like justice and friendship, are central features of our daily lives. Traditionally, abstract concepts are distinguished from other concepts in that they cannot be directly experienced through the senses. As such, they pose a challenge for strongly embodied models of semantic representation that assume a central role for sensorimotor information. There is growing recognition, however, that it is possible for meaning to be ‘grounded’ via cognitive systems, including those involved in processing language and emotion. In this article, we focus on the specific proposal that social significance is a key feature in the representation of some concepts. We begin by reviewing recent evidence in favour of this proposal from the fields of psycholinguistics and neuroimaging. We then discuss the limited extent to which there is consensus about the definition of ‘socialness’ and propose essential next steps for research in this domain. Taking one such step, we describe preliminary data from an unprecedented large-scale rating study that can help determine how socialness is distinct from other facets of word meaning. We provide a backdrop of contemporary theories regarding semantic representation and social cognition and highlight important predictions for both brain and behaviour.
This article is part of the theme issue ‘Concepts in interaction: social engagement and inner experiences’.

Full text

royalsocietypublishing.org/journal/rstb
Review
Cite this article: Pexman PM, Diveica V,
Binney RJ. 2022 Social semantics: the
organization and grounding of abstract
concepts. Phil. Trans. R. Soc. B 378: 20210363.
https://doi.org/10.1098/rstb.2021.0363
Received: 25 November 2021
Accepted: 23 June 2022
One contribution of 23 to a theme issue
‘Concepts in interaction: social engagement
and inner experiences’.
Subject Areas:
neuroscience, cognition
Keywords:
semantic memory, social brain, embodied
cognition, abstract concepts, socialness
Author for correspondence:
Richard J. Binney
e-mail: r.binney@bangor.ac.uk
Social semantics: the organization and
grounding of abstract concepts
Penny M. Pexman
1
, Veronica Diveica
2
and Richard J. Binney
2
1
Department of Psychology and Hotchkiss Brain Institute, University of Calgary, Calgary, Canada, T2N 1N4
2
School of Human and Behavioural Sciences, Bangor University, Bangor LL57 2AS, UK
PMP, 0000-0001-7130-0973; VD, 0000-0002-5696-8200; RJB, 0000-0003-4474-2922
Abstract concepts, like justice and friendship, are central features of our daily
lives. Traditionally, abstract concepts are distinguished from other concepts
in that they cannot be directly experienced through the senses. As such,
they pose a challenge for strongly embodied models of semantic represen-
tation that assume a central role for sensorimotor information. There is
growing recognition, however, that it is possible for meaning to be
‘grounded’via cognitive systems, including those involved in processing
language and emotion. In this article, we focus on the specific proposal
that social significance is a key feature in the representation of some concepts.
We begin by reviewing recent evidence in favour of this proposal from the
fields of psycholinguistics and neuroimaging. We then discuss the limited
extent to which there is consensus about the definition of ‘socialness’and
propose essential next steps for research in this domain. Taking one such
step, we describe preliminary data from an unprecedented large-scale
rating study that can help determine how socialness is distinct from other
facets of word meaning. We provide a backdrop of contemporary theories
regarding semantic representation and social cognition and highlight
important predictions for both brain and behaviour.
This article is part of the theme issue ‘Concepts in interaction: social
engagement and inner experiences’.
1. Introduction
You are mistaken, Mr Darcy, if you suppose that the mode of your declaration affected
me in any other way than as it spared me the concern which I might have felt in refus-
ing you, had you behaved in a more gentleman-like manner.
–Jane Austen, Pride and Prejudice [1]
This brief extract from Pride and Prejudice, a classic tale in the importance of per-
sonal character, integrity and morality, is rich with references to concepts of a
social nature (e.g. manner,gentleman and refuse). Indeed, a large portion of
even the most everyday vocabulary is occupied by abstract words imbued
with a sense of socialness. Arguably, this reflects the vital role of social concep-
tual knowledge in navigating our interpersonal world. After all, humans are
intrinsically and uniquely social. We exhibit a natural propensity to cooperate,
coordinate and learn from one another, and to a very large extent, this is done
though the medium of language. It is argued that our advanced social cognitive
and emotional abilities, as well as the evolution of language, are an adaptation
to, and thus a direct consequence of life lived in groups [2,3]. By extension, this
suggests there could be a fundamental nature to the social qualities of words.
Recent work in the field of cognitive science has begun to elucidate the ways
in which socialness impacts the structure of concepts and the representation of
semantic knowledge in the human brain, and this work will be the subject of
the first two parts of this paper. In Part A, we will begin with a brief overview
of general theories of semantic memory, with a particular emphasis on what is
known as the grounding problem and the difficulties it poses for representing
© 2022 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution
License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original
author and source are credited.

abstract word knowledge. Then we will introduce nascent
theories that posit social experience as a mechanism for
grounding conceptual knowledge, together with a review of
recent semantic feature generation/ratings studies that ident-
ify socialness as an important factor for distinguishing among
different ‘types’of concepts. In Part B, we will review a set of
neuroimaging studies that have approached socialness from a
different methodological perspective, exploring if and how
socialness of concepts is represented at the level of macroscale
brain anatomy. This includes evidence that is in line with
claims that social concepts have a special, or even privileged
status over other types of concepts, and suggests socialness
drives the functional organization of neurobiological systems.
Moreover, a key aim of this paper is to highlight major
outstanding questions, and this includes one very fundamen-
tal issue that arises from the work described in both Parts A
and B: what is it exactly that defines a word as being ‘social’?
In Part C, we will discuss the limited extent to which there is
consensus on the kinds of semantic features that amount to
‘socialness’and the degree to which it has been established
as a valid and meaningful construct. Consequently, we
argue that to further progress theory, the field must first
establish a clearer working definition of socialness. To this
end, we describe preliminary data from a large-scale rating
study in which Diveica et al. [4] provided participants with
an inclusive definition of socialness and asked them to collec-
tively rate over 8000 English words. This includes findings
that appear to confirm that these ratings capture aspects of
word meaning that are distinct from those measured via
other semantic variables like concreteness.
The issue of whether social concepts are a distinct type,
either from other forms of abstract concept or even more gen-
erally speaking (i.e. such that this extends to concrete social
concepts), has important theoretical implications regarding
the fundamental organizational principles underpinning
semantic representation.
1
In turn, it has implications for our
understanding of the configuration of brain systems, includ-
ing those responsible for language and social cognition.
These implications extend to applied areas of research
where an improved framework for understanding the way
social and affective concepts are learned, represented and
impaired, could have important implications for educational
and clinical practice (see [5,6]).
2. Part A –Abstract word representation: a role
for socialness?
There are now numerous theories of semantic knowledge,
which vary in the extent to which sensorimotor information
(e.g. visual, auditory or tactile experience) plays a role in
the representation and processing of word meaning. At one
end of the spectrum, amodal theories posit that semantic
knowledge is represented symbolically, distinct from the
ways we experience the world (e.g. [7,8]). At the other end,
strongly embodied theories posit that knowledge is rep-
resented by sensory and motor systems (e.g. [9,10]).
Between these poles lie multimodal or multiple represen-
tation theories, which posit that semantic knowledge is
represented in many ways (e.g. via language, emotion, intro-
spection and sensorimotor systems), and some versions of
those theories include an intermediary supramodal hub
(e.g. [11,12]). The hub accounts position language
information as one of many types of knowledge connected
to the hub. Further, the multiple representation accounts
assume that different kinds of information are important
for different types of concepts (e.g. [13,14]).
Proponents of semantic theories that include reliance on
sensorimotor systems have argued that these theories have
the advantage of addressing the grounding problem [15,16].
The grounding problem asks, in essence, if knowledge is rep-
resented as symbols, then how do those symbols map to the
world? Embodied theories solve that problem by proposing
that cognition engages modal systems (e.g. those used for
perception, action) to represent semantic knowledge.
Strongly embodied theories, however, run into difficulty
explaining representation of abstract words. The meanings
of abstract words, by definition, cannot be learned or experi-
enced through sensorimotor systems, so they cannot be
accounted for by embodied theories. To explain knowledge
of abstract words, other means of learning and representation
must be considered. Barsalou et al. (e.g. [17,18]) have noted
that too many approaches to abstract concepts emphasize
what they do not contain (sensorimotor information) and
that a more positive approach is needed to explore what
they do contain. To that end, Barsalou & Wiemer-Hastings
[18] (see also [19]) used a property generation task to com-
pare the features of a small set of abstract and concrete
words. They found that abstract words were notably different
in that their meanings were mainly associated with introspec-
tions and, in particular, social aspects of situations, such as
people, communication and social institutions.
In addition, work has begun to identify concept ‘types’
within the abstract realm. Much of this work is inspired by
multiple representation views and considers multiple sources
of grounding beyond the sensorimotor, including the potential
contributions of action, language, interoception, emotion,
cognition and other internal states. Notably, Borghi and col-
leagues have proposed the Words as Social Tools (WAT)
account, which focuses particularly on the acquisition and rep-
resentation of abstract word meaning [14,20]. They argue that
abstract words are associated with richer linguistic, inner and,
importantly for present purposes, social experience, than are
concrete words (also [21]). Further, they suggest that there
could be different types of abstract concepts which vary in
their reliance on these different types of information. They
suggested that these types of abstract concepts might include
institutional, temporal, mental states, emotional, numerical
and social concepts.
In related empirical work, researchers have explored the
features or properties of abstract word meanings in order to
derive potential clusters or types. For instance, Harpaintner
et al. [22] examined the features listed for 296 abstract words
and found that they fell into three clusters. The largest cluster
was primarily distinguished by a higher proportion of sensor-
imotor features, with some social features. A second cluster
was distinguished by a high proportion of internal/emotional
features and more social features than either of the other clus-
ters. The third, smallest cluster was distinguished by a high
proportion of verbal association features. Similarly, Troche
et al. [23] investigated the organization of abstract and concrete
English nouns by asking participants to rate 200 concrete and
200 abstract words on 12 dimensions. They analysed the rat-
ings and identified three latent semantic factors: affective
association/social cognition, perceptual salience and magni-
tude (also see [24]). Abstract word meanings relied more
royalsocietypublishing.org/journal/rstb Phil. Trans. R. Soc. B 378: 20210363
2

heavily on affective association/social cognition than did con-
crete meanings. Villani et al.[25] asked participants to rate 425
abstract Italian nouns on 15 dimensions and identified four
clusters: philosophical/spiritual concepts; physical, spatio-
temporal and quantitative concepts; emotional/inner state
concepts; and self and sociality concepts. Additional analyses
showed that the involvement of the dimension they called
social metacognition (defined as a reliance on other people
to understand the meaning) distinguished abstract from con-
crete words, with more abstract words tending to have
higher ratings of social metacognition. In addition, ratings
on a dimension that they termed social valence (defined as
evocative of social situations) were associated with emotion
ratings, and with ratings of mouth movement and hearing.
These latter relationships were attributed to the important
role that language is assumed to play in representing abstract
concepts, and to the importance of mouth movement and
hearing to language.
Similar conclusions about the existence of types of
abstract concepts were drawn from an fMRI study reported
by Vargas & Just [26]. They investigated the clustering of
28 abstract words in terms of neural signatures after partici-
pants were scanned while thinking of properties of each
word. Results showed that there tended to be commonalities
across participants in terms of the neural signatures of each
word, and the authors identified three latent factors includ-
ing verbal representation, externality/internality and social
content (also see [27,28]).
Thus, there is evidence from some property-generation
and feature-rating studies that social words may be a distinct
type of abstract word, consistent with assumptions of the
WAT theory and other multimodal accounts. Each of these
studies, however, has involved a relatively small sample of
abstract words, many fewer than people actually know. There-
fore, it is possible that the results couldbe specific to the words
tested and may not generalize to a larger set. Thus, there is a
need to explore socialness at a much larger scale and right
along the concreteness continuum. There is also a need to
investigate behavioural effects of socialness. That is, if social
words are a distinct type, then one might expect that a
word’s degree of socialness would be reflected in some way
in behavioural measures of lexical-semantic processing, as
much as semantic dimensions like valence [29] and concrete-
ness [30] are related to such processing (e.g. [31–33]). One
might also expect behavioural responses to social abstract
words to be different to those given to other types of abstract
words (see [34] for an example of this approach). And yet,
comparisons between social abstract words and other abstract
word types have still to be made in the context of larger scale
behavioural studies. However, they have been contrasted in
the neuroimaging literature, reviewed next.
3. Part B –Socialness and the brain
A review of neuroimaging literature concerning the repre-
sentation of abstract concepts identified a small number of
papers that treat social concepts a priori as a discrete ‘cat-
egory’[35]. Most of these studies contrasted social words
with a more general class of abstract or concrete words and
set out to identify common activity, and/or that which is
uniquely associated [36–39]. The earliest of these studies gen-
erated a hypothesis that social concepts are a class of concepts
with a special, or even privileged status over other types of
conceptual knowledge [36,37]. In this context, social concep-
tual knowledge has been broadly defined as person-specific
knowledge [40], but also knowledge about interpersonal
relationships, social behaviours and of more abstract social
concepts such as truth and liberty [36,41]. These early studies
revealed a patch of anterior temporal association cortex that
the authors claimed is selectively involved in processing
semantic information of a social nature [40,41].
The ‘social knowledge hypothesis’[36,41] can be likened
to other forms of ‘multiple semantics’views [42–44]in
which the semantic system is composed of multiple indepen-
dent stores that are differentiated by their link to distinct
sensorimotor modalities. Of course, the difference is that
the social distinction is based on domain-specificity rather
than modality. To understand how this hypothesis formed
the starting point for this particular set of neuroimaging
studies, one can look to the broader social neuroscience
field from which they stemmed. The emergence of this field
was triggered, at least in part, by the ‘social brain hypothesis’
[45,46], which states that the expansion of frontal and tem-
poral neocortices across primate species in the human
evolutionary lineage is explained by their increasingly high
levels of sociality (see [47] for a related review). This created
a pervasive assumption, sometimes implicit, that there is a
circumscribed set of brain regions that are dedicated to,
and, by inference, support specialized processes for social
perception and cognition [45,48]. The extent to which
domain-specificity of systems for processing social infor-
mation exists is hotly debated [49–52], but there is evidence
for the existence of brain regions or pathways that are sensi-
tive to socialness, particularly at the level of perceptual
processes [53–55]. This includes visual cortex with ostensibly
selective engagement by faces [56], bodies [57] and social
interactions [58]. Whether this putative specialization cas-
cades downstream to higher-order cognitive systems
(e.g. memory; executive function) is a more contentious
issue [52,59–61].
To date, the leading candidate in terms of a locus for a
selective social semantic system lies within the dorsolateral
aspects of the anterior temporal lobe (ATL), specifically the
anterior superior temporal gyri/sulci [41,62,63]. These ATL
subregions exhibit elevated blood-oxygen-level-dependent
responses when semantic judgements made on socially rel-
evant stimuli are compared to those made on non-social
stimuli [36–39,64,65]. The dorsolateral ATL also appears to
increase its response in line with an accumulation of social
meaning across connected text [66]. The role of the ATL in
representing social knowledge has been ascribed with a
right lateralization within some accounts [67], although indi-
vidual fMRI studies [37,38,62,65,68] and meta-analyses
[62,63] indicate bilateral involvement (also see [69,70]).
More recent neuroimaging studies have attempted to dis-
entangle the socialness effect driving some ATL activations
from other potentially confounding variables. For example,
it is possible that the social concepts explored in neuroima-
ging studies are, on average, more abstract than more
general classes of concepts. However, studies have shown
that preferential left hemisphere dorsolateral/polar ATL
activation cannot be easily explained by differences in concre-
teness, or at least imageability, between social words and
control words [38,64,65,71], nor by differing degrees of mul-
tiplicity of single word meanings (sometimes referred to as
royalsocietypublishing.org/journal/rstb Phil. Trans. R. Soc. B 378: 20210363
3

‘semantic diversity’[72]) [38]. A putative involvement of
these regions in combinatorial conceptual processes does
also not appear to explain differential engagement by social
and non-social words [73,74]. Many of these studies have
also been careful to rule out explanations in terms of funda-
mental lexical properties such as word frequency or
syllable/word length [38,64,65,68]. Another semantic factor
that could covary with socialness, and account for some
preferential activations, is emotional valence. Indeed, one
study has shown that social-emotional stimuli elicit
stronger responses in the left dorsolateral/polar ATL than
other social words, which in turn activate the region more
strongly than stimuli lacking any social meaning [75].
However, Wang et al. [71] demonstrated at least partially
dissociable responses across left lateral ATL subregions
to the socialness, valence and abstractness dimensions
underlying word meanings. Overall, this collection of neuro-
imaging studies suggests that the socialness of a concept
makes a unique contribution to driving differential recruit-
ment of the brain regions involved in processing meaning
(for neuropsychological evidence, see [76,77]). Moreover,
they are to some extent compatible with the claim that
there is a semantic system dedicated to the representation
of social conceptual knowledge, and that this is located in
the left dorsolateral/polar ATL [36,40,41].
Of course, an alternative to the notion of a ‘social brain’
or, more precisely, that there are networks or subsystems
specialized for social processes, is that social cognition is
underpinned by a set of domain-general systems [49,51,52].
As alluded to above, from a strong version of this perspec-
tive, socialness effects at the levels of brain and behaviour
reflect variations among more general properties of stimuli
and/or task demands, rather than socialness per se.Froma
more compromising perspective, it is argued that social inter-
action could draw on an array of neurocognitive systems in
something of a unique way, but, fundamentally, those sys-
tems are built for more generalized processes (e.g. [51,52]).
For example, an alternative to domain-specific accounts of
ATL function like the social knowledge hypothesis is the
‘graded semantic hub’account proposed by Binney et al.
[38,78–80]. According to this framework, the whole ATL com-
prises a unified semantic representational space, all of which
is engaged by the encoding and retrieval of concepts, and by
concepts of any kind. At the centre of this space lies the ven-
trolateral ATL, which has a supramodal semantic function,
meaning that its engagement during semantic processing is
invariant to, for example, idiosyncratic task features, includ-
ing the modality through which concepts are accessed.
Near the edges of this space, however, there are connec-
tivity-driven gradual shifts in semantic function toward
subspecializations for processing certain types of semantic
features (for a computational exploration of this general
hypothesis, see [81]). This might include, for example, at
the dorsolateral aspects, a specialization for processing
socio-emotional semantic features [38], which could arise
from greater proximity and connectivity to the limbic
system [78,81,82] (also see below). Consistent with this
account are a series of neuroimaging studies by Binney
et al. which show that, when care is taken to ensure that
fMRI signal can be acquired from across the whole ATL, it
becomes clear that the ventrolateral ATL activates strongly
and equivalently during semantic judgements made on
social and non-social stimuli [38,39] (also see [80]). This
same ventrolateral ATL region is implicated in general
semantic processing in several neuropsychological, neurosti-
mulation, neuroimaging and electrophysiological studies
that have used a variety of verbal and nonverbal tasks/
stimuli [83–90]. Critical for this graded hub account is the
additional fact that the omni-category response of the ventro-
lateral ATL is much greater in magnitude than that of the
social-selective response of the dorsolateral ATL [38,39].
Therefore, these latter studies suggest that, at least within
the ATL, differences in the way the brain is engaged by
social and non-social concepts are small, or subtle, compared
to the similarities. This is consistent with the claim that,
rather than there being distinct systems for social and general
semantic representation, there is a single domain-general con-
ceptual system and parts of this system are dynamically and
differentially engaged by different types of meaningful
stimuli and semantic task demands (cf. the Social Semantics
framework outlined by Binney & Ramsey [52]; also [11]).
The graded hub hypothesis is an extension of the hub-
and-spoke model of semantic representation proposed by
Patterson et al. [11,12]. According to this framework, the
ATL sits at the heart of a spoked semantics architecture com-
prised of association regions involved in modality-specific
sensorimotor processing, as well as affective and linguistic
processes. The hub-and-spoke model emphasizes that seman-
tic representation arises from the conjoint action of modal
systems and an intermediary supramodal hub [11]. It offers
a reconciliation between distributed-only embodied accounts
in which concepts are dependent upon systems involved in
sensory and motor processing [91–94] and neuropsychologi-
cal and computational modelling data that point towards
the existence of a hub (e.g. [95,96]). A fuller discussion sur-
rounding the necessity of a hub is beyond the scope of this
review, and for a starting point we refer the reader to
Lambon Ralph et al.[11], as well as Meteyard et al.[44]. How-
ever, we have chosen to raise this broader hub-and-spoke
proposal here because it is a neurobiologically constrained
model that, like multimodal or multiple representation
views, acknowledges sources of semantic information
beyond sensorimotor experience, including contributions
from language, emotion and other internal states [11]. More-
over, like some of the multimodal views described in the
previous section (e.g. [13,14]), it hypothesizes that different
types of concepts (e.g. tools) can vary in their reliance on
different sources of information (e.g. object affordances and
kinematics), which will be reflected in differential engage-
ment of spoke regions [97,98] (also see [99]). This notion
lends one interpretation to neuroimaging studies that investi-
gate social concept representation and implicate brain regions
outside of the ATL. For example, two recent studies have
demonstrated an apparent selective engagement of the precu-
neus, a region associated with visual-spatial imagery [100],
during the processing of abstract social words [26,101]. This
could reflect a tendency for social concepts to draw differen-
tially upon systems that capture visual or spatial elements of
interpersonal contexts [26].
4. Part C –What is ‘socialness’?
In the sections above we have provided a brief overview
of two parallel literatures among which socialness has
begun to emerge as an important organizational principle
royalsocietypublishing.org/journal/rstb Phil. Trans. R. Soc. B 378: 20210363
4

underpinning semantic representation. In Part A, we
described property generation and feature rating studies
that have explored the attributes of abstract words and
have extracted socialness as a latent factor that distinguishes
abstract from concrete words, and even distinguishes differ-
ent types of abstract words. In Part B, we reviewed a
literature that has emerged in parallel, describing a set of neu-
roimaging studies that have probed socialness as a predictor
of differential patterns of brain activation evoked during
semantic processing. In contrast to property-generation
research, most of these neuroimaging studies approached
social concepts as an a priori discrete type of concept. This
has, thus far, been fruitful in that this brain-based evidence
points to socialness being independent of more general
semantic properties, such as abstractness, emotional valence
and other facets of single-word meaning. There is now a bur-
geoning debate regarding the relative size of the contribution
that socialness makes to semantic representation. On one
hand, it has been argued that social words are a distinct
type and, moreover, that there are specialized neural
systems underpinning social semantics. On the other hand,
socialness can be framed as one of many dimensions that
coexist to define a single representational space underlying
general semantics.
However, we assert that, while these lines of research are
both intriguing and promising, the conclusions and discus-
sions that have transpired from them are mostly premature
because the ostensive evidence has accumulated in the
absence of clear boundaries between what is social and
what is not. This is true both at the level of theory and in
the empirical measures. Without agreeing on this definition,
at least to some extent, it will not be possible to compare the-
ories and evaluate evidence in support of them. So, what is
socialness actually?
Socialness as a construct has been characterized variably
in terms of behavioural descriptiveness, and social con-
cepts have been distinguished from non-social concepts on
divergent sets of criteria. To illustrate this point, we have
collected examples in table 1 (also see [35]). Many of these
studies focused on a word’s reference to social interaction
by measuring, for example, the extent to which a word
refers to relationships between people [23,24], or how
often its referent involves interaction between people
[64,65,71,74]. By contrast, other definitions emphasize
specific aspects of social experience, such as how well a
word describes social behaviours [36], or the degree to
which word meanings relate to the relationship between
self and others [104].
Following a review of the material presented in table 1,
we suggest that there are two distinct emerging approaches
to the construct of socialness. On one hand, there are social
measures designed to capture contextual information, such
as the degree to which a word meaning evokes a set of
social circumstances [25], or whether it applies to social as
opposed to individual contexts [102]. On the other hand,
there are measures probing specific social features of word
meaning, such as the scale of interaction/number of agents
implicated [68] and the degree to which a referent has
human-like intentions, plans or goals [103]. This distinction
might reflect different representational frameworks for mean-
ing, such as those based on features/similarity and those
based on association [107,108], and it could be an important
avenue for future research into the mechanisms by which
socialness is attributed to concepts. However, the heterogen-
eity in definitions across this set of studies is striking,
highlights theoretical inconsistencies, and hinders our ability
to compare findings across studies. Certainly, it imposes
grave limitations on the conclusions that can be made
presently about socialness as a neurobiologically and/or
behaviourally relevant principle.
We argue that, to further progress theory, the field must
first establish a clearer working definition of socialness.
Further, the field would be advanced if large-scale norms of
rated socialness were available, much as they have been
made available for thousands of words for semantic variables
like concreteness [30], emotional valence [29] and others. We
believe this can best be achieved, at least initially, by adopting
a broad definition of socialness. To aid this endeavour, we
recently obtained ratings for 8388 English words by asking
participants to rate socialness according to the following
definition [4]:
the extent to which each word has social relevance by describing
or referring to a social characteristic of a person or group of
people (e.g. ‘trustworthy’), a social behaviour or interaction (e.g.
‘to fight’), a social role (e.g. ‘teacher’), a social space (e.g. ‘pub’),
a social institution (e.g. ‘hospital’) or system (e.g. ‘nation’), a
social value (e.g. ‘righteousness’) or ideology (e.g. ‘feminism’),
or any other socially-relevant concept.
To our knowledge, the resulting norms are the largest set of
openly available word socialness ratings. We believe that
employing an inclusive definition was a crucial next step
for understanding the construct of socialness. This allowed
us to test the extent to which socialness is reliably perceived
as a broad construct, and as applicable to various types of
words/parts of speech. Initial explorations of the ratings
reveal that, when broadly defined, socialness ratings have
good reliability and validity. We have begun exploring to
what extent these new socialness ratings capture aspects of
word meaning that are distinct from those measured via
other semantic variables, such as concreteness and emotional
valence (figure 1). Results showed that socialness is nega-
tively correlated with concreteness [30], but also that the
two variables share only a modest 10% of variance. Another
key observation was that words rated as high in socialness
spanned the entire concreteness dimension, from concrete
concepts like people and festival to abstract ones like democracy
and cooperate. As might be expected [23], socialness was posi-
tively associated with valence extremity (the absolute
difference between the valence rating and the neutral point
of the original valence scale [29]), but it shared only 4.8%
of variance. We provide more extensive description and
exploration of the socialness norms in Diveica et al. [4] but,
in summary, our preliminary analyses indicated that this
socialness measure captures a distinct psycholinguistic
construct.
5. Conclusion and future directions
The research we have reviewed here suggests that socialness,
broadly construed, is a dimension of word meaning that can
be distinguished from other dimensions such as concreteness
and valence. Moreover, there is some evidence that socialness
is reflected within the organization of neural systems that
support semantic processing. It remains to be seen whether
this is indicative of social words being a distinct type, or
whether socialness is just one of many dimensions that
royalsocietypublishing.org/journal/rstb Phil. Trans. R. Soc. B 378: 20210363
5

Table 1. Definitions used to measure socially relevant semantic constructs in previous studies.
publication name of construct type definition
Arioli et al.[102] sociality dimension how much a word inherently refers to information concerning social as
opposed to individual contexts
Binder et al.[103] social dimension the degree to which one thinks of a thing as an activity or event that involves
an interaction between people
human dimension the degree to which one thinks of a thing as having human or human-like
intentions, plans or goals
communication dimension the degree to which one thinks of a thing as a thing or action that people
use to communicate
self dimension the degree to which one thinks of a thing as related to your own view of
yourself, a part of your self-image
Catricalà et al.[70] social dimension how much a word is linked to a social situation or to an interaction among
people, both in terms of inclusion and exclusion
Crutch et al.[104] social interaction dimension the degree to which concepts relate to the relationships between self
and others
Diveica, et al. [4] socialness dimension the degree to which a word’s meaning has social relevance by describing or
referring to a social characteristic of a person or group of people, a social
behaviour or interaction, a social role, a social space, a social institution or
system, a social value or ideology, or any other socially relevant concept
Harpaintner et al. [22] social constellation category a feature or a situation that describes the coexistence of different persons or
which implies an interaction between at least two different persons
Lin et al.[68] sociality dimension the number of people involved in an event to which a verb refers
biological motion dimension the extent to which the meaning of a verb brings to mind biological motion
Lin et al.[64,65,74],
Wang et al. [71]
sociality dimension how often the meaning of a word/the use of an object involves an interaction
between people
Mellem et al.[75] social content category referring to people either by a proper name or the name of an
occupation/title
Roversi et al. [105] institutional objects category an artefact that performs its function via the collective acceptance
displayed by a given community (status function) and not in virtue of its
physical features
social objects category an entity that presupposes the existence of at least two agents engaged in
some form of common activity and that does not have a clear status
function attached to it
Troche et al. [23,24] social interaction dimension the degree to which the word relates to relationships between people
morality dimension the degree to which the word relates to morality, rules, or anything that
governs one’s behaviour
Vargas & Just [26] social content dimension the degree to which the concept involves social interaction or self-perception
as affected by social interaction
Villani et al.[25] social metacognition dimension how much others were needed to understand the meaning of the word
social valence dimension the degree to which the concept evokes social circumstances
Villani et al.[106] pure institutional objects category entities constituted by formalized rules in a social framework
meta-institutional objects category concepts that are necessary to define the content of institutions but are not
defined by those institutions
Zahn et al.[36] behaviour descriptiveness dimension how well a word describes a detailed specific set of social behaviours
of persons
social category breadth dimension how many different kinds of social behaviours of persons a word can apply to
Zhang et al.[66] social semantic richness dimension the extent to which the word/sentence/narrative is related to interactions
between people
royalsocietypublishing.org/journal/rstb Phil. Trans. R. Soc. B 378: 20210363
6

define a unified domain-general semantic space. At present,
there remain two key shortcomings in this exciting area of
research. First, researchers need to begin agreeing on terms
and definitions of ‘socialness’so that we are better able to
compare theories and evaluate evidence in support of them
[109]. Second, there is very little research on the behavioural
consequences of socialness and behavioural relevance is, of
course, a gold standard for psychological theory. In terms
of refining models of semantic representation, we believe
that there are four key avenues for future research, and they
have been made possible by the availability of the new social-
ness ratings [4] described above. We will outline these
research questions in the paragraphs below.
First, there are testable predictions that can be derived
from WAT and other multiple representation theories. For
instance, WAT proposes that social experience is key to learn-
ing and representing abstract concepts [14]. In line with this
proposal, one could predict that (i) socialness facilitates the
acquisition of abstract words, (ii) socialness contributes to
the acquisition of abstract words more than to that of concrete
words and (iii) abstract words are associated with more social
content than are concrete words. In addition, WAT proposes a
close link between linguistic and social experience. Consistent
with this, Villani et al.[25] found that, in a sample of abstract
words, the more the words evoked social circumstances, the
more they relied on auditory experiences, and on mouth
motor system activation. These relationships could be further
evaluated to understand how linguistic and social infor-
mation jointly support acquisition and representation of
abstract words.
Second, Diveica et al. [4] characterized socialness in a
broad and inclusive way and found this to be a useful and
meaningful starting point. However, subsequent research is
needed to more thoroughly explore the nature of the infor-
mation captured by the socialness dimension and to
evaluate whether there are important distinctions that it
does not capture. In future research, it will be helpful to con-
sider narrower definitions to explore whether there are
clusters or subtypes of social words. Moreover, it remains
to be seen what aspects of the social experience, such as
those measured by the more specific socially relevant dimen-
sions listed in table 1, are most related to lexical-semantic
processing in terms of both behaviour and brain. It is possible
that there are sub-types of social words that rely on different
kinds of information. To some degree, this could mirror the
more general concrete-abstract distinction, possibly in terms
of how concepts rely differentially upon qualitatively differ-
ent representational frameworks, such as those based on
features/similarity and those based on association (cf. the
proposal outlined by Crutch and Warrington [107,108]). For
example, Roversi et al. [105] investigated the properties
associated with two potential sub-types of social concepts.
They found that ‘social objects’(defined in table 1), such as
choir, elicited mainly contextual situations (e.g. concert),
while institutional artefacts, such as marriage, evoked a
higher proportion of normative relations (e.g. commitment).
Further, the abstract-concrete distinction was more marked
for social objects compared to institutional artefacts. Social
objects that are concrete were associated with thematic/situa-
tional relations, while those that are abstract elicited more
mental associations. In a related study, Villani et al.[106]pro-
posed a further distinction between pure institutional
concepts (e.g. marriage) that relied more on exteroceptive
information, and meta-institutional concepts (e.g. duty) for
which interoceptive, affective and metacognitive information
was more important. Future research that applies a data-
driven approach across a large sample of abstract and con-
crete words will shed light on more specific socialness
constructs and the way in which individual social word
meanings potentially cluster together into sub-types.
Third, there are several implications for neuroimaging
research into the representation of social concepts, and we
have some recommendations. Now that large-scale socialness
ratings are available and their independence from measures
of concreteness and emotional valence has been more
firmly established [4], researchers are better positioned to
comprehensively disentangle the neural correlates of social-
ness from other semantic variables. Indeed, right across the
line of neuroimaging research reviewed in Part B, there is a
need for greater integration of the kind of property gener-
ation, feature rating and behavioural research we reviewed
in Part A. It will be instrumental for driving the next set of
key questions, including those regarding the neural correlates
of different types of concepts, and a putative privileged status
afforded by socialness [52]. At present, there is a lack of clear
evidence in favour of an absolute boundary between social
concepts and other types of concepts, and this suggests that
there is going to be considerable overlap in the systems that
represent them [38,39]. In this case, it will be important to
use experimental designs and analytical techniques that
allow for detecting more graded distinctions. To date, social-
ness has only been explored using univariate, magnitude-
based approaches, whereas information-based approaches,
including multivariate pattern analysis and repetition sup-
pression paradigms, will be essential, particularly for
understanding overlapping activation, which could reflect
either shared processes or tightly yet separately packed cog-
nitive functions that only dissociate when investigated at
higher spatial resolutions [110,111]. Moreover, a key
5
4
concreteness
3
2
1
1234
socialness
valence extremity
0123
567
Figure 1. The relationship between socialness ratings [4] and concreteness
ratings [30] for 8388 English words is illustrated and highlighted by the
loess line. The colour of the dots represents valence [29] extremity—the
darker the colour, the more valenced the word. The density distributions
of the socialness and concreteness dimensions are plotted on the top and
right of the graph, respectively. The graph shows that words with high
mean socialness ratings span the entire concreteness dimension, and that
the socialness measure captures information distinct from valence.
royalsocietypublishing.org/journal/rstb Phil. Trans. R. Soc. B 378: 20210363
7

methodological determinant for obtaining a complete picture
of the neural basis of social concepts will be the use of neuroi-
maging techniques that maximize the signal obtained from
across the entirety of key brain regions. This includes the
anterior temporal lobe, of which some subregions are invis-
ible to standard fMRI [38,80,112].
Fourth, it is worth noting that concepts are not static and
that their representation depends on ongoing task contexts as
well as prior experience [113,114]. For example, it has been
shown that concepts are to some degree influenced by culture
and the language spoken [115,116]. Given that cultural
environments are intrinsically linked to our social experi-
ences, social concepts might be particularly susceptible to
cultural influences. Moreover, a variety of socially relevant
characteristics (e.g. race, gender) impact our social experi-
ences, which can consequently lead to between-individual
variability in the representation of social concepts. In line
with this, Mazzuca et al.[117] showed that the features par-
ticipants associated most strongly with the social concept
gender depended on their gender identity and sexual orien-
tation. This potential variability could be investigated in
future research and might manifest in various ways. For
instance, some abstract words, including those high in
social content, might place greater demands upon cognitive
control processes because their exact meaning is dependent
on context. This might be reflected in differential engagement
of regions associated with controlled semantic selection and
retrieval, such as the left inferior frontal gyrus (IFG; see
[118–120] for related discussions). Consistent with this,
some individual neuroimaging studies reported greater acti-
vation of semantic control regions (the IFG) during the
processing of social, as compared to non-social words/sen-
tences [38,75] (also see [121]). In addition, there is some
limited behavioural evidence that implicit semantic proces-
sing of social words compared to non-social words slows
reaction times in a Stroop task in adults [102] and in a selec-
tive attention task in children [122], indicating greater
demand for cognitive control. However, more research is
needed to understand what task contexts and concept
features might drive an increased need for regulatory
mechanisms when processing social concepts.
In summary, in the present paper, we have outlined the
ways in which two different literatures have explored the
idea that social concepts might be a special type and have
offered suggestions for integrating and advancing these
research efforts. Further, we presented initial psycholinguistic
explorations of a new and openly available set of socialness
ratings for over 8000 words (see [4] for a detailed descrip-
tion). These suggest that socialness is indeed a distinct
aspect of word meaning and one that should be incorporated
in theories of semantic representation. Social words, like
manner,gentleman and refuse, convey information about our
relationships with people and inform our understanding of
their actions. Socialness gives words salience and gives
meaning to the interactions and events that make up sources
like Pride and Prejudice, and that occur in the personal and
interpersonal complexities of our everyday lives.
Data accessibility. The data illustrated in figure 1 are available via the
Open Science Framework at https://osf.io/2dqnj/.
Authors’contributions. P.M.P.: conceptualization, funding acquisition,
investigation, methodology, project administration, resources, super-
vision, writing—original draft, writing—review and editing; V.D.:
conceptualization, data curation, formal analysis, funding acqui-
sition, investigation, methodology, writing—original draft,
writing—review and editing; R.J.B.: conceptualization, funding
acquisition, investigation, methodology, project administration,
resources, supervision, writing—original draft, writing—review and
editing.
All authors gave final approval for publication and agreed to be
held accountable for the work performed therein.
Conflict of interest declaration. We declare we have no competing interests.
Funding. This work was supported by the Economic and Social
Research Council (ESRC) Wales Doctoral Training Partnership in
the form of a PhD studentship (grant no. ES/P00069X/1), a joint
award from UK Research and Innovation (UKRI) and Mitacs under
the UK-Canada Globalink Doctoral Exchange Scheme (grant no.
NE/T014180/1) (both awarded to V.D. and R.J.B.; PhD student:
V.D.), and a Social Sciences and Humanities Research Council
(SSHRC) of Canada Insight Grant (awarded to P.M.P.).
Acknowledgements. All authors contributed equally to the writing of this
article.
Endnotes
1
There is some debate about whether the conceptual system is separ-
ate from the lexical-semantic system. We take the position that the
conceptual system is not separate from the meaning accessed
during language processing.
References
1. Austen J. 2019 Pride and prejudice, 3rd edn. Oxford,
UK: Oxford University Press.
2. Tomasello M. 2009 Why we cooperate. Cambridge,
MA: MIT Press.
3. Tomasello M. 2020 The adaptive origins of uniquely
human sociality. Phil. Trans. R. Soc. B 375,
20190493. (doi:10.1098/RSTB.2019.0493)
4. Diveica V, Pexman PM, Binney RJ. In press.
Quantifying social semantics: an inclusive definition
of socialness and ratings for 8388 English words.
Behav. Res. Methods (doi:10.3758/s13428-022-
01810-x)
5. Nook EC, Stavish CM, Sasse SF, Lambert HK, Mair P,
McLaughlin KA, Somerville LH. 2020 Charting the
development of emotion comprehension and
abstraction from childhood to adulthood using
observer-rated and linguistic measures. Emotion 20,
773–792. (doi:10.1037/EMO0000609)
6. Binney RJ, Zuckerman B, Reilly J. 2016 A
neuropsychological perspective on abstract word
representation: from theory to treatment of acquired
language disorders. Curr. Neurol. Neurosci. Rep. 16,
1–8. (doi:10.1007/S11910-016-0683-0/FIGURES/1)
7. Quillian MR. 1969 The teachable language
comprehender: a simulation program and theory of
language. Commun. Assoc. Comput. Mach. 12,
459–476. (doi:10.1145/363196.363214)
8. Pylyshyn ZW. 1984 Computation and cognition:
toward a foundation for cognitive science.
Cambridge, MA: MIT Press.
9. Glenberg AM. 2015 Few believe the world is flat:
how embodiment is changing the scientific
understanding of cognition. Can. J. Exp. Psychol. 69,
165–171. (doi:10.1037/CEP0000056)
10. Glenberg AM, Gallese V. 2012 Action-based
language: a theory of language acquisition,
comprehension, and production. Cortex
48, 905–922. (doi:10.1016/J.CORTEX.2011.
04.010)
11. Lambon Ralph MA, Jefferies E, Patterson K, Rogers
TT. 2017 The neural and computational bases of
semantic cognition. Nat. Rev. Neurosci. 18,42–55.
(doi:10.1038/nrn.2016.150)
12. Patterson K, Nestor PJ, Rogers TT. 2007 Where do
you know what you know? The representation of
royalsocietypublishing.org/journal/rstb Phil. Trans. R. Soc. B 378: 20210363
8

semantic knowledge in the human brain. Nat. Rev.
Neurosci. 8, 976–987. (doi:10.1038/nrn2277)
13. Barsalou LW. 2008 Grounded cognition. Annu. Rev.
Psychol. 59, 617–645. (doi:10.1146/ANNUREV.
PSYCH.59.103006.093639)
14. Borghi AM, Barca L, Binkofski F, Castelfranchi C,
Pezzulo G, Tummolini L. 2019 Words as social tools:
language, sociality and inner grounding in abstract
concepts. Phys. Life Rev. 29, 120–153. (doi:10.1016/
J.PLREV.2018.12.001)
15. Harnad S. 1990 The symbol grounding problem.
Physica D 42, 335–346. (doi:10.1016/0167-
2789(90)90087-6)
16. Searle JR. 1980 Minds, brains, and programs. Behav.
Brain Sci. 3, 417–424. (doi:10.1017/
S0140525X00005756)
17. Barsalou LW. 2020 Challenges and opportunities for
grounding cognition. J. Cogn. 3,1–24. (doi:10.
5334/JOC.116)
18. Barsalou LW, Wiemer-Hastings K. 2005 Situating
abstract concepts. In Grounding cognition: the role of
perception and action in memory, language, and
thinking (eds D Pecher, RA Zwaan), pp. 129–164.
Cambridge, UK: Cambridge University Press.
19. Wiemer-Hastings KK, Xu X. 2005 Content differences
for abstract and concrete concepts. Cogn. Sci. 29,
719–736. (doi:10.1207/S15516709COG0000_33)
20. Borghi AM, Binkofski F. 2014 Words as social tools:
an embodied view on abstract concepts. New York,
NY: Springer.
21. Zdrazilova L, Pexman PM. 2013 Grasping the
invisible: semantic processing of abstract words.
Psychon. Bull. Rev. 20, 1312–1318. (doi:10.3758/
s13423-013-0452-x)
22. Harpaintner M, Trumpp NM, Kiefer M. 2018 The
semantic content of abstract concepts: a property
listing study of 296 abstract words. Front. Psychol.
9, 1748. (doi:10.3389/fpsyg.2018.01748)
23. Troche J, Crutch S, Reilly J. 2014 Clustering,
hierarchical organization, and the topography of
abstract and concrete nouns. Front. Psychol. 5, 360.
(doi:10.3389/FPSYG.2014.00360)
24. Troche J, Crutch SJ, Reilly J. 2017 Defining a
conceptual topography of word concreteness:
clustering properties of emotion, sensation,
and magnitude among 750 english words.
Front. Psychol. 8, 1787. (doi:10.3389/fpsyg.
2017.01787)
25. Villani C, Lugli L, Liuzza M, Borghi AM. 2019
Varieties of abstract concepts and their multiple
dimensions. Lang. Cogn. 11, 403–430. (doi:10.
1017/LANGCOG.2019.23)
26. Vargas R, Just MA. 2020 Neural representations of
abstract concepts: identifying underlying
neurosemantic dimensions. Cereb. Cortex 30,
2157–2166. (doi:10.1093/CERCOR/BHZ229)
27. Huth AG, de Heer WA, Griffiths TL, Theunissen FE,
Gallant JL. 2016 Natural speech reveals the semantic
maps that tile human cerebral cortex. Nature 532,
453–458. (doi:10.1038/nature17637)
28. Huth AG, Nishimoto S, Vu AT, Gallant JL. 2012 A
continuous semantic space describes the
representation of thousands of object and action
categories across the human brain. Neuron 76,
1210–1224. (doi:10.1016/J.NEURON.2012.10.014)
29. Warriner AB, Kuperman V, Brysbaert M. 2013 Norms
of valence, arousal, and dominance for 13,915
English lemmas. Behav. Res. Methods 45,
1191–1207. (doi:10.3758/s13428-012-0314-x)
30. Brysbaert M, Warriner AB, Kuperman V. 2014
Concreteness ratings for 40 thousand generally
known English word lemmas. Behav. Res. Methods
46, 904–911. (doi:10.3758/s13428-013-0403-5)
31. Newcombe PI, Campbell C, Siakaluk PD, Pexman
PM. 2012 Effects of emotional and sensorimotor
knowledge in semantic processing of concrete and
abstract nouns. Front. Hum. Neurosci. 6, 275.
(doi:10.3389/fnhum.2012.00275)
32. Kousta ST, Vigliocco G, Vinson DP, Andrews M, Del
Campo E. 2011 The representation of abstract
words: why emotion matters. J. Exp. Psychol. Gen.
140,14–34. (doi:10.1037/a0021446)
33. Pexman PM, Yap MJ. 2018 Individual differences in
semantic processing: insights from the Calgary
Semantic Decision Project. J. Exp. Psychol. Learn.
Mem. Cogn. 44, 1091–1112. (doi:10.1037/
XLM0000499)
34. Muraki EJ, Sidhu DM, Pexman PM. 2020
Heterogenous abstract concepts: is ‘ponder’different
from ‘dissolve’?Psychol. Res. (doi:10.1007/s00426-
020-01398-x)
35. Conca F, Borsa VM, Cappa SF, Catricalà E. 2021 The
multidimensionality of abstract concepts: a
systematic review. Neurosci. Biobehav. Rev. 127,
474–491. (doi:10.1016/J.NEUBIOREV.2021.05.004)
36. Zahn R, Moll J, Krueger F, Huey ED, Garrido G,
Grafman J. 2007 Social concepts are represented in
the superior anterior temporal cortex. Proc. Natl
Acad. Sci. USA 104, 6430–6435. (doi:10.1073/pnas.
0607061104)
37. Ross LA, Olson IR. 2010 Social cognition and
the anterior temporal lobes. Neuroimage 49,
3452–3462. (doi:10.1016/j.neuroimage.2009.
11.012)
38. Binney RJ, Hoffman P, Lambon Ralph MA. 2016
Mapping the multiple graded contributions of the
anterior temporal lobe representational hub to
abstract and social concepts: evidence from
distortion-corrected fMRI. Cereb. Cortex 26,
4227–4241. (doi:10.1093/cercor/bhw260)
39. Rice GE, Hoffman P, Binney RJ, Lambon Ralph MA.
2018 Concrete versus abstract forms of social
concept: an fMRI comparison of knowledge about
people versus social terms. Phil. Trans. R. Soc. B
373, 20170136. (doi:10.1098/rstb.2017.0136)
40. Simmons WK, Reddish M, Bellgowan PSF, Martin A.
2010 The selectivity and functional connectivity of
the anterior temporal lobes. Cereb. Cortex 20,
813–825. (doi:10.1093/CERCOR/BHP149)
41. Olson IR, Mccoy D, Klobusicky E, Ross LA. 2013
Social cognition and the anterior temporal lobes: a
review and theoretical framework. Soc. Cogn. Affect.
Neurosci. 8, 123–133. (doi:10.1093/scan/nss119)
42. Shallice T. 1987 Impairments of semantic
processing: multiple dissociations. In The cognitive
neuropsychology of language (eds M Coltheart, G
Sartoni, R Job), pp. 111–127. Hove, UK: Lawrence
Erlbaum Associates, Inc.
43. Warrington EK, McCarthy RA. 1994 Multiple
meaning systems in the brain: a case for visual
semantics. Neuropsychologia 32, 1465–1473.
(doi:10.1016/0028-3932(94)90118-X)
44. Meteyard L, Cuadrado SR, Bahrami B, Vigliocco G.
2012 Coming of age: a review of embodiment and
the neuroscience of semantics. Cortex 48, 788–804.
(doi:10.1016/J.CORTEX.2010.11.002)
45. Brothers L. 1990 The social brain: a project for
integrating primate behavior and neurophysiology
in a new domain. Concepts Neurosci. 1,27–51.
46. Dunbar RIM. 1998 The social brain hypothesis. Evol.
Anthropol. 6, 178–190. (doi:10.1002/(SICI)1520-
6505(1998)6:5)
47. Braunsdorf M, Blazquez Freches G, Roumazeilles L,
Eichert N, Schurz M, Uithol S, Bryant KL, Mars RB.
2021 Does the temporal cortex make us human? A
review of structural and functional diversity of the
primate temporal lobe. Neurosci. Biobehav. Rev.
131, 400–410. (doi:10.1016/J.NEUBIOREV.2021.08.
032)
48. Frith CD. 2007 The social brain? Phil. Trans. R. Soc. B
362, 671–678. (doi:10.1098/RSTB.2006.2003)
49. Amodio DM. 2019 Social cognition 2.0: an
interactive memory systems account. Trends Cogn.
Sci. 23,21–33. (doi:10.1016/j.tics.2018.10.002)
50. Kilner JM. 2011 More than one pathway to action
understanding. Trends Cogn. Sci. 15, 352–357.
(doi:10.1016/J.TICS.2011.06.005)
51. Spunt RP, Adolphs R. 2017 A new look at
domain specificity: insights from social
neuroscience. Nat. Rev. Neurosci. 18, 559–567.
(doi:10.1038/nrn.2017.76)
52. Binney RJ, Ramsey R. 2020 Social semantics:
the role of conceptual knowledge and cognitive
control in a neurobiological model of the social
brain. Neurosci. Biobehav. Rev. 112,28–38. (doi:10.
1016/j.neubiorev.2020.01.030)
53. Adolphs R. 2010 Conceptual challenges and
directions for social neuroscience. Neuron 65,
752–767. (doi:10.1016/j.neuron.2010.03.006)
54. Kanwisher N. 2010 Functional specificity in the
human brain: a window into the functional
architecture of the mind. Proc. Natl Acad. Sci. USA
107, 11 163–11 170. (doi:10.1073/PNAS.
1005062107/-/DCSUPPLEMENTAL)
55. Pitcher D, Ungerleider LG. 2021 Evidence for a third
visual pathway specialized for social perception.
Trends Cogn. Sci. 25, 100–110. (doi:10.1016/J.TICS.
2020.11.006)
56. Kanwisher N, Yovel G. 2006 The fusiform face area:
a cortical region specialized for the perception of
faces. Phil. Trans. R. Soc. B 361, 2109–2128.
(doi:10.1098/RSTB.2006.1934)
57. Downing PE, Jiang Y, Shuman M, Kanwisher N.
2001 A cortical area selective for visual processing
of the human body. Science 293, 2470–2473.
(doi:10.1126/SCIENCE.1063414)
58. Isik L, Koldewyn K, Beeler D, Kanwisher N. 2017
Perceiving social interactions in the posterior
superior temporal sulcus. Proc. Natl Acad. Sci. USA
royalsocietypublishing.org/journal/rstb Phil. Trans. R. Soc. B 378: 20210363
9

114, E9145–E9152. (doi:10.1073/PNAS.
1714471114/-/DCSUPPLEMENTAL)
59. Adolphs R. 2003 Cognitive neuroscience of human
social behaviour. Nat. Rev. Neurosci. 43, 165–178.
(doi:10.1038/nrn1056)
60. Mars RB, Neubert F-X, Noonan MP, Sallet J, Toni I,
Rushworth MFS. 2012 On the relationship between
the ‘default mode network’and the ‘social brain’.
Front. Hum. Neurosci. 6,1–9. (doi:10.3389/fnhum.
2012.00189)
61. Rushworth MFS, Mars RB, Sallet J. 2013 Are there
specialized circuits for social cognition and are they
unique to humans? Curr. Opin. Neurobiol. 23,
436–442. (doi:10.1016/J.CONB.2012.11.013)
62. Rice GE, Lambon Ralph MA, Hoffman P. 2015 The
roles of left versus right anterior temporal lobes in
conceptual knowledge: an ALE meta-analysis of 97
functional neuroimaging studies. Cereb. Cortex 25,
4374–4391. (doi:10.1093/cercor/bhv024)
63. Arioli M, Gianelli C, Canessa N. 2021 Neural
representation of social concepts: a coordinate-
based meta-analysis of fMRI studies. Brain Imaging
Behav. 15, 1912–1921. (doi:10.1007/S11682-020-
00384-6)
64. Lin N, Wang X, Xu Y, Wang X, Hua H, Zhao Y, Li X.
2018 Fine subdivisions of the semantic network
supporting social and sensory–motor semantic
processing. Cereb. Cortex 28, 2699–2710. (doi:10.
1093/CERCOR/BHX148)
65. Lin N, Xu Y, Wang X, Yang H, Du M, Hua H, Li X.
2019 Coin, telephone, and handcuffs: neural
correlates of social knowledge of inanimate objects.
Neuropsychologia 133, 107187. (doi:10.1016/J.
NEUROPSYCHOLOGIA.2019.107187)
66. Zhang G, Xu Y, Zhang M, Wang S, Lin N. 2021 The
brain network in support of social semantic
accumulation. Soc. Cogn. Affect. Neurosci. 16,
393–405. (doi:10.1093/SCAN/NSAB003)
67. Gainotti G. 2015 Is the difference between right and
left ATLs due to the distinction between general
and social cognition or between verbal and non-
verbal representations? Neurosci. Biobehav. Rev. 51,
296–312. (doi:10.1016/J.NEUBIOREV.2015.02.004)
68. Lin N, Bi Y, Zhao Y, Luo C, Li X. 2015 The theory-of-
mind network in support of action verb
comprehension: evidence from an fMRI study. Brain
Lang. 141,1–10. (doi:10.1016/J.BANDL.2014.11.
004)
69. Pobric G, Lambon Ralph M, Zahn R. 2016
Hemispheric specialization within the superior
anterior temporal cortex for social and nonsocial
concepts. J. Cogn. Neurosci. 28, 351–360. (doi:10.
1162/JOCN_A_00902)
70. Catricalà E, Conca F, Fertonani A, Miniussi C, Cappa
SF. 2020 State-dependent TMS reveals the
differential contribution of ATL and IPS to the
representation of abstract concepts related to social
and quantity knowledge. Cortex 123,30–41.
(doi:10.1016/J.CORTEX.2019.09.018)
71. Wang X, Wang B, Bi Y. 2019 Close yet independent:
dissociation of social from valence and abstract
semantic dimensions in the left anterior temporal
lobe. Hum. Brain Mapp. 40, 4759–4776. (doi:10.
1002/HBM.24735)
72. Hoffman P, Lambon Ralph MA, Rogers TT. 2013
Semantic diversity: a measure of semantic
ambiguity based on variability in the contextual
usage of words. Behav. Res. Methods 45, 718–730.
(doi:10.3758/s13428-012-0278-x)
73. Lin N, Yang X, Li J, Wang S, Hua H, Ma Y, Li X. 2018
Neural correlates of three cognitive processes
involved in theory of mind and discourse
comprehension. Cogn. Affect. Behav. Neurosci. 18,
273–283. (doi:10.3758/s13415-018-0568-6)
74. Lin N, Xu Y, Yang H, Zhang G, Zhang M, Wang S,
Hua H, Li X. 2020 Dissociating the neural correlates
of the sociality and plausibility effects in simple
conceptual combination. Brain Struct. Funct. 225,
995–1008. (doi:10.1007/S00429-020-02052-3)
75. Mellem MS, Jasmin KM, Peng C, Martin A. 2016
Sentence processing in anterior superior temporal
cortex shows a social-emotional bias.
Neuropsychologia 89, 217–224. (doi:10.1016/J.
NEUROPSYCHOLOGIA.2016.06.019)
76. Catricalà E, Della Rosa PA, Plebani V, Vigliocco G,
Cappa SF. 2014 Abstract and concrete categories?
Evidences from neurodegenerative diseases.
Neuropsychologia 64, 271–281. (doi:10.1016/J.
NEUROPSYCHOLOGIA.2014.09.041)
77. Catricalà E et al. 2021 Different types of abstract
concepts: evidence from two neurodegenerative
patients. Neurocase 27, 270–280. (doi:10.1080/
13554794.2021.1931345)
78. Binney RJ, Parker GJM, Lambon Ralph MA. 2012
Convergent connectivity and graded specialization in
the rostral human temporal lobe as revealed by
diffusion-weighted imaging probabilistic
tractography. J. Cogn. Neurosci. 24, 1998–2014.
(doi:10.1162/jocn_a_00263)
79. Rice GE, Hoffman P, Lambon Ralph MA. 2015
Graded specialization within and between the
anterior temporal lobes. Ann. N. Y. Acad. Sci. 1359,
84–97. (doi:10.1111/nyas.12951)
80. Balgova E, Diveica V, Walbrin J, Binney RJ. 2021 The
role of the ventrolateral anterior temporal lobes in
social cognition. bioRxiv.(doi:10.1101/2021.09.10.
459496)
81. Plaut DC. 2002 Graded modality-specific
specialisation in semantics: a computational account
of optic aphasia. Cogn. Neuropsychol. 19, 603–639.
(doi:10.1080/02643290244000112)
82. Jung JY, Cloutman LL, Binney RJ, Lambon Ralph
MA. 2017 The structural connectivity of higher order
association cortices reflects human functional brain
networks. Cortex 97, 221–239. (doi:10.1016/j.
cortex.2016.08.011)
83. Marinkovic K, Dhond RP, Dale AM, Glessner M, Carr
V, Halgren E. 2003 Spatiotemporal dynamics of
modality-specific and supramodal word processing.
Neuron 38, 487–497. (doi:10.1016/S0896-
6273(03)00197-1)
84. Chan AM, Baker JM, Eskandar E, Schomer D, Ulbert
I, Marinkovic K, Cash SS, Halgren E. 2011 First-pass
selectivity for semantic categories in human
anteroventral temporal lobe. J. Neurosci. 31,18
119–18 129. (doi:10.1523/JNEUROSCI.3122-11.2011)
85. Pobric G, Jefferies E, Lambon Ralph MA. 2007
Anterior temporal lobes mediate semantic
representation: mimicking semantic dementia by
using rTMS in normal participants. Proc. Natl Acad.
Sci. USA 104, 20 137–20 141.
86. Pobric G, Jefferies E, Lambon Ralph MA. 2010
Amodal semantic representations depend on both
anterior temporal lobes: evidence from repetitive
transcranial magnetic stimulation. Neuropsychologia
48, 1336–1342. (doi:10.1016/J.
NEUROPSYCHOLOGIA.2009.12.036)
87. Binney RJ, Embleton KV, Jefferies E, Parker GJM,
Lambon Ralph MA. 2010 The ventral and
inferolateral aspects of the anterior temporal lobe
are crucial in semantic memory: evidence from a
novel direct comparison of distortion-corrected fMRI,
rTMS, and semantic dementia. Cereb. Cortex 20,
2728–2738. (doi:10.1093/cercor/bhq019)
88. Binney RJ, Lambon Ralph MA. 2015 Using a
combination of fMRI and anterior temporal lobe
rTMS to measure intrinsic and induced activation
changes across the semantic cognition network.
Neuropsychologia 76, 170–181. (doi:10.1016/J.
NEUROPSYCHOLOGIA.2014.11.009)
89. Shimotake A et al. 2015 Direct exploration of the
role of the ventral anterior temporal lobe in
semantic memory: cortical stimulation and local
field potential evidence from subdural grid
electrodes. Cereb. Cortex 25, 3802–3817. (doi:10.
1093/CERCOR/BHU262)
90. Visser M, Lambon Ralph MA. 2011 Differential
contributions of bilateral ventral anterior temporal
lobe and left anterior superior temporal gyrus to
semantic processes. J. Cogn. Neurosci. 23,
3121–3131. (doi:10.1162/JOCN_A_00007)
91. Gallese V, Lakoff G. 2007 The brain’s concepts: the
role of the sensory-motor system in conceptual
knowledge. Cogn. Neuropsychol. 22, 455–479.
(doi:10.1080/02643290442000310)
92. Pulvermüller F. 2001 Brain reflections of words and
their meaning. Trends Cogn. Sci. 5, 517–524.
(doi:10.1016/S1364-6613(00)01803-9)
93. Vigliocco G, Vinson DP, Lewis W, Garrett MF. 2004
Representing the meanings of object and action
words: the featural and unitary semantic space
hypothesis. Cogn. Psychol. 48, 422–488. (doi:10.
1016/J.COGPSYCH.2003.09.001)
94. Barsalou LW. 2009 Simulation, situated
conceptualization, and prediction. Phil. Trans. R. Soc.
B364, 1281–1289. (doi:10.1098/RSTB.2008.0319)
95. Rogers TT, Lambon Ralph MA, Garrard P, Bozeat S,
McClelland JL, Hodges JR, Patterson K. 2004
Structure and deterioration of semantic memory: a
neuropsychological and computational investigation.
Psychol. Rev. 111, 205–235. (doi:10.1037/0033-
295X.111.1.205)
96. Lambon Ralph MA, Sage K, Jones RW, Mayberry EJ.
2010 Coherent concepts are computed in the
anterior temporal lobes. Proc. Natl Acad. Sci. USA
107, 2717–2722. (doi:10.1073/pnas.0907307107)
royalsocietypublishing.org/journal/rstb Phil. Trans. R. Soc. B 378: 20210363
10

97. Pobric G, Jefferies E, Lambon Ralph MA. 2010
Induction of semantic impairments using rTMS:
evidence for the hub-and-spoke semantic theory.
Behav. Neurol. 23, 217–219. (doi:10.3233/BEN-
2010-0299)
98. Chiou R, Lambon Ralph MA. 2019 Unveiling the
dynamic interplay between the hub- and spoke-
components of the brain’s semantic system and its
impact on human behaviour. Neuroimage 199,
114–126. (doi:10.1016/J.NEUROIMAGE.2019.05.059)
99. Wilson-Mendenhall CD, Simmons WK, Martin A,
Barsalou LW. 2013 Contextual processing of abstract
concepts reveals neural representations of
nonlinguistic semantic content. J. Cogn. Neurosci.
25, 920–935. (doi:10.1162/JOCN_A_00361)
100. Bzdok D, Heeger A, Langner R, Laird AR, Fox PT,
Palomero-Gallagher N, Vogt BA, Zilles K, Eickhoff
SB. 2015 Subspecialization in the human posterior
medial cortex. Neuroimage 106,55–71. (doi:10.
1016/J.NEUROIMAGE.2014.11.009)
101. Leshinskaya A, Contreras JM, Caramazza A, Mitchell
JP. 2017 Neural representations of belief concepts: a
representational similarity approach to social
semantics. Cereb. Cortex 27, 344–357. (doi:10.1093/
CERCOR/BHW401)
102. Arioli M, Basso G, Poggi P, Canessa N. 2021 Fronto-
temporal brain activity and connectivity track
implicit attention to positive and negative social
words in a novel socio-emotional Stroop task.
Neuroimage 226, 117580. (doi:10.1016/j.
neuroimage.2020.117580)
103. Binder JR, Conant LL, Humphries CJ, Fernandino L,
Simons SB, Aguilar M, Desai RH. 2016 Toward a
brain-based componential semantic representation.
Cogn. Neuropsychol. 33, 130–174. (doi:10.1080/
02643294.2016.1147426)
104. Crutch SJ, Williams P, Ridgway GR, Borgenicht L.
2012 The role of polarity in antonym and synonym
conceptual knowledge: evidence from stroke
aphasia and multidimensional ratings of abstract
words. Neuropsychologia 50, 2636–2644. (doi:10.
1016/J.NEUROPSYCHOLOGIA.2012.07.015)
105. Roversi C, Borghi AM, Tummolini L. 2013 A
marriage is an artefact and not a walk that we take
together: an experimental study on the
categorization of artefacts. Rev. Philos. Psychol. 4,
527–542. (doi:10.1007/S13164-013-0150-7)
106. Villani C, D’Ascenzo S, Borghi AM, Roversi C, Benassi
M, Lugli L. 2021 Is justice grounded? How expertise
shapes conceptual representation of institutional
concepts. Psychol. Res. 1,1–17. (doi:10.1007/
S00426-021-01492-8)
107. Crutch SJ, Warrington EK. 2005 Abstract and
concrete concepts have structurally different
representational frameworks. Brain 128, 615–627.
(doi:10.1093/BRAIN/AWH349)
108. Crutch SJ, Ridha BH, Warrington EK. 2007 The
different frameworks underlying abstract and
concrete knowledge: evidence from a bilingual
patient with a semantic refractory access dysphasia.
12, 151–163. (doi:10.1080/13554790600598832)
109. Gray K. 2017 How to map theory: reliable methods
are fruitless without rigorous theory. Perspect.
Psychol. Sci. 12, 731–741. (doi:10.1177/
1745691617691949)
110. Henson R. 2006 Forward inference using functional
neuroimaging: dissociations versus associations.
Trends Cogn. Sci. 10,64–69. (doi:10.1016/j.tics.
2005.12.005)
111. Humphreys G, Lambon Ralph M, Simons J. 2020 A
unifying account of angular gyrus contributions to
episodic and semantic cognition. PsyArXiv.(doi:10.
31234/OSF.IO/R2DEU)
112. Embleton KV, Haroon HA, Morris DM, Lambon Ralph
MA, Parker GJM. 2010 Distortion correction for
diffusion-weighted MRI tractography and fMRI in
the temporal lobes. Hum. Brain Mapp. 31,
1570–1587. (doi:10.1002/hbm.20959)
113. Pexman PM. 2020 How does meaning come to
mind? Four broad principles of semantic processing.
Can. J. Exp. Psychol. 74, 275. (doi:10.31234/osf.io/
scwau)
114. Yee E, Thompson-Schill SL. 2016 Putting concepts
into context. Psychon. Bull. Rev. 23, 1015–1027.
(doi:10.3758/S13423-015-0948-7)
115. Malt BC, Majid A. 2013 How thought is mapped
into words. Wiley Interdiscip. Rev. Cogn. Sci. 4,
583–597. (doi:10.1002/WCS.1251)
116. Thompson B, Roberts SG, Lupyan G. 2020 Cultural
influences on word meanings revealed through
large-scale semantic alignment. Nat. Hum. Behav.
4, 1029–1038. (doi:10.1038/s41562-020-0924-8)
117. Mazzuca C, Majid A, Lugli L, Nicoletti R, Borghi AM.
2020 Gender is a multifaceted concept: evidence
that specific life experiences differentially shape the
concept of gender. Lang. Cogn. 12, 649–678.
(doi:10.1017/langcog.2020.15)
118. Hoffman P, Tamm A. 2020 Barking up the right
tree: univariate and multivariate fMRI analyses of
homonym comprehension. Neuroimage 219,
117050. (doi:10.1016/j.neuroimage.2020.117050)
119. Hoffman P, Binney RJ, Lambon Ralph MA. 2015
Differing contributions of inferior prefrontal and
anterior temporal cortex to concrete and abstract
conceptual knowledge. Cortex 63, 250–266. (doi:10.
1016/j.cortex.2014.09.001)
120. Diveica V, Koldewyn K, Binney RJ. 2021 Establishing
a role of the semantic control network in social
cognitive processing: a meta-analysis of functional
neuroimaging studies. Neuroimage 245, 118702.
(doi:10.1016/J.NEUROIMAGE.2021.118702)
121. Satpute AB, Badre D, Ochsner KN. 2014 Distinct regions
of prefrontal cortex are associated with the controlled
retrieval and selection of social information. Cereb.
Cortex 24,1269–1277. (doi:10.1093/cercor/bhs408)
122. Pérez-Edgar K, Fox NA. 2007 Temperamental
contributions to children’s performance in an
emotion-word processing task: a behavioral and
electrophysiological study. Brain Cogn. 65,22–35.
(doi:10.1016/J.BANDC.2006.10.010)
royalsocietypublishing.org/journal/rstb Phil. Trans. R. Soc. B 378: 20210363
11