Research
How do we understand a sentence? For decades the standard answer was: by decoding words into propositions and reasoning over them. A growing body of work tells a richer story. Meaning is also built by the body and by the imagination (Lakoff & Johnosn 1980; Turner 1996; Gallese 2025) – and what happens in a second language reveals this with particular clarity (Monaco et al. 2019).
Classical pragmatics, built on Grice’s notion of speaker meaning (Grice 1957), treats communication as the recovery of propositions. The listener infers what the speaker intended, and that intended content has a clear, sentence-like shape. This picture works well for everyday exchanges. But much of language resists this neat treatment.
A vivid description, a line of poetry, or a fresh metaphor does more than transmit a proposition: it evokes images, sensations and feelings that are hard to paraphrase (Pilkington 2000; Wilson & Carston 2019; De Saussure & Wharton 2022).
Two theoretical currents have pushed this point. Embodied Cognition argues that understanding words reactivates the same sensory and motor brain systems we use to act and perceive (Gallese & Lakoff 2005; Garello et al. 2024a). Work on mental imagery argues that comprehension can summon quasi–perceptual representations – we seem to “see” what words describe (Nanay 2023; Garello 2024). Metaphor is the ideal test case: it uses concrete, bodily words to convey abstract ideas – we grasp a concept, we break someone’s heart. As Aristotle observed in the Rhetoric, metaphor “puts things before the eyes” (Cuccio et al. 2013). If any use of language draws on the body and the imagination, metaphor should (Ricoeur 1975).
The first study we review (Garello et al. 2024a, 2024b) tested this directly. Forty native Italian speakers judged whether sentences made sense immediately after watching a short video of a hand performing an action – either matching the verb or not. The sentences came in three conditions: literal (“the father breaks bread”), metaphorical (“the girl breaks her boyfriend’s heart”) and idiomatic (“to break a lance in favour of someone”). For literal and metaphorical sentences, a matching action video sped up comprehension. For idiomatic sentences, it made no difference. The interpretation is that literal and metaphorical sentences keep a live link to the motor meaning of the verb, so a related action primes them; idioms have hardened into single units whose meaning no longer touches the verb’s bodily origin.
This graded picture – strong motor involvement for concrete language, weaker for abstract, none for frozen idioms – was further confirmed by fMRI data: the same decrease from literal to metaphorical to idiomatic sentences appears in patterns of motor cortex activation (Ferroni et al. under review). Strikingly, novel metaphors showed a subtler effect: a matching video actually slowed comprehension of fresh expressions, as if the motor system stayed active too long while the mind worked out a new meaning. And in a related fMRI study (Ferroni et al. under review), novel metaphor comprehension was associated with activation of occipital, visual areas – a signature of mental imagery being recruited when processing is effortful (Carston 2018). Together, the behavioural and neuroimaging data converge on the same picture: bodily simulation and imagery are most visible precisely when language is new, poetic or affectively dense. These findings point toward what we call a multimodal pragmatics: the mind integrates several formats of representation – propositional, motor, perceptual, affective – and the question is no longer whether non–propositional content matters, but under what conditions and through which channels it shapes understanding (Garello 2025).
If meaning is partly built by the body, a natural question follows: what happens in a language we learned later, in a classroom rather than through lived experience? Monaco et al. (2019) survey the evidence on embodiment in a second language. Their headline findings are both reassuring and nuanced. Across behavioural and neurophysiological studies, the sensorimotor system is involved when people process their L2. The second language is embodied – but in most studies, to a lesser degree or in a different pattern than the first. A first language is grounded in years of bodily experience; a second language is often learned through words alone, mediated by the first.
We replicated our original experiment with the same Italian speakers processing sentences in English (all upper–intermediate, B2 learners who had acquired English after age ten in class). Three results stand out. Every condition was slower in English – the expected cost of a non-native language. More interestingly, the benefit of a matching action video was larger in English than in Italian for literal sentences: visual and motor cues help more when linguistic proficiency is lower. And the idiomatic condition behaved differently across languages: in Italian, idioms were unaffected by the video, stored as frozen units; in English, the same type of idioms were sped up by a matching action, because they had not yet frozen and were being processed compositionally, word by word. L2 speakers tended to treat idiomatic, metaphorical and literal sentences alike – parsing them all somewhat literally.
The emotional dimension adds a further layer. Ifantidou et al. (2019) found that emotionally loaded, imagistic paraphrases supported L2 metaphor comprehension just as effectively as plain literal restatements. Drawing on Carston’s theory, the authors argue that imagery and feeling are not decorative side–effects: when the literal meaning of a word is not fully grasped, they become a working route to meaning. Learners faced with an unknown word can lean on the mental picture and the emotional tone a sentence evokes, and arrive at a plausible interpretation (Rinaldi & Gallese 2025; Rinaldi et al. under review).
A question this body of work naturally raises is whether the context of acquisition – not only the level of proficiency – shapes the degree of embodiment. Our ongoing study addresses this directly by comparing two groups of learners with equivalent proficiency (Ferroni et al. in prep.; Garello et al. in prep.): one that acquired the second language in a naturalistic context (living in the target-language country) and one that acquired it in a formal classroom setting. Using the same experimental design as the original priming study, we find that the two groups show measurably different embodiment effects. Learners who acquired the language in context are significantly more sensitive to action priming – their comprehension is more robustly modulated by motor cues – than those who learned through classroom instruction alone. This dissociation suggests that what matters is not proficiency in the abstract, but the richness and variety of the embodied experiences through which a language was encountered. A word learned while performing an action, navigating a market, or participating in a social situation carries a different bodily trace than a word encountered on a page. The implications are substantial: two speakers at the same formal level may understand language through quite different underlying processes (Ibanez et al. 2023; Ferroni & Garello in prep.).
This last finding connects naturally to a broader philosophical point. Ludwig Wittgenstein argued in his later work that language does not float free of the world: words acquire their meaning through Sprachspiele – language games – embedded in Lebensformen, forms of life. To understand an expression is not merely to know its propositional content but to be at home in the practices, gestures and situations in which it lives. A word like break, in this view, means something slightly different depending on whether you have ever broken bread together with others, broken a machine with your hands, or only encountered the word in a grammar exercise. Our experimental data can be read as an empirical echo of this insight. The greater sensitivity to action priming among context–immersed learners is precisely what Wittgenstein’s picture would predict: language learned within a form of life retains the bodily and situational texture of that life (Carapezza 2019). Classroom learners, by contrast, have acquired a kind of linguistic knowledge that is more detached from lived practice – competent, but thinner in its sensorimotor grounding. Empirically, in this sense, we are measuring something Wittgenstein described philosophically: the degree to which a speaker has become a participant in the language game, not merely a decoder of its rules.
Read together, these studies tell a coherent story. Understanding language is not a single, purely verbal computation. It is a dynamic interplay of words, bodily simulation, mental images and emotions – and the weight of each ingredient shifts with the material, the speaker and the history through which that speaker came to know the language (Rinaldi et al. under review). The less proficient, or the less contextually grounded, the speaker, the more the non–propositional channels carry the load.
That last point is the bridge to practice. If a second language is real but only partly embodied, and if images and emotions help most exactly when words fall short, then teaching can be designed to supply what the learner’s still-thin linguistic knowledge cannot yet provide. Pairing words with action and gesture, using vivid imagery for abstract and metaphorical vocabulary, exploiting emotionally engaging material, treating idioms and metaphors differently, and above all creating opportunities for language to be encountered within genuine situations rather than in isolation – these are not stylistic choices but interventions grounded in what we now know about how the body reads language. The underlying message is simple: language learning is not only a matter of memorizing words and rules. It is also a matter of grounding meaning in the body, the senses and the world. A teaching practice that takes the embodied, imaginative side of comprehension seriously gives learners more ways in – and gives them the most help precisely when they need it most.
References:
Carapezza, M. (2019) “The Language Game of lost meaning: Using literal meaning as a metalinguistic resource” in Intercultural Pragmatics, vol. 16, n. 3, pp. 305-318.
Carston, R. (2018) “Figurative Language, Mental Imagery and Pragmatics”, in Metaphor and Symbol, 33, n. 3, 1-46.
Cuccio, V., Carapezza, M., Gallese, V. (2013) “Metafore che risuonano. Linguaggio e corpo tra filosofia e neuroscienze” in Senso e Sensibile. Prospettive tra estetica e filosofia del linguaggio, vol. VII, n. 17, pp. 69-74.
De Saussure L. & Wharton T. (2022) Pragmatics and emotion. Cambridge UP.
* Ferroni F., Garello S., Arcuri E., Ardizzi M., Gallese V. (under review) “Do I break the bread or the heart? An fMRI study on language comprehension with a Representational Similarity Analysis Technique”
* Ferroni F., Garello S., Gallese V. (in prep.) “Frontiers in Embodied Cognition and Language” in Trends in Cognitive Science.
* Ferroni F., Garello S., Rinaldi L., Gallese V. (in prep.) A tDCS study on second language comprehension: an Embodied Cognition perspective.
Gallese, V. (2025) “Language and bodily multimodality. The role of Embodied Simulation” in Sistemi intelligenti. Rivista quadrimestrale di scienze cognitive e di intelligenza artificiale (2), 165-182.
Gallese, V., Lakoff, G. (2005) “The Brain’s Concepts: The Role of the Sensory-Motor System in Conceptual Knowledge”. Cognitive Neuropsychology 22, 3-4 pp. 455–79.
Garello S. (2024) “The visibility of speech: an investigation into the relationship between metaphor and visual imagery” in Pragmatics & Cognition, 30 (2), 354–377 – John Benjamins.
Garello S. (2025) “Metafore, Embodied Simulation e Mental imagery: verso una pragmatica multimodale” in Rivista Italiana di Filosofia del Linguaggio (eds. M. De Palo, F. Diodato).
Garello S., Ferroni F., Gallese V., Cuccio V., Ardizzi M. (2024a) “From breaking bread to breaking hearts: Embodied simulation and action language” in Language, Cognition and Neuroscience 39 (4), 489–500 – Taylor and Francis.
Garello S., Ferroni F., Gallese V., Cuccio V., Ardizzi M. (2024b) “Embodied Cognition in L1 and L2: Towards an Embodied and Multimodal Perspective” in Scientific Reports 14, 1–10 – Nature.
* Garello S., Mezzadri M., Paita M.P., Ferroni F. (in prep.) “Towards an embodied approach to second language learning” in Brain & Language.
Grice, H. P. (1957) “Meaning”. The Philosophical Review 66: 377-88.
Ibanez et al. 2023
Ifantidou E. & Hatzidaki A. (2019) “Metaphor comprehension in L2: Meaning, images and emotions” in Journal of Pragmatics, 1491, 78-90.
Lakoff, G., Johnson, M. (1980a) Metaphors we Live by. University of Chicago Press. Monaco et al. 2019
Nanay B. (2023) Mental imagery. Oxford UP.Pilkington, A. (2000) Poetic Effects: A Relevance Theory Perspective. Amsterdam: John Benjamins.
Ricoeur, P. (1975) La métaphore vive. Paris: Seuil.
Rinaldi L. & Gallese V. (2025) “How the mouth became symbolic” in Trends in Cognitive Sciences.
* Rinaldi L., Ferroni F., Garello S., Gallese V. (under review) “From the hand to the mouth: iconicity, phonosymbolism and language” in Experimental Brain Research.
Turner M. (1996) The literary mind: the origins of thought and language. Oxford UP.
Wilson, D. & Carston, R. (2019) “Pragmatics and the challenge of non-propositional effects” in Journal of Pragmatics, vol. 145, pp. 31-38.
Wittgenstein, L. (1953) Ricerche Filosofiche. A cura di Trinchero, M. Torino: Einaudi (2009).
We indicated with “ * ” the papers published as a result of the Imminent Research Grant.
Stefana Garello
PhD in Philosophy of Language
Stefana Garello received her PhD in philosophy of language from the University of Palermo in 2022. She was a post-doctoral researcher at the Unit of Neuroscience of the University of Parma and received a research fellowship at the Centre for Philosophical Psychology of the University of Antwerp. In the past, she was visiting researcher at Leiden University and University College London. Her research interests include metaphors, pragmatics and the relationship between language and the non-propositional dimension of cognition.
Francesca Ferroni
Neurobiologist
Francesca Ferroni obtained her Master's degree in Neurobiology from the University of Pavia (Italy) in 2016. In 2021, she earned her PhD in Neuroscience from the University of Parma, Italy, where she began investigating how the brain constructs the experience of the body in space—known as peripersonal space—through the integration of multisensory information. She is currently a Postdoctoral Researcher at the Department of Medicine and Surgery, University of Parma. Her research focuses on individual differences in the representation and plasticity of peripersonal space in both healthy individuals and patients with schizophrenia. More broadly, her work aims to elucidate how alterations in multisensory integration contribute to disturbances of the bodily self, providing new insights into the neurobiological mechanisms underlying these phenomena.