The Convergence of Federated Learning, Knowledge Graphs, and Large Language Models for Language Learning: A Scoping Review

Large Language Models (LLMs) in Intelligent Computer-Assisted Language Learning (iCALL) offer personalization potential but introduce critical challenges in pedagogical grounding, data privacy, and pedagogical validity. While Knowledge Graphs (KGs) and Federated Learning (FL) address these concerns individually, systematic integra-tion of all three technologies remains absent or insufficiently addressed in current re-search. This scoping review maps the FL–KG–LLM convergence landscape in educa-tional contexts. Following PRISMA-ScR guidelines, we searched six databases and screened 51 papers published between 2019 and 2025 using automated extraction. Our findings reveal a pronounced convergence deficit: no papers integrate all three domains, while 58.8% of approaches operate within isolated technological silos. Criti-cal reporting gaps emerge across the corpus, with an average “Not Reported” (NR) rate of 84.5%, particularly in privacy mechanisms (92.2%), validation metrics (90.2%), and Common European Framework of Reference for Languages (CEFR) alignment (88.2%). Domain-specific analysis reveals two distinct patterns: inter-domain gaps (disciplinary silos resulting in expected CEFR absence in single-domain papers) and intra-domain gaps (failure to report domain-critical variables, including 100% parameter NR in FL studies, 86.7% validation NR in KG studies, and 100% CEFR NR in convergence pa-pers). We identify two pillars of pedagogical grounding: a Grounding Pillar, which con-strains LLM outputs via Knowledge Graph rules, and a Validation Pillar, which con-cerns how authoritative source frameworks are mapped onto Knowledge Graph schemas. The latter remains completely unaddressed in the reviewed literature, re-vealing what we term the Integrity Gap—a systematic disconnection between techno-logical innovation and pedagogical grounding in iCALL. By framing pedagogical alignment as an upstream control and validation problem, this review offers insights relevant to the design of user-facing automated systems where trust, transparency, and human oversight are critical.