Kalamang language

Kalamang, sometimes also called Karas, is a divergent Trans–New Guinea language spoken on the biggest of the Karas Islands off the Bomberai Peninsula, that is part of the West Bomberai family. It is spoken in Antalisa and Mas villages on Karas Island.^[2]

Kalamang
Region	West Papua
Native speakers	(100 cited 2000)[1]
Language family	Trans–New Guinea Berau Gulf West Bomberai Kalamang
Language codes
ISO 639-3	kgv
Glottolog	kara1499
ELP	Karas
Karas is classified as Severely Endangered by the UNESCO Atlas of the World's Languages in Danger
Kalamang Show map of Western New Guinea Kalamang Show map of Indonesia Kalamang Show map of Southeast Asia
Coordinates: 3.47°S 132.68°E / -3.47; 132.68

Phonology

Consonants^[3]

	Labial	Alveolar	Dorsal
Plosive	p b	t d	k g
Fricative	(f)	s	(h)
Nasal	m	n	ŋ
Approximant	w	r, l	j

• The consonants /f/ and /h/ are marginal.

Vowels^[3]

	Front	Central	Back
High	i		u
Mid	e		o
Low		a

• The vowels /a e i/ are reduced to [ə] in unstressed syllables in fast or casual speech.

Additionally, the following diphthongs are present: /ei/, /oi/, /ou/, /ui/.

Pronouns

Cowan (1953) records the following pronouns for Karas.

		singular	dual	plural
1st person	exclusive	aan	inir	piridok
	inclusive			aantemu (?)
2nd person		kame	?	kijumene
3rd person		mame	mjeir	mubameir

Visser (2020) records the following pronouns for Karas of Maas village:

Free nominative singular dual plural 1st person exclusive an in-ier in inclusive pi-er pi 2nd person ka ki-er ki 3rd person ma m-ier mu

Free possessive singular plural 1st person exclusive aŋ-gon pi-n inclusive iŋ-gon 2nd person ka-in ki-n 3rd person ma-in mu-in

Possessive suffix singular plural 1st person exclusive -an -pe, -p-in inclusive -un 2nd person -tʃa -tʃe 3rd person -un -un

The free possessives and possessive suffixes can occur together.^[4]

Machine Translation from One Book

In 2023, Kalamang was used by machine learning researchers for a benchmark called "Machine Translation from One Book". It was chosen because of its negligible presence in the Internet and because field research materials were collected by Eline Visser, who published "A grammar of Kalamang" as her PhD thesis. Although Kalamang is primarily oral language, it can be written in the Indonesian alphabet. Researchers used all existing materials (grammar book, short dictionary, and small set of Kalamang-English sentences) to test how large language models (LLM) can learn a language from a single source, and tested the quality of translations.^[5] In 2024, researchers from Google showed that their latest LLM, Gemini 1.5, can translate English to Kalamang with similar quality to a human who learned from the same resources.^[6]