Empires of the Word

Empires of the Word: A Language History of the World
by Nicholas Ostler
HarperCollins, 2005

Reviewed 2007

Sometimes you find wonderful books in the unlikeliest of places.

Nobody who’s had the dubious privilege of flying through Newark airport would ever describe it as a bibliophile’s paradise. Indeed, despite being an inveterate browser of airport bookstores, these days I don’t even bother to visit EWR’s sad few examples. But on a recent trip I found myself stranded with no reading material. More in hope than in expectation, I wandered over to the Relay store at EWR terminal B. And I’m glad I did.

“Empires of the Word” is, quite simply, the best non-fiction book I have read in years. It’s certainly one of the top 3 I’ve read in the last decade, and probably one of the top 5 I’ve read in all my life1. It’s that good. If you’re at all interested in languages, history, and the history of languages, you should stop reading this review right now, and buy the book. You won’t regret it.

Still with me? Oh ye of little faith! I suppose I had better do some reviewing, then.

The book is subtitled “A Language History of the World”, and that’s precisely what it is: it’s a history of the languages of the world. Not just one, or ten, or even fifty: all of them. How languages emerge, how they change, how they die. How languages spread from place to place and from people to people. How languages can be influenced externally (by empires, war and conquest), and internally (by their own grammatical or notational characteristics).

The above paragraph may seem overly abstract and analytical, but nothing could be further from the truth. For Ostler fills his book with immense, indeed incredible detail, and it’s this detail that makes it such a rewarding read.

Consider, for example, the Fertile Crescent. Over the millennia, this relatively small parcel of land has seen Sumerian, Akkadian, Babylonian, Egyptian, Hittite, Punic, Hebrew, Aramaic, Greek, Roman, Parthian, Persian, Arabic, Turkic and Turkish spoken extensively (not to mention dozens more minor tongues). And Ostler describes every one of them! He describes how they were linked linguistically2; he describes how they were linked politically3; he describes how they rose and fell, ebbed and flowed. It’s fascinating.

And he does this for a dozen, nay, several score Fertile Crescents, all around the world. Want to know why classical Latin splintered into a motley of Romance languages while classical Chinese retained its unity? (Hint: non-phonetic script).

Want to know why Sanskrit grammar is considered “programmatic”? Want to know how and why the aesthetics of Attic Greece came to form the core of the English public school curriculum even today? Want to know what languages were spoken in South America before the Spanish conquest, and why their distribution mattered? Want to know more about Russian, about Japanese, about Bantu, about the Celtic and Germanic languages, about Quechua and Maori, about Indo-European and Afro-Asiatic? It’s all there in this marvellous book.

I could go on and on about “Empires of the Word”, but I won’t. Suffice to say it’s a book I have longed to find (I’ve always been interested in historical linguistics), and now that I’ve found it, it has met and indeed surpassed all my expectations. I cannot recommend it highly enough.

Buy it!

P.S. Okay, so I can’t resist adding another beautiful little tidbit from the book.

Sanskrit is the punster’s dream language. First, words tend to have multiple literal meanings (not 2 or 3 as is common in English, but 6 or 7 or more). Second, words often have many more allusive (“literary”) meanings. Third and perhaps most important, words can be combined, and the combination can be then decomposed in multiple ways, leading to even more meanings.

An English analogy for the technique would be something like this. Consider the word CONTRACTION. This can be split into CON + TRACTION, also CONTRA + ACTION, also CONTRACT + ION – three different “meanings”.

So far so good. Now for the (amazing, incredible, mindblowing) tidbit. A certain Sanskrit poet dude (whose name, sadly, I forget) used the language’s flexibility to write new verse versions of the Ramayana and the Mahabharata – simultaneously! So, the very same text, read using one set of word meanings, gives the Ramayana; read using another set, gives the Mahabharata. This is not merely a thought experiment; the text exists.

No further comment. I think my ingenuity meter just burned out.

  1. In case anybody cares, my all-time favourite non-fiction books: Cosmos, Godel Escher Bach, The Worst Journey in the World, The Great Game, and now Empires of the Word. ↩

  2. For instance: languages from the Hamito-Semitic family (such as Hebrew, Aramaic and Arabic) all share a common characteristic: word roots are defined by consonants, and the choice of intervening vowels determines the specific sense of the word. So the root SLM (meaning roughly “peace, rest”) can be found in words such as SaLaaM (“peace be with you” in Arabic), ShaLoM (the same, in Hebrew), SaLMu (“sunset” in Aramaic) and “iSLaM” (“submission” in Arabic).

    Furthermore, these languages tend to omit vowels in their written form, which leads to another interesting piece of trivia. Standard Arabic script has only the consonants written in the “main” word; the vowels are indicated by diacritical marks (the little squiggles above and below characters that you see in calligraphy or printed text; they tend to be omitted in cursive writing). But certain consonant sounds (“-t” at the end of a word, and any “double” consonant) are also represented by diacritical marks, when you would expect them to be part of the main word. Why is this so?

    The answer is religious. Apparently the Prophet Mohammed spoke a (non-standard) dialect of Arabic which omitted these consonant sounds, so the original transcription of the Koran did not have them. When the time came to spread the Koran to speakers of standard Arabic, it became necessary to include these consonants. But altering the Koran is blasphemy. What to do? An ingenious solution was devised: add the consonants, but as diacritical marks, so that the main text is unchanged. And ever since then, standard Arabic has had diacritical marks to represent these consonants and these consonants alone. Isn’t that beautiful? ↩

  3. For instance: Babylonian imperial policy involved the transportation of thousands of refugees from conquered lands to various other parts of the empire; this had the interesting consequence that certain peripheral languages – specifically, Aramaic, a nomadic desert tongue – became the common parlance of the empire, and eventually supplanted Babylonian itself. Contrast British imperial policy, which saw English become the common tongue of the periphery (what language am I writing this in?). Or Greek (strictly speaking Macedonian) imperial policy, where the Greek language was confined to the administrative class, and hence disappeared when the Greek empire did. ↩