Languages Without Sibilants

Over at a tear in the fabric of spacetime, Rachel mentions reading the claim that all languages have a sibilant consonant, and realizing that this isn't true―Bardi is a counterexample. I was able to find several more counterexamples using the UPSID database.

UPSID stands for UCLA Phonological Segment Inventory Database. It contains the phoneme inventories of 451 languages, with each phoneme described by a set of distinctive features that can be queried on. So, without too much trouble I was able to produce a list of languages that are in the database, but have no phonemes that are [+sib]. Here's the list of languages and their genetic classifications:

ALAWA: Australian, Maran
ARRERNTE: Australian, Pama-Nyungan
AUCA: S. American, Andean
BARDI: Australian, Nyulnyulan
BURARRA: Australian, Burarran
DERA: Papuan, Trans-New Guinea
DINKA: Nilo-Saharan, Eastern Sudanic, Nilotic, Dinka-Nuer
DIYARI: Australian, Pama-Nyungan
DYIRBAL: Australian, Pama-Nyungan
EKARI: Papuan, Trans-New Guinea
GADSUP: Papuan, Trans-New Guinea
GARAWA: Australian, Garawan
GUGU-YALANDYI: Australian, Pama-Nyungan
HAWAIIAN: Austro-Tai, Austronesian, E. Malayo-Polynesian, Oceanic, Polynesian
KALKATUNGU: Australian, Pama-Nyungan
KOIARI: Papuan, Trans-New Guinea
MALAKMALAK: Australian, Daly
MARANAO: Austro-Tai, West Malayo-Polynesian
MAUNG: Australian, Yiwaidjan
MBABARAM: Australian, Pama-Nyungan
MURINHPATHA: Australian, ungrouped
NASIOI: Papuan, East Papuan
NGARINJIN: Australian, Wororan
NGIYAMBAA: Australian, Pama-Nyungan
NUNGGUBUYU: Australian, ungrouped
NYAH KUR: Austro-Asiatic, Mon
RORO: Austro-Tai, E. Malayo-Polynesian
ROTOKAS: Papuan, East Papuan
SENTANI: Papuan, Trans-New Guinea
TIWI: Australian, ungrouped
WARAY: Australian, ungrouped
WESTERN DESERT: Australian, Pama-Nyungan
WIK-MUNKAN: Australian, Pama-Nyungan
YANYUWA: Australian, Ungrouped
YIDINY: Australian, Pama-Nyungan
YOLNGU: Australian, Pama-Nyungan

Looks like a lot of Australian languages, some Papuan and Austro-Tai languages, and a couple of odds and ends. That's an areal feature if I've ever seen one.

Comments

[entangledbank]

A selective database like UPSID even disguises this. It's virtually universal for all Australian languages, of all families. No fricatives or affricates. (The palatal stop [tj] may be affricate-like in some.)

Polynesian languages have either [s] (e.g. Samoan) or [h], e.g. Samoan island Savai'i = Hawaiian Hawai'i = Maori legendary Hawaiki.

[Claire]

I looked at spectrograms of Bardi palatal stops a while ago because the word-final ones were sounding like affricates. They aren't, though. They have a closure but the noise after the release isn't an s, the energy peak is about 1500Hz too high. It looks more like heavy aspiration.

The other place people occasionally claim there's an s is word-final trills. That's just wrong (at least for Bardi). They're devoiced trills. Helped my Turkish accent no end...