A portion of the dramatic painting of the biblical Tower of Babel

What we know from 150,000 years of human language

There was likely once a single language, first developed 150,000 years ago. That grew to as many as 100,000 languages, before we developed farming. Today there are 6,000 and by 2100, that might be back to as few as 500.

Along the way, languages have emerged, influenced each other and continue to change.

That comes from the 2003 book “Power of Babel,” the third consecutive book I read by linguist John McWhorter, which I finished early this summer. In the last six months, I’ve become quite a big fan of his — having read his 2016 book on language evolution and his 2009 book on the lesser-known stories of the English language history, I seem to be working through his language books in reverse chronological order. (Read the Guardian’s review of this book here.)

The title of the book is, of course, a reference to the biblical story in Genesis of the Tower of Babel. Following the Great Flood, humans speaking a single language plan to build a tower that can reach heaven. God destroys it, sets humans into an array of languages and spreads them across the world to keep them from conspiring to do something like that ever again.

Find my notes from the book below.

Here are some of my favorite lessons:

  • So-called “primitive languages” (ones that are not written and spoken in smaller, less economically connected societies) are actually more complex than ones spoken more widely, as languages get simpler with more nonnative speakers. (p. 6)
  • Linguists put the development of verbal communication at 150,000 years ago  because we believe that mutation happened while we were still in Africa.
  • The (wrongly) presumed degradation of language is like former Congressman Patrick Moniyhan’s “defining deviance downward” concept on normalizing crime from 1993. (20)
  • Linguists use history and sense to chart back language (including poetry to track rhyming) (44)
  • Media and writing in America and elsewhere now slow dialectic change (62) this began to develop our sense of one true language not a collection of dialects. Consider the cat metaphor: there isn’t one colored cat that is the official one but we now do expect a standard German, an historic anomaly.  That sense of language standardization transformed with the rise of nationalism (64)
  • Like “Fratlin” earlier (the generations-long transition period between Latin and French), language transitions slowly. In 1293 Dante wrote in Italian over Latin for first time (67), a language that “even housewives can converse,” noting the idea of writing was that it was for the higher-minded.
  • Geopolitical and culture determine language v dialect (85)
  • People use multiple dialects like formal writing and casual at home (88) That’s diglossia
  • Most (maybe 99 percent) of English words started as loan words (Latin, French, Germanic) not Old English (95), though those that remain are heavily used ones (like “and,” “but,” “father.” “love,” “fight,” and “from”)’
  • The three language invasions to the Germanic root of English mean many of our compound words lose their sensibility (Some 1,000 Old Norse words came from the 787 Viking raids; the 1066 Norman Invasion brought 7,500 words still used today and heavy usage of Latin by the elite and clergy) (96) so our compound words are often not that to us. That’s why French and Latin vocabularies are easier for English speakers but German grammar is. (98)
  • Though English is Germanic in root (and grammar), it’s heavy reliance on vocabulary from the Romance languages and influences from many others is one reason English learners are not good second language learners (100)
  • The words “castle” and “chateau” are both the same word brought into English from specific dialects at specific times in French’s language journey. (103) It’s also why we have warranty and guarantee.
  • A language annihilation is about loss of grammar not an infusion of words but we aren’t conscience of grammar so we focus on infusions of “outside words” when we complain. (118)
  • In 1561, John Clarke wrote that “English should remain “unmangled by borrowing of other tongues”
  • Many countries have so many languages that simple pidgins develop not for reading literature but for communicating. For example, there are 800 languages in Papua New Guinea and 250 in Nigeria (132). Look at Russenorsk for an extreme example.
  • The difference between a pidgin (not a full language) and a creole (a new, fully formed language) has the same gray area of “blue-green crayons in the Crayola box” (162)
  • Creoles are mostly about forced migration (having to learn pidgin as primary language ) so they’re largely associated with displaced or enslaved brown people but there are exceptions like Afrikaans.
  • Linguists have specific characteristics that define a creole, like simplifying past tense verb endings, reducing gendered nouns and using subject-verb-object sentence construction (164)
    Creoles can be disagreed: Guyana and Jamaica would say they speak English but they have semicreole local elements (166)
  • AAVE is dialect not semicreole for clear reasons (169)
  • An infix is like “fan-fucking-tastic.”
  • Note that the “have found” vs “he found” is past perfect versus past, Russian and other languages don’t have this but we do. (178)
  • English doesnt have evidential markers like many Amerindian languages (where verbs grammaticized into suffixes tell the listener the source of information, like “saw” or “heard” it (180) and things like inalienable possession in which languages like Navajo have markers to show you “have” something integral (like a body part) vs. a possession like a couch.
  • “Je me lave” in French is an example of reflexivity in which you must say you’re bathing yourself (as opposed to “I bathe” and “I bathe myself” being the same in English) (183)
  • Many language elements are overgrowth like traits selected alongside others that are adapted. Not all language elements are explicitly for communication but appear ornamental (190) These are spandrels, like in evolutionary biology.
  • It bears repeating that less industrialized civilizations tend to have more complicated languages, be they having had less foreign entanglement for second language learners (201) almost like “inside jokes” with a small group of people, it’ll get more elaborate over time.”
  • Creoles are the world’s simplest languages because they are younger than 500 years old and built on simplified pidgins
  • Major languages have developed far more slowly in last 500 years than before and that’s because of standardization and literacy (221) but remember how new writing is and not how most languages are consumed today and certainly not most in human history
  • Written language holds a formality from an older French and lags behind spoken language. He gave two great examples that were totally new to me:
    1. I learned the standard French “Nous” for the first person plural but really native continental French speakers always use the impersonal, indefinite pronoun “On” in those settings, though “nous” is commonly written and used formally.
    2. In casual spoken French, to create the negative it is far more common to drop the “ne,” which is the first in the negative creation. (224)
  • The concept of freezing language rules in place started with early grammarians like Lowth and then Murray.
  • Writing keeps forms in lock. For example, “you were” for singular is because we dropped “thou” for singular but kept plural verb conjugation (230)
  • Written and spoken English are different altogether (238). For example, we are more redundant and use more conjunctions when speaking with shorter bits.
    Writing in the 1400s still mirrored talking but by the 1640s English had a standard writing tone that was distinct and more flowery (241).
  • Spoken languages often don’t have subordinate clauses, as we don’t speak em. Likely writing invented them so the Proto Indo European language likely had none (248) Similarly, native Alaskan languages and Somali developed them only after starting to write.
  • Speaking is communication; Writing is art. (251)
  • Of the 20 most commonly spoken languages, 7 are in India (257) and 96 percent of people on the world speak one of those 20 languages.
  • A language is dying every two weeks (258), which will bring us from 6,00 languages today to 500 in 2100. At peak linguistic equilibrium some 11,000 years ago, there were perhaps as many as 100k languages spoken.
  • Language consolidation may be inevitable and our belief in sustaining more native languages in an active sense might be wishing people don’t join the global economy (274) Linguist Peter Ladegofed argued this in a prominent article in the Language journal, if a small remote village speaker wants to join the global economy with a major language: “who am I to argue?”
  • McWhorter does highlight the important work many linguists are doing to preserve language in writing and recording but is somewhat fatalistic about the idea that we’ll end up with far fewer languages (like having a Chinese restaurant in your city, which isn’t quite like having the country of China still vibrant). For an example, read this article on the final generations of the once powerful Aramaic language.
  • McWhorter notes his work isn’t Chomskyan, the dominant thread of linguistics focused on its development (282) but does shoutout a Steven Pinker book called “The Language Instinct.”
  • He doesn’t believe it’s presently possible for linguists to believably reconstruct the first language (language reconstruction) because 150,000 years back represents too much movement. (290)

(The photo is a portion of the painting “The Tower of Babel” by Pieter Bruegel the Elder (1563) via Wikimedia)