Answers are grounded in this paper and cite their sources.
Recent experiment-based psycho- and neuro-linguistic research brought new insights into language processing mechanisms and meaning representation in the brain. More specifically, it highlighted the dynamic nature of brain connections and a constant interplay between distributed neuronal circuits during meaning processing. These developments led to a shift from an amodal view, which perceives conceptual information activation as parallel to and independent from adjoining neural activation in sensorimotor circuits (Mahon & Caramazza, 2008; Meteyard et al., 2012), to the embodied cognition view (EC) that highlights the role of sensorimotor experience in the formation of flexible, distributed conceptual representations encompassing features acquired via different perceptual modalities (Barsalou, 2010; Fischer & Zwaan, 2008). EC, therefore, suggests that conceptual knowledge and, consequently, semantic knowledge are grounded in bodily experience and situated actions (Glenberg et al., 2008; Pulvermüller, 2013) . A broad range of behavioral, physiological, and neuroimaging data demonstrating co-activation of language- and action-related brain areas support this claim with regard to concrete language (Binder et al., 2005; Pulvermüller et al., 2005; Barsalou, 2008; Hauk et al., 2008). However, the data are less conclusive with regard to figurative expressions, which constitute a significant part of language. One of the reasons is that figurative language subsumes a wide variety of heterogeneous phenomena (metonymy, idioms, metaphors, proverbs, hyperbole, irony) which differ syntactically (from phrasal verbs to compounds and even sentences), as well as in their properties (familiarity, ambiguity, transparency, compositionality, salience, predictability) and essential features (although both irony and hyperbole are based on cognitive contrast, it is a contrast of kind for irony and a contrast in magnitude for hyperbole; Hsiao & Lily, 2010). This diversity and complexity of non-literal language types does not allow for clear-cut and strictly defined boundaries; it led to a distinction of non-literal phenomena not dichotomously, but along a conventionality continuum (Cacciari & Papagno, 2012). Secondly, the linguistic phenomena, embraced by the broad term “non-literal language” have been analyzed to different degrees of detail. Specifically, the study of metaphor has dominated the field: metaphors being a powerful cognitive device guided by environmental experiences enable metaphor framing to profoundly influence not only linguistic communication per se, but also judgments, reasoning, intentions, and actions (Elmore & Luna-Lucero, 2017; Robins & Mayer, 2000; Marin et al., 2014; Thibodeau & Boroditsky, 2011; 2013; Hauser & Schwarz, 2015; Landau et al., 2014; Slepian et al., 2010; Thibodeau et al., 2017). Idioms, on the other hand, were for a long time considered secondary to metaphors and it was debated whether idiom processing is independent from constant recourse to conceptual metaphors (Owens, 2016). However, idioms are not uniform, they diverge both syntactically and compositionally (Caillies & Butcher, 2007) and are intrinsically different from metaphors. The main distinction is that, although some idioms are still partially motivated by conceptual mappings between domains (Gibbs, 1992), in general idioms are fixed and conventionalized expressions, the meaning of which should be learned, while metaphoric meanings are more flexible and intricate, and therefore require online construction (Bambini et al., 2016). Distinctive neural correlates for processing of idioms (left MTG and left IFG, involved in selection-inhibition operations) and metaphors (left precentral gyrus (BA 6), linking concrete and an abstract domains and the left inferior parietal lobe (IPL), executing higher-order cognitive motor functions; Fogassi & Luppino, 2005) also argue against conflating them.Disregarding these principal differences between the two linguistic forms results in their interchangeable use (e. g., Aziz-Zadeh et al., 2006), which, in turn, affects the findings, as comprehending these intrinsically different figurative devices engages dissimilar cognitive mechanisms: based on both semantic and structural analysis of meaning and retrieval from semantic memory during idiom processing, and focused on the conceptual models and templates underlying metaphor meaning construction.Recent studies on processing non-literal expressions with action-related semantics reported activation of motoric brain areas during either literal (Raposo et al., 209), metaphoric (Desai et al., 2011) or only during metaphor but not idiom processing (Cacciari et al., 2011; Desai et al., 2013). Only limited publications present evidence of sensorimotor engagement during idiomatic meaning comprehension (Boulenger et al., 2009; 2012). Overall, the studies emphasize the role of context in meaning disambiguation and suggest that an increase in abstractness of the language stimuli leads to a decrease in the sensorimotor system’s involvement (especially in case of idioms). However, these results could be interpreted not in favor of idiom disembodiment, but as a demonstration of different processing schemas that idioms and metaphors employ: the dual-reference idiomatic nature enables engagement of a hybrid processing mechanism that encompasses both compositional and holistic context-based analysis during idiom comprehension (Caillies & Butcher, 2007; Boulenger et al., 2012; Cacciari & Pesciarelli, 2013). Metaphors, which retain stronger links to the original meaning of the constituent words, rely more on online mental simulation to compute complex, flexible meanings. Engaging different processing mechanisms may result in spatially and temporally different patterns of neurocognitive involvement (Yang & Shu, 2016; Rapp et al., 2012). For example, Cacciari et al. (2011) reported no motor engagement in idiom processing, but single-pulse TMS applied at sentence ends to collect meaning-induced MEPs could be inefficient to record idiom-induced motor activation, since idioms are processed online (mentally simulated) only till the idiom is recognized and then a switch to the non-compositional mode (retrieval from semantic memory) occurs. Lack of motor engagement during idiom comprehension can be explained by heterogeneity of idioms: e.g., Raposo et al. (2009) used highly familiar and opaque idioms, which minimized the need for mental simulation during their processing and consequently may have reduced the level of sensorimotor cortical activation. Therefore, idiomatic meaning may be less embodied, compared to metaphoric meaning, but not totally disembodied.This evidence highlights the need for a more profound exploration of properties specific to figurative language types and subtypes of each phenomenon, which could considerably benefit present-stage figurative language research and promote a better understanding of the mechanisms the human brain employs for their acquisition, production and processing. This will provide an integrative theoretical model that can more comprehensively and consistently outline the cognitive mechanisms and neural circuitry underlying processing of heterogeneous and multifaceted figurative language. Taken together, it will inform the development of more precise neuro-cognitive models, support AI applications and enhance understanding of language processing in general.
No takes yet. Share an insight, caveat, or question.
Kulkova et al. (2019) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: