cs.CLSep 30, 2026

Dissonant ballerinas and crafty carrots: a comparative multi-modal analysis of Italian brain rot

Authors: Anca Dinu, Andra-Maria Florescu, Marius Micluta-Campeanu, Stefana-Arina Tabusca, Claudiu Creanga, Andreiana Mihail

Organizations: University of Bucharest, Bd. Regina Elisabeta no. 4-12, Bucharest · Faculty of Foreign Languages and Literatures, Str. Edgar Quinet no. 5-7, Bucharest · Interdisciplinary School of Doctoral Studies, Bd. Mihail Kogălniceanu no. 36-46, Bucharest · Faculty of Mathematics and Computer Science, Str. Academiei no. 14, Bucharest

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

This paper presents a comparative multi-modal analysis of Italian and Romanian brain rot memes, investigating the factors that contribute to its appeal and the linguistic and cultural distinctions between the two versions. To conduct this analysis, we introduce a multi-modal brain rot dataset named CRIB (Collection of Romanian and Italian Brain rot), a manually curated collection of 240 TikTok videos stratified by language (Italian, Romanian) and popularity, on which we examine textual, acoustic, and visual features. Our findings indicate that popularity is not significantly correlated with textual elements like sentiment, absurdity, or rhyme, or acoustic elements such as vocal features or sentiment of the sound. Instead, in Romanian language, video-level dynamics, specifically faster cutting speeds and a more rapid overall pace, are strong predictors of a video's success. The cross-linguistic analysis reveals significant differences. Italian brain rot is textually more negative, exhibits higher perplexity, and uses more rhyme, while its sound is characterized by higher melodic range and loudness. Romanian audio is spectrally brighter with more erratic pitch variations.

Figures & tables

Figure 1 : Example of Italian brain rot character Ballerina Cappuccina
Figure 2 : Example of Romanian brain rot character Morcoveață
Figure 3 : Sentiment scores (Negative, Neutral, Positive) for each language and popularity group of the brain rot texts
Figure 4 : Continuous‐sentiment distributions (Negative, Neutral, Positive) for each language and popularity group of the brain rot audios
Figure 5 : Faster video editing speeds show a stronger correlation with popularity than slower editing speeds
Appendix figures & tables1 asset

Supplementary material from the paper’s appendix.

Appendix

pairMANOVAMann-Whitney for neg.Mann-Whitney for neu.Mann-Whitney for pos.
Ro pop. - Ro unpop.λ\lambda = 0.99, F(3, 116) = 0.55, p = 0.65p = 0.79 (M: 0.38 vs. 0.36)p = 0.57 (M: 0.39 vs. 0.40)p = 0.990 (M: 0.24 vs. 0.23)
It pop. - It unpop.λ\lambda = 0.97, F(3, 116) = 1.15, p = 0.33p = 0.36 (M: 0.62 vs. 0.67)p = 0.54 (M: 0.19 vs. 0.17)p = 0.39 (M: 0.19 vs. 0.16)
It - Roλ\lambda = 0.40, F(3, 116) = 56.74, p ¡ 0.001p ¡ 0.0001 (M: 0.38 vs. 0.16)p ¡ 0.0001 (M: 0.39 vs. 0.17)p ¡ 0.0001 (M: 0.67 vs. 0.24)
Appendix
Table 1 : Statistical tests for Sentiment Analysis of the texts

Explore similar work

CardsList

  1. RoIt-XMASA: Multi-Domain Multilingual Sentiment Analysis Dataset for Romanian and Italian

    Apr 18, 2026Andrei-Marius Avram, Aureliu Valentin Antonie, Cosmin-Mircea Croitoru +2Adversarial Training

    National University of Science and Technology POLITEHNICA Bucharest, Bucharest, Romania

  2. Brain-LLM Alignment Tracks Training Data, Not Typology

    May 21, 2026Dongxin Guo, Jikun Wu, Siu Ming YiuCross-Lingual ConsistencyTypology

    The University of Hong Kong Hong Kong, China · Stellaris AI Limited Hong Kong, China

  3. Multilingual Emotion Neurons in Large Audio-Language Models

    Aug 9, 2026Xiutian Zhao, Philipp Koehn, Björn Schuller +1Cross-Lingual ConsistencyLarge Audio Language Models

    Center for Language and Speech Processing (CLSP), Johns Hopkins University, USA · Group on Language, Audio & Music (GLAM), Imperial College London, UK