REVIEW: Our Magnificent Bastard Tongue, by John McWhorter

Our Magnificent Bastard Tongue: The Untold History of English, John McWhorter (Avery, 2009).

Knew you that John McWhorter has a day job? When he writes not his column for the New York Times, nor podcasts with Glenn Loury, nor opines about race and wokeness in America, he is a professional linguist.

Okay, so that was a weird-looking couple of sentences. I’m sure you could understand what I was saying, but it felt bizarre, stilted, unnatural — not really like English at all, in fact, while not being technically incorrect. And all I did to create that strange impression was to leave out two key — but highly unusual — features of our language: the meaningless “do” and the progressive construction for the standard present tense.

Did you know English is one of the very few languages that can say things like “did you know”? We’re forever sticking the verb “to do” into questions and negative statements: “Do you have a giraffe in your back yard?” “No, I do not have a giraffe, because giraffes do not live in this climate.” Certainly none of English’s Germanic cousins would put it that way: Germans would say Hast du eine Giraffe?, Swedes Har du en giraff?, the Dutch Heb je een giraffe?.¹ And even when we step outside the comforting confines of our own language family, we find that practically no one else goes around “doing” unless they’re actually, you know, doing. The French would never say je ne fais pas avoir une girafe, any more than English-speakers would say “I have not a giraffe.” (Some of the Germanic languages do negate the noun rather than the verb for something that translates more or less as “I have no giraffe,” which sounds more natural in English but still isn’t how any native speaker would answer the question.) But it turns out there is a language family that also employs this unusual meaningless “do,” and moreover it’s the same one that has the other weird feature I omitted to produce my strange English up above.

If someone walked into the room right now and asked what I was doing, a chart of English verb forms might tell you I ought to say “I write.” After all, that’s the present tense! (Yesterday “I wrote” and tomorrow “I will write.”) But obviously I wouldn’t say that: I would say “I’m writing.” And again, English is odd (especially among Germanic languages) for this, because most languages describe present action with the bare verb (Ich schreibe, j’écris, escribo) and use the progressive (Ich bin am schreiben, je suis en traîn d’écrire, estoy escribiendo) only when emphasizing that an action is presently underway. In English I have to add extra words to say that, like “I’m in the middle of writing (so I’ll get you some cheerios in just a second).” Of course, we do use the bare verb in English, but when I say “I write” I am describing² something I generally do: “I write during naptime but today I woke up early so I’m writing in the morning.”

English is the only Germanic language, and one of the few languages period, that can produce a sentence like “Did you see what he was doing?” And this idiosyncrasy certainly didn’t come from the Anglo-Saxons: the Old English they brought to Britain in the 5th century, and were writing down by the 7th, was a bog-standard Germanic language where, like German or Danish or Frisian, you would instead say “Saw you what he did?” But the island to which they brought their Germanic grammar was already inhabited, and its inhabitants spoke the only languages in the world that are full of meaningless “do” and progressive “-ing”: the Celtic tongues.

McWhorter’s discussion of “-ing” gets a little bit in the grammatical weeds about gerunds and participles, but meaningless “do” is a quick and easy example. Welsh is our best model here: of the other two Celtic languages of mainland Great Britain, the last native Cornish speaker died in 1891; Gaelic up north still has some 50,000 native speakers, but their language only arrived in Argyll with the Scotti about the same time the Anglo-Saxons were settling in southeastern England, so I don’t think they’re terribly relevant.³ Here’s how Welsh does it (nes means “did”):

• Did I write? = Nes i ysgrifennu?

• I did not write. = Nes i ddim ysgrifennu.

• I wrote. = Nes i ysgrifennu.⁴

You’ll note that, unlike modern English, Welsh employs meaningless “do” in normal affirmative sentences as well as in questions or negation. Of course we do say “do” in the affirmative, but only — as just now — for emphasis: I did write, even if Substack ate my post. As recently as the Elizabethan period, though, a Welsh-style affirmative “do” was common. Shakespeare is full of doths and dids and dosts. Consider Ariel’s song in The Tempest, which I picked at random just because I like it:

Full fathom five thy father lies;
Of his bones are coral made;
Those are pearls that were his eyes:
Nothing of him that doth fade
But doth suffer a sea-change
Into something rich and strange.

This seems like a pretty open-and-shut case: a bunch of guys speaking a normal Germanic language with neither “-ing” nor “do” show up in a place full of Celtic-speakers with both, and a few hundred years later the descendants of both groups are speaking the only Germanic language that has “-ing” and “do.” Gosh, what could have happened? And yet it’s not the standard story of English, and McWhorter dedicates 80% of his chapter on the Celtic influence on English to exploring why that is. He offers a number of suggestions, most of which really boil down to the fact that people who write histories of English specialize in language change — the ways languages develop over time when left to their own devices, full of exciting usage tables and statistical analyses — rather than language contact. (Have I mentioned that John McWhorter is a language contact specialist? If you move in the right circles, his theories about creoles are more contentious than his thoughts on affirmative action.) But one of his suggestions interested me a great deal, because it answered a question I’ve always wondered about.

Shakespeare wrote about five hundred years ago, and even aside from the frequency of meaningless “do” in normal sentences, it’s clear that our language has changed since his day. But it hasn’t changed that much. Much less, for example, than English changed between Beowulf (probably written in the 890s AD)⁵ and The Canterbury Tales (completed by 1400), another five hundred year gap. Just compare this:

Hwæt. We Gardena in geardagum, þeodcyninga, þrym gefrunon, hu ða æþelingas ellen fremedon.⁶

to this:

Whan that Aprille with his shoures soote,
The droghte of March hath perced to the roote,
And bathed every veyne in swich licóur
Of which vertú engendred is the flour…⁷

That’s a huge change! That’s way more than some extraneous verbs, the loss of a second person singular pronoun (thou knowest what I’m talking about), or a shift in some words’ definition.⁸ That’s practically unrecognizable! Why did English change so much between Beowulf and Chaucer, and so little between Shakespeare and me?

There’s a two part answer to this, and I’ll get to the real one in a minute (the changes between Old English and Middle English really are very interesting), but actually I must first confess that it was a trick question, because my dates are way off: even if people wrote lovely, fancy, highly-inflected Old English in the late 9th century, there’s no real reason to think that’s how they spoke.

On one level we know this must be true: after all, there were four dialects of Old English (Northumbrian, Mercian, Kentish, and West Saxon) and almost all our written sources are in West Saxon, even the ones from regions where that can’t have been the lingua franca.⁹ But it goes well beyond that: in societies where literacy is not widespread, written language tends to be highly conservative, formal, and ritualized. Take, for example, the pre-Reformation West, where all educated people used Latin for elite pursuits like philosophical disputatio or composing treatises on political theory but spoke French or Italian or German or English in their daily lives. It wasn’t quite Cicero’s Latin (though really whose is), but it was intentionally constructed so that it could have been intelligible to a Roman. Similarly, until quite recently Sanskrit was the written language of India even though it hadn’t been spoken for centuries. This happens in more modern and broadly literate societies as well: before the 1976 linguistic reforms, Greeks were deeply divided over “the language question” of whether to use the vernacular (dimotiki) or the elevated literary language (Katharevousa).¹⁰ And modern Arabic-speaking countries have an especially dramatic case of this: the written language is kept as close to the language of the Quran as possible, but the spoken language has diverged to the point that Moroccan Arabic and Saudi Arabic are mutually unintelligible.

Linguists call this phenomenon “diglossia.” It can seem counter-intuitive to English speakers, because we’ve had an unusually long tradition of literature in the vernacular, but even for those of us who use only “standard” English there are still notable differences between the way we speak and the way we write: McWhorter points out, for example, that if all you had was the corpus of Time magazine, you would never know people say “whole nother.” Obviously the situation is far more pronounced for people who speak non-standard dialects, whether AAVE or Hawaiian Pidgin (actually a creole) or Cajun English. (Even a hundred years ago, the English-speaking world had many more local dialects than it does today, so the experience of diglossia would have been far more widespread.)¹¹

Anyway, McWhorter suggests that Old English seems to have changed very little because all we have is the writing, and the way you wrote wasn’t supposed to change. That’s why it’s so hard to date Beowulf from linguistic features: the written language of 600 is very similar to the written language of 1000! But despite all those centuries that the written language remained the perfectly normal Germanic language the Anglo-Saxons had brought to Britain, the spoken language was changing behind the scenes. As an increasing number of wealhs adopted it (because we now have the aDNA proof that the Anglo-Saxons didn’t displace the Celts), English gradually accumulated all sorts of Celtic-style “do” and “-ing”…which, obviously, no one would bother writing down, any more than the New York Times would publish an article written the way a TikTok rapper talks.

And then the Normans showed up.

The Norman Conquest had remarkably little impact on the grammar of modern English (though it brought a great deal of new vocabulary),¹² but the replacement of the Anglo-Saxon ruling class more or less destroyed English literary culture. All of a sudden anything important enough to be written down in the first place was put into Latin or French, and by the time people began writing in English again two centuries later nothing remained of the traditional education in the conservative “high” Old English register. There was no one left who could teach you to write like the Beowulf poet; the only way to write English was “as she is spoke,” which was Chaucer’s Middle English.

So that’s one reason we don’t see the Celtic influence, with all its “do” and “-ing,” until nearly a thousand years after the Anglo-Saxons encountered the Celts. But there are a whole lot of other differences between Old English and Middle English, too, which are harder to lay at the Celtic languages’ door, and for those we have to look to another set of Germanic-speaking newcomers to the British Isles: the Vikings.

Grammatically, English is by far the simplest of the Germanic languages. It’s the only Indo-European language in Europe where nouns don’t get a gender — la table vs. le banc, for instance — and unlike many other languages it has very few endings. It’s most obvious with verbs: in English everyone except he/she/it (who gets an S) has a perfectly bare verb to deal with. None of this amō, amās, amat rigamarole: I, you, we, youse guys, and they all just “love.” (In the past, even he/she/it loses all distinction and we simply “loved.”) In many languages, too, you indicate a word’s role in the sentence by changing its form, which linguists call case. Modern English really only does this with our possessive (the word’s role) and our pronouns¹³ (“I see him” vs. “he sees me”); we generally indicate grammatical function with word order and helpful little words like “to” and “for.” But anyone learning Latin, or German, or Russian — probably the languages with case markings most commonly studied by English-speakers — has to contend with a handful of grammatical cases. And then, of course, there’s Hungarian.

As I keep saying, Old English was once a bog-standard Germanic language: it had grammatical gender, inflected verbs, and five cases (the familiar nominative, genitive, dative, and accusative, plus an instrumental case), each indicated by suffixes. Now it has none. Then, too, in many European languages, and all the other Germanic ones, when I do something that concerns only me — typically verbs concerning moving and feeling — I do it to myself. When I think about the past, I remember myself. If I err in German, I mistake myself. When I am ashamed in Frisian, I shame me, and if I go somewhere in Dutch I move myself. English preserves this in a few archaic constructions (I pride myself on the fact that my children can behave themselves in public, though I now run the risk of having perjured myself by saying so…), but Old English used it all the time, as in Beseah he hine to anum his manna (“Besaw he himself to one of his men”).

Another notable loss is in our direction words: in modern English we talk about “here,” “there,” or “where,” but not so long ago we could also discuss someone coming hither (“to here”) or ask whence (“from where”) they had gone. Every other Germanic language still has its full complement of directional adverbs. And most have a useful impersonal pronoun, like the German or Swedish man: Hier spricht man Deutsch.¹⁴ We could translate that as “one speaks German here” if we’re feeling pretentious, or perhaps employ the parental “we” (as in “we don’t put our feet in our mouths”), but English mostly forces this role on poor overused “you” (as in “you can’t be too careful”) because, again, we’ve lost our Old English man.

In many languages — including, again, all the other Germanic languages — you use the verb “be” to form the past perfect for words having to do with state or movement: “I had heard you speak,” but “I was come downstairs.” (This is the bane of many a beginning French student who has to memorize whether each verb uses avoir or être in the passé composée.) Once again, Old English did this, Middle English was dropping it, and modern English does it not at all. And there’s more, but I am taken pity on you…

So how did we lose all this? Some of the other Germanic languages have lost one or another of these features — although Icelandic, a language so conservative that its speakers can read the Old Norse sagas their ancestors wrote a thousand years ago with little trouble, has kept them all — but none have dropped as many as English.¹⁵ And they dropped first in the north, where the Vikings settled most densely. Northumbrian Old English was losing its inflected endings by the end of the tenth century — with the exception of the dative plural ending, which happens to be the same in Old English and Old Norse and hung around a lot longer — whereas they held on in the south well into the era of Middle English. (All that “-eth” and “-est” in Shakespeare is very southern.) In fact, the rural dialect of Dorset maintained grammatical gender into the 19th century.¹⁶

McWhorter argues that this is just what happens to languages when they’re suddenly picked up by a large number of adult learners. It’s hard to learn a new language as an adult — you rarely get it all quite right — and when there are enough new speakers in one place, the language itself begins to drop the most complicated and challenging features. This is also his argument about creoles, although he never comes right out and says that English ought to be regarded as one.¹⁷

But it’s interesting to note the difference between the Celtic influence and the Norse influence on Old English: the Celts lived alongside the Anglo-Saxons, maintaining their old language for centuries while also learning the new one, and gradually peppered that new one with their own familiar grammatical constructions. This is termed “linguistic equilibrium,” and it’s a common phenomenon worldwide: nearby languages from unrelated families often accumulate similar grammatical features. (Gaston Dorren’s Lingo: Around Europe in Sixty Languages has an entertaining explanation of the surprising grammatical similarities the unrelated Balkan languages share with one another but not with their “cousin” languages.) The Vikings, on the other hand, came in relatively small numbers, mostly intermarried with the English, and stopped speaking Norse within a generation or two. Their Old English didn’t accumulate Nordicisms from the two languages being spoken in parallel for centuries; it just got messed up by adult learners making the sorts of mistakes adult learners often make, and in large enough numbers that their descendants preserved the mistakes.

This is an oddly unfocused book, really more a series of semi-connected essays about things John McWhorter thinks are interesting about English than any sort of concerted argument. The story I’ve told you above is drawn from the first and third chapters (of five), which deal with language contact and which I find the most interesting. The rest of the book is devoted to a chapter-long takedown of linguistic prescriptivism, interesting for its list of things we say now that people used to find objectionable (apparently it was once vulgar to say “all the time” instead of “always,” “the first two” instead of “the two first,” or “being built” instead of “building,” and “standpoint” was a flat no) and extended criticism of the Sapir-Whorf hypothesis.

But insofar as McWhorter has an argument, it’s that language is always changing so it’s silly for us to criticize anything as “grammatically incorrect.” After all, we speak a bastardized, miscegenated language, one that has lost many of its original Germanic features and gained Celtic “-ing” and “do” — who are we to say that “Timmy and me went to the store” is, in some sense, “wrong”?¹⁸ But that’s an easy answer. It’s the same one I give when my daughters ask why they need to brush their hair if it will just get tangled again, and it’s the one I can imagine Ealdwine the tenth-century scribe giving when he explains why you have to use the proper endings and no, we’re not writing down text full of meaningless “do”: this is how you show the world that someone cared enough to teach you to do it right.

But wait! I said there were five chapters and I’ve only told you what was in four of them! And that’s because the fifth and final chapter is totally weird.

McWhorter has spent most of the book arguing that English is weird among world languages. Now he takes a step back in time and says that in fact all the Germanic languages are — that Proto-Germanic is as much a bizarre outlier among its Indo-European siblings as English is among the Germanic languages.

You may have heard of Grimm’s Law (yes, those Grimms), which describes how Proto-Indo-European’s P, T, and K sounds became Proto-Germanic’s F, Th, and H. (Latin pater and English father are cognate, for instance, both deriving from PIE *ph₂tḗr.) These shifts aren’t totally unheard-of in linguistic development — there are other languages where plosives become fricatives — but they’re unusual among Indo-European languages. Why did this happen? Well, maybe it’s just chance, but McWhorter, being a language contact guy and having just written 80% of a book about why you’d understand English better if you looked at it with your language contact lenses, thinks that, just maybe, it was language contact. And the most plosive-heavy languages a speaker of Proto-Germanic could have come across were the Semitic languages.

But wait! There’s more! Because the Germanic languages have another weird feature: strong verbs. These are the ones where you make the past tense not by adding “-ed” to the end but by changing the vowel, and which it takes a child years to master: no, you didn’t “drinked” your milk, you “drank” it. Again, this is something foreign to most Indo-European languages, but familiar to Semitic ones (sort of — the Semitic triconsonantal root system is much more complicated than “you change the vowel to make it past,” but does include that).¹⁹ And hey, just like English has way fewer endings and cases than any other Germanic language, Proto-Germanic has way fewer endings and cases than any other contemporary Indo-European language! And Proto-Germanic has all kinds of ocean-related borrowings that don’t appear to be cognate with words in other Indo-European languages…

So obviously it was the Phoenicians.

And “Baldur,” a god who shows up in Germanic mythology, is etymologically Phoenician Baal.

At least, this is the case of Theo Vennemann of the University of Munich, who argues that the Phoenicians could have ended up in the northern shore of Europe, somewhere around what’s now the border of Germany and Denmark, around 500 BC, and that the influence of Phoenician could account for all the oddness of Proto-Germanic. (McWhorter adds only that maybe they came from Carthage, because why not.)

It’s a fun, if totally conjectural, theory, but McWhorter advances it in service of his broader argument: “The lesson: the idea that there was once an English somehow pristine, a pure issuance, is false. … People today bemoan the eclipse of whom’s marking of the accusative, unaware that Proto-Germanic speakers let go of four of the cases that Proto-Indo-European speakers used. The world kept turning. You don’t like nucular? Well, how do you think the likes of ‘fopcorn’ sounded to a Proto-Germanic speaker watching that kind of pronunciation spread?”

But come on, man. No one says the decline and fall of “whom” is a sign of the End of Days; we just think it’s a nice way of indicating grammatical case and sounding, you know, like, educated. Is it Correct in some deep ontological sense? No, obviously not, no more than Latin quem or however you render the relative pronoun in Chinese. But a thing doesn’t need to be Correct in some deep ontological sense to be worth keeping around. It’s just neat to speak a language that can do things! Because not all languages can do the same things, and it’s a shame to lose what’s unique about ours. Of course it didn’t fall pristine from the heavens, any more than the genes of any currently existing human populations have always existed in their current configurations: people and languages are always on the move, always mixing and changing. But the fascinating story of how we got here, the recognition that this isn’t where we’ll stay forever, doesn’t mean we can’t love and appreciate where we are right now.

1. Afrikaans split from Dutch long enough ago that it preserves the archaic “camelopard” (Boers would say Het jy ‘n kameelperd?). I was hoping the same was true of Yiddish, which split from Middle High German nearly a thousand years ago, but apparently you do say דזשעראַף (jeraf). This probably comes down to the difference between people who are discussing something from a book vs. the linguistic conservatism of people who might actually have a giraffe in the back yard.

Not that anyone asked, but “giraffe” comes to English from French and thence from Arabic زَرَافَة (zarāfa), which itself derives from the Persian zurnāpā, a compound of the word for “flute” and the word for “leg.” Personally I prefer “camelopard.”

2. Look! The progressive construction again!

3. The language of the Picts, whom they replaced, was probably another Celtic tongue more closely related to the Brittonic languages to the south.

4. And in case you’re wondering (I did), ysgrifennu derives from the Proto-Brythonic *ɨskrivenn, which is just the Latin scribendum. Which I guess makes sense: if someone a cool sword and a red crest on his helmet shows up on your island doing a thing you’ve never heard of before, you’ll probably adopt his word for it.

5. This is extremely contentious. The poem is known to us from only one manuscript, which was produced sometime near the turn of the tenth/eleventh century, and scholars disagree vehemently both about whether its composition was contemporary with the manuscript or much earlier and about whether it was passed down through oral tradition before being written. J.R.R. Tolkien (who also had a day job, in his case as a scholar of Old English — the Rohirrim are more or less the Anglo-Saxons) was a strong proponent of the 8th century view. Personally I don’t have a strong opinion; my rhetorical point here could be just as clearly made with an Old English document of unimpeachably eleventh century composition, but Beowulf is more fun.

6. Old English orthography is not always obvious to a modern reader, so you can find a nice video of this being read aloud here. It’s a little more recognizable out loud, but not very.

7. Here’s the corresponding video for Middle English, which I think is actually harder to understand out loud.

8. Of course words shift their meanings all the time. I’m presently reading Mansfield Park and giggling every time Fanny gets “knocked up” by a long walk.

9. Curiously, modern English derives much more from Mercian and Northumbrian (collectively referred to as “Anglian”) than from the West Saxon dialect that was politically dominant in the Anglo-Saxon period. Meanwhile Scots (the Germanic language, not to be confused with the Celtic language of Scots Gaelic or whatever thing that kid wrote Wikipedia in) has its roots in the Northumbrian dialect.

10. This is a more interesting and complicated case, because when the Greeks were beginning to emerge from under the Ottoman yoke it seemed obvious that they needed their own language (do you even nationalism, bro?) but spoken Greek was full of borrowings from Italian and Latin and Turkish, as well as degenerate vocabulary like ψάρι for “fish” when the perfectly good ιχθύς was right there. Many educated Greeks wanted to return to the ancient language but recognized that it was impractical, so Katharevousa (lit. “purifying,” from the same Greek root as “Cathar”) was invented as a compromise between dimotiki and “proper” Ancient Greek. Among other things, it was once envisioned as a political tool to entice the newly independent country’s Orthodox neighbors, who used Greek for their liturgies, to sign on to the Megali Idea. It didn’t work.

The word ψάρι, by the way, derives from the Ancient Greek ὀψάριον, meaning any sort of little dish eaten with your bread but often containing fish; see Courtesans and Fishcakes: The Consuming Passions of Classical Athens for more. Most of the places modern Greek uses different vocabulary than the ancient tongue have equally fascinating etymologies. I think my favorite is άλογο, which replaced ίππος as the word for horse. See here for more.

11. Diglossia is such a big deal in so many societies that I’ve always thought it would be fun to include in my favorite genre, fantasy fiction, but it would be hard to represent in English. Anyone who’s bounced off Dickon’s dialogue in The Secret Garden or Edgar’s West Country English in King Lear knows how difficult it is to understand most of the actually-existing nonstandard dialects; probably the only one that’s sufficiently familiar to enough readers would be AAVE — but that would produce a very specific impression, and probably not the one you want. So I think the best alternative would be to render the “low” dialect in Anglish, a constructed vocabulary that uses Germanic roots in place of English’s many borrowings from Latin and French. (“So I think the best other way would be to give over the ‘low’ street-talk in Anglish, a built wordhoard that uses Germanic roots in spot of English’s many borrowings…”) It turns out Poul Anderson did something similar, because of course he did.

12. My favorite is food, because of course it is: our words for kinds of meat all derive from the French name for the animal (beef is boeuf, pork is porc, mutton is mouton) while our words for the animal itself have a good Germanic roots: cow, pig, sheep. Why? Well, think about who was raising the animal and who was eating it…

13. And even this is endangered; how many people do you know, besides me, who say “whom” aloud?

14. Yes, this is where Heidegger gets das Man.

15. Did you notice that little “in Europe” proviso up there? That’s because Afrikaans actually has dropped grammatical gender, but you do still remember yourself, come hither, and have a better word than “you” to describe the hypothetical individual who’s doing all this.

16. It divided things into “personal” and “impersonal” genders, though “personal” encompassed not only people but all living things and also tools. A tree is “he” or “thëase tree” while water is “it” or “this water.”

17. He might address this further in his more academic Language Interrupted: Signs of Non-Native Language Acquisition in Standard Language Grammars, but that’s $60 used so someone else will have to read it and tell me.

18. I agree with him on the impersonal “they,” as in “whoever left their dirty socks on the floor is in big trouble!” because of its long pedigree. My objection to the singular personal “they” isn’t grammatical pedantry but insistence is that human sex is binary and we should use words that reflect reality.

19. Very short version: the basic form of a word in a Semitic language is three consonants, and you add vowels and sometimes other consonants in regular patterns to create whatever word you need. So say your root is K - T - B (or V), meaning “write” — add one arrangement of vowels and you produce kātib, “writer”; add a different one and you get kitāb, “book.” With a root meaning “drink” you could follow the First Consonant - ā - Middle Consonant - i - Last Consonant pattern to produce the word “drinker.” Sometimes you run into a confusing combination of root and pattern, like KTB + a pattern that means the passive emphatic — what does it mean to be really really written? Well, according to the dictionary, “engraved.”