The grammatical encoding of expressivity and emotions
EMO launch workshop — September 11, 2026
On September 11, 2026, the EFL project is launching the EMO working group with a one-day workshop on the grammatical encoding of expressivity and emotions. Seven talks, in French and English, examine how languages express fear, anger, surprise or irony through syntactic constructions, discourse markers and interjections — from political discourse to Internet memes.
Organiser: Agnès Celle (agnes.celle@u-pariscite.fr) Venue: Room 720, Olympe de Gouges building, Université Paris Cité
Program
Morning
| Time | Speaker | Title |
|---|---|---|
| 11:00–11:10 | Agnès Celle | Introduction |
| 11:10–11:40 | Carlota Piedehierro | The humorous semantic construction of Internet memes: an “emotional utterance” of great argumentative efficiency |
| 11:40–12:10 | Fatima-Zahra Aklalouch | Direct fear-related words in Trump’s and Johnson’s Covid-19 discourses: A comparative analysis |
| 12:10–12:40 | Fernando Casanova Martínez | Expressive interjections in Broca’s aphasia: a pragmatic analysis of emotional and cognitive expression |
Lunch break
Afternoon
| Time | Speaker | Title |
|---|---|---|
| 13:40–14:10 | Georgeta Cislaru | Illocutionary intentions and perlocutionary effects of some polemical discourse markers |
| 14:10–14:40 | Natacha Marjanovic | ‘What kind of a stupid argument is that?’: identifying an expressive construction |
| 14:40–15:10 | Efrat Levant and Cameron Morin | Expressive Meaning in Construction – Sarcastic Interrogatives as a Case Study |
Break
| Time | Speaker | Title |
|---|---|---|
| 15:20–15:50 | Anaïs Robin | Mirative turn out as a verb marking truth revelation: a comparison between two constructions |
| 15:50–16:20 | Agnès Celle | Reported information as a source of expressivity |
| 16:20–16:30 | Lisa Brunetti and Agnès Celle | Conclusion and outlook |
Abstracts
Carlota Piedehierro (ALTAE, Université Paris Cité)
The humorous semantic construction of Internet memes: an “emotional utterance” of great argumentative efficiency
Research on memes has shown that these digital contents — made up of an image, a video, a text, or a combination of these elements — fulfil far-reaching social functions: they shape public discourse, reflect and spread currents of opinion, feed public political debate, or serve as digital marketing mechanisms.
Humour is one of the most effective means by which memes fulfil these functions. This talk analyses the semantic construction underlying the humorous mechanism of certain memes. Drawing on tools from Jean-Claude Anscombre’s theory of stereotypes, the hypothesis put forward is that memes operate a diversion of shared knowledge — a common principle invoked by a collective or anonymous voice (an “ON-speaker”) and embedded in a linguistic community’s heritage.
The example analysed diverts a well-known Spanish proverb (Al que madruga, Dios le ayuda, “The early bird gets God’s help”) while preserving its formal mould and its implicit semantic mould, producing a humorous effect through subversion of the general rule it states. The image accompanying the diverted text (Pikachu) plays an anchoring role in Barthes’s sense, substituting the authority of the online community for that of the linguistic community. The analysis thus shows that memes build an “emotional utterance” which, through diverted structures of meaning, achieves effective argumentative force.
Fatima-Zahra Aklalouch
Direct fear-related words in Trump’s and Johnson’s Covid-19 discourses: A comparative analysis
This study examines how direct fear-related words are linguistically and discursively constructed in the Covid-19 discourses of Donald Trump and Boris Johnson. Focusing on explicit emotion terms such as fear, panic, anxiety, scare, the analysis investigates how fear is grammatically organized, attributed, intensified, mitigated, and rhetorically negotiated in political crisis communication.
The items were identified using Docuscope, a corpus-rhetorical analysis tool, then analyzed qualitatively in their immediate co-text. At the micro-level, the study examines argument structure, participant roles, aspectual distinctions, and the distribution of experiencers, sources, and agents of fear. At the macro-level, it analyzes discourse-pragmatic functions: reassurance, blame attribution, emotional regulation, leadership performance, and the negotiation of public fear during the pandemic.
The comparison shows two markedly different strategies. In Trump’s discourse, panic is the most frequent term, associated with negative constructions and causation/responsibility attribution (e.g. “I don’t want to create panic”). In Johnson’s discourse, anxiety is more prominent and linked to identifiable social groups through possessive forms (“your anxiety”, “their anxieties”). These patterns suggest two contrasting orientations: control and displacement of panic on one hand, recognition and institutional management of shared anxiety on the other — showing how direct emotion words organize political responsibility, leadership ethos, and crisis communication strategies.
Fernando Casanova Martínez (University of Alicante)
Expressive interjections in Broca’s aphasia: a pragmatic analysis of emotional and cognitive expression
Aphasia is an acquired language disorder that affects language production and comprehension, substantially limiting everyday communication. Broca’s aphasia in particular is characterised by marked impairments in speech production, which highlights the importance of investigating the linguistic resources that remain available for communication. Expressive interjections, used to externalise speakers’ mental states, have received little attention in aphasic discourse, particularly regarding the pragmatic functions they perform.
This study has two objectives: (i) identifying the emotional and cognitive states conveyed through expressive interjections during spontaneous discourse, and (ii) examining the pragmatic functions these interjections fulfil in interaction. The analysis is based on the AphasiaBank discourse protocol and includes twenty-five English-speaking participants with Broca’s aphasia (Aphasia Quotient between 51 and 75), discussing language self-perception, stroke memories, personal experiences, and rehabilitation strategies.
The findings are expected to contribute to a better understanding of how expressive interjections support communication in aphasic discourse, offering new insight into their communicative and pragmatic role in clinical interaction.
Mots-clés : Broca’s aphasia; expressive interjections; pragmatics; emotional and cognitive expression; clinical linguistics
Georgeta Cislaru
Intentions illocutoires et effets perlocutoires de quelques marqueurs de discours polémique
Natacha Marjanovic
‘What kind of a stupid argument is that?’ : identification d’une construction expressive
L’expression des émotions dans le langage est une tâche délicate, souvent limitée à l’identification de marqueurs lexicaux comme les noms d’émotions ou les métaphores d’émotions. Pourtant, de nombreux travaux récents soulignent le potentiel expressif des interrogatives, en lien avec la surprise mais aussi avec des émotions négatives comme la colère. Alors que ces interrogatives expressives sont généralement identifiées par leur fonction pragmatique — questions rhétoriques ou autres usages non canoniques — cette présentation propose d’enquêter sur une construction expressive de la colère formellement marquée : What kind of X.
À partir d’une annotation des interrogatives exprimant la colère, grâce à une adaptation de la théorie de la pragmatique affective (TAP) de Scarantino, dans un corpus tiré de subreddits féministes et masculinistes, cette communication décrit les formes et usages de la construction What kind of X et suggère qu’elle présente un potentiel expressif certain.
Efrat Levant et Cameron Morin
Expressive Meaning in Construction – Sarcastic Interrogatives as a Case Study
Construction Grammar has long allowed meanings beyond truth-conditional content to be associated with grammatical form, yet expressive and attitudinal meanings remain under-theorized, especially when they arise from recurrent grammatical patterns rather than lexical items. This talk addresses the question through a case study of sarcastic interrogatives, asking whether sarcasm should be treated only as an inferential effect of context, or whether it can become conventionalized as part of a stored form–meaning pairing.
The analysis focuses on three interrogative constructions in American English — alternative questions, disjunctive polar questions, and split interrogatives — drawn from three corpora (COCA, the Internet Argument Corpus, Switchboard). Each token was annotated as literal or sarcastic, and sarcastic tokens further classified as context-independent or context-dependent.
Three main findings emerge. First, sarcastic interpretations are significantly more likely to be context-independent than context-dependent. Second, the three interrogative types differ significantly in their association with sarcasm: split interrogatives are the most sarcasm-prone, a pattern interpreted as constructionally motivated. Third, the interrogative type most strongly associated with sarcasm carries the fewest additional expressive markers (focus particles, general extenders, emojis), while intensifiers, discourse markers, quotation marks and capitalization show no comparable association. Together, these findings motivate a gradient model of expressive constructionalization, in which grammar can conventionalize speaker attitude to different degrees rather than merely hosting pragmatic inference.
Keywords: expressive meaning; Construction Grammar; sarcasm; interrogatives; conventional pragmatics; use-conditional meaning
Anaïs Robin
Mirative turn out as a verb marking truth revelation: a comparison between two constructions
This presentation compares two syntactic structures containing the phrasal verb turn out, alongside the verb prove: the raising construction and the extraposed impersonal construction. Prove was selected because it is attested in the raising construction only. The two structures form part of an alternation and convey similar meaning, but previous studies have attempted to identify their differences on several linguistic levels — semantics, discourse and pragmatics. Turn out and prove are labelled as Evidential and Mirative Raising Verbs in the literature. The comparison sheds light on how truth revelation is grammatically encoded and on the discursive factors that condition the choice between the two constructions.
Agnès Celle (Université Paris Cité, ALTAE)
Reported information as a source of expressivity
This talk looks at the relationship between reported information and expressivity, focusing on the discourse marker askip, a truncated form of à ce qu’il paraît (“apparently”) that emerged from social-media communication. (il) paraît and askip share similarities as reportative markers, but differ notably in their position: (il) paraît is typically used in right-detached position, whereas askip is frequently found in left-detached position (e.g. “Askip j’suis moche mais j’suis dans son tel en favori” — “Apparently I’m ugly but I’m on his phone as a favourite”), where it conveys a distancing stance that can amount to rejecting the proposition.
Drawing on annotated data from the French TenTen 23 corpus, the talk shows that askip has extended to medial and final position. Yet askip seems to blur the source of information: illocutionarily, it oscillates between confirming and rejecting an assertion. The talk argues that askip primarily negotiates the interaction on an intersubjective level rather than specifying the source of information — confirming the expressive evolution of markers originally associated with reported speech. What is at stake is not the source of information but the speaker’s stance towards a proposition, with initial position emerging as the privileged site for the development of this metatextual function.