Speaking of Words
By MICHAEL FERBER

        In earlier columns I tried to say something reasonable about what the oldest living language might be (there isn’t one), and how language in general got started (we don’t know, but it may have needed a genetic mutation or two).  In this column I’d like to consider a related question: Given the many thousands of languages spoken today, and records of hundreds of extinct ones, do we have enough evidence to reconstruct “Proto-World,” the ancestor of all natural languages?

        Before we can begin to look at the evidence, however, we need to consider whether human language arose only once.  I think that is a reasonable assumption, because the possession of language would give a great advantage to one group of humans over another; on Darwinian grounds we can see how those with the great tool of speech would outcompete and replace those without it.  If it arose fairly suddenly in one human clan, however, why couldn’t it have arisen in a second clan, or a third?  Does natural selection work so fast that the first group that acquired speech would eliminate all the others before the same mutations could be repeated?  We cannot answer this question with any certainty, so it is possible that all human languages go back to two mother tongues, or three, or four.  If that is so, then it is hopeless to seek a single common ancestor.

        Another preliminary question is the time-depth we are dealing with.  Some anthropologists have argued that language is about 50,000 years old, because only modern humans of the Cro-Magnon type had a vocal tract that would allow clear discrimination of vowels, or because it is about 50,000 years ago that certain cultural artifacts such as representational paintings can be dated, or because the dateable settlements around the world of human groups of African origin go back that far.  But an article by geneticists appearing in Frontiers in Psychology this March claims that genomic studies show that the division of Homo sapiens from the “original stem,” after which all human groups developed language, took place 135,000 years ago.  By that date our ancestors had the capacity for language, and they probably realized that capacity by 100,000 years ago.

        Whether it is 100,000 years or half that long, the time depth is forbidding.  Over the last 250 years, such brilliant and exacting work has been done on reconstructing Proto-Indo-European (PIE) that, if we could take a time machine to visit the Yamnaya people on the Pontic-Caspian steppe, we could probably talk to them.  But that would be only 5000 years ago.  Recent genetic and linguistic research has postulated a precursor tongue, Proto-Indo-Anatolian, spoken in the Caucasus about 7000 years ago.  It has probable links to Caucasian languages still spoken today, and to both the Afro-Asiatic (including Semitic) and the Uralic families, which might take us back another 4000 years or more to something now posited as “Nostratic.”  With each step back, however, the ancestral language, if there was one, looks more and more vague and disputable. 

        Most historical linguists think that the horizon for reliable reconstruction is at most 10,000 years: though we can see similarities among languages farther back than that, reconstructing an ancestor descends into misty speculation.  Further work may clear up a few things, but if we want to go all the way back we are faced with a length of time longer by an order of magnitude than 10,000 years.

        A linguist named Merritt Ruhlen has recently presented data to show what certain words in Proto-World sounded like.  He thinks the word for “who” was ku, “what” was ma, and “hair” was sum.  Following other researchers he came up with these by reconstructing the likeliest word in PIE, Proto-Afro-Asiatic, Proto-Amerind, and several other major families.  Many of these words are highly speculative themselves and would have to be 20,000 years old.  He then looked to see what similarities there might be among these postulated words.  There is probably no other way to proceed, but Ruhlen was generous in detecting resemblances, and many linguists think his whole project is a waste of time.

        The example I found most wondrous is Ruhlen’s proto-word for “water”: akwa.  It’s amazing that the Romans and their ancestors somehow preserved this word for 50,000 years, or twice that, unchanged.  Italians still say acqua!  The trouble is, nothing like akwa was the PIE word for “water.”  The main one was something like wodr-, and was indeed the ancester of our word water, German Wasser, Greek hydor (whence hydrant, hydrate, etc.), Latin unda (“wave, sea”), and—farther back—Hittite watar.  The akwa-word may well have been picked up by Proto-Latin from another language family, and by Proto-Germanic as well, as we see in some placenames and in an older form of island.  So it may be an old word but it is a newcomer into our extended family.

        And look what has happened to aqua in just 2000 years.  Italian (or Tuscan) preserves it nicely (acqua), and Spanish preserves it fairly well but with a very soft consonant (agua).  In Romanian it is apa, in one dialect of Sardinian it is abba, in Provençal it is aigo, and in French it has shrunk to one syllable, eau (pronounced as a long o).  If a word can be nearly obliterated by internal changes in two millennia, how likely is it that any word can remain stable for fifty millennia, or a hundred?

        Or take the word for “dog.”  In Proto-Indo-European it was probably kuwon, and it appears much like that in Greek (kuon, root kun-).  In Latin it is canis (root can-), in Old Irish cu, but in Sanskrit shva, in Lithuanian shuo, and in German Hund.  Though these are readily traceable to PIE kuwon, the initial consonant has branched into three quite different sounds.  And in English, though we have the word hound, it is no longer our general term.  Instead we have dog, whose source is a complete mystery, as it is with the similar words frog, hog, pig, and stag.

        This sort of slippage in vocabulary in a relatively brief period should make us wary of large claims about “Proto-World.”  And it is not just words that slip and slide.  There are starkly different grammatical features among the world’s languages.  Words are all single syllables in some languages, and in others they often have more than ten syllables and act as whole sentences or clauses.  In some tongues the verb normally comes first, in others last in yet others in the middle.  Some languages have no genders, some have two, some three, some quite a few.   Some languages, such as Greek, attach prefixes and suffixes to indicate tense, aspect, mood, voice, person, number, case, and gender; others, such as English, do much of this with word order or auxiliary verbs.

        So the prospects are gloomy, alas, for figuring out what the words in the first language were like.  For all we know, “water” might have been a prefix or a suffix and not a word at all! 

        I am happy to hear from readers with questions or comments: mferber@unh.edu.

Michael Ferber moved to New Hampshire in 1987 to join the English Department at UNH, from which he is now retired. Before that he earned his BA in Ancient Greek at Swarthmore College and his doctorate in English at Harvard, taught at Yale, and served on the staff of the Coalition for a New Foreign Policy in Washington, DC. In 1968 he stood trial in Federal Court in Boston for conspiracy to violate the draft law, with the pediatrician Benjamin Spock and three other men. He has published many books and articles on literature, and has a deep interest in linguistics. He is married to Susan Arnold; they have a daughter in San Francisco.