AI-assisted Historical Phonology Reconstruction of Pre-Qashqai Oghuz Dialects (12th–13th Century)

This record documents a research and methodological framework for the AI-assisted reconstruction of historical phonology in endangered Oghuz dialects, with a specific focus on pre-Qashqai linguistic forms dated approximately to the 12th–13th centuries. Due to the absence of historical audio recordings, this work does not claim to reproduce definitive or authentic historical pronunciations. Instead, it applies established methods from comparative historical linguistics, phonological reconstruction, and dialectology, supported by probabilistic AI modeling under continuous human linguistic and cultural supervision. The project integrates contemporary Qashqai dialect recordings, comparative data from related Oghuz languages (including Old Anatolian Turkish, Old Azerbaijani, and Turkmen), and historically documented sound shifts. Artificial intelligence tools are used strictly as assistive instruments to simulate, annotate, and document multiple plausible phonological scenarios, each explicitly marked with uncertainty indicators. All generated outputs are labeled as:“Historically reconstructed pronunciation – AI-assisted, linguistically supervised.” This work is part of the QashqAI Voice initiative and aims to contribute to responsible AI research, endangered language preservation, and ethically grounded cultural heritage documentation. The framework is published as open-access to enable scholarly review, institutional collaboration, and transparent reuse under appropriate attribution.