Syllable Structure

Approaching the historical phonology of three highly eroded Sino-Tibetan languages: Naxi, Na and Laze

by Alexis Michaud

G. Jacques and A. Michaud (2011). “Approaching the historical phonology of three highly eroded Sino-Tibetan languages: Naxi, Na and Laze.” Diachronica 28 (4): 468-498 plus online Appendix containing the reconstructed vocabulary.

Naxi, Na and Laze are three languages whose position within Sino-Tibetan is controversial. We propose that these... more Naxi, Na and Laze are three languages whose position within Sino-Tibetan is controversial. We propose that these languages are descended from a common ancestor ("Proto-Naish"). Unlike conservative languages of the family, such as Rgyalrong and Tibetan, which have consonant clusters and final consonants, Naxi, Na and Laze share a simple syllabic structure (consonant+glide+vowel+tone) due to phonological erosion. This raises the issue of how the regular phonological correspondences between these three languages should be interpreted, and which phonological structure should be reconstructed for Proto-Naish. The regularities revealed by the comparison of the three languages are interpreted in light of potentially cognate forms in conservative languages. This comparison brings out numerous cases of phonetic conditioning of the vowel by the place of articulation of a preceding consonant or consonant cluster. Overall, these findings warrant a relatively optimistic conclusion concerning the feasibility of unraveling the phonological history of highly eroded language subgroups within Sino-Tibetan.

The increasing of the number of syllables and the chronology of anaptyxis and prosthesis in West European loan-words of Ottoman Turkish

by Marek Stachowski

Continuous speech recognition using syllables

by Rhys Jones

With Simon Downey and John Mason. First author. Published in the proceedings of Eurospeech, 1997, vol. 3, pp. 1171-1174.

The vast majority of work in continuous speech recognition uses phoneme-like units as the basic recognition component.... more The vast majority of work in continuous speech recognition uses phoneme-like units as the basic recognition component. The work presented here investigates the practicability of syllable-like units as the building blocks for recognition. A phonetically annotated telephony database is analysed at the syllable level, and a set of syllable-based Hidden Markov Models (HMMs) are built. Refinements including the introduction of syllable-level bigram probabilities, word- and syllable-level insertion penalties, and the investigation of different model topologies are found to improve recogniser performance. It is found that the syllable-based recogniser gives recognition accuracies of over 60%, which compares with 35% as the baseline accuracy for monophone recognition. It is envisaged that practical applications of syllable recognition could be in a hybrid system, where the most common syllable HMMs would be used in conjunction with whole-word and phoneme models.

¿En qué se parecen el árabe, el coreano y el castellano medieval?

by Antonio Fábregas

Actas del II Congreso de la Asociación Coreana de Hispanistas. María Ángeles Álvarez Martínez (ed.), Alcalá de Henares, Universidad de Alcalá, pp. 53-63 (2003).