Grammar as Control: Modular Language Generation for the Long Tail

Large language models (LLMs) can, in principle, bootstrap language technologies for long-tail languages due to their pattern recognition capabilities. Yet in practice, without structured guidance, they produce narrow, unrepresenta-tive samples that fail to cover the morphosyn-tactic space of typologically underrepresented languages. We propose Modular Typology-Informed Generation (mTIG), a prompting framework that transforms descriptive grammars into explicit control mechanisms that guide LLMs to generate typologically balanced synthetic data for downstream training. mTIG decomposes grammars into modular grammar slices , each targeting a specific morphosyntactic phenomenon (e.g., passive voice, causative morphology). Across three low-resource languages, mTIG improves typological entropy by up to 19% and yields a “student-beats-teacher” effect, where distilled models outperform the source LLM by up to +20 chrF in machine translation. These findings show that grammar-as-control can construct training corpora wherever formal linguistic descriptions exist.