Open research questions in Language, Metaphor, and Cognition

What the literature leaves open

• approach to behavior. In Handbook of Cognitive Science 159–187 (Elsevier, 2008). 46. Allen, M. & Friston, K. J. From cognitivism to autopoiesis: towards a computational framework for the embodied mind. Synthese 195, 2459–2482 (2018). 47. Chemero, A. Radical embodied cognitive science. Rev. Gen. Psychol. 17, 145–150 (2013). 48. García-Barriocanal, E., Sicilia, M. A. & Lytras, M. D. Methodology for the definition of a glossary in a collaborative research project and its application to a European network of excellence. Int. J. Knowl. Based Intell. Eng. Syst. 15, 213–222 (2011). 49. Beatty, A. S. & Kaplan, R. M. (eds) Ontologies in the Behavioral Sciences: Accelerating Research and the Spread of Knowledge (National Academies Press, 2022). 50. Stute, K. et al. The fNIRS glossary project: a consensus-based resource for functional near-infrared spectroscopy terminology. Neurophotonics 12, 027801 (2025). 51. Grys, M., Casciano, R. & Pieniazek, I. MSR56 Comparison of AI- enhanced tools for automating scientific literature reviews. Value Health 28, S284 (2025). 52. Devi, A., Barooah, P. K. & Ahmed, Z. Enhancing literature review through AI-based research tools: a comparative study of SciSpace and Semantic Scholar. TIJER Int. Res. J. 11, 779–785 (2024). 53. Stein, S. et al. Toward more ethical engagements between Western and Indigenous sciences. Facets 9, 1–14 (2024). 54. Butterworth, B. The development of arithmetical abilities. J. Child Psychol. Psychiatry 46, 3–18 (2005). 55. Penner-Wilger, M. & Anderson, M. L. The relation between finger gnosis and mathematical ability: why redeployment of neural circuits best explains the finding. Front. Psychol. 4, 877 (2013). 56. Armitage, K. L. & Gilbert, S. J. The nature and development of cognitive offloading in children. Child Dev. Perspect. 19, 108–115 (2025). 57. Barrocas, R., Roesch, S., Gawrilow, C. & Moeller, K. Putting a finger on numerical development—reviewing the contributions of kindergarten finger gnosis and fine motor skills to numerical abilities. Front. Psychol. 11, 1012 (2020). 58. Moeller, K., Martignon, L., Wessolowski, S., Engel, J. & Nuerk, H. C. Effects of finger counting on numerical development—the opposing views of neurocognition and mathematics education. Front. Psychol. 2, 328 (2011). 59. Ranzini, M. et al. Unveiling the sensorimotor basis of numerical processing: a functional near-infrared spectroscopy (fNIRS) study. NeuroImage 324, 121576 (2025). 60. Alibali, M. W. & DiRusso, A. A. The function of gesture in learning to count: more than keeping track. Cogn. Dev. 14, 37–56 (1999). 61. Butterworth, B. A head for figures. Science 284, 928–929 (1999). 62. Soylu, F., Lester, F. K. Jr & Newman, S. D. You can count on your fingers: the role of fingers in early mathematical development. J. Numer. Cogn. 4, 107–135 (2018). zdzb2014), the Shenzhen Science and Technology Innovation Commission Foundation (grant no. 202307313000096), the Social Science Foundation from China’s Ministry of Education (grant no. 23YJA880093), the Science and Technology Innovation Projects of the State General Administration of Sports (23KJCX057), the China Postdoctoral Science Foundation (grant no 2022 M711174), the National Center for Mental Health Foundation (grant no. Z014), Research Excellence Scholarships of Shenzhen University (grant no. ZYZD2305), Research Funding for Society of Sport Science (grant no PT2023030), the Natural Science Foundation of Shenzhen University (grant no. 000311), and the Guangdong Youth Health Research Fund (grant no 2024WT006). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

  Toward a unified vocabulary for embodiment research challenges and solutions · 2026 · DOI

• A shared vocabulary removes that barrier, letting us focus on the science itself. In summary, the ideas encapsulated by these many terms are too important to remain siloed and semantically entangled. The recognition that cognition is not just “in the head” but in dynamic concert with body and world is transforming multiple disciplines. To accelerate this transforma- tion, we must also advance scientific communication. Let us strive for the three Cs—clear, concise, consistent terminology—so that when different experts discuss “embodiment”, we can translate across disciplinary dialects and state clearly what we mean. In a decade when interdisciplinary colla- boration and public engagement are highly valued, having a shared, inter- operable language for these concepts, including to be transparent about underlying theoretical commitments, is not just an academic ideal, but also a imperative for progress in theory and practice.

  Toward a unified vocabulary for embodiment research challenges and solutions · 2026 · DOI

• audience (e.g., educators or the general public). For instance, a technical term such as “embodied pedagogical dynamics” can be accompanied by an explanatory gloss (e.g., “how bodily action and perception are intentionally leveraged during instruction to support learning”), while keeping “embodiment” for the construct-level claim it often carries (e.g., grounding of cognition in perception–action systems rather than mere movement). This two-layer approach supports translation beyond academia while preserving conceptual commitments. For instance, defining ‘embodiment’ as an umbrella term for all body-mind integration, with sub-terms like ‘embodied (body-focused)’ and ‘embedded (environmentfocused)’, builds on the 4E cognition framework. Concise labels. Science often deals with complex ideas, but our labels should aim for brevity without changing the construct. Conciseness should not come at the expense of precision, especially when a term signals theoretical assumptions (e.g., whether embodiment is treated as grounded in perception–action and neural systems, versus describing physical activity in general). A useful compromise is to (1) choose shorter, self-descriptive labels within scholarly discourse (e.g., “embodimentinformed instruction” instead of “embodied pedagogical dynamics”), and (2) provide a secondary, practice-facing descriptor when communicating with practitioners (e.g., “learning supported by structured gesture and action”), while explicitly linking that descriptor to the technical construct in the manuscript. This preserves theoretical nuance for researchers while improving accessibility for non-specialist audiences. Consistent usage. Use terms consistently: define them at first use, align with established frameworks, and avoid scope drift across papers. Rely on shared glossaries and clear author guidelines to stabilize meaning and train newcomers. When theories truly differ, give them distinct labels rather than overloading the term “embodied cognition” may be appropriate. Implementation can be supported at the journal level. To improve discoverability and evidence synthesis, journals could also encourage authors to align title/abstract keywords with the defined terms used in the manuscript. For interdisciplinary submissions, journals could require a brief “Key Terms Definition” table (or boxed glossary) that lists how central terms (e.g., embodiment, grounded cognition, enactivism, ecological psychology) are defined and operationalized in the manuscript, with notes for any intended scope (umbrella term vs. specific framework). Such a requirement would not police theoretical positions; rather, it would make authors’ usage explicit, help reviewers evaluate construct validity, and reduce unintentional scope drift across papers. We suggest that adopting the 3 Cs in this field can make it easier to line up studies side by side to see what patterns emerge.

  Toward a unified vocabulary for embodiment research challenges and solutions · 2026 · DOI

• Box 1 | Finger use in mathematics as a cross-disciplinary case study A single learning behavior—using fingers to support early mathematics— illustrates how terminological siloing obscures both mechanisms and educational implications. Across fields, the phenomenon is indexed under different keywords that rarely cross-reference: (1) Neuroscience: “finger gnosis,” “somatosensory activation,” and parietal–sensorimotor networks implicated in number processing54,55; (2) Developmental psy- chology: “sensorimotor grounding,” “off-loading,” or the “extended mind” (fingers as external cognitive supports56); (3) Education: “finger counting,” “manipulatives,” and classroom debates about whether finger-based strategies should be encouraged or phased out57,58. These labels often point to complementary questions. Neuroscience work asks how finger representation and parietal–sensorimotor circuitry relate to numerical cognition across development59. Developmental psychology asks when fingers reduce working-memory demands, scaffold strategy acquisition, or shape internal number representations60,61. Education research asks what teachers should do: allow finger use, teach structured finger patterns, or discourage it to promote “mental” calculation62. When these literatures are separated by vocabulary, synthesis suffers. A review of ‘embodied math’ may under-sample neuroscientific evidence on developmental constraints, while a meta-analysis of “finger counting interventions” may miss mechanistic work that clarifies why effects vary by age, task type, or learner profile. A shared vocabulary—or a mapped keyword taxonomy—would make it easier to integrate these perspec- tives into testable predictions (what should change, for whom, and when) and into actionable guidance (when to encourage finger use, how to scaffold transition, and which outcomes to monitor). progress. This deceleration is particularly critical given the field’s trajectory. As Fig. 1 indicates, embodiment research is not a passing trend but has established a steady state of high-volume output over the last decade. Consequently, the terminology barrier is not just a nuisance for a fledgling field, but a structural brake on a permanent, maturing area of scientific inquiry. Scholars may spend effort debating definitions or defending ter- minological turf instead of building a clear terminology on each other’s insights. This is also evident when scholars operating outside mainstream Western academic traditions are not offered an appropriate participation in the conversation, for instance, indigenous knowledge systems, such as Māori or African philosophies, where body-environment integration is central and could offer important insights28. If we treat “consensus” as mere simplification, we risk erasing precisely these subtle, yet important, termi- nological differences. Thus, any effort to develop a shared vocabulary should be deliberately inclusive and explicit about the epistemic provenance of terms. The rise of AI (artificial intelligence)-assisted scholarship heightens this urgency29. As LLM (large language model)-based search, recommendation, and summarization tools increasingly mediate what researchers and stu- dents read30, fragmented vocabularies can distort the evidence base in two itself becomes biased31: term-dependent queries ways. First, retrieval (whether typed by humans or generated by AI agents) are more likely to surface work from the user’s home discipline, while omitting adjacent lit- eratures that use different keywords for the same construct. Second, we hypothesize that when models are asked to ‘compare’ or ‘contrast’ research fields, their documented sensitivity to meaning-preserving wording choices may lead them to overinterpret superficial naming differences—producing spurious oppositions between near-synonyms or treating historical labels as distinct mechanisms rather than overlapping32 traditions33. This has practical consequences for several research fields, especially the science of learning. A student preparing a literature review, a teacher exploring a movement-based intervention, or a policymaker scanning ‘what works’ reports may receive AI-generated syntheses that under-integrate neuroscience, psychology, and education—not because the evidence is incompatible, but because it is indexed inconsistently. A shared vocabulary or, at a bare minimum, transparent thesaurus mapping of keywords to shared parent concepts can be seen as a worthwhile infrastructure inter- vention that improves human discoverability while also providing machine- readable anchors that help future AI tools synthesize across silos more faithfully34.

  Toward a unified vocabulary for embodiment research challenges and solutions · 2026 · DOI

• Fig. 2 | Terminological spread across eight scientific disciplines in embodiment- related research (2015–2025). Searches were conducted in Web of Science and Scopus, and term inclusion was based on qualitative synthesis rather than a fixed frequency threshold. Terms were selected based on their theoretical significance and prevalence in foundational texts, recent reviews, and widely-cited studies within each discipline, as identified through qualitative synthesis by the multidisciplinary author team. when proceeding with an ambiguous usage of terminology, we risk building parallel silos of knowledge that do not connect—a form of academic “fragmentation” that leaves the full picture of embodiment obscure and hinders the unfolding of its promising potential in different, highly relevant application scenarios, such as education or public health. Taken together, the examples above show that inconsistent labels do more than slow discovery and complicate meta-analysis; they also blur the mechanisms the field is trying to explain. Work described as “grounded cognition,” “sensorimotor recycling,” “motor cognition,” or “embodied learning” often converges on a shared neurocognitive story: Circuits ori- ginally evolved for perception and action—particularly sensorimotor and parietal networks—are recruited during abstract conceptual processing, supporting what many frameworks describe as neural reuse23,24. When different scientific vocabulary challenge the efficient synthesis of research findings, it becomes harder (i) to treat corresponding observations as cumulative evidence, (ii) to compare candidate pathways (e.g., motor imagery versus online action planning), and (iii) to test when embodiment is a causal driver of learning rather than a by-product of task demands. This matters for practically relevant application scenarios, such as education, because mechanistic clarity is what allows researchers and practitioners to specify what an “embodied” intervention is intended to change, predict boundary conditions (e.g., age, prior knowledge, modality, and classroom constraints), and select measures that link observable classroom behavior to underlying neurocognitive processes. Over time, a shared terminology could also support the development of biomarkers or neurocognitive proxies for effective embodied pedagogy, helping distin- guish interventions that truly engage targeted sensorimotor resources from those that merely increase attention, novelty, or arousal. For example, semantic integration approaches that map standardized clinical measures to shared ontologies have enabled deep phenotyping and supported identifi- cation of known biomarkers in electronic health record data25. Figure 2 visualizes how closely related ideas disperse across fields under different labels, setting up the concrete cross-disciplinary tracer example that follows. To make this terminological spread tangible, and to show why it is more than a search inconvenience, Box 1 traces one familiar learning phenomenon, finger use in mathematics, as it appears under different keywords across neuroscience, psychology, and education. The case study highlights how the same behavior can be treated as a neural mechanism, a cognitive strategy, or an instructional practice depending on disciplinary entry point, and how those framing can be productively aligned when the scientific vocabulary allows cross-communication. Seen this way, vocabu- lary is not merely descriptive; it is a coordination tool. When terms diverge, researchers may unknowingly reinvent methods, compare non-equivalent interventions, or misinterpret null results as theoretical disagreement rather than definitional mismatch26. In the next section, we argue that these local inefficiencies compound into system-level slowdowns for interdisciplinary progress.

  Toward a unified vocabulary for embodiment research challenges and solutions · 2026 · DOI

• The findings of this study highlight significant implications for the development and implementation of the NCTLL. Teachers’ metaphorical perceptions reveal both strengths and weaknesses that can guide curriculum developers and policymakers. To address the disconnect between the curriculum and teachers’ expectations, a more flexible approach should be adopted, incorporating teacher input to enhance engagement. The complexity of the curriculum underscores the need for more professional development and practical resources to support teachers in its effective implementation. Additionally, many teachers perceive little difference between the new and old curriculum, suggesting that future revisions should introduce more noticeable innovations. The metaphor of a “building without solid foundations” points to the necessity for better resources, including updated teaching materials and technology integration. Lastly, establishing regular feedback channels will allow teachers to share their insights, fostering the curriculum’s adaptability and improving its overall effectiveness. The findings suggest several directions for further research into teachers’ metaphorical perceptions of the NCTLL. However, metaphors reflect only specific aspects of a concept, limiting their ability to capture its full scope (Lakoff & Johnson, 1980). Therefore, the current study’s results should be understood within this context. Future studies could involve larger samples and diverse sampling methods to improve generalizability. Longitudinal research could explore how teachers’ perceptions evolve over time, offering insights into the curriculum’s long-term effectiveness. Additionally, examining the relationship between teachers’ metaphorical perceptions and their pedagogical practices could reveal how these perceptions impact classroom implementation. Comparative studies across different subjects or curricula could further inform curriculum development by broadening our understanding of how teachers interpret various curriculum frameworks.

  Metaphorical Perceptions of Turkish Language and Literature Teachers on the Renewed Curriculum: The Century of Türkiye Education Model · 2026 · DOI

• This study innovates methodologically by integrating elicited metaphor with CVT in an under-researched policy context (Kurikulum Merdeka), using age-appropriate prompts and dual coding to detect ambivalence.

  ‘Learning English is like being swallowed by a whale … ': a CVT-metaphor approach to Indonesian children’s motivation · 2026 · DOI

• Future research could address these limitations by including native speakers as a control group, recruiting participants from diverse L1 backgrounds, incorporating neuroscientific methods to examine the neural basis of embodied semantic processing in foreign-language learners, and increasing sample sizes and the number of items to improve sensitivity to higher-order interactions.

  The role of vocabulary concreteness in modulating embodied semantic representations in learners of Chinese as a foreign language · 2026 · DOI

• We recognize that the size of the dataset can be seen as a limitation: the manually curated subset consists of 300 instances, while the automatically generated multimodal subset includes only 30 oxymorons. This reflects a deliberate choice to prioritize qual- ity over quantity, with the primary goal of providing a replicable methodology for systematically constructing oxymorons and investigating an underexplored linguistic phenomenon, and it is consistent with previous work on antonyms and oxymorons. Nonetheless, the limited size remains a constraint. In future work, we plan to address this by extending the dataset, evaluating a larger number of examples, and ultimately moving toward the creation of a large-scale training resource. Our experiments are conducted with a single proprietary LLM (GPT-4o-mini). This choice was motivated by computational constraints, but we acknowledge that including a broader mix of open-source and proprietary models with varying parameter sizes and language coverage would yield a more comprehensive evaluation. At the same time, we believe that the proposed methods should be robust across model families, sizes, and languages. We highlight both limitations as concrete directions for future work. 32 l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / c o l i / l a r t i c e - p d f / d o i / . 1 0 1 1 6 2 C O L / I . . / a 5 8 6 2 5 7 1 0 2 8 / c o l i . . a 5 8 6 p d . f I I b y B B L O T E C A F E D E R A T A M E D C N A u s e r o n 0 9 F e b r u a r y 2 0 2 6 I I Computational Linguistics Just Accepted MS. https://doi.org/10.1162/COLI.a.586 © 2025 Association for Computational Linguistics Published under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) license Di Palma & Rizzi et al.

  Multimodal OXYmorons: A Comprehensive Introduction and Computational Analysis Using a Dataset of Oxymoronic Memes in Italian and Spanish · 2026 · DOI

• Organization-level entities within the ontology may bias the model toward metonymic interpretation even when usage remains literal, suggesting edge-type sensitivity as a potential research avenue.

  Graph convolution-based techniques for pragmatic Arabic figurative language classification · 2026 · DOI

• Originality: Few studies have explored how second language learners represent concepts that are encoded differently across languages, or whether advanced learners retain native-like conceptual patterns after attaining high L2 proficiency.

  Conceptual transfer in the mind of Chinese English learners: Spatial metaphorical concepts of kin terms · 2025 · DOI

• Although the role of the DLPFC in idiom comprehension has been extensively studied, ambiguous idioms—those with more than one acceptable meaning—have been relatively understudied in the neuroscientific literature.

  Comprehension of Ambiguous Idioms in Prefrontal Cortex: Evidence from rTMS · 2024 · DOI

• Abstract There have been limited studies comparing Chinese and American news media in their descriptions of vaccines, especially their use of metaphors in the context of COVID-19.

  The use of metaphors to construct crisis discourses in describing COVID-19 vaccines in the Chinese and the American news media · 2023 · DOI

• The authors bring new knowledge to the international community from these understudied languages, which may become inaccessible in the not too distant future, particularly those from oral sources, given Ghana’s political embrace of neoliberal global flows of people, goods, and information which expands the reaches of language shift.

  Re-thinking everyday metaphors through Indigenous Ghanaian languages · 2021 · DOI

• By adopting this approach, I contribute, first, to the field of literary linguistics, relatively under-explored in connection with Arab American and Anglophone Arab fiction, and, second, chart a new pathway towards decolonizing trauma studies by examining its relationships with multilingualism, war, and nostalgia.

  Multilingualism, Trauma, and Liminality in The Bullet Collection: Contact Zones, Checkpoints, and Liminal Points · 2021 · DOI

• Overall, data from three framing experiments (N = 1114) in Germany and the United States indicate limited evidence on the effectiveness of the tested militaristic metaphors.

  “It's a war! It's a battle! It's a fight!”: Do militaristic metaphors increase people's threat perceptions and support for <scp>COVID</scp>‐19 policies? · 2021 · DOI

• Further study of address forms may be promising in their interrelation with other means expressing the category of speech addressability, in revealing their genre specificity and characteristic features of other stylistic devices actualization in the vocative position.

  Semantic contrast and inequality in the vocative position (based on English fiction) · 2020 · DOI

• This means that euphemisms can be discussed as a type of language use by means of which the speaker can impose a preferred interpretation of meaning, and, as far as the motives for euphemization are concerned, they should be examined more closely in accordance with the specificities of context and discourse in which euphemisms are used.

  EUPHEMISMS THROUGH TIME: THE RHETORICAL POWER OF PALLIATION · 2017 · DOI

• It is noted that further study of the role of intonation in the creation of exclamation sentence is necessary because existing facilitate systematization of the results of scientific research and strengthening thoughts on intonation as situational motivated, dynamic broadcast unit, aimed at expressing the communicative intent of the speaker.

  Intonation as a means of expression of exclamation sentences · 2016 · DOI

• This research focuses on conceptual TIME metaphors in Kristijonas Donelaitis’s The Seasons in order to supplement a so far scarce research in Lithuania of conceptual metaphors in translation, as time metaphors constitute fundamental aspects of culture and communication (Lakoff, Turner 1989).

  Conceptual metaphors of TIME in translation: The Seasons by Kristijonas Donelaitis · 2015 · DOI

• This paper presents an overview of the emerging field of Cultural Linguistics and argues, by presenting examples from several varieties of English, that world Englishes need to be examined from the perspective of Cultural Linguistics in order for us to gain a better understanding of how English is used by communities of speakers around the world to express their cultural conceptualisations, including their world views.

  Cultural Linguistics and world Englishes · 2015 · DOI

• A more extended review of the problems of reliability and validity is beyond the scope of this paper, but we should note that we can strengthen a given measurement approach through a systematic tinkering with the measurement instruments themselves, if combined with a rigor- ous norm development and validation program.

  The colossal metaphor: Uses of measurement in communication research · 1970 · DOI

• T h e mature mataphor, on the other hand, consists of a concrete analog of a general proposition, an analog not limited to the clearly figurative terms but including their relational context as well, making for a statement whose literal meaning, at least, is fairly unequivocal.

  A Developmental Analysis of Proverb Interpretations · 1959 · DOI

• Future research should explore the implications of digital literacy and emerging communication technologies on traditional language-literary relationships.

  From Linguistic Foundations to Literary Art: Exploring the Interdependent Relationship Between Language and Literature Through Statistical Analysis · 2026 · DOI

• While the euphemism treadmill is a testament to the human desire for empathy and face-saving, its effectiveness is limited by the “gravity” of the concepts it tries to escape.

  EUPHEMISM TREADMILL IN MODERN CORPORATE AND SOCIAL DISCOURSE · 2026 · DOI

Most-cited papers in Language, Metaphor, and Cognition

• A Contrastive Study on the Thematic Structure Functions of English and Arabic Short Stories · World Journal of English Language · 2024 · 189 citations
• The experience of emotion: Directions for tourism design · Annals of Tourism Research · 2020 · 144 citations
• Language is primarily a tool for communication rather than thought · Nature · 2024 · 132 citations
• Categorization and metaphor understanding. · Psychological Review · 1992 · 104 citations
• Use of the Proverbs Test for differentiating schizophrenics from normals. · Journal of Consulting Psychology · 1956 · 101 citations
• Cross-modal iconicity: A cognitive semiotic approach to sound symbolism · Sign Systems Studies · 2010 · 96 citations
• The metaphor in public address · Speech Monographs · 1962 · 72 citations
• Cultural Linguistics and world Englishes · World Englishes · 2015 · 71 citations
• Conceptual engineering: when do we need it? How can we do it? · Inquiry · 2021 · 63 citations
• Metaphors in critical Internet and digital media studies · New Media & Society · 2021 · 61 citations

Most recent work

• Composition-Layer Adoption of the Orthonymic Configuration: A Field Observation of Framework Capture in Google AI Mode, 7 June 2026, with the Personal-Recognition Asymmetry as Control Case (v1.0) · Zenodo (CERN European Organization for Nuclear Research) · 2026
• Metaphorical landscapes of happiness: A cognitive semantic analysis in Hindi and Bangla · Russian Journal of Linguistics · 2026
• Exploring the Use of Visual Metaphor in App Icons of Social Media · International Journal of Humanities and Social Science · 2026
• The one–new–idea constraint, functor–argument metaphors, and holistic speech processing · Functions of Language · 2026
• Articulating the essence: developing and naming themes in descriptive phenomenological research · Qualitative Research Journal · 2026
• Graph convolution-based techniques for pragmatic Arabic figurative language classification · Frontiers in Artificial Intelligence · 2026
• COLOR TERMS IN ENGLISH PHRASEOLOGY AND PAREMIOLOGY · Advances in Science and Humanities · 2026
• Beyond “Taiji Diagram”: how multimodal metaphor deciphers Yin Xu and Yang Xu through Traditional Chinese Medicine science communication short videos · Frontiers in Public Health · 2026
• Brain oscillations during naturalistic tutoring of science concepts using metaphors · JoCN Forum · 2026
• Multimodal OXYmorons: A Comprehensive Introduction and Computational Analysis Using a Dataset of Oxymoronic Memes in Italian and Spanish · Computational Linguistics · 2026

Related topics in Psychology