When a language disappears, we lose a valuable window into our own biological endowment — language.

Much of the literature on language endangerment and death focuses on our understanding of language variation. While this is usually not the main concern of the language communities in question, it is of interest to linguists who study grammatical and sound patterns across languages.

Modern linguistics is founded upon the notion that the structure of language and the limits of its variation are set by our innate abilities to process and produce language. If this is correct, then understanding our linguistic capabilities as humans can only proceed through understanding the commonalities across all human languages. On this approach, when a language disappears, we lose a valuable window into our own biological endowment.

To take two examples from our own work at ELA, linguist Joseph Greenberg proposed the following apparent universal constraint on human languages: if an auxiliary (words like “can”, “will”, “do” in English) precedes the main verb (as in English “He will eat”), then the verb will precede the object (as in English “He eats pineapple”). Conversely, if the auxiliary follows the main verb, the verb will follow the object (as in Japanese, Hindi, Turkish, and myriad other “verb-final” languages). The Garifuna language counter-exemplifies this putative universal as the verb precedes the object but auxiliaries must follow the verb. To take another example, interrogative words like “who”, “what”, “where”, etc. typically either stay in their expected position in questions, e.g. “You said WHAT?”, or move to the beginning of the sentence, e.g. “What did you say?”. It is safe to say that these two surface patterns describe how questions are formed in well over 95% of the world’s languages. Two languages we have had the good fortune of being able to work with in New York show far more interesting patterns, however. In Ossetian, the interrogative must immediately precede the verb in questions. Whereas the regular word order in a statement is Subject-Object-Verb, e.g. “He me saw”, when questioning the subject, it would be Object-Subject-Verb, e.g. “Me who saw?”. An even rarer pattern is found in the Bantu language Ikota. Here, the interrogative must come at the end of the clause, an apparent mirror image of English. In Ikota the regular word order in statements is Subject-Verb-Object, e.g. “She saw me”, but when questioning the subject this order becomes Verb-Object-Subject, e.g. “Saw me who?”.

Such uncommon patterns are of great importance to the goal of identifying language universals. We are only beginning to explore the consequences of these particular patterns for our understanding of universal grammar.