Varieties of Abstract Concepts and Their Grounding in Perception Or Action

Home , Abstract and concrete, Motor cognition

Open Psychology 2020; 2: 119–137

Review Article

Markus Kiefer*, Marcel Harpaintner Varieties of abstract concepts and their grounding in perception or action

https://doi.org/10.1515/psych-2020-0104 Received January 28, 2020; accepted May 26, 2020

Abstract: For a very long time, theorizing in the cognitive sciences was dominated by the assumption that abstract concepts, which lack a perceivable referent, can only be handled by amodal or verbal linguistic representations. In the last years, however, refined grounded cognition theories emphasizing the importance of emotional and introspective information for abstract concepts, in addition to verbal associations and sensorimotor information, have received increasing support. Here, we review theoretical accounts of the structure and neural basis of conceptual memory and evaluate them in light of recent empirical evidence with regard to the processing of concrete and abstract concepts. Based on this literature review, we argue that abstract concepts should not be treated as a homogenous conceptual category, whose meaning is established by one single specific type of representation. Instead, depending on the feature composition, there are different subgroups of abstract concepts, including those with strong relations to vision or action, which are represented in the visual and motor brain systems similar to concrete concepts. The reviewed findings with regard to concrete and abstract concepts can be accommodated best by hybrid theories of conceptual representation assuming an interaction between modality-specific, multimodal and amodal hub areas.

Keywords: semantic memory; language; grounded cognition; abstract concepts

Introduction

Concepts held in semantic long-term memory are important for various aspects of cognition: They support problem solving, object recognition, action planning and verbal communication, because they are basic units of cognition and constitute the meaning of objects, events or abstract ideas (Humphreys, Riddoch, & Quinlan, 1988; Kiefer & Pulvermüller, 2012; Levelt, Roelofs, & Meyer, 1999). It is generally accepted that concepts provide factual knowledge generalizing across exemplars and situations in a categorical fashion (Humphreys et al., 1988; Kiefer & Pulvermüller, 2012; Tulving, 1972). Concepts refer to concrete objects or actions, but also have referents, which cannot be directly perceived, like mental or emotional states, abstract ideas, social constellations and scientific theories. The latter types of concepts, which do not refer to physical entities, are typically termed “abstract concepts”. Abstract concepts are a challenge for all theories of conceptual representations, because their semantic content is highly variable and dependent on the context (Borghi et al., 2017; Dove, 2016; Hoffman, 2016). Consider the abstract concept “JUSTICE”:

Article note: This article is part of Special Issue Experimental Cognitive Psychology

*Corresponding author: Markus Kiefer, University of Ulm, Department of Psychiatry, Section for Cognitive Electrophysiology, Leimgrubenweg 12, 89075 Ulm, Germany, E-mail: [email protected] Marcel Harpaintner, University of Ulm, Department of Psychiatry, Section for Cognitive Electrophysiology, Leimgrubenweg 12, 89075 Ulm, Germany

There might be quite different definitions and interpretations of the meaning of this abstract concept across individuals and contexts compared to the concrete concept “TABLE”. Abstract concepts are furthermore a particular challenge for grounded cognition or embodied cognition theories, which assume that conceptual representations are essentially rooted in modal experiential information (Barsalou, Simmons, Barbey, & Wilson, 2003; Bermeitinger & Kiefer, 2012; Kiefer & Pulvermüller, 2012; Kiefer & Spitzer, 2001; Lakoff & Johnson, 1999; Pulvermüller, 2005). Although the terms “grounded cognition” and “embodied cognition” are frequently exchangeably used, these terms occasionally refer to slightly different classes of theories, depending on the precise theoretical stance1. For the purpose of this article, we use the somewhat broader term “grounded cognition theories” to refer to theories, which emphasize the relevance of modal experiential information for cognition. Before returning to the research on abstract concepts, we first introduce two competing classes of current theories aimed at explaining conceptual representations in more detail: amodal theories vs. grounded cognition theories.

Theories of conceptual representations: amodal vs. grounded cognition approaches

In the modern cognitive sciences, concepts have been traditionally viewed as abstract mental entities, different from the perceptual or motor brain systems (Anderson, 1983; Mahon & Caramazza, 2009; Pylyshyn, 1984; Quillian, 1969; Tyler & Moss, 2001): Sensory or motor features of objects and events are assumed to be transformed into a common amodal representation format, detached from modality-specific information. According to this traditional view, conceptual information is represented in an amodal fashion (Machery, 2007; Mahon & Caramazza, 2009; Weiskopf, 2010). Popular examples for such amodal theories are semantic network models (Collins & Loftus, 1975; Quillian, 1969) or some forms of connectionist network models (Caramazza, Hillis, Rapp, & Romani, 1990; Devlin, Gonnerman, Andersen, & Seidenberg, 1998; McClelland & Rogers, 2003; Rogers et al., 2004; Tyler & Moss, 2001). In these amodal models, all kind of information is stored in a modality-independent form. At an anatomical level, amodal conceptual representations are assumed to be held in heteromodal association cortex such as the anterior temporal (McClelland & Rogers, 2003; Rogers et al., 2004) or posterior temporal cortex (Hoffman, Pobric, Drakesmith, & Ralph, 2012), so-called semantic hubs. Although some traditional theories do not deny the involvement of the sensory and motor systems in conceptual tasks, they assume that activation of modality-specific representations during language comprehension or conceptual thinking is only an auxiliary or concomitant process after the amodal concept has been accessed, due to imagery (Machery, 2007) or passive spreading of activation to input or output levels (Mahon, 2015a, 2015b). In contrast to this classical view, grounded cognition approaches propose that concepts are essentially rooted in perception and action (Barsalou et al., 2003; Gallese & Lakoff, 2005; Kiefer & Barsalou, 2013; Kiefer & Pulvermüller, 2012; Kiefer & Spitzer, 2001; Lakoff & Johnson, 1999; Martin & Chao, 2001; Pulvermüller, 2005; Warrington & McCarthy, 1987). Distributed modality-specific representations of external (perception) and internal states (proprioception, emotion and introspection), as well as actions, are hypothesized to establish conceptual representations and thus play a functional role in conceptual cognition. At a neuroanatomical level, it is assumed that activation of cell assemblies in distinct brain areas responsible

1 While some embodied cognition theories confine experiential information to the sensory-motor systems, grounded cognition theories are typically broader and also consider the relevance of experiential information based on introspection and emotional states. Furthermore, strong embodied cognition theories assume a complete functional and neuroanatomical overlap between sensory-motor and conceptual processing, including mandatory activation of primary sensory-motor brain areas (for a discussion, see Meteyard, Cuadrado, Bahrami, & Vigliocco, 2012). In contrast, weaker embodied cognition theories and grounded cognition theories propose a partial overlap between sensory-motor and conceptual processing (for further differentiations between grounded and embodied cognition theories, see Kiefer & Barsalou, 2013). Accumulating evidence speaks against strong variants of embodied cognition theories, but support weaker embodied or grounded cognition theories (see discussions in Meteyard et al., 2012; Popp, Trumpp, & Kiefer, 2019a; Popp, Trumpp, Sim, & Kiefer 2019b). We return to this issue later in this article in more detail. Varieties of abstract concepts and their grounding in perception or action 121 for sensory, motor, introspective and emotional processes constitute the concept. These modality-specific systems are situationally recruited as a function of task demands, resulting in a flexible composition of features contributing to the concept (Hoenig, Sim, Bochev, Herrnberger, & Kiefer, 2008; Popp et al., 2019a; Pulvermüller, 2018; van Dam, van Dijk, Bekkering, & Rueschemeyer, 2012). Note that, besides top-down processes triggered by task demands, factors related to bottom-up processes such as word class (Perani et al., 1999), conceptual category (Kiefer, 2005; Trumpp, Traub, & Kiefer, 2013b) or psycholinguistic variables (Lewis & Frank, 2016) have been found to influence conceptual processing. These effects have been observed in shallow conceptual tasks such as lexical decision (Kiefer, Sim, Herrnberger, Grothe, & Hoenig, 2008) or passive viewing (Trumpp et al., 2013b) as well as under visual masking conditions (Trumpp, Traub, Pulvermüller, & Kiefer, 2014), which prevent visual awareness of the stimuli. This indicates that modality-specific representations are automatically activated in a bottom-up fashion. Recent research in the area of unconscious cognition indicates that even automatic processes are susceptible to top-down influences (Kiefer, 2019; Kiefer & Martens, 2010). Hence, automatic bottom-up activation of modality-specific conceptual features and modulation of this activity by top-down factors are not contradictory. Forced by empirical evidence described below, hybrid frameworks combine assumptions of amodal and grounded cognition theories (Hoffman, McClelland, & Ralph, 2018; Kiefer & Pulvermüller, 2012; Patterson, Nestor, & Rogers, 2007). According to these hybrid models, conceptual representations depend on an interaction between modality-specific, multimodal and amodal conceptual hub areas, the latter serving supramodal semantic binding and integration. These hybrid theories are also called hub and spokes models (Patterson & Ralph, 2016; Ralph, Jefferies, Patterson, & Rogers, 2017). We return to these hybrid models in the next sections.

The easier case: Processing of concrete concepts

The functional neuroanatomical architecture of the conceptual system has been intensively investigated in the past using concrete concepts as stimuli because their semantic content can be more clearly characterized (Barsalou & Wiemer-Hastings, 2005). Concrete concepts are also easier to investigate because conceptual taxonomies, as well as categories within a taxonomy such as tools or animals, can be more unequivocally defined (e.g., Rosch, 1973, 1975) compared to abstract concepts.

Evidence for amodal conceptual hubs

Within this research on concrete concepts, initial evidence for the existence of an amodal semantic hub in anterior temporal cortex comes from neuropsychological studies in patients with Semantic Dementia, a relatively focal brain damage in the temporal pole and its vicinity due to a degenerative disease (Patterson et al., 2007; Rogers et al., 2004). Patients with Semantic Dementia exhibit a general loss of conceptual knowledge across all conceptual categories (including animals, tools etc.) or conceptual feature types (e.g., visual, auditory, action-related). The functional importance of anterior temporal areas for semantic processing has been replicated in healthy volunteers with transcranial magnetic stimulation (TMS). TMS to anterior temporal areas deteriorated performance in semantic tasks for pictures and words similar to the deficit seen in Semantic Dementia (Pobric, Jefferies, & Lambon Ralph, 2010). As the impairment in Semantic Dementia generalized across conceptual information derived from different modalities, it has been proposed that regions within anterior temporal cortex hold amodal conceptual representations (McClelland & Rogers, 2003; Patterson et al., 2007; Rogers et al., 2004). Meta-analyses of functional magnetic resonance imaging (fMRI) and positron emission tomography studies (PET) in healthy volunteers provided evidence for further hub regions, which were activated independent of conceptual category and feature content (Binder, Desai, Graves, & Conant, 2009). Inferior parietal, posterior middle temporal and medial prefrontal regions are thought to serve as convergence zones, which code amodal conceptual representations (Binder, 2016), whereas inferior prefrontal areas are seen to be part of a conceptual control network supporting 122 M. Kiefer, M. Harpaintner controlled retrieval of conceptual information (Thompson-Schill, D’Esposito, & Kan, 1999; Ulrich, Hoenig, Grön, & Kiefer, 2013; Wagner, Paré-Blagoev, Clark, & Poldrack, 2001). Recent evidence suggests that some hub regions such as inferior parietal cortex do not necessarily hold amodal representations but may retain modal feature content (Kuhnke, Kiefer, & Hartwigsen, 2020). These semantic hubs should therefore be better considered as multimodal, rather than amodal regions. Most likely, a hierarchy of processing circuits ranging from lower-level modality-specific cortex over bi-, tri- or multimodal regions up to top-level amodal areas in heteromodal cortex, presumably indexing increasing levels of abstraction, establishes conceptual representations (Kuhnke et al., 2020; Popp et al., 2019b). We return to this issue later in this article in more detail.

Evidence for modality-specific conceptual systems

However, there is also converging evidence in support of grounded cognition theories demonstrating differential involvement of brain areas in the processing of words and concepts of different concrete conceptual categories (for an overview, see Kemmerer & Gonzalez-Castillo, 2010; Kiefer & Barsalou, 2013; Kiefer & Pulvermüller, 2012; Pulvermüller & Fadiga, 2010). PET and fMRI studies in healthy participants (e.g., Hoenig et al., 2008; Martin, Wiggs, Ungerleider, & Haxby, 1996; Simmons, Martin, & Barsalou, 2005) demonstrated activation in sensory-motor brain areas in a broad range of conceptual tasks. In the most convincing neuroimaging, but also behavioral, patient and brain stimulation experiments, a theory- driven experimental approach is used (e.g., Hauk, Johnsrude, & Pulvermüller, 2004; Kiefer et al., 2008; Pulvermüller, Hauk, Nikulin, & Ilmoniemi, 2005; Trumpp, Kliese, Hoenig, Haarmaier, & Kiefer 2013a). In a first step, researchers determined the semantic content of concrete concepts, for instance, by using feature ratings or property listings, and selected sets of concepts with a high relevance of a given feature type originating from a specific modality (e.g., motor features or acoustic features). In the second step, such a priori specified conceptual categories are presented in the main experiment, including neurophysiological recordings and/or behavioral performance measures. Several fMRI studies investigating the processing of motor- (Hauk et al., 2004), visual- (Simmons et al., 2007), sound- (Kiefer et al., 2008), gustatory- (Simmons et al., 2005) and olfactory-related (Gonzalez et al., 2006) concepts found increased activity in corresponding sensory or motor brain regions. This shows that the modal representation of conceptual features in the corresponding sensorimotor brain areas is a general principle that applies to many modalities. Event-related potential studies, which track the time course of conceptual processing with a high temporal resolution showed differential ERPs in response to conceptual categories starting at about 150 ms after target onset (Hauk & Pulvermüller, 2004; Hoenig et al., 2008; Kiefer, Sim, Helbig, & Graf, 2011). The early onset of differential effects speaks in favor of early lexico-semantic processes in conceptual processing and contradicts a postconceptual explanation such as imagery, semantic elaboration or spreading activation. Neurophysiological studies only provide correlational information about conceptual processing in the perception and action systems. In order to test a functional contribution of modality-specific brain areas, lesion, behavioral interference or TMS studies are unavoidable. Such studies indicated that visual (Vermeulen, Corneille, & Niedenthal, 2008), auditory (Trumpp et al., 2013a) and action information (Helbig, Steinwender, Graf, & Kiefer, 2010; Pulvermüller et al., 2005) plays a functional role in the processing of concrete concepts. Using mapping of sensory and motor areas with functional localizer tasks, recent fMRI studies (Kuhnke et al., 2020; Popp et al., 2019b) indicated that not only modality-specific brain areas as defined by localizer tasks (e.g., acoustic localizer: listening to sounds; motor localizer: moving the hands) but also adjacent higher-level multimodal regions responded to concepts with a high relevance of a given feature type (e.g., acoustic or action features). Activity in both modality-specific and multimodal regions was modulated by task demands indicating conceptual flexibility at various levels of the conceptual processing hierarchy. Varieties of abstract concepts and their grounding in perception or action 123

Hybrid models: Interplay between modality-specific, multimodal and amodal processing circuits

Evidence obtained with concrete concepts reviewed above suggests an interplay between modality-specific, bi- or trimodal, multimodal and amodal semantic hub regions. Modality-specific and multimodal regions presumably represent conceptual feature content, whereas amodal semantic hubs code conceptual information in an overarching supramodal fashion. We, therefore, propose that a hierarchy of processing circuits ranging from lower-level modality-specific cortex over multimodal regions up to top-level amodal areas in heteromodal cortex, presumably indexing increasing levels of abstraction, establishes conceptual representations. The available evidence is thus most consistent with hybrid models of conceptual representations combining modality-specific and multimodal circuits with amodal conceptual hubs (Binder, 2016; Fernandino, Humphries, Conant, Seidenberg, & Binder, 2016; Garagnani & Pulvermüller, 2016; Hoffman et al., 2018; Kiefer & Pulvermüller, 2012; Kiefer, Schuch, Schenck, & Fiedler, 2007a; Kuhnke et al., 2020; Patterson et al., 2007; Popp et al., 2019b; Pulvermüller et al., 2010; Simmons & Barsalou, 2003). Activity within this processing hierarchy is modulated as a function of task or context, providing the neural basis of conceptual flexibility (Kemmerer, 2015a; Kuhnke et al., 2020; Popp et al., 2019b).

Abstract concepts: The new frontier

In contrast to the well-investigated concrete concepts, the representation of abstract concepts, such as abstract ideas or scientific theories, imposes challenges for all classes of theories and opens a new frontier in the research on the nature of concepts for the following reasons (see also, Dove, 2016): Abstract concepts are more complex and ambiguous than concrete concepts because they apply to more and rather heterogeneous situations (Barsalou & Wiemer-Hastings, 2005; Hoffman, Ralph, & Rogers, 2013). Therefore, all models of conceptual representations have to deal with a high degree of conceptual flexibility. Furthermore, by definition, abstract concepts do not refer to physical objects that can be directly experienced by the senses. For that reason, their semantic content is less obvious compared to concrete concepts, which refer to a physical object or to a well-defined event. Abstract concepts are a particular challenge for grounded cognition theories because at the first glance, it is hard to imagine how concepts without a referent, which can be perceived or acted upon, could be grounded in the sensory and motor brain systems (Dove, 2009, 2016). Hence, the mere existence of abstract concepts appears to falsify grounded cognition theories and seems to point to an amodal symbolic or verbal representation.

Representational dualism and the Dual Coding Theory

In fact, past research was dominated for a long time by the view that abstract concepts require amodal, symbolic (Mahon & Caramazza, 2009) or verbal representations (Paivio, 1986), the latter view is also present in current theoretical considerations (Dove, 2016). Although amodal or verbal representations have in principle sufficient computational power to code the meaning of abstract concepts, this explanation appears somewhat ad-hoc: Given that converging evidence points to a modality-specific representation of concrete concepts, as described in the previous section, the request for an amodal or verbal representation of abstract concepts establishes a representational dualism for abstract vs. concrete concepts but does not lead to an integrated mechanistic theory, capable of explaining the representation of both types of concepts (Kiefer & Pulvermüller, 2012). Such a representational dualism dates back to the influential work by Paivio on his Dual Coding Theory (Paivio, 1986). Paivio (1986) assumed that abstract concepts are stored in a verbal-symbolic code, while concrete concepts rely on both a visual imaginary and a verbal-symbolic code. Early studies were predominantly inspired by the Dual Coding Theory. They investigated differences in the processing between abstract and concrete concepts at a behavioral and neural level. In this line of research, abstract concepts were treated as a homogenous conceptual category, solely defined by a referent with a lack of unique 124 M. Kiefer, M. Harpaintner physical features. This research yielded a processing advantage for words referring to concrete concepts in comparison to words referring to abstract concepts, the so-called “concreteness effect”: Concrete words are remembered better (Marschark & Paivio, 1977) and recognized faster than abstract words (James, 1975). This effect has been explained by dual coding for concrete concepts (visual and verbal-symbolic codes) and the exclusively verbal-symbolic coding of abstract concepts. It has, however, been argued that concrete and abstract words only differ with regard to the availability of contextual information, e.g., situations in which the concept is encountered, but do not differ with regard to representational codes. This argument is based on the observation that the concreteness effect is reduced when target words were presented within a sentence context (Schwanenflugel & Akin, 1994). Research on the neural substrate of concrete and abstract words yielded inconsistent results. According to the Dual Coding Theory, abstract concepts rely on a verbal-symbolic code and should predominantly activate language regions in the left hemisphere. Concrete concepts, in contrast, due to dual coding, should activate both the left hemisphere language network as well as the visual imagery network in the right hemisphere. Some neuroimaging studies found a greater left hemisphere activation in language regions for abstract words and greater right hemisphere activation for concrete words as predicted by the Dual Coding Theory (e.g., Binder, Westbury, McKiernan, Possing, & Medler, 2005). Other studies observed left hemisphere activity for both word classes (e.g., Sabsevitz, Medler, Seidenberg, & Binder, 2005) with greater activity within language regions for abstract concepts and greater activity within the visual system for concrete concepts (e.g., Desai, Binder, Conant, Mano, & Seidenberg, 2011; Sakreida et al., 2013). Others obtained activity within the sensory-motor system for both abstract and concrete words (Pexman, Hargreaves, Edwards, Henry, & Goodyear, 2007), sometimes complemented by activity in left hemisphere language regions (Desai et al., 2011), or even greater right hemisphere involvement for abstract words (Kiehl et al., 1999). A meta-analysis (Wang, Conder, Blitzer, & Shinkareva, 2010) yielded a greater engagement of left hemisphere perceptual areas for concrete concepts, whereas abstract concepts activated left temporal language regions, particularly the middle and superior temporal gyri, and the left inferior frontal gyrus more strongly than concrete concepts. Although abstract concepts most consistently activated left hemisphere language regions, suggesting relatively greater importance of verbal information for this conceptual category (Dove, 2014), findings were remarkably variable. This suggests that abstract concepts, like concrete concepts, are quite heterogeneous. Most likely, they depend on different types of conceptual feature information and not exclusively on verbal information. We discuss these issues in the next sections of this article.

Beyond representational dualism: Grounded cognition theories and their refinements

When applied to the representation of abstract concepts, grounded cognition theories are confronted with the apparent paradox to explain the representation of a mental entity, which lacks a clearly perceivable referent, by means of its foundations in action and perception. One of the first attempts to apply the grounded cognition theory to abstract concepts was the Conceptual Metaphor Theory (CMT) (Lakoff & Johnson, 1980). CMT assumes that perceptual and action information constitutes the meaning of abstract concepts through metaphoric relations to concrete concepts. It is argued that abstract concepts are understood by mapping them to a concrete knowledge domain, which provides their grounding in perception and action. CMT has stimulated many studies yielding support for concepts mapped to space such as “TIME”, “POWER” or “DIVINITY” (Pecher, Boot, & Van Dantzig, 2011). However, it is unlikely that metaphoric relations can be established for all abstract concepts. Barsalou and Wiemer-Hastings (2005) therefore propose that abstract concepts receive a direct grounding in sensorimotor experience because they refer to concrete situations to which they apply. Representations underlying abstract concepts might be formed by collections of concrete situations that share the abstract concept. Abstract concepts are thus thought to be grounded in situations and can be simulated in modality-specific brain systems. Thinking about the abstract concept “JUSTICE”, for instance, Varieties of abstract concepts and their grounding in perception or action 125 might reactivate sensory and motor memory traces of a situation, in which one saw four children evenly sharing twelve candies. In a property generation study, Barsalou and Wiemer-Hastings (2005) asked participants to produce associations with regard to abstract (e.g., “FREEDOM”), concrete (e.g., “CAR”) and intermediate (e.g., “FARM”) concepts. Although situational experiences were generated for all types of concepts, abstract concepts were mainly associated with introspections, mental states and social aspects of situations. This indicates that, in addition to perception and action, other types of information contributes to the meaning of abstract concepts. The Affective Embodiment Account (AEA) emphasizes the importance of affective information for establishing the meaning of abstract concepts (Kousta, Vigliocco, Vinson, Andrews, & Del Campo, 2011). Likewise, in expanding the grounded cognition account, we and others have suggested that abstract concepts might not only be grounded in the perception of external events such as situations, but also in the introspection of internal mental states and in mentalizing social constellations (Barsalou & Wiemer- Hastings, 2005; Borghi & Binkofski, 2014; Harpaintner, Trumpp, & Kiefer 2018; Kiefer & Barsalou, 2013). In line with the proposed important role of emotions (Vigliocco et al., 2014), mental states (Wilson- Mendenhall, Simmons, Martin, & Barsalou, 2013) and social interactions (Wilson-Mendenhall et al., 2013) for the meaning of abstract concepts, activity in the corresponding neural circuits has been observed. In congruency with the consistent involvement of left hemisphere language regions in abstract concept processing, several theories stress the greater significance of linguistic information for the meaning of abstract concepts compared to concrete concepts. It has been argued that verbal associations based upon the statistical co-occurrence of words in language provide an important knowledge base for the meaning of abstract concepts due to their reduced grounding in sensorimotor information (Borghi & Binkofski, 2014; Connell, 2019; Dove, 2009, 2014; Hoffman, 2016; Louwerse, 2011). Verbal associations might provide a rapid short-cut to the meaning of a word when it takes too much time to simulate sensorimotor information as suggested by the Language and Situated Simulation Theory (LAS, Barsalou, Santos, Simmons, & Wilson, 2008). Furthermore, the Words as Social Tools Theory (WAT) proposes that abstract concepts are frequently acquired via verbal communications and are therefore grounded in the language system, including motor areas required for articulation (Borghi & Binkofski, 2014; Borghi & Zarcone, 2016). Interestingly, congruent with the notion that verbal representations are more important for conceptual processing, when a direct sensorimotor grounding is reduced, deaf individuals recruited language areas more strongly than hearing individuals in a conceptual task, presumably as compensation for the lacking auditory sensory channel (Trumpp & Kiefer, 2018). In conclusion, empirical evidence and theories reviewed in this section indicated that grounded cognition theories have been refined, in order to account for the representation of abstract concepts. Besides experiential sensorimotor information due to metaphoric mapping or due to relations to classes of situations, the relevance of emotional, introspective, social and linguistic information has been stressed (for a discussion of multiple representations, see also Borghi et al., 2017). Some theories emphasize specific knowledge aspects such as emotional (e.g., Kousta et al., 2011) or linguistic information (e..g, Borghi & Binkofski, 2014; Dove, 2014) as exclusive basis for the representation of abstract concepts. However, most likely, these different knowledge aspects contributing to abstract concepts should not be considered as mutually exclusive, but rather as complementary (see also the discussion in Borghi et al., 2017). As we argue in the next section, abstract concepts, similar to concrete concepts, should not be treated as homogenous conceptual category. Instead, we propose that the semantic content of abstract concepts is highly heterogeneous and based on a multitude of conceptual knowledge types depending on the individual concept (Harpaintner et al., 2018).

Varieties of abstract concepts

As described above, several lines of research indicate that abstract concepts not only depend on the verbal system, but also on a variety of modal systems, including perception, action, emotion and introspection. Furthermore, inconsistent findings with regard to the involvement of modal systems in the processing of abstract concepts indicate that the different conceptual feature types are not equally important for the 126 M. Kiefer, M. Harpaintner meaning of all abstract concepts. Nevertheless, some scientists still treat abstract concepts as a homogenous category, simply characterized by the lack of a physical referent object, and contrast them with concrete concepts (e.g., Dove, 2014; Wang et al., 2010). However, the semantic content of abstract concepts might be rich and highly heterogeneous (Ghio, Vaghi, & Tettamanti, 2013; Harpaintner et al., 2018; Hoffman, 2016; Kiefer & Pulvermüller, 2012; Wiemer- Hastings, Krug, & Xu, 2001). For instance, differential patterns of conceptual relations for abstract emotional and non-emotional concepts were found (Barca, Mazzuca, & Borghi, 2017). Although a bimodal distribution of concreteness ratings indicated a categorization of abstract and concrete concepts into two large clusters, there was also a large variance within these clusters (Harpaintner et al., 2018; Wiemer-Hastings et al., 2001). This suggests that similar to the research on concrete concepts (Kiefer & Barsalou, 2013), differential classes of abstract concepts and their relation to conceptual features types (sensorimotor, emotional, introspective, social, verbal, etc.) have to be determined. It is therefore unlikely that the meaning of all abstract concepts is established by predominantly one single feature type (e.g., emotional or verbal). To test the notion of multiple feature types contributing to abstract concepts, Barsalou and Wiemer- Hastings (2005) asked their participants to generate properties and associations for a small set of concepts differing with regard to their level of concreteness/abstractness. Besides social, mental and introspective content, content about physical settings has also been found to play a role in abstract concepts, albeit to a somewhat smaller extent compared to concrete concepts. The semantic variability of abstract concepts was further demonstrated by a rating study (Ghio et al., 2013), which indicated that clearly distinguishable subcategories of abstract concepts related to mental states, emotions and mathematics exist (for neuroimaging evidence see Ghio, Vaghi, Perani, & Tettamanti, 2016). Given the probable heterogeneity of abstract concepts, we characterized the semantic content of a large set of 296 abstract concepts in a property listing study (Harpaintner et al., 2018). Participants were asked to write down properties such as features, situations and associations coming into mind for 296 abstract concepts. These properties were categorized by a coding-scheme making a classification into modality-specific and verbal contents – namely into sensorimotor features, features describing social constellations, internal states and emotions as well as verbal associations – possible. Hierarchical cluster analyses were used to shed light on the heterogeneity of abstract concepts. Descriptive analyses (Figure 1) revealed that participants generated a substantial proportion of introspective, emotional and social properties, in addition to verbal associations. In terms of quantity, however, sensory and motor properties played the most crucial role in this study. This is quite remarkable, given that sensorimotor features are thought to be only associated with concrete concepts, but not with abstract concepts (Dove, 2014; Paivio, 1986; Wang et al., 2010). More recent studies, however, also indicated the significance of sensory and motor features even for the representation of abstract concepts (Dreyer et al., 2015; Lynott & Connell, 2013; Mason & Just, 2016; Moseley, Carota, Hauk, Mohr, & Pulvermuller, 2012; van Dantzig, Cowell, Zeelenberg, & Pecher, 2011; Vukovic, Feurra, Shpektor, Myachykov, & Shtyrov, 2017). We return to this issue in the next section. Hierarchical cluster analyses of the generated properties (Figure 2) demonstrated the existence of several subgroups of abstract concepts (Harpaintner et al., 2018). Similar to concrete concepts, abstract concepts could be divided into different subcategories according to the dominance of certain conceptual features. The largest cluster (131 words) was characterized by a large proportion of sensory and motor- related features (e.g., OBSERVATION, FITNESS). In the second-largest cluster (113 words), high proportions of generated properties referred to internal states, emotions and social constellations (e.g., NIGHTMARE, ARGUMENT). In the smallest cluster (42 words), verbal associations were the most dominant feature type (e.g., THEORY, DIGNITY). The broad diversity in participants’ listings is consistent with refined grounded cognition theories by showing that the semantic content of abstract concepts – just like that of concrete concepts – includes introspective, affective, social, sensory and motor-related features (Barsalou & Wiemer-Hastings, 2005; Harpaintner et al., 2018; Kiefer & Barsalou, 2013; Kiefer & Pulvermüller, 2012; Kousta et al., 2011). These results also suggest that only one relatively small subgroup of abstract concepts is predominantly related to verbal associations, although several theories (Borghi & Binkofski, 2014; Connell, 2019; Dove, 2014; Varieties of abstract concepts and their grounding in perception or action 127

Hoffman, 2016; Louwerse, 2011) emphasize the role of the language system for all kinds of abstract concepts including Dual Coding Theory (Paivio, 1986). The results of this property listing study, of course, do not necessarily imply that the identified modal feature types are also represented in the corresponding modal brain areas, as claimed by grounded cognition theories. This specific assumption of grounded cognition theories has to be tested in behavioral and neuroscientific studies.

Figure 1: Descriptive statistics of feature types, generated in the property listings of 296 abstract concepts, and corresponding boxplots. (M = mean, SD = standard deviation, x0.1/0.9 = first and ninth decile. SM = sensorimotor feature, IS / E = internal state / emotion, SC = social constellation, VA = verbal association. ExHigh / ExLow depicts exemplary abstract concepts with a high/low portion of generated features in the respective categories. Modified after Harpaintner and colleagues (2018).

Figure 2: Results of hierarchical cluster analyses of feature types, derived from property listings of 296 abstract concepts. (A) Dendrogram visualizing the k = 3 cluster solution. The different clusters are marked by different colors. (B) Boxplots depicting generated features per cluster (SM = sensorimotor feature, IS / E = internal state / emotion, SC = social constellation, VA = verbal association). Modified after Harpaintner and colleagues (2018). 128 M. Kiefer, M. Harpaintner

Abstract concepts and their relation to the sensory and motor systems

As reviewed above, property listing (Barsalou & Wiemer-Hastings, 2005; Harpaintner et al., 2018) and rating studies (Lynott & Connell, 2013; van Dantzig et al., 2011) highlight the importance of sensory and motor features also for the semantic content of abstract words, as predicted by grounded cognition theories. However, studies assessing the subjective content of abstract concept do not address the question, whether processing of abstract concepts depends on the sensory and motor systems. This information is provided by behavioral and neuroimaging studies as well as by studies in brain-damaged patients. While the validity of early behavioral studies (e.g., Strack, Martin, & Stepper, 1988) suggesting an involvement of the motor system in the processing of abstract concepts has been seriously questioned by replication studies (Wagenmakers et al., 2016), results from several more recent studies were more convincing. For instance, suppression of spontaneous facial expressions interfered with emotion recognition (Havas, Glenberg, Gutowski, Lucarelli, & Davidson, 2010) and semantic processing of emotion words (Foroni & Semin, 2009) demonstrating the functional significance of the motor system for the processing of abstract emotional concepts. Furthermore, understanding the meaning of concrete and abstract transfer events has been shown to induce a corresponding motor action: Participants were faster to verify abstract transfer sentences such as “Liz told you the story” with responses toward themselves, compared with making hand movements away from themselves. This action compatibility effect (ACE), which has been originally established using concrete transfer sentences, was observed, even though the described situation did not involve any specific hand movement (Glenberg et al., 2008). Finally, mouth-related actions facilitated recognition memory of non-emotional abstract concepts (Mazzuca, Lugli, Benassi, Nicoletti, & Borghi, 2018), in line with the notion that language is frequently used to convey the meaning of abstract words (Borghi & Binkofski, 2014), while for emotional abstract concepts an inconsistent result pattern was found. Although these behavioral studies suggest a functional role of the motor system in the processing of some types of abstract concepts similarly to concrete concepts, behavioral compatibility effects might arise at a response stage outside the conceptual system. Neuroscientific studies are, therefore, necessary to demonstrate the direct involvement of modal systems in conceptual processing. The importance of modal cortex in the processing of abstract concepts has been demonstrated in several neuroimaging and patient studies: Pexman and colleagues (2007) observed activation within the sensorimotor system for both abstract (e.g., “REASON”) and concrete (e.g., “EARTH”) concepts. Understanding the meaning of the abstract concept “OBSERVE” elicited activity in visual and auditory cortex (Wilson-Mendenhall, Barrett, Simmons, & Barsalou, 2011), while abstract emotion words (e.g., “FEAR”) activated the primary motor cortex (Dreyer & Pulvermüller, 2018; Moseley et al., 2012), a region which is crucial for the communication of internal emotional states by gestures and facial expressions. A study in brain-lesioned patients indicated a causal role of the motor cortex in the processing of abstract emotion words (Dreyer et al., 2015). Even abstract physical concepts elicited activations in wide-spread modal brain regions (Mason & Just, 2016). For instance, physical concepts related to periodicity (e.g., “frequency”) activated postcentral and parietal brain regions, regions found to be active when performing rhythmic movements (Chen, Zatorre, & Penhune, 2006). Enhanced activity in the motor cortex was furthermore observed in the case of number concepts such as NINE (Tschentscher, Hauk, Fischer, & Pulvermuller, 2012), which are typically associated with finger counting movements (Domahs, Moeller, Huber, Willmes, & Nuerk, 2010). While the relevance of linguistic information, emotions, mental states and social interactions for the meaning of abstract concepts has been well investigated (see the previous sections), grounding of abstract concepts in the sensory and motor systems lacks systematical investigation and has been implied only by a few studies as indicated above. Furthermore, interpretation of findings from previous studies is limited: Subjective measures might not necessarily reflect the semantic content of abstract words in a valid fashion because not all aspects of semantics might be consciously accessible (Harpaintner et al., 2018; Kemmerer, 2015b). Moreover, sometimes only single abstract concepts served as stimuli (Wilson-Mendenhall et al., 2011). In addition, sensorimotor feature composition of the concepts was typically not determined a priori but only inferred post-hoc from the activation pattern. Hence, implicit measures of conceptual processing, Varieties of abstract concepts and their grounding in perception or action 129 e.g., derived from neuroscientific methods (but also from reaction time measurements), should be applied to a large set of abstract words characterized a priori with regard to their sensorimotor content. We therefore systematically examined the relation of abstract concepts to the sensory and motor brain systems in a larger set of abstract nouns (N = 64) in an fMRI study (Harpaintner, Sim, Trumpp, Ulrich, & Kiefer, 2020). We adopted a theory-driven approach frequently used for investigating concrete concepts (e.g., Kiefer et al., 2008; Kiefer et al., 2012) to the domain of abstract concepts and compared brain activation to well-defined subtypes of abstract concepts with a known feature content. Words have been a priori selected on the basis of sensory and motor features obtained in a previous property listing study (Harpaintner et al., 2018). Based on the available property listings, we selected 32 abstract words strongly linked to motor properties (e.g., “fight”) and other 32 abstract words strongly linked to visual conceptual properties (e.g., “beauty”). Motor and visual abstract words were presented among pseudowords within a lexical decision task enabling an implicit access to conceptual word meaning. Participants performed a go/ no-go task, which did not require a motor response to the critical word stimuli. This avoids interference of the overt motor response with conceptual processing, in particular within the motor system. To test whether conceptual processing of motor and visual abstract concepts activates corresponding sensorimotor brain regions similar to action and perception, motor and visual localizer tasks were administered. In these tasks, participants performed real hand movements (motor localizer) or observed object pictures passively (visual localizer). Motor abstract concepts activated bilateral fronto-parietal brain areas and subcortical structures including bilateral precentral and postcentral gyrus, left precuneus, left posterior cingulate cortex and left thalamus. Activity to motor abstract concepts overlapped with activations during the execution of real movements in the left precentral and postcentral gyrus (Figure 3). The processing of visual abstract concepts (Figure 4), in contrast, was associated with enhanced brain activity in occipital and temporal brain regions, the insula and the limbic lobe, partially overlapping with activations during the observation of real pictures, particularly in the fusiform and lingual gyrus, i.e. in parts of the ventral visual stream. Specific activation to visual abstract concepts was less robust and only observed at a more lenient statistical threshold. Conjunction analysis revealed activity in the posterior temporal gyri common to both types of abstract concepts indicating the involvement of an amodal hub in posterior temporal cortex (Binder et al., 2009; Gold et al., 2006). Results of this study thus show that the subjectively reported motor and visual feature content of abstract concepts is associated with activity in corresponding modal brain areas (Harpaintner et al., 2020). Most likely, encoded memories of situational perceptions are simulated when accessing conceptual knowledge for abstract concepts (Barsalou & Wiemer-Hastings, 2005; Harpaintner et al., 2020; Kiefer & Pulvermüller, 2012). Furthermore, metaphoric relations between abstract and concrete concepts (Lakoff & Johnson, 1980) might lead to a grounding of abstract concepts in modal brain systems. Overall, our fMRI study (Harpaintner et al., 2020) suggests that the representation of abstract concepts does not differ essentially and necessarily from that of concrete concepts, at least with respect to the investigated subgroups of visual and motor abstract concepts (Harpaintner et al., 2020). Representation of such abstract concepts is most likely established by an interplay between modality-specific areas and amodal hubs, as suggested by hybrid models of conceptual cognition.

Conclusions and future research directions

For a very long time, theorizing in the cognitive sciences was dominated by the assumption that abstract concepts can only be handled by amodal or verbal linguistic representations. It was not conceivable how concepts without a direct perceivable referent should be grounded in modal systems such as perception, action, emotion and introspection. In the last years, refined grounded cognition theories, emphasizing the importance of emotional and introspective information for abstract concepts, in addition to verbal associations and sensorimotor information, have received increasing support. As already indicated above, it is unlikely that abstract concept processing is established by predominantly one feature type such as 130 M. Kiefer, M. Harpaintner

Figure 3: Relation of abstract concepts to the motor system: (A) Greater activation to motor versus visual abstract concepts (depicted in green) during the lexical decision task, overlaid with activations during the motor localizer task (depicted in red). Overlapping activations are depicted in yellow. The contrast motor vs. visual abstract concepts was voxel thresholded at p < .001, cluster-corrected at pFWE < .05. Results of the motor localizer task are reported at a voxel threshold of pFWE < .05 corrected for multiple comparisons with a minimum cluster size of 10 voxels. Note that the cluster of overlapping voxels was only determined by quantitatively overlaying the activation patterns of the localizer and lexical decision tasks for visualization purposes, without formal statistical analyses (B) Results of formal inclusive masking analysis to reveal common activation for motor abstract concepts and the motor localizer task. Significant overlap was found in left pre- and postcentral gyri. Inclusive masking analysis (depicted in blue) were voxel thresholded at p < .001, cluster-corrected at pFWE < .05. The right bar graph depicts the magnitude of neural activation of the whole cluster as a function of motor vs. visual abstract concepts, as represented by averaged parameter estimates. Color range bars indicate T-scores. Error bars represent SEM. Cereb = cerebellum; OG = occipital gyri; opIFG = opercular inferior frontal gyrus; pre/postcentral = pre- and postcentral gyrus; SFG = superior frontal gyrus; SMA = supplementary motor area; SMG = supramarginal gyrus; mot = motor abstract concept; vis = visual abstract concept; L/R = left/right hemisphere; a.u. = arbitrary unit. Modified after Harpaintner and colleagues (2020). Varieties of abstract concepts and their grounding in perception or action 131

Figure 4: Relation of abstract concepts to the visual system. (A) Greater activation to visual versus motor abstract concepts (depicted in green) during the lexical decision task visually overlaid with activations during the visual localizer task (depicted in red). Overlapping activations are depicted in yellow. Activations below the brain surface are visualized by a cutout. (B) Mul- tislice image of overlapping activations to visual versus motor abstract concepts and activations obtained during the visual localizer task in the right fusiform gyrus. The right bar graph depicts the magnitude of neural activation of the whole cluster as a function of visual vs. motor abstract concepts, as represented by averaged parameter estimates. The contrast visual vs. motor abstract concepts was voxel thresholded at p < .005 (uncorrected for multiple comparisons) with a minimum cluster size of 10 voxels. Results of the visual localizer task are reported at a voxel threshold of pFWE < .05 corrected for multiple comparisons with a minimum cluster size of 10 voxels. Color range bars indicate T-scores. Error bars represent SEM. FuG = fusiform gyrus; Hippo = hippocampus; OG = occipital gyri; opIFG = opercular inferior frontal gyrus; pre/postcentral = pre- and postcentral gyrus; SMA = supplementary motor area; SFG = superior frontal gyrus; tPole = temporal pole; triIFG = triangular inferior frontal gyrus; mot = motor abstract concept; vis = visual abstract concept; L/R = left/right hemisphere; a.u. = arbitrary unit. Modified after Harpaintner and colleagues (2020). sensorimotor features. Instead, we propose that abstract concept representation depends on a variety of experiential channels like emotions, introspections, mental states, action and perception. As reviewed in this article, accumulating evidence indicates that the meaning of abstract concepts is represented in a variety of modal systems similar to concrete concepts. However, concrete and abstract concepts most likely differ with regard to their precise feature composition. While the meaning of concrete concepts is assumed to be predominantly established by sensory and motor features (McRae & Cree, 2002), the feature composition of abstract concepts seems richer 132 M. Kiefer, M. Harpaintner and highly heterogeneous (Ghio et al., 2013; Harpaintner et al., 2018; Hoffman, 2016; Kiefer & Pulvermüller, 2012; Wiemer-Hastings et al., 2001). As shown in the previous section, there are subgroups of abstract concepts with strong relations to vision or action, which are represented in the visual and motor systems similar to concrete concepts. However, abstract concepts with a strong link to emotions (e.g., “FEAR”) might be represented in brain structures associated with affective processing (Herbert et al., 2009; Vigliocco et al., 2014). Abstract concepts with a strong link to social constellations and mental states (e.g., “CONVINCE”) might be represented in brain areas associated with social cognition and mentalizing (Wilson-Mendenhall et al., 2013). Concepts at the extreme end of the abstractness scale (e.g., “DIGNITY” or scientific concepts like “GRAVITY”) with a strong link to verbal associations might predominantly depend on language regions or amodal hub regions (Wang et al., 2010). In any case, abstract concepts should not be treated as a homogenous conceptual category, whose meaning is established by simply one specific conceptual feature type (Borghi et al., 2017; Harpaintner et al., 2018). Instead, the richness and heterogeneity of the semantic content of abstract concepts has to be recognized and empirically determined, as shown above (e.g., Harpaintner et al., 2018). Neuroimaging or behavioral studies can then investigate the processing and representation of specific abstract subcategories (e.g., sound-related abstract concepts or abstract concepts specifically related to emotions, social constellations and verbal associations) in a theoretically and empirically informed manner. The findings with regard to concrete and abstract concepts, which were reviewed in this article, can be accommodated best by hybrid theories of conceptual representation assuming an interaction between modality-specific, multimodal and amodal hub areas (Binder, 2016; Fernandino et al., 2016; Garagnani & Pulvermüller, 2016; Hoffman et al., 2018; Kiefer & Pulvermüller, 2012; Kiefer et al., 2007a; Kuhnke et al., 2020; Patterson et al., 2007; Popp et al., 2019b; Pulvermüller et al., 2010; Simmons & Barsalou, 2003). Modality-specific systems include the sensorimotor systems, but also the modal systems involved in the processing of emotions, introspections, mentalizing, social constellations and language. The latter modal systems are probably more important for abstract than for concrete concepts. Of course, as studies using neuroimaging techniques can only provide correlational evidence, methods allowing for causal conclusions, as transcranial magnetic stimulation (Pulvermüller et al., 2005; Vukovic et al., 2017), investigation of brain-lesioned patients (Dreyer et al., 2015; Trumpp et al., 2013a) or behavioral interference paradigms (Shebani & Pulvermüller, 2013; Vermeulen et al., 2008), are mandatory in order to demonstrate a functional role of the modal systems for the processing of abstract concepts. The meaning of abstract concepts is quite ambiguous referring to a broad range of situations or constellations (Hoffman et al., 2013). The feature compositions of abstract concepts are therefore presumably highly variable and situation-dependent, displaying a high degree of conceptual flexibility (Dove, 2016). The notion of a high conceptual flexibility has just recently been supported by an ERP-study from our group (unpublished data) indicating that abstract concept processing is modulated by task demands (implicit vs. explicit task). Future research could complement these results by investigating the dependency of abstract concept processing on different contexts, situational factors, task sets, habitual preferences and differential attentional foci at hand, similarly as it already has been done in the case of concrete concepts (e.g., Popp et al., 2019a). Furthermore, the semantic content of abstract concepts and its representation in the modal systems most likely reflects intraindividual and interindividual experience with and exposure to specific situations and constellations. A grounding of concepts in modal systems depending on individual sensorimotor experience, as predicted by grounded cognition theories (Kiefer & Barsalou, 2013; Kiefer & Pulvermüller, 2012), has been previously demonstrated mainly for concrete object concepts (Hoenig et al., 2011; Kiefer, Sim, Liebich, Hauk, & Tanaka, 2007b; Trumpp & Kiefer, 2018), but recently also for abstract number concepts (Bechtold, Bellebaum, Egan, Tettamanti, & Ghio, 2019). It would be an interesting field of future research to systematically investigate experience-dependent plasticity of abstract concepts at the functional and neural level as a function of training or expertise.

Financial statement: This research was supported by a grant of the German Research Community (DFG Ki 804/7-1) to MK. Varieties of abstract concepts and their grounding in perception or action 133

Ethics statement: not applicable.

Conflict of interest: Professor Markus Kiefer is Editor-in-Chief of Open Psychology although had no involvement in the final decision taken solely by the editor.

References

Anderson, J. R. (1983). The architecture of cognition. Hillsdayle, NJ: Lawrence Erlbaum Associates, Inc. Barca, L., Mazzuca, C., & Borghi, A. M. (2017). Pacifier overuse and conceptual relations of abstract and emotional concepts. Frontiers in Psychology, 8. doi: Artn 2014 10.3389/Fpsyg.2017.02014 Barsalou, L. W., Santos, A., Simmons, W. K., & Wilson, C. D. (2008). Language and simulation in conceptual processing. In M. De Vega, A. M. Glenberg & A. C. Graesser (Eds.), Symbols, embodiment, and meaning. Oxford: Oxford University Press. Barsalou, L. W., Simmons, W. K., Barbey, A. K., & Wilson, C. D. (2003). Grounding conceptual knowledge in modality-specific systems. Trends in Cognitive Sciences, 7(2), 84-91. Barsalou, L. W., & Wiemer-Hastings, K. (2005). Situating abstract concepts. In D. Pecher & R. Zwaan (Eds.), Grounding cognition: The role of perception and action in memory, language, and thought (pp. 129-163). New York: Cambridge University Press. Bechtold, L., Bellebaum, C., Egan, S., Tettamanti, M., & Ghio, M. (2019). The role of experience for abstract concepts: Expertise modulates the electrophysiological correlates of mathematical word processing. Brain and Language, 188, 1-10. doi: 10.1016/j.bandl.2018.10.002 Bermeitinger, C., & Kiefer, M. (2012). Embodied concepts. In S. C. Koch, T. Fuchs, M. Sumna & C. Müller (Eds.), Body memory, metaphor and movement (pp. 121-140). Amsterdam: John Benjamins Publishing Company. Binder, J. R. (2016). In defense of abstract conceptual representations. Psychonomic Bulletin & Review, 23(4), 1096-1108. doi: 10.3758/s13423-015-0909-1 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(12), 2767-2796. doi: 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(6), 905-917. Borghi, A. M., & Binkofski, F. (2014). Words as social tools: An embodied view on abstract concepts. New York, USA: Springer- Verlag New York. Borghi, A. M., Binkofski, F., Castelfranchi, C., Cimatti, F., Scorolli, C., & Tummolini, L. (2017). The challenge of abstract concepts. Psychological Bulletin, 143(3), 263-292. doi: 10.1037/bul0000089 Borghi, A. M., & Zarcone, E. (2016). Grounding abstractness: Abstract concepts and the activation of the mouth. Frontiers in Psychology, 7. doi: Artn 1498 10.3389/Fpsyg.2016.01498 Caramazza, A., Hillis, A. E., Rapp, B. C., & Romani, C. (1990). The multiple semantics hypothesis: Multiple confusions? Cognitive Neuropsychology, 7(3), 161-189. Chen, J. L., Zatorre, R. J., & Penhune, V. B. (2006). Interactions between auditory and dorsal premotor cortex during synchro- nization to musical rhythms. Neuroimage, 32(4), 1771-1781. doi: 10.1016/j.neuroimage.2006.04.207 Collins, A. M., & Loftus, E. F. (1975). A spreading–activation theory of semantic processing. Psychological Review, 82, 407–428. Connell, L. (2019). What have labels ever done for us? The linguistic shortcut in conceptual processing. Language, Cognition and Neuroscience, 34(10), 1308-1318. doi: 10.1080/23273798.2018.1471512 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(9), 2376-2386. doi: 10.1162/jocn.2010.21596 Devlin, J. T., Gonnerman, L. M., Andersen, E. S., & Seidenberg, M. S. (1998). Category-specific semantic deficits in focal and widespread brain damage: A computational account. Journal of Cognitive Neuroscience, 10(1), 77-94. Domahs, F., Moeller, K., Huber, S., Willmes, K., & Nuerk, H. C. (2010). Embodied numerosity: Implicit hand-based representations influence symbolic number processing across cultures. Cognition, 116(2), 251-266. Dove, G. (2009). Beyond perceptual symbols: A call for representational pluralism. Cognition, 110(3), 412-431. doi: 10.1016/j. cognition.2008.11.016 Dove, G. (2014). Thinking in words: Language as an embodied medium of thought. Topics in Cognitive Science, 6(3), 371-389. doi: 10.1111/tops.12102 Dove, G. (2016). Three symbol ungrounding problems: Abstract concepts and the future of embodied cognition. Psychonomic Bulletin & Review, 23(4), 1109-1121. doi: 10.3758/s13423-015-0825-4 134 M. Kiefer, M. Harpaintner

Dreyer, F. R., Frey, D., Arana, S., von Saldern, S., Picht, T., Vajkoczy, P., & Pulvermuller, F. (2015). Is the motor system necessary for processing action and abstract emotion words? Evidence from focal brain lesions. Frontiers in Psychology, 6. doi: 10.3389/Fpsyg.2015.01661 Dreyer, F. R., & Pulvermüller, F. (2018). Abstract semantics in the motor system? - An event-related fMRI study on passive reading of semantic word categories carrying abstract emotional and mental meaning. Cortex, 100, 52-70. doi: 10.1016/j. cortex.2017.10.021 Fernandino, L., Humphries, C. J., Conant, L. L., Seidenberg, M. S., & Binder, J. R. (2016). Heteromodal cortical areas encode sensory-motor features of word meaning. Journal of Neuroscience, 36(38), 9763-9769. doi: 10.1523/Jneurosci.4095- 15.2016 Foroni, F., & Semin, G. R. (2009). Language that puts you in touch with your bodily feelings: The multimodal responsiveness of affective expressions. Psychological Science, 20(8), 974-980. Gallese, V., & Lakoff, G. (2005). The brain’s concepts: The role of the sensory-motor system in conceptual knowledge. Cognitive Neuropsychology, 22(3-4), 455-479. Garagnani, M., & Pulvermüller, F. (2016). Conceptual grounding of language in action and perception: a neurocomputational model of the emergence of category specificity and semantic hubs. European Journal of Neuroscience, 43(6), 721-737. doi: 10.1111/ejn.13145 Ghio, M., Vaghi, M. M. S., Perani, D., & Tettamanti, M. (2016). Decoding the neural representation of fine-grained conceptual categories. Neuroimage, 132, 93-103. doi: 10.1016/j.neuroimage.2016.02.009 Ghio, M., Vaghi, M. M. S., & Tettamanti, M. (2013). Fine-grained semantic categorization across the abstract and concrete domains. Plos One, 8(6). doi: ARTN e67090 10.1371/journal.pone.0067090 Glenberg, A. M., Sato, M., Cattaneo, L., Riggio, L., Palumbo, D., & Buccino, G. (2008). Processing abstract language modulates motor system activity. Quarterly Journal of Experimental Psychology, 61(6), 905-919. doi: 10.1080/17470210701625550 Gold, B. T., Balota, D. A., Jones, S. J., Powell, D. K., Smith, C. D., & Andersen, A. H. (2006). Dissociation of automatic and strategic lexical-semantics: Functional magnetic resonance imaging evidence for differing roles of multiple fronto- temporal regions. The Journal of Neuroscience, 26(24), 6523-6532. Gonzalez, J., Barros-Loscertales, A., Pulvermüller, F., Meseguer, V., Sanjuan, A., Belloch, V., & Avila, C. (2006). Reading cinnamon activates olfactory brain regions. Neuroimage, 32(2), 906-912. Harpaintner, M., Sim, E. J., Trumpp, N. M., Ulrich, M., & Kiefer, M. (2020). The grounding of abstract concepts in the motor and visual system: An fMRI study. Cortex, 124, 1-22. doi: 10.1016/j.cortex.2019.10.014 Harpaintner, M., Trumpp, N. M., & Kiefer , M. (2018). The semantic content of abstract concepts: A property listing study of 296 abstract words. Frontiers in Psychology, 9, 1748. doi: 10.3389/fpsyg.2018.01748 Hauk, O., Johnsrude, I., & Pulvermüller, F. (2004). Somatotopic representation of action words in human motor and premotor cortex. Neuron, 41(2), 301-307. Hauk, O., & Pulvermüller, F. (2004). Neurophysiological distinction of action words in the fronto-central cortex. Human Brain Mapping, 21(3), 191-201. Havas, D. A., Glenberg, A. M., Gutowski, K. A., Lucarelli, M. J., & Davidson, R. J. (2010). Cosmetic use of botulinum toxin affects processing of emotional language. Psychological Science, 21(7), 895-900. doi: 10.1177/0956797610374742 Helbig, H. B., Steinwender, J., Graf, M., & Kiefer, M. (2010). Action observation can prime visual object recognition. Experimental Brain Research, 200, 251-258. Herbert, C., Ethofer, T., Anders, S., Junghöfer, M., Wildgruber, D., Grodd, W., & Kissler, J. (2009). Amygdala activation during reading of emotional adjectives - An advantage for pleasant content. Social Cognitive and Affective Neuroscience 4(1), 35-49. Hoenig, K., Müller, C., Herrnberger, B., Spitzer, M., Ehret, G., & Kiefer, M. (2011). Neuroplasticity of semantic maps for musical instruments in professional musicians. NeuroImage, 56(3), 1714–1725. Hoenig, K., Sim, E.-J., Bochev, V., Herrnberger, B., & Kiefer, M. (2008). Conceptual flexibility in the human brain: Dynamic recruitment of semantic maps from visual, motion and motor-related areas. Journal of Cognitive Neuroscience, 20(10), 1799-1814. Hoffman, P. (2016). The meaning of ‘life’ and other abstract words: Insights from neuropsychology. Journal of Neuropsy- chology, 10(2), 317-343. doi: 10.1111/jnp.12065 Hoffman, P., McClelland, J. L., & Ralph, M. A. L. (2018). Concepts, control, and context: A connectionist account of normal and disordered semantic cognition. Psychological Review, 125(3), 293-328. doi: 10.1037/rev0000094 Hoffman, P., Pobric, G., Drakesmith, M., & Ralph, M. A. L. (2012). Posterior middle temporal gyrus is involved in verbal and non-verbal semantic cognition: Evidence from rTMS. Aphasiology, 26(9), 1119-1130. doi: 10.1080/02687038.2011.608838 Hoffman, P., Ralph, M. A. L., & Rogers, T. T. (2013). Semantic diversity: A measure of semantic ambiguity based on variability in the contextual usage of words. Behavior Research Methods, 45(3), 718-730. doi: 10.3758/s13428-012-0278-x Humphreys, G. W., Riddoch, M. J., & Quinlan, P. T. (1988). Cascade processes in picture identification. Cognitive Neuropsy- chology, 5, 67-103. James, C. T. (1975). The role of semantic information in lexical decisions. Journal of Experimental Psychology: Human Perception & Performance, 1, 130-136. Varieties of abstract concepts and their grounding in perception or action 135

Kemmerer, D. (2015a). Are the motor features of verb meanings represented in the precentral motor cortices? Yes, but within the context of a flexible, multilevel architecture for conceptual knowledge. Psychonomic Bulletin & Review, 22(4), 1068-1075. doi: 10.3758/s13423-014-0784-1 Kemmerer, D. (2015b). Are we ever aware of concepts? A critical question for the Global Neuronal Workspace, Integrated Information, and Attended Intermediate-Level Representation theories of consciousness. Neuroscience of Consciousness,, 2015(1), 1-10. doi: https://doi.org/10.1093/nc/niv006 Kemmerer, D., & Gonzalez-Castillo, J. (2010). The Two-Level Theory of verb meaning: An approach to integrating the semantics of action with the mirror neuron system. Brain and Language, 112(1), 54-76. Kiefer, M. (2005). Repetition priming modulates category-related effects on event-related potentials: Further evidence for multiple cortical semantic systems. Journal of Cognitive Neuroscience, 17(2), 199-211. Kiefer, M. (2019). Cognitive control over unconscious cognition: Flexibility and generalizability of task set influences on subsequent masked semantic priming. Psychological Research, 83(7), 1556-1570. doi: 10.1007/s00426-018-1011-x Kiefer, M., & Barsalou, L. W. (2013). Grounding the human conceptual system in perception, action, and internal states. In W. Prinz, M. Beisert & A. Herwig (Eds.), Action science: Foundations of an emerging discipline (pp. 381-407). Cambridge: MIT Press. Kiefer, M., & Martens, U. (2010). Attentional sensitization of unconscious cognition: Task sets modulate subsequent masked semantic priming. Journal of Experimental Psychology: General, 139(3), 464-489. Kiefer, M., & Pulvermüller, F. (2012). Conceptual representations in mind and brain: Theoretical developments, current evidence and future directions. Cortex, 48, 805-825. Kiefer, M., Schuch, S., Schenck, W., & Fiedler, K. (2007a). Mood states modulate activity in semantic brain areas during emotional word encoding. Cerebral Cortex, 17(7), 1516-1530. Kiefer, M., Sim, E.-J., Helbig, H. B., & Graf, M. (2011). Tracking the time course of action priming on object recognition: Evidence for fast and slow influences of action on perception. Journal of Cognitive Neuroscience, 23(8), 1864–1874. Kiefer, M., Sim, E.-J., Herrnberger, B., Grothe, J., & Hoenig, K. (2008). The sound of concepts: Four markers for a link between auditory and conceptual brain systems. The Journal of Neuroscience, 28(47), 12224–12230. Kiefer, M., Sim, E.-J., Liebich, S., Hauk, O., & Tanaka, J. W. (2007b). Experience-dependent plasticity of conceptual representations in human sensory-motor areas. Journal of Cognitive Neuroscience, 19(3), 525-542. Kiefer, M., & Spitzer, M. (2001). The limits of a distributed account of conceptual knowledge. Trends in Cognitive Sciences, 5(11), 469-471. Kiefer, M., Trumpp, N. M., Herrnberger, B., Sim, E.-J., Hoenig, K., & Pulvermüller, F. (2012). Dissociating the representation of action- and sound-related concepts in middle temporal cortex. Brain and Language, 122(2), 120-125. 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(4), 225-233. 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(1), 14-34. doi: 10.1037/a0021446 Kuhnke, P., Kiefer, M., & Hartwigsen, G. (2020). Task-dependent recruitment of modality-specific and multimodal regions during conceptual processing. Cerebral Cortex, in press. doi: doi: 10.1093/cercor/bhaa010 Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago, USA: University of Chicago Press. Lakoff, G., & Johnson, M. (1999). Philosophy in the flesh: The embodied mind and its challenge to Western thought. New York: Basic Books. Levelt, W. J., Roelofs, A., & Meyer, A. S. (1999). A theory of lexical access in speech production. Behavioral and Brain Sciences, 22(1), 1-38; discussion 38-75. Lewis, M. L., & Frank, M. C. (2016). The length of words reflects their conceptual complexity. Cognition, 153, 182-195. doi: 10.1016/j.cognition.2016.04.003 Louwerse, M. M. (2011). Symbol interdependency in symbolic and embodied cognition. Topics in Cognitive Science, 3(2), 273-302. doi: 10.1111/j.1756-8765.2010.01106.x Lynott, D., & Connell, L. (2013). Modality exclusivity norms for 400 nouns: The relationship between perceptual experience and surface word form. Behavior Research Methods, 45(2), 516-526. doi: 10.3758/s13428-012-0267-0 Machery, E. (2007). Concept empiricism: A methodological critique. Cognition, 104, 19-46. Mahon, B. Z. (2015a). The burden of embodied cognition. Canadian Journal of Experimental Psychology-Revue Canadienne De Psychologie Experimentale, 69(2), 172-178. doi: 10.1037/cep0000060 Mahon, B. Z. (2015b). What is embodied about cognition? Language Cognition and Neuroscience, 30(4), 420-429. doi: 10.1080/23273798.2014.987791 Mahon, B. Z., & Caramazza, A. (2009). Concepts and categories: A cognitive neuropsychological perspective. Annual Review of Psychology, 60, 27-51. Marschark, M., & Paivio, A. (1977). Integrative processing of concrete and abstract sentences. Journal of Verbal Learning & Verbal Behavior, 16, 217-231. Martin, A., & Chao, L. L. (2001). Semantic memory and the brain: Structure and processes. Current Opinion in Neurobiology, 11, 194-201. 136 M. Kiefer, M. Harpaintner

Martin, A., Wiggs, C. L., Ungerleider, L. G., & Haxby, J. V. (1996). Neural correlates of category-specific knowledge. Nature, 379, 649-652. Mason, R. A., & Just, M. A. (2016). Neural representations of physics voncepts. Psychological Science, 27(6), 904-913. doi: 10.1177/0956797616641941 Mazzuca, C., Lugli, L., Benassi, M., Nicoletti, R., & Borghi, A. M. (2018). Abstract, emotional and concrete concepts and the activation of mouth-hand effectors. Peerj, 6. doi:10.7717/Peerj.5987 McClelland, J. L., & Rogers, T. T. (2003). The parallel distributed processing approach to semantic cognition. Nature Reviews Neuroscience, 4(4), 310-322. McRae, K., & Cree, G. S. (2002). Factors underlying category-specific semantic deficits. In E. M. E. Forde & G. W. Humphreys (Eds.), Category-specificity in mind and brain (pp. 211-249). East Sussex, U.K.: Psychology Press. Meteyard, L., Cuadrado, S. R., Bahrami, B., & Vigliocco, G. (2012). Coming of age: A review of embodiment and the neuroscience of semantics. Cortex, 48(7), 788-804. doi: 10.1016/j.cortex.2010.11.002 Moseley, R., Carota, F., Hauk, O., Mohr, B., & Pulvermuller, F. (2012). A role for the motor system in binding abstract emotional meaning. Cerebral Cortex, 22(7), 1634-1647. doi: 10.1093/cercor/bhr238 Paivio, A. (1986). Mental representations: A dual coding approach. New York: Oxford University Press. 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(12), 976-987. Patterson, K., & Ralph, M. A. L. (2016). The hub-and-spoke hypothesis of semantic memory. In G. Hickok & S. L. Small (Eds.), Neurobiology of Language (pp. 765-775). Burlington: Academic Press. Pecher, D., Boot, I., & Van Dantzig, S. (2011). Abstract concepts: Sensory-motor grounding, metaphors, and beyond. In B. Ross (Ed.), The psychology of learning and motivation (Vol. 54, pp. 217-248). Burlington: Academic Press. 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, 122(Pt 12), 2337-2344. 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(8), 1407-1419. Pobric, G., Jefferies, E., & Lambon Ralph, M. A. (2010). Amodal semantic representations depend on both anterior temporal lobes: Evidence from repetitive transcranial magnetic stimulation. Neuropsychologia, 48(5), 1336-1342. Popp, M., Trumpp, N. M., & Kiefer, M. (2019a). Processing of action and sound verbs in context: An fMRI study. Translational Neuroscience, 10, 200-222. doi: 10.1515/tnsci-2019-0035 Popp, M., Trumpp, N. M., Sim, E. J., & Kiefer , M. (2019b). Brain activation during conceptual processing of action and sound verbs. Advances in Cognitive Psychology, 15(4), 236-255. doi: 10.5709/acp-0272-4 Pulvermüller, F. (2005). Brain mechanisms linking language and action. Nature Reviews: Neuroscience, 6(7), 576-582. Pulvermüller, F. (2018). Neurobiological mechanisms for semantic feature extraction and conceptual flexibility. Topics in Cognitive Science, 10(3), 590-620. doi: 10.1111/tops.12367 Pulvermüller, F., Cooper-Pye, E., Dine, C., Hauk, O., Nestor, P., & Patterson, K. (2010). The word processing deficit in Semantic Dementia: All categories are equal but some categories are more equal than others. Journal of Cognitive Neuroscience, 22(9), 2027-2041. Pulvermüller, F., & Fadiga, L. (2010). Active perception: Sensorimotor circuits as a cortical basis for language. Nature Reviews Neuroscience, 11(5), 351-360. Pulvermüller, F., Hauk, O., Nikulin, V. V., & Ilmoniemi, R. J. (2005). Functional links between motor and language systems. European Journal of Neuroscience, 21(3), 793-797. Pylyshyn, Z. W. (1984). Computation and cognition: Towards a foundation for cognitive science. Cambridge: MIT Press. Quillian, M. R. (1969). The teachable language comprehender. Communications of the ACM, 12, 459-476. Ralph, M. A. L., Jefferies, E., Patterson, K., & Rogers, T. T. (2017). The neural and computational bases of semantic cognition. Nature Reviews Neuroscience, 18(1), 42-55. doi: 10.1038/nrn.2016.150 Rogers, T. T., Lambon Ralph, M. A., Garrard, P., Bozeat, S., McClelland, J. L., Hodges, J. R., & Patterson, K. (2004). Structure and deterioration of semantic memory: A neuropsychological and computational investigation. Psychological Review, 111(1), 205-235. Rosch, E. (1973). Natural categories. Cognitive Psychology, 4, 328-350. Rosch, E. (1975). Cognitive representations of semantic categories. Journal of Experimental Psychology: General, 104, 192-233. Sabsevitz, D. S., Medler, D. A., Seidenberg, M., & Binder, J. R. (2005). Modulation of the semantic system by word imageability. Neuroimage, 27, 188-200. 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, doi: 10.3389/ fnhum.2013.00125. Schwanenflugel, P. J., & Akin, C. E. (1994). Developmental trends in lexical decisions for abstract and concrete words. Reading Research Quarterly, 29, 251-264. Shebani, Z., & Pulvermüller, F. (2013). Moving the hands and feet specifically impairs working memory for arm- and leg-related action words. Cortex, 49(1), 222-231. doi: DOI 10.1016/j.cortex.2011.10.005 Varieties of abstract concepts and their grounding in perception or action 137

Simmons, W. K., & Barsalou, L. W. (2003). The similarity-in-topography principle: Reconciling theories of conceptual deficits. Cognitive Neuropsychology, 20, 451-486. Simmons, W. K., Martin, A., & Barsalou, L. W. (2005). Pictures of appetizing foods activate gustatory cortices for taste and reward. Cerebral Cortex, 15(10), 1602-1608. Simmons, W. K., Ramjee, V., Beauchamp, M. S., McRae, K., Martin, A., & Barsalou, L. W. (2007). A common neural substrate for perceiving and knowing about color. Neuropsychologia, 45(12), 2802-2810. doi: 10.1016/j.neuropsychologia.2007.05.002 Strack, F., Martin, L. L., & Stepper, S. (1988). Inhibiting and facilitating conditions of the human smile - a nonobtrusive test of the facial feedback hypothesis. Journal of Personality and Social Psychology, 54(5), 768-777. doi: Doi 10.1037/0022-3514.54.5.768 Thompson-Schill, S. L., D’Esposito, M., & Kan, I. P. (1999). Effects of repetition and competition on activity in left prefrontal cortex during word generation. Neuron, 23, 513-522. Trumpp, N. M., & Kiefer, M. (2018). Functional reorganization of the conceptual brain system after deafness in early childhood. PLoS ONE, 13(7). doi: ARTN e0198894 10.1371/journal.pone.0198894 Trumpp, N. M., Kliese, D., Hoenig, K., Haarmaier, T., & Kiefer , M. (2013a). Losing the sound of concepts: Damage to auditory association cortex impairs the processing of sound-related concepts. Cortex, 49, 474-486. doi: 10.1016/j.cortex.2012.02.002 Trumpp, N. M., Traub, F., & Kiefer, M. (2013b). Masked priming of conceptual features reveals differential brain activation during unconscious access to conceptual action and sound information. PLoS ONE 8(5), e65910. doi: 10.1371/journal.pone.0065910 Trumpp, N. M., Traub, F., Pulvermüller, F., & Kiefer, M. (2014). Unconscious automatic brain activation of acoustic and action-related conceptual features during masked repetition priming. Journal of Cognitive Neuroscience, 26(2), 352-364. doi: 10.1162/ jocn_a_00473 Tschentscher, N., Hauk, O., Fischer, M. H., & Pulvermuller, F. (2012). You can count on the motor cortex: Finger counting habits modulate motor cortex activation evoked by numbers. Neuroimage, 59(4), 3139-3148. doi: 10.1016/j.neuroimage.2011.11.037 Tulving, E. (1972). Episodic and semantic memory. In E. Tulving & W. Donaldson (Eds.), Organization of memory (pp. 381-403). New York: Academic Press. Tyler, L. K., & Moss, H. E. (2001). Towards a distributed account of conceptual knowledge. Trends in Cognitive Sciences, 5, 244-252. Ulrich, M., Hoenig, K., Grön, G., & Kiefer, M. (2013). Brain activation during masked and unmasked semantic priming: Commonalities and differences. Journal of Cognitive Neuroscience, 25(12), 2216-2229. doi: 10.1162/jocn_a_00449 van Dam, W. O., van Dijk, M., Bekkering, H., & Rueschemeyer, S. A. (2012). Flexibility in embodied lexical-semantic representations. Human Brain Mapping, 33(10), 2322-2333. doi: 10.1002/hbm.21365 van Dantzig, S., Cowell, R. A., Zeelenberg, R., & Pecher, D. (2011). A sharp image or a sharp knife: norms for the modality-exclusivity of 774 concept-property items. Behavior Research Methods, 43(1), 145-154. doi: 10.3758/s13428-010-0038-8 Vermeulen, N., Corneille, O., & Niedenthal, P. M. (2008). Sensory load incurs conceptual processing costs. Cognition, 109(2), 287-294. 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(7), 1767-1777. doi: 10.1093/cercor/bht025 Vukovic, N., Feurra, M., Shpektor, A., Myachykov, A., & Shtyrov, Y. (2017). Primary motor cortex functionally contributes to language comprehension: An online rTMS study. Neuropsychologia, 96, 222-229. doi: 10.1016/j.neuropsychologia.2017.01.025 Wagenmakers, E. J., Beek, T., Dijkhoff, L., Gronau, Q. F., Acosta, A., Adams, R. B., . . . Zwaan, R. A. (2016). Registered replication report: Strack, Martin, & Stepper (1988). Perspectives on Psychological Science, 11(6), 917-928. doi: 10.1177/1745691616674458 Wagner, A. D., Paré-Blagoev, E. J., Clark, J., & Poldrack, R. A. (2001). Recovering meaning: Left prefrontal cortex guides controlled semantic retrieval. Neuron, 31, 329-338. 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(10), 1459-1468. doi: 10.1002/hbm.20950 Warrington, E. K., & McCarthy, R. (1987). Categories of knowledge. Brain, 110, 1273-1296. Weiskopf, D. A. (2010). Embodied cognition and linguistic comprehension. Studies In History and Philosophy of Science Part A, 41(3), 294-304. Wiemer-Hastings, K., Krug, J., & Xu, X. (2001). Imagery, context availability, contextual constraint and abstractness. In J. D. Moore & K. Stenning (Eds.), Proceedings of the Annual Conference of Cognitive Science Society (pp. 1106-1111). Hillsdale, NJ: Lawrence Erlbaum Associates. Wilson-Mendenhall, C. D., Barrett, L. F., Simmons, W. K., & Barsalou, L. W. (2011). Grounding emotion in situated conceptualization. Neuropsychologia, 49(5), 1105-1127. 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(6), 920-935. doi: 10.1162/jocn_a_00361