Waking a sleeping language
A. When the last fluent speaker of a language dies, linguists usually record the event with a date and a note of regret. The language is then classed as extinct, and the matter is considered closed. Yet a small but growing number of communities have refused to accept this verdict. Instead of mourning, they have set about the unlikely task of bringing their ancestral speech back into daily use, a process that specialists prefer to call revitalisation rather than resurrection, since what returns is never quite what was lost.
B. The raw material for such projects is rarely generous. Typically it consists of a few old word lists compiled by travellers, some grammar notes written by an enthusiastic but imperfectly trained amateur, a handful of scratchy recordings and the half-remembered phrases of elderly relatives who heard the language as children but never spoke it. Reconstructing a sentence from these fragments resembles restoring a mosaic when most of the tiles are missing: the pattern can be guessed, but the guesses cannot be checked. Pronunciation is the greatest difficulty, because written records of the past preserve spelling conventions far better than they preserve sounds, and even a confident reconstruction may owe more to the imagination of its author than to the speech of his ancestors.
C. The case of Tarvani, spoken historically on the island of Kelmar, is instructive. By the 1950s its last native speakers were in their eighties, and the island's only school taught nothing but the mainland language. In 1994 a retired surveyor named Idris Holm began collecting every Tarvani word he could find, and within a decade he had assembled a dictionary of some six thousand entries. His critics, including several professional linguists, complained that he had filled gaps with invention, borrowing missing words from a related language and bending them to fit Tarvani sound patterns. Holm replied that a language with holes in it is of no use to a child who wants to ask for something at the breakfast table. The argument was conducted with more courtesy than such arguments often are, but it was never settled.
D. The dispute points to a deeper tension. Scholars are trained to value accuracy and to distinguish sharply between what is attested and what is merely conjectured. Speakers, on the other hand, need fluency, and fluency requires vocabulary for ordinary things, which the historical record, preoccupied with ritual and trade, tends to omit. Nobody wrote down the Tarvani word for a microwave oven, or indeed for a kettle. A revived language therefore has to grow, and growth means choices that no surviving source can authorise. Some communities settle the question by committee; others leave it to usage, accepting that whatever young people actually say will prevail in the end.
E. Perhaps the most striking finding concerns who ends up speaking the revived language best. In Kelmar, adults who attended evening classes mostly remained hesitant, translating in their heads from the mainland language and apologising for their mistakes. Children raised from infancy in Tarvani-speaking homes, by contrast, developed an ease that their parents never achieved, and they introduced forms that appear in no grammar book. One such form, a simplified way of marking past time, has now spread to nearly every speaker under thirty. Older enthusiasts regard it as an error; the linguists who study the island regard it as the first sign that the language is genuinely alive, since only living languages change without being told to.
F. Economic arguments are seldom made in favour of these efforts, and with good reason. Producing teaching materials for a few hundred learners is expensive, and the practical benefit of speaking Tarvani is slight when every islander already speaks the mainland language. Supporters tend to defend the work on other grounds: that a language carries local knowledge, such as names for tides, winds and seabirds that have no tidy translation, and that belonging to a community is felt more keenly when that community has a voice of its own. Whether such benefits can be measured is doubtful, and the researchers who try are accused of reducing something intimate to a statistic.
G. It would be unwise to conclude that revival is easy or that it always succeeds. Many projects collapse when a single dedicated teacher moves away, and some produce a language that survives only in ceremonies, much as Latin survives in certain formal phrases. What the Kelmar experience suggests is more modest: that a language need not have an unbroken chain of native speakers to be spoken again, and that the line between extinct and sleeping may owe less to linguistics than to the determination of the people who inherit the silence.