A few days ago, someone asked me a question about a common situation that's rarely discussed: How can an adult learn to communicate in a language they don't know, without access to courses and books and instructors? And what if the problem isn't just lack of foresight and preparation, because no courses or books or instructors exist for the language or dialect in question?
This question's background is an international development project, where many of the people to be reached are illiterate speakers of undocumented and unwritten languages, and are also often not fluent in the local lingua franca.
Some people may be skeptical of various aspects of the premise. But let's grant it and try to address the question.
Refugees, migrants, explorers, traders, missionaries and conquerors have been dealing with versions of this problem for millennia. But I don't know of any systematic discussion of solutions before the mid-20th century, when Ken Pike developed and taught his monolingual elicitation techniques.
Some aspects of Pike's approach assume an interest in linguistic analysis, and a basic practical understanding of the issues involved. He elicits a few nouns (rock, leaf, …), a few counting words (one, three, …), and a few words for qualities (white, black, red, green, ...), and then he explores questions like the nature of pluralization, the difference between modifying and predicating adjectives, the existence of modifier-head agreement, and so on — of course leaving open the possibility that these categories and relations are not quite as he expects them to be. Along the way, he develops and tests hypotheses about the phonological inventory, syllable structures and phonotactics, morphology, word order typology, and so forth.
Traditional approaches to this problem are no doubt generally less analytic, and also in most cases less monolingual — usually there are locals with at least some knowledge of another language shared with the outsiders. And the local language may also be closely-enough related to a known language for cognates to be helpfully available. Most linguistic fieldworkers similarly also take advantage of resources beyond monolingual elicitation, even though they share Pike's analytic interests and methods. (And access to polyglot helpers seems especially likely in the case of a modern economic development project.)
But where does this leave us in answering the original question?
I've watched talented linguistic fieldworkers sit down with a speaker of unknown Language X, and after a few hours they're chatting happily away — in a semi-pidgin variety of X, but still…
Can this ability be taught to non-linguists? (Or even to most linguists?) If so, what should the instruction be like? How much linguistics should be part of the instruction? How crucial is prior knowledge of a related language?
What about the role of individual differences in aptitude? Can only a few super-talented individuals succeed at the task of quickly acquiring minimal communicative competence in an unwritten language? Or is this something that almost everyone can do, even if some people go further faster than others?
Update — I've heard positive reviews of Pike's monolingual demonstrations, but here's some testimony indicating that he was mortal:
I had heard about Pike's monolingual elicitation from the time I was s student. In the mid 1980s we finally were able to invite him to <University> to do the demonstration. Your experience (or perhaps a video you may have seen) sounds like it was more successful than what we saw. It was as you describe (he came in with a branch, a rock etc.), but he made very little progress in the one hour. The language was Kinande. At the end he said a few things about how the language worked (some of which was quite wrong), and announced, "I don't think it's Bantu."