A l’occasion de sa venue à Paris, du 25 au 31 janvier 2026, le professeur Yeshes Vodgsal Atshogs ཡེ་ཤེས་འོད་གསལ་ཨ་ཚོགས། , éminent spécialiste de linguistique tibétaine et d’études sino-tibétaines et directeur du Centre de recherche sur les langues sino-tibétaines à l’université de Nankai (Chine), prononcera deux conférences en anglais, coorganisées par le CRLAO et l’Inalco, les 27 et 28 janvier:
- Mardi 27 janvier 2026, 18h-19h
Titre : From Sanskrit-Tibetan Classical Prosody to Modern Prosodic Phonology of Tibetan: An Analysis of Stress and Poetic Meter
Lieu : Salle 4.05, INALCO, 65 rue des Grands Moulins, 75013 Paris
- Mercredi 28 janvier 2026, 14h-16h
Titre : New Light on an Old Puzzle: The ‘Phags-pa Script Voicing Flip-flop in Yuan Chinese Documented by the Menggu Ziyun
Lieu : Salle 0.015, Bâtiment de recherche sud, 5 Cours des Humanités, Campus Condorcet, 93322 Aubervilliers Cedex
Vous trouverez ci-dessous les résumés de ces deux interventions.
- Conférence à l’Inalco, mardi 27 janvier 2026, 18h-19h
Titre : From Sanskrit-Tibetan Classical Prosody to Modern Prosodic Phonology of Tibetan: An Analysis of Stress and Poetic Meter
Résumé:
Tibetan history has a long-standing tradition of research into prosody. Since Sakya Pandita first systematically introduced and adapted Indian prosodic theories, over twenty Tibetan treatises have been produced, forming a extensive body of literature. However, by the 15th century, when the great translator Taktsang Lotsawa explicitly asserted that « Tibetan does not require a prosodic light-heavy distinction, » Tibetan prosody became more clearly defined as a discipline primarily focused on Sanskrit prosodic theory and its poetic meters. For convenience, we refer to this tradition circulating in Tibet as Sanskrit-Tibetan Classical Prosody (STCP), while contemporary research on the stress and prosody of Archaic and Modern Tibetan is termed Modern Prosodic Phonology of Tibetan (MPPT).
Prosodic analysis in STCP is primarily based on syllable weight, forming two metrical systems: vṛtta (བྲྀཏྟ), based on syllable count, and jāti (ཛཱ་ཏི), based on moraic count. However, both Archaic and Modern Tibetan are quantity-insensitive languages. The Sanskrit system of prosodic analysis—which determines light-heavy by calculating Mātrā (Tibetan: ཕྱི་མོ; equivalent to the mora)—is entirely incompatible with the Tibetan language. Therefore, within the context of STCP, Taktsang Lotsawa’s conclusion that Tibetan does not require light-heavy analysis was perfectly accurate.
Nevertheless, through the application of modern linguistic methodologies, it can be argued that Tibetan possesses its own prosodic system based on stress analysis. Significant progress has been made starting from early scholars such as Sprigg (1966, 2002), Bielmeier (1985, 1988), Sun (1986, 2003), and Haller (2004), to more recent research by myself (Atshogs 2003/2004, 2005, 2012, 2020), Caplow (2009), and Duan Haifeng (2012). Our fundamental analysis concludes that, overall, Tibetan stress can be analyzed at the lexical level as a fixed stress system that does not distinguish lexical meaning, though it manifests differently across historical stages and dialects.
Simultaneously, the specific acoustic correlates are quite complex. On one hand, Tibetan exhibits characteristics of a pitch accent (or musical stress) language, marked by pitch prominence; on the other hand, it displays characteristics of a stress or stress accent (dynamic stress) language, marked by duration and vowel quality (segmental changes). These two features can even co-occur within a single disyllabic word: manifesting as « initial stress » (dynamic) and « final high-pitch » (pitch accent). At the phenomenal level, this can be succinctly summarized as « initial-stress, final-high ».
Both stress phenomena adhere to the principles of relative prominence in modern metrical phonology, as well as attributes such as culminativity, demarcative function, and rhythmicity. From the perspective of prosodic phonology, this seemingly « contradictory » phenomenon of dual-stress within a single word should actually be assigned to different levels of the prosodic hierarchy: The Foot level and the Prosodic Word level. The « initial-stress, final-high » pattern can be traced back to Proto-Tibetan and is widely observed across modern Tibetan dialects, serving as a common rhythmic template that governs tone sandhi patterns in tonal dialects. Furthermore, this prosodic system is extensively reflected in ancient and modern literature—from scholarly poetry to folk song lyrics. Our analysis of the representative six-syllable and seven-syllable works, such as the Love Songs of the Sixth Dalai Lama and the Sakya Lekshe (Elegant Sayings), confirms this. Such « meter » reflects not a conscious creative template used by poets, but rather a latent rhythmic system followed subconsciously.
- Conférence au CRLAO, mercredi 28 janvier 2026, 14h-16h
Titre : New Light on an Old Puzzle: The ‘Phags-pa Script Voicing Flip-flop in Yuan Chinese Documented by the Menggu Ziyun
Résumé:
For a century, the problem of the « voicing flip-flop » in the ‘Phags-pa script’s transcription of Chinese consonant initials in the Yuan Dynasty’s Menggu Ziyun has long confounded the academic community. It has been regarded as a « quite surprising, » « incredible, » and « puzzling » « mystery » (Zhongw Shen 2008: 115). Eastern and Western scholars, such as Karlgren (1915), Dragunov (Александр Драгунов 1930), and Hattori (服部四郎1946), as well as Shen (2008), Chen (陈鑫海 2015), Wang (王文敏 2021), and Mai (麦耘 2022), have offered a variety of very different interpretations. While each of these explanations has its merits, no definitive conclusion has yet been reached.
By comparing Northern Chinese phonology of the Yuan Dynasty with Tibetan phonology of the Sakya Dynasty era, and considering the characteristics of the ‘Phags-pa script, this study proposes a brand-new line of reasoning to explain the mystery of the « voicing flip-flop » in the Menggu Ziyun. That is to say, in both Yuan Dynasty Chinese and Sakya-era Tibetan, voiced stops and affricate consonants had already undergone devoicing; in practice, the issue of consonant voicing contrast no longer existed. Previous scholars, based on the assumption that ‘Phags-pa letters represented « voicing » in Old Tibetan, hypothesized various special « phonation type » features for Yuan Chinese consonants. In reality, the single-consonant letters for stops and affricates in the Sakya-era Tibetan no longer reflected consonant voicing, but rather tonal pitch (high vs. low). The Menggu Ziyun precisely used ‘Phags-pa letters that reflected the « Voiceless-High / Voiced-Low » register of Sakya Tibetan to match the « Yin-Low / Yang-High » tones of Yuan Chinese (i.e., « high sounds follow Yang, low sounds follow Yin » (高声从阳，低声从阴) as described in Yu Ji (虞集)’s Preface to Zhongyuan Yinyun). This ultimately resulted in « voiced » letters being paired with Yin tones and « voiceless » letters being paired with Yang tones, creating a « voicing flip-flop » effect for modern observers.
In essence, this is an illusion arising from a chronological mismatch—applying the phonological rules of ancient scripts (voicing) to the phonetic realities of Early Modern Chinese (tonal pitch). This ‘judging the modern by the ancient’ represents a profound misunderstanding of how the ‘Phags-pa script was used to precisely align the tonal registers between Tibetan and Chinese.
At the origin of Chinese tones, as in Tibetan, the pattern was « Voiceless-High / Voiced-Low » (similar to modern Wu dialects where the ancient voicing contrast and its tonal correlates are still preserved). The reason Northern Chinese in the Yuan Dynasty exhibited a « Yin-Low / Yang-High » pattern is that Northern Chinese had undergone a complex « Circular tonal change » (Hirayama Hisao 1984, 1991), and tonal pitch had already undergone a massive change; the Yuan Dynasty was precisely in an « inverted » state of « Yin-Low / Yang-High » relative to the origin.
Furthermore, the secret of why the voicing of fricatives in the ‘Phags-pa script of Menggu Ziyun is not flipped lies in the fact that in Sakya-era Tibetan, while voiced stops and affricates had devoiced, the voicing contrast in fricatives still persisted. From this, it can also be seen that the ‘Phags-pa transcription of Chinese in the Menggu Ziyun was conducted by someone highly proficient in Tibetan phonology—that is, the creator of the phonetic notation or the principles of the Menggu Ziyun was, in all probability, ‘Phags-pa himself.