The Universal Atlas of Phonetic Evolution
Back to Matrix [h] → [s] Edit this Data

[s]→[h]Certainty: 5/5Commonality: 4/5

Voiceless alveolar fricative to Voiceless glottal fricative

In the process of Tonogenesis, the debuccalization of a final sibilant [s] into a glottal fricative [h] represents a critical intermediate stage. According to Haudricourt's model, this Laryngeal Coda originally triggered a lowering of the vowel's Pitch Contour, which eventually became phonemicized as a falling tone (Qù sheng or sắc/nặng).

Phonetic Effects

Debuccalization, Tonogenesis, Pitch Fall, Voice Loss, Falling Tone

Language Examples

*pas→bá /baː˧˥/— Historical final -s in Mon-Khmer was first debuccalized to -h and then elided, giving rise to the 'Sắc' (high-falling) tone in Vietnamese.
Historical *-s→*-h— A primary stage in Haudricourt's model where final sibilants weakened to laryngeal fricatives.
*pɑs→bà /pɑ˥˩/— The 'Departing' category (Qù sheng) in Middle Chinese is reconstructed as having a final [s] (later [h]) which influenced the Pitch Contour of the syllable.
Old Chinese *-s→*-h— Led to the development of the Qù (去) or Departing tone.
*kraːs→khrâa /kʰraː˥˩/— Tai-Kadai languages show a similar pattern where final spirants debuccalized, contributing to the development of the falling tone register.
Proto-Tupi *s→Guarani /h/— One of the most significant phonetic innovations distinguishing Guarani from Tupi is the systematic replacement of sibilant /s/ with glottal fricative /h/. E.g., Proto-Tupi *so → Guarani ha (color), *se → Guarani he (this), *sipo → Guarani hipo (snake). This represents a regular sound law in the Tupi-Guarani divergence, not driven by tonogenesis but by debuccalization of the anterior fricative.
Proto-Indo-Iranian *s (word-initial before vowel)→Iranian *h— One of the most diagnostic Iranian sound laws: Proto-Indo-Iranian word-initial *s → Proto-Iranian *h before vowels. Classic example: PII *sapta 'seven' → Avestan hapta, Persian هفت haft. Sanskrit preserves *s: saptá. Similarly PII *sṷana- 'dog' → Avestan spā (via *h-), Old Persian *ha-. The change was blocked before non-nasal stops and after i, u, r, k.
PII *sūra- 'sun/hero'→Avestan hūra- / Old Persian hvar-— PII *s → Iranian *h word-initially. Compare Sanskrit sūra- 'sun' which preserves *s. This isogloss defines the Iranian branch of Indo-Iranian alongside the satem-like sibilant shifts.
Proto-Mongolic *s (before non-i vowels)→Buryat /h/— In Buryat, the inherited sibilant *s debuccalized to /h/ before vowels other than *i. E.g., Proto-Mongolic *sang 'storehouse' → Buryat han(g); *sira 'yellow' → Buryat shara (via intermediate *h-palatalization). The change is blocked before *i, where *s is preserved or merges with palatal outcomes. This parallels the spirantization of *c→s in the same language.

Related Transformations

Spotted a mistake or have a better example?

EchoDrift is community-driven. You can suggest a change directly through GitHub.

Propose an Edit on GitHub