Wired for Culture: Origins of the Human Social Mind

by Mark Pagel

  Linguists and linguist-missionaries have translated this list into thousands of languages around the world. The frequently used words in nearly every language are typically I, you, he, she, it, and other pronouns; the verb to be; number words like two, three, four, five; and who, what, where, why, and when. Other words in the Swadesh list such as scratch, guts, stick, throw, and dirty are typically used much less often, no matter what culture is studied. Putting this together with the results from the Oxford English Dictionary, we not only seem to be using language the same way around the world, but for the same reasons—principally to talk about each other, to refer to quantities of things, and to what people are doing, when they did it, to whom, and how much. These are also the words that we would expect to be used if language is for promoting and monitoring social relations, making exchanges, and advertising and assessing reputations. Because this information comes from a worldwide sample of languages, it is a reasonable step from there to guess that this will have been true throughout the history of human language use.

  The genes you pass on to your children will have been replicated or copied only a small number of times between you acquiring them from your parents and then your children inheriting them from you. Even so, each time a gene is copied there is a chance that a mutation or error will creep in, and over long periods of time genes can change out of all recognition. But for a word to last for even a generation it will have been spoken—its sound replicated—thousands or even millions of times, and of course this will be especially true of the frequently used words. If words could not be stably transmitted from mind to mind, then something I told you about someone else might get corrupted by the time it makes it way around the tribe—like what happens in a game of Whispers or Telephone. And something my mother said in her youth might not be intelligible to me by the time I am old enough to appreciate it. If this happened, our cultures would erode and decay.

  But this is not what happens. In fact, our languages can demonstrate quite extraordinary degrees of stability over long periods of time, far longer than is necessary for us to be able to communicate with each other throughout our lifetimes. Here, for example, is 1,000 years of the evolution of the familiar Lord’s Prayer, spanning Old to Modern English:

  Fæder ure þu þe eart on heofonum; Si þin nama gehalgod to becume þin rice gewurþe ðin willa on eorðan swa swa on heofonum.

  OLD ENGLISH—11TH C.

  Oure fadir that art in heuenes, halewid be thi name; thi kyndoom come to; be thi wille don in erthe as in heuene.

  MIDDLE ENGLISH—1380

  Our father which art in heauen, hallowed be thy name. Thy kingdome come. Thy will be done, in earth, as it is in heauen.

  EARLY MODERN ENGLISH: King James Bible—1611

  Our Father, who art in heaven, Hallowed be thy Name. Thy kingdom come. Thy will be done, On earth as it is in heaven.

  LATE MODERN ENGLISH Book of Common Prayer—1928

  Most of us have great difficulty now in reading the Old English, but what we see from one version of the Lord’s Prayer to the next is the gradual process that Darwin called “descent with modification.” We saw this earlier as a description of how species, over long periods of time, give rise to somewhat different daughter or descendant species. Darwin appreciated that a similar process was true of languages, saying in The Descent of Man, “The formation of different languages and of distinct species, and the proofs that both have been developed through a gradual process, are curiously parallel… .” Ancestral languages evolve to give rise to somewhat different daughter languages, which in turn do the same. Each of these daughter languages retains some of the features of its ancestral or mother tongue, but differences creep in as people fan out to occupy new areas.

  When this process is played out over a large area, and among different sets of ancestral and descendant languages, entire family trees of related languages evolve. This was recognized as early as the late eighteenth century, nearly a hundred years before Darwin, by an English judge, Sir William Jones, working in colonial India during the reign of George III. To process court papers, Sir William found it necessary to learn Sanskrit, and in doing so he became aware of curious parallels between Sanskrit, Latin, and Greek. Jones described these to a meeting of the Asiatic Society in Calcutta on February 2, 1786, noting that the Sanskrit language bears “a stronger affinity… [to Greek and Latin] . . . both in the roots of verbs and in the forms of grammar, than could possibly have been produced by accident; so strong, indeed, that no philologer could examine them all three, without believing them to have sprung from some common source, which, perhaps, no longer exists.” For instance, the Sanskrit word for the English “brother” is bhratar; the ancient Greeks used the word phrater (ϕρᾱτηρ) to mean something akin to a fraternity or brotherhood; and frater is the Latin word for “brother.” Another example is the Sanskrit word for “three” or tri, which is tria (τρία) in Greek and Latin. The familiar-sounding khanda in Sanskrit is the French word candi and the English “candy.”

  What Jones had identified would later be recognized as the Indo-European language family, and it includes the languages currently spoken all over Europe, parts of Central Asia, and the Indian subcontinent. The archaeologist Colin Renfrew in his Archaeology and Language links the spread of these languages to the origin of agriculture sometime around 9,000–10,000 years ago in the region of the world known as the Fertile Crescent (roughly present-day Turkey and Iraq). Farming reset the world’s carrying capacity to a higher level, allowing a greater number of people to survive in a given area. The growing populations meant that farmers and their ideas spread out in all directions from the Fertile Crescent. Those that went north and west formed what we recognize today as the Greek, Germanic, and Romance or Latinate languages of Western Europe; those that went north and east produced the Slavic languages; and those that went south gave rise to the languages of Iran, Afghanistan, Pakistan, and India. The Basque language of northern Spain, although in Europe, is not an Indo-European language. In fact, it might be an isolated remant of the languages that were spoken by the hunter-gatherers in Europe before farming and farmers arrived. Russell Gray and his colleagues were later able to confirm Renfrew’s arguments by applying dating techniques to the Indo-European languages, confirming that this family probably did arise sometime around 9,000 years ago. And it is because of descent with modification within the separate branches of this family that today we recognize similarities in the Romance languages of Spanish, French, Italian, and Portuguese, just as we recognize similarities in the Germanic languages of German, Danish, Dutch, and English.

  But how long can the words causing these similarities last, and are there some in particular that turn up over and over in different languages as ones that get retained for the longest periods of time? It turns out that it is the small subset of words that we use most frequently in our everyday speech—and especially those we have suggested are related to social relations—that can show truly startling resistance to changing, sometimes being conserved in a related form across all Indo-European languages. For example, linguists recognize that the number word for two of something is probably derived in all Indo-European languages from the same shared ancestral sound that has been conserved for thousands of years. Thus, in Spanish the word is dos, it is twee in Dutch, deux in French, due (doo-ay) in Italian, dois in Portuguese, duo (δύο) in Greek, di in Albanian, do in Hindi and Punjabi, and Caesar would have said duo. This leads to the proposal that the original or proto-Indo-European word was also two-like in its sound, and indeed some scholars suggest duwo or duoh.

  A handful of other words, including the words for three and five, who, I, and you, are also highly conserved like the word for two. For example, the English three is tre in Swedish and Danish, drei in German, tre in Italian and tres in Spanish, tria in Greek, teen in Hindi and tin in Panjabi, and tri in Czech. The proto-Indo-European word for three might have been trei. These conserved words are closely followed by other pronouns such as h
e and she, and by the what, where, and why words, all of which can show a striking degree of similarity among many Indo-European languages. Remembering that the Indo-European languages all derive from a common ancestral language, this tells us these words have been retained in separately evolving branches of this family tree, each one of which represents up to 9,000 years of language change. If we add up the time in each of these branches, we see that some words have not substantially changed their forms for what is, in effect, hundreds of thousands of “language-years” (thus, for example, the sound for three has been evolving separately in the Germanic, Romance, Slavic, and Indian languages).

  This degree of conservation tells us that the people alive in the parts of the world where the Indo-European languages arose might have been using forms of these slowly changing words that we could still recognize today. Their words would not have been identical to our modern words, but close enough that if we travelled back in time and encountered a gang of three of our proto-Indo-European ancestors, we might be able to point to ourselves and say, “I, one,” then pointing to them, say, “You, three,” and be understood. Now, this might not be a wise or even very imaginative thing to do, and you might think it a limited conversation anyway. But consider that if you are of Indo-European origins and you are reading this book on an international flight, you might have less in common linguistically with the person seated next to you than you would with a linguistic ancestor who lived close to 10,000 years ago.

  How far back in time can we go with words? The existence of a set of highly conserved words raises the possibity that we might be able to find some evidence for what our mother tongue, or the language of the very first humans, was like. The linguist Merrit Ruhlen has proposed twenty-seven “global etymologies,” or words left over from our original or proto-language, citing evidence that they are found in many language families from all over the world. Ruhlen’s list includes words for who, what, two, water, and finger or one. For instance, Ruhlen points out that the sounds tok, tik, dik, or tak surface repeatedly in these language families as a word for the number one or toe or a digit. Ruhlen’s proposals have always been highly controversial among linguists, but it should not escape our attention that he includes words we have seen are among those that are both frequently used and highly conserved. They are also those we might have expected if indeed we have used language to monitor and manipulate social relations throughout our history.

  Whether or not we can ever reconstruct the human mother tongue, we should be astonished that words can be retained and conserved over thousands of years and potentially millions of speakers. For a word to be transmitted, a sound I make must travel through the air and enter your ear, where it is turned (transduced) into an electrical signal that travels to your brain. Then, at a later time when you want to use the word, your brain must somehow send messages to your mouth and lungs to get you to produce the same sound. That sound will then travel to someone else’s ear, where the process will be repeated. The opportunities for corruption or loss of signal are many. We should also remember that, unlike for genes, there is seldom any necessary connection between a word’s form (its sound) and its meaning—I might just as easily call a tree a table, and vice versa. Where a gene’s chemical form is directly related to the protein it makes, this “form-function” connection is generally only true in language of the so-called onomatopoeic words that imitate sounds, like bang, meow, moo, woof, or pop.

  We might have suspected that the more a word is used, the more likely it is to acquire mistakes or errors as it passes out of successive speakers’ mouths, into others’ minds via their ears, and then out again. Instead, it is the words we use less often that are more prone to changing. Consider that in contrast to a highly conserved word like two, what English speakers call a bird the Germans call a vogel, the Spanish say pajaro, the Italians ucello, the French oiseau; to Aristotle and modern Greeks after him, the word is pouli, and Caesar would have said avis. Words like dirty and guts are even more variable. For instance, compared to the conservation of sounds for two of something, there are at least forty-five different words for dirty among the Indo-European languages (four of these are the German schmutzig, Dutch vuil, French sale, and Spanish sucio). In fact, a given word for dirty gets replaced by some new sound every 800 years or so, compared to the many thousands of years a word like two can last.

  Are these infrequently used words more prone to change because they are somehow less important and so it doesn’t matter so much if someone comes up with a new word or mispronounces an existing one? This might be part of the answer, but only indirectly. When we think about it, we realize that because there is seldom any connection between a word as a mere sound and that word’s meaning, different words (sounds) have to compete with each other for the privilege of carrying a particular meaning. We can pick and choose among them, and this competition for our affections is natural selection acting on words, sieving out the winners from the losers. Among English speakers, nearly everyone uses the word kitchen to describe the place where they prepare their food. That sound has won in competition with other sounds we might produce—such as scullery or galley—to describe that particular meaning. In other cases, we can see the competition for space in our minds going on in front of us in common words like sofa and couch or living room, sitting room, reception room, or parlor.

  Which of these forms will win? There are no simple answers, but precisely because there is seldom any necessary connection between a sound and its meaning, the competition often focuses on characteristics of the sounds themselves, and we can expect the competition to be more intense the more we use a word. One of the most common ways that frequently used words adapt to our minds is to get shorter. Do you say automobile or car? Refrigerator or fridge? Cannot or can’t? The Harvard linguist George Kingsley Zipf had recognized this relationship between frequent use and shorter words by the 1930s and it is now enshrined in the eponymous Zipf’s law. We give people “diminutives,” or shorter names, for names like Alexander (Alex) or William (Will, Bill), Arthur (Art), Richard (Dick), David (Dave), Michael (Mike), but we seldom if ever lengthen a name. We see this too in the contractions—can’t, won’t, don’t, and in familiar usages like g’morning or g’day. So strong is the effect Zipf’s law describes that none of the words that English-language speakers use more than a few thousand utterances in every million is longer than around five letters (words like the, and, and is might be used 10,000 to 30,000 times per million). If a sound “wants” to be highly used as a word, it can’t afford to be long. When we say someone uses “big words” to mean he or she uses words we don’t know, we are, without knowing it, recognizing that the words we don’t know are typically ones we don’t hear very often, and so natural selection has not pared them down to size.

  We learn from this that out of all the possible sounds we could make as words, only a far smaller number has ever been used. The Oxford English Dictionary catalogues around 250,000 English words, of which around 50,000 are considered extinct or no longer used. This might sound like a large number, but even 250,000 words is just a minuscule subset of all possible words. Consider just for sake of discussion that English conventionally defines five vowels and twenty-one consonants. Let’s consider how many six-letter words we could make from this set of twenty-six letters, restricting ourselves to words with two vowels and four consonants, as in a word like letter or friend. If words are unrestricted, then each of the four consonants could be any one of the twenty-one that we can choose from and each of the two vowels can be any one of five. This leads to 21×21×21×21×5×5 = 4,862,025 possible six-letter words with two vowels. Let’s now move to seven-letter words also with two vowels: say, a word like letters or friends. There are 21×4,862,025 = 102,102,525 possible varieties of these!

  This simple exercise tells us that the survivors among all the possible sounds that could have become words—that is, the words we actually use—are a highly rarefied set of winners of an overwhelming competition to o
ccupy our minds. For example, there are no words in English that begin with “ng” and words that begin with “q” are nearly always followed by “u.” Adding to this already intense competition among sounds is that our minds are remarkably good at remembering words, meaning we can compare them with ease. In fact, so good are our minds at remembering words that we easily recognize words that are used less than once per million utterances in normal speech, even if we cannot always remember their meanings. Adumbrate, dilatory, feculent, parvenu, fractious, and traduce are all used perhaps one to five times per million utterances. Eponymous—mentioned earlier—is another. How many do you recognize? How about demanate? Don’t recognize it? Good, because it is not recognized as an English word. If you got all of these right, or even most of them, then your mind can routinely discriminate words it hears at only very low frequencies from so-called words like demanate that it should have never heard.

  We are beginning to get an answer to our question of why frequently used words might be so stable. It is not necessarily their importance to communication per se. Rather, the frequently used words might be stable over long periods of time because they have become so highly adapted to our minds that it becomes difficult for a new form to arise to outcompete or dislodge them. It is no accident, then, that all of the highly used and highly conserved words that we use in talking about social relations—who, what, where, why, when, you, me, I, he, she, we, it, and the number words from one to ten—are short, mostly monosyllabic, and easy to pronounce. This isn’t just a property of English or Indo-European languages more generally. Listen to the numbers one to nine for the New Guinean language of Mangareva: tahi, rua, toru, ha, rima, ono, hitu, varu, iva. The frequently used words also evolve to be distinct from each other and therefore less prone to being misunderstood. Try saying the words mail and nail, fail and veil, or bale and pail over the phone to someone, and there is a good chance you will be misunderstood. Now, imagine fail means two and veil means ten, and you are using these words to order shirts over the phone from a clothing retailer: chances are if you wanted two, your bill might be higher than you expected. But this sort of thing seldom happens with the number words because we don’t mishear them—they have evolved to be distinct. And this is also true of the words we have said are related to social relations more generally. Humans have domesticated language to make it efficient at communicating what they want to talk about the most.