Platforms, protocols, and open data

CadLin’s latest publications arrive on three fronts. One group of articles examines language in digital environments: platforms, browsers, comment moderation, podcasts, generative artificial intelligence, and public discourse on technology. Another moves into methodology, with a Registered Report protocol testing the Montreal Forced Aligner on Brazilian Portuguese data. A third turns to data circulation, with the presentation of the dossier on the sharing of linguistic data and the publication of an open-access corpus of child speech. These materials meet around a shared concern: the conditions under which language research is carried out today.

CadLin’s endorsement of the Stockholm Declaration and the new theme issue call supported by CIPL add an editorial and institutional dimension to this update. Together, these items raise a common question: how should knowledge about language be produced, evaluated, shared, and published in a research environment marked by digital platforms, automation, and new demands for scientific openness?

Thiago César da Costa Carneiro (UFPE) opens one path into this debate in How much is it worth to be a subject? Subject effects in the constitution of the archive of/on X. The article compares answers given by Grok to the question “what is a dictatorship?” across different X profiles and across the free and paid versions of the platform. The variation is not treated simply as a matter of bias. It becomes a way of examining the subject-user produced by the platform and the non-neutral archive constituted by X.

Thomas Falconi and Lucas Alves Selhorst (UNISUL) take browsers themselves as objects of analysis in Subject-avatar in computerized enunciative spaces: between technical normalizations and resistance practices. Chrome and Tor are compared as enunciative spaces with distinct regimes of individuation. Chrome is associated with an extractivist logic of data circulation; Tor makes possible other forms of technical individuation, with discursive consequences.

Aude Seurrat (Université Paris-Est Créteil) and María Inés Laitano (Université Paris 8) publish, in French, Réseaux socio-numériques et frontières du lire/écrire dans la modération des commentaires, a comparative study of comment moderation on X, YouTube, Twitch, and Instagram. The article treats moderation as a way of regulating reading and writing. Each platform combines algorithmic automation and human judgment differently, producing its own conditions for online interaction.

Letícia Moraes (UFPB) and Silvia Maria de Sousa (UFF) bring discursive semiotics to the problem of datafication in Generative AI and meaning-making: a semiotic approach from the Global South. The article examines how generative AI models participate in meaning production and argues for an analysis grounded in an enunciative position from the Global South, rather than one derived from dominant centres of theoretical production.

Evandra Grigoletto (UFPE), Silmara Dela Silva (UFF), and Solange Maria Leda Gallo (UNISUL) analyse the Brazilian federal government’s July 2024 proposal on artificial intelligence in “AI for the good of all”? Discourse, subject, contradiction. From the perspective of Pecheutian Discourse Analysis, the phrase “AI for the good of all” is read as a discursive formulation marked by contradictions in the relation between technology, subject, and public interest.

Ramon Miranda and Ana Maria Sá Martins (UEMA), in The discursive representation of Yglésio Moyses (PRTB) in the podcast “Abrindo o Verbo”: an analysis through CDA, apply Critical Discourse Analysis to a political interview podcast. The regional focus — a politician from Maranhão — shows how public images and political representations are built in media circuits outside the traditional press.

These six articles form part of the theme issue Media, Technologies, and Artificial Intelligence, organized by Silmara Dela Silva, Evandra Grigoletto, and Solange Maria Leda Gallo, which remains open to new submissions.

A related discussion appears in the journal’s next theme issue, Languages, Race, Gender, and Intersectionality in the Struggle for Rights. In Algorithmic textuality and necroalgorithmization, Júlio Araújo (UFC) argues that exclusions produced by AI systems around race, gender, and sexuality are part of how those systems operate. To analyse them, he proposes the Discursive Triangle of Algorithmic Textuality, distinguishing text-prompt, text-algorithm, and text-response as instances with their own regimes of production and reading.

Livia Oushiro (UNICAMP) publishes Reanalysis of mid pretonic vowels with the MFA forced aligner, the third Registered Report protocol in the dossier Reproducibility in Linguistics. It follows the protocols by Godoy, Ananias, and Sidhu on lexical iconicity and by Alves et al. on automatic prosodic segmentation with machine learning.

The study will replicate earlier analyses of pretonic /e/ and /o/ in the speech of Northeastern migrants in São Paulo, using the Montreal Forced Aligner instead of manual segmentation. The protocol is registered publicly from the start. The full analysis and results will be published in the next stage.

The theme issue Sharing of Linguistic Data, organized by Juliana Bertucci Barbosa, Marcia dos Santos Machado Vieira, and Raquel Meister Ko Freitag, now has its editorial presentation. The text treats data sharing as a matter of scientific infrastructure: metadata, licensing, documentation, long-term access, and reuse.

Pedro Perini-Santos (UFVJM) contributes to the dossier with Child Verb Timeline: 21 months of verbal expression from a Brazilian child. The corpus, developed within the CIL project — Longitudinal Child Corpus — records 587 verbal occurrences across 210 excerpts of interaction among a child, her mother, and her grandmother in a city in northeastern Minas Gerais, from March 2016 to February 2018. Each occurrence is categorized according to a taxonomy of verbal diatheses in Brazilian Portuguese. The data are available under an open licence on Zenodo.

CadLin has joined the group of journals endorsing the Stockholm Declaration, published by Bernhard Sabel and Dan Larhammar in Royal Society Open Science. The document proposes action on four fronts: returning control of scientific publishing to academia, valuing quality over quantity, creating independent mechanisms for detecting fraud, and strengthening scientific integrity policies.

These principles are aligned with practices already adopted by CadLin: diamond open access, open peer review, preprint deposit, and mandatory statements on data, materials, and code.

The call Scientific Associations in Linguistics: history, impact, and current perspectives is open, with the endorsement of CIPL — Comité International Permanent des Linguistes.

The theme issue invites presidents, former presidents, and other formal representatives of associations, societies, academies, committees, and research networks to submit texts in the Critical Perspective format. Contributions should offer analytical accounts of these organizations’ histories, forms of action, and present-day difficulties.

Submissions are open until October 30, 2026, at cadernos.abralin.org.

CadLin is also receiving submissions for the following calls:

• Diachronic Construction Grammar

• Cognitive Linguistics in the Real World

• From the Roots to New Directions: The Trajectory of Brazilian Dialectology

• Media, Technologies, and Artificial Intelligence

• Contemporary Historiographies of Linguistic Studies and Language Teaching

Deadlines and guidelines are available at cadernos.abralin.org.

CadLin follows the Diamond Open Access model: authors do not pay to submit or publish, and readers do not pay to access articles. Learn how to submit.