Core-periphery patterns in knowledge graphs reveal digital visibility hierarchies of South American languages

We analyze the digital representation of indigenous South American languages through a statistical-physics perspective, applying onion decomposition to a semantically enriched knowledge graph built from Wikipedia, Wikidata, and Glottolog. This multi-scale method reveals a pronounced core-periphery hierarchy with three distinct regions: i) a compact core of highly connected languages dominating digital visibility, ii) intermediate layers with strong family-level clustering, and iii) a sparse periphery of linguistically isolated languages. Quantitatively, core languages have, on average, ten times the degree of peripheral ones and are seven times more likely to possess complete genealogical metadata. The results provide measurable evidence of systemic biases in the digital documentation of cultural heritage. Beyond diagnosis, the analysis identifies strategic targets for intervention: intermediate-layer languages as potential bridges to enhance peripheral visibility, and completely isolated cases as priorities for urgent preservation. This approach offers a transferable framework for quantifying representational inequalities in cultural knowledge systems.