Tambora was not Austronesian. This single fact makes it extraordinary. The Austronesian language family is one of the largest on Earth, stretching from Madagascar to Easter Island, and it dominates the Indonesian archipelago almost completely. Yet Tambora belonged to the Papuan language group -- the non-Austronesian languages found primarily in New Guinea and scattered islands to its west. It was the westernmost known Papuan language, an outlier separated by hundreds of kilometers of ocean from its nearest relatives. Linguist Mark Donohue identified a plausible connection in the word taintu, meaning hand, which resembles words in the Papuan languages of Timor and Alor: tang in Abui, tana in Oirata, tan in Kui. If the connection holds, Tambora was a surviving thread of a linguistic web that once stretched more broadly across eastern Indonesia before the Austronesian expansion swept most of it away.

Working from a vocabulary of fewer than fifty words is like reconstructing a cathedral from a doorknob. Yet linguists have teased out surprising structural insights. Body-part words share a common prefix pattern: saing'ore for eye, saing kome for nose, sontong for teeth, sumore for belly. The initial syllable appears to be a possessive marker with a nasal element that assimilates to the consonant following it. The numbers reveal even more. Several begin with a sV- prefix, a pattern common in Austronesian languages where the word for "one" shrinks into a counting prefix. Donohue suggests that sarone, meaning ten, paired with sisarone for twenty, may preserve traces of a vigesimal counting system -- base twenty rather than base ten -- possibly derived from sa- meaning one and doh meaning person, a common way of expressing twenty across the region. If so, the decimal system arrived later, imposed by trade with partners who counted differently.

What makes Tambora unusual among Papuan languages is the kind of society that spoke it. Papuan languages are typically associated with small-scale communities in mountainous or remote terrain. Tambora was the language of a maritime trading state, a kingdom whose pottery has been matched to Vietnamese styles and whose merchants exported sandalwood, honey, and sappan wood across the East Indies. The Malay loanword makan -- to eat -- appears in the surviving vocabulary, evidence that Tambora speakers operated in a multilingual commercial world. Only two other Papuan-language trading states existed at the time: Ternate and Tidore, the spice-trade sultanates off the coast of Halmahera. Tambora's linguistic isolation, combined with its economic sophistication, makes it one of the most unusual cultural configurations in the pre-colonial Indonesian world.

One word in the list carries a particular resonance: morihoh, meaning God. The term echoes similar words found across eastern Indonesia, possibly derived from Sanskrit, carried eastward by centuries of Hindu-Buddhist cultural influence. But in Tambora, morihoh also resembles homori, the word for father. Whether the resemblance is coincidence or reflects a theological concept linking the divine to the paternal is impossible to determine from a single word list. Donohue notes that neither word can be confidently assigned as native Tambora vocabulary. This ambiguity captures the fundamental challenge of studying the language: every interpretation rests on a foundation so thin that certainty is unattainable. What is certain is that morihoh was the last word anyone would ever speak for the concept of God in the Tambora language.

Among the surviving words, a pattern emerges that no one will ever fully explain. Several words end in -kong or -ong: ingkong for sun, kongkong for day, mang'ong for moon, kingkong for star. The semantic clustering is striking -- all relate to celestial bodies or time. Is -kong a suffix meaning something like "light" or "sky-thing"? Or is the resemblance accidental, a trick played by a sample too small to distinguish pattern from noise? Another set of words ends in -ore, though their meanings are less clearly grouped. These dangling threads -- suggestive but unresolvable -- are the permanent condition of Tambora linguistics. Raffles preserved just enough to prove that something remarkable once existed, and not nearly enough to say what it was.