The Tonogenesis Continuum in Tibetan: A Computational Investigation

TL;DR

This work investigates tonogenesis in Tibetan languages by quantifying the functional load of pitch using pitch-flattening and ASR performance. It trains separate XLS-R 300m models on six Tibetan varieties spanning Amdo, Khams, and Ü-Tsang, and evaluates word/character error rates on original versus $f_0$-flattened data, revealing a continuum from atonal to fully tonal speech. The results show Amdo varieties are minimally affected by pitch removal, Ü-Tsang varieties exhibit strong pitch dependence, and Khams varieties lie in between, suggesting a gradual reweighting of segmental and suprasegmental cues during tonogenesis. These findings demonstrate the utility of computational methods to capture fine-grained stages of sound change and argue for more nuanced functional-load metrics and dialect-aware ASR systems that account for evolving pitch dependence, with implications for cross-dialect phonology and speech technology.$f_0$ is used as a primary lexical cue in tonal systems, and CER/WER changes quantify its shifting role across dialects.

Abstract

Tonogenesis-the historical process by which segmental contrasts evolve into lexical tone-has traditionally been studied through comparative reconstruction and acoustic phonetics. We introduce a computational approach that quantifies the functional role of pitch at different stages of this sound change by measuring how pitch manipulation affects automatic speech recognition (ASR) performance. Through analysis on the sensitivity to pitch-flattening from a set of closely related Tibetan languages, we find evidence of a tonogenesis continuum: atonal Amdo dialects tolerate pitch removal the most, while fully tonal U-Tsang varieties show severe degradation, and intermediate Kham dialects fall measurably between these extremes. These gradient effects demonstrate how ASR models implicitly learn the shifting functional load of pitch as languages transition from consonant-based to tone-based lexical contrasts. Our findings show that computational methods can capture fine-grained stages of sound change and suggest that traditional functional load metrics, based solely on minimal pairs, may overestimate pitch dependence in transitional systems where segmental and suprasegmental cues remain phonetically intertwined.

Figures (2)

• Figure 1: Approximate distribution of major Tibetan dialect groups. Ü-Tsang (red), Khams (green), and Amdo (purple) show distinct historical trajectories in tonogenesis.
• Figure 2: A schematic illustrating how Tibetan script maps to Wylie, then to acoustic features such as pitch and voicing, which the ASR model weighs during decoding.