This repository contains the materials, stimuli, and tasks for the ML One Network project — a many-languages extension of the SPAML semantic priming project.
01-Translation — Stimuli selection, translation pipeline, and finalized word pairs for each language. Start here to understand how cue-target pairs were chosen and translated.
02-Stimuli — Per-language norming files and scripts for creating the final stimuli sets used in the experiment.
03-Tasks — JATOS experiment templates and build scripts for each task, organized by study.
04-Ethics — IRB application, approval, and ethics packet.
05-Power — AIPE-based sample-size power simulations for the semanticprimeR norming variables, run against the finalized SPAML datasets.
06-Data — Collected data (stored externally; see folder for access details).
R environment: This project uses renv for reproducible package management. Run renv::restore() after cloning to install all packages. sylly.fr is sourced from a custom repository (https://undocumeantit.github.io/repos/l10n/) and is included in the lockfile.
subs2vec downloads: The stimuli creation pipeline downloads word frequency and vector files from subs2vec. As of June 2026 these require a login and cannot be run automatically. Three languages (Japanese, Chinese Simplified, Chinese Traditional) also required hand-created vector files not available via subs2vec. See 01-Translation/README.md for details.