We are all aware of similarities in words between different languages, and also how much some words vary between cultures separated both in time and location. Often it is the most common words in a language that retain a strong similarity to their origin languages: for instance, the English word brother and the French frère are derived from words in ancient languages: the Sanskrit bhrātr and the Latin frāter. It is obvious, therefore, that the distinctive sound of a word can remain associated with the same meaning for thousands of years. But how far back in time can we go to find common words?
A 2012 PNAS paper, "Ultraconserved words point to deep language ancestry across Eurasia", attempted to answer that question, with surprising results. Researchers noted that these early 'root words'...
...can be predicted from information independent of their sounds. We showed in a sample of Indo-European languages that the frequency with which a word is used in everyday speech, along with its part of speech, can predict how rapidly words evolve, with frequently used words on average retained for longer periods of time
We have recently extended this result to include speakers from the Uralic, Sino-Tibetan, Niger-Congo, Altaic, and Austronesian families, in addition to Indo-European, plus the isolate Basque and the Creole Tok Pisin. Even in languages as widely divergent as these, we found that a measure of the average frequency of use predicted rates of lexical replacement as estimated in the Indo-European languages.
The study uncovered 23 "ultraconserved words" that "point to deep language ancestry across Eurasia". How deep? This subset of words, the researchers believe, "have remained associated with their particular meanings independently in separate branches of this superfamily since the end of the last ice age."
The 23 words identified by researchers included obvious ones (in terms of common usage) - “I,” “we.” “who,” “not,” “that,” “mother,” "man" - but also less commonly used words today which nevertheless were likely very important some 15,000 years ago, such as "fire," "ashes,", “bark,” and "worm".
Researchers noted that their unique approach in predicting these words independently of their sound correspondences "dilutes the usual criticisms leveled at such long-range linguistic reconstructions, that proto-words are unreliable or inaccurate, or that apparent phonetic similarities among them reflect chance sound resemblances."
So if you're planning on time travel back to the end of the last Ice Age any time soon, it might be worth brushing up on those 23 words...
...
Sources and further reading