The first chapter in the new history of English is that bastardization I mentioned.

German, Dutch, Swedish, and the gang are, by and large, variations on what happened to Proto-Germanic as it morphed along over three thousand years. They are ordinary rolls of the dice. English, however, is kinky. It has a predilection for dressing up like Welsh on lonely nights.

The Kinks

Did you ever notice that when you learn a foreign language, one of the first things you have to unlearn as an English speaker is the way we use do in questions and in negative statements? Take Did you ever notice . . . ? for example. Or I did not notice. We’re used to this do business, of course. But it’s kind of strange if you think about it. In this usage, do has no meaning whatsoever. It’s just there, but you have to use it. One cannot, speaking English, walk around saying things like Noticed you ever? or I not notice. English has something we will call meaningless do.

Most languages, unsurprisingly, have no interest in using the word do in a meaningless way. If you’ve studied Spanish, you quickly learned that to put a verb in the past, you do not stick in a past form of the verb for doDid she talk? is not Hago ella hablar? Nor do you jam in do to make a sentence negative—She does not talk is not Ella no hace hablar but Ella no habla. Nor is it Elle ne fait parler in French, or Ona ne delaet govorit’ in Russian, Hi lo osa ledaber in Hebrew, or—you get the picture.

Did she walk? feels utterly conventional to us, when if you step beyond English, you look for do used that way and come up short.1 None of the other Germanic languages use do the way English, ever the wayward one, does.

Then there is this -ing thing. We are given a tacit sense that tense marking in English works like this:


But if you think about it, I write is not really present tense. Imagine you’re at your laptop writing an e-mail and someone asks what you’re doing and you say “I write.” It’s impossible to imagine that said by anyone without a foreign accent, and one imagines that the e-mail such a person would write would be full of mistakes. “I write” would be, quite simply, incorrect. Your answer would be “I’m writing.”

“I write,” on the other hand, is what you would say to express something more specific: that it’s something you do on a regular basis. I write, usually, from about ten A.M. to one P.M. The present tense, in English, is expressed not with a bare verb, but with the progressive -ing. The bare verb has a different meaning, which linguists call habitual.2

Yet once again, that’s not the way it is in any other language you learn. In Spanish, your answer if asked what you were doing would be “Escribo.” The French person would answer “J’écris.” Sure, both of these languages and many, many others have ways of calling attention to the fact that you are in the process of writing the letter at this very instant: in Spanish, Estoy escribiendo, in French, Je suis en traîn d’écrire. Germanic languages do, too, like German’s Ich bin am schreiben, which comes out as “I am on the writing.” But it’s the decidedly peculiar individual who is given to stressing for every one of their actions that they are indeed in the process of accomplishing it at this very instant . In a normal language, you use a progressive construction when there’s a reason to. Otherwise, to answer “I write” sounds perfectly fine in most languages. But in English, it sounds vaguely funereal, and -ing is the ordinary way to use the present tense.

English, then, is the only Germanic language out of the dozen in which there could be a sentence like Did you see what he is doing? rather than Saw you what he does? Since none of the other offshoots of Proto-Germanic seems to have sprouted oddities like these, one might ask whether there is a reason that English has.

And if one asks that, presumably it will strike one as germane that there happen to be languages with precisely the same oddities spoken right on the same island where English arose, long before English got there.

Yet most scholars of English’s history find this neither germane nor, really, even interesting. Why?

Uninteresting Likenesses

The languages in question belong to another Indo-European subfamily, Celtic. There are only a few of these languages today, although they once held sway across vast swatches of Western Europe. Irish Gaelic is one of them (and its variant Scottish Gaelic, an export to Britain, another). But the ones most of interest to us are those of Britain: Welsh in Wales and Cornish in Cornwall.3 The last native speaker of Cornish died in 1891, but there is a hardy revival movement for it today.

If English is the odd one out as Germanic languages go, Celtic languages are odd ones out as Indo-European languages go. Verbs sometimes coming last in German strikes us as weird enough, although it is actually ordinary worldwide. But in Celtic, verbs come first in a sentence, which is less ordinary worldwide, and downright freaky within Indo-European languages. There are other features in which Celtic marches to the beat of its own drum, and two of them are the way it uses do and -ing.

Take a look at this in Welsh. Nes means “did.” Welsh puts words in a different order than English, and so nes is always first. What’s interesting is that it is there, just as in English:


Welsh uses do in the same meaningless way that English does. Do just sits there taking up space, not contributing any meaning to the sentence.

Note that Welsh is different from English in one way: it uses do in “normal” sentences, affirmative ones as well, as we see in that third sentence. When a Welshman states Nes i agor, they are using the words that come out in English as “I did open,” but not with an emphatic meaning as in our did open. They mean it as if we were speaking English in the Elizabethan period and said, “Since it was so hot out, I did open a window for you.”

But in that people still said things like that then, English was more like Welsh than it is today. Even further back in Middle English, one might say for “You wept” Thou dudest wepe. Our sense that to speak fake “Olde English” means sticking “dosts” and “doths” all over the place corresponds to a Middle English reality, which persisted for centuries afterward. Here is Gertrude in Hamlet, addressing same:

Alas, how is’t with you

That you do bend your eye on vacancy,

And with th’incorporal air do hold discourse? (III, iv, 120-22)

Upon which he answers (147-48):

My pulse as yours doth temperately keep time, 

And makes as healthful music.

English has gone its own way since and dropped this do usage in affirmative (“neutral”) sentences, keeping it in the negative and question contexts. But there was a time when English was even more like Welsh on this score than it is now.

Or how about English’s progressive construction, as in Mary is singing. In English, -ing leads a double life. In one guise, it makes a verb into a gerund, which means that it makes the verb into a noun. One sings, and one may enjoy that which is known as singing, a noun: Singing is fun. As a matter of fact, gerunds are sometimes called “verb-nouns.”

Then, -ing has a second identity, when it is used in the progressive construction: Mary is singing. Here, singing is not a verb-noun—Mary is singing does not mean “Mary embodies the act of rendering song.” Singing in Mary is singing is just a verb, specifically what is called the present participle form of a verb. Our -ing is two things.

The important point is the fact that in English, as we have seen, this progressive Mary is singing construction is our present tense. If someone asks you what you’re doing as you warble “Just the Way You Are,” your answer must be “I’m singing,” not “I sing.” Interestingly, in Welsh as well, to answer that question you must use a progressive construction: Welsh and other Celtic languages have the same -ing fetish as English. Remembering that Celtic word order is odd to our eyes and ears, in Welsh, if someone asks, “What’s our Mary doing?”, the answer is not “Mary sings” but “Mary is in singing”:

Mae Mair yn canu. 

is Mary in singing

Canu is the verb-noun for sing in Welsh: “Mary is in the act of singing.”

Now, to be sure, in Welsh the present is expressed with a verb-noun progressive, while in English’s Mary is singing progressive, singing is not a verb-noun but a participle. However, the participle is just a latter-day morphing of what started as a verb-noun. Just as with meaningless do, what English does now is a drifting from what first was even more like Celtic.

It went like this. In Old English one could say “I am on hunting” to mean that you were hunting. This was, obviously, just like the Welsh “Mary is in singing.” Then in Middle English, the on started wearing down and one might say “I am a-hunting,” just as we now say “Let’s go” instead of “Let us go.”

Then before long, the a- was gone completely—as in the way we casually say “ ‘Tsgo” for “Let’s go”—and hence, just “I am hunting.” Ladies and gentlemen, the birth of a present participle. Celtic was English’s deistic God—it set things spinning and then left them to develop on their own. But that first spin—I am on hunting—was key. It was just like Welsh.

And it’s not just Welsh. Cornish down south has the same kinks about do and -ing. The way to say I love is:

Mi a wra cara. 

I at do love

Sort of an Elvisesque “I’m a-doing loving,” except it is a perfectly normal sentence in Cornish, and its do is used the same way with negative sentences and in questions:

Gwra cara? 

do-you love

Then it has the same mysterious drive to use verb-nouns for the present tense. Here is She is buying vegetables, in which the word for buy is a verb-noun:

Yma hi ow prena hy losow. 

is she at buying her vegetables

So: the Angles, Saxons, and Jutes brought a language to Britain in which a sentence like Did you see what he is doing? would have sounded absurd. The people already living in Britain spoke some of the very, very few languages in the world—and possibly the only ones—where that sentence would sound perfectly normal. After a while, that kind of sentence was being used in English as well.

And yet specialists in the history of English sincerely believe that English started using do and -ing by itself, and that it is irrelevant, or virtually so, that Welsh and Cornish have the same features. You can page through countless books and articles on The History of English, and even on specifically the history of meaningless do or the -ing present, and find Celtic either not mentioned at all, actively dismissed, or, at best, mentioned in passing as “a possible influence” (read: of no significant bearing upon the issue).

There is clearly something strange about this, but it is not that legions of scholars are incompetent, stubborn, bigoted against Celts, or anything else of the sort. Rather, they come at the issue with certain established assumptions, reasonable in themselves, which if held, understandably leave one comfortable treating such close correspondences between English and Celtic as accidents.

Those assumptions, however, are mistaken.

Assumption Number One: The Celts All Just Died

The first assumption is that after their arrival in England in A.D. 449, the Germanic invaders routed the Celts in more or less a genocide, leaving mere remnants huddling on the southwesterly fringes of the island. From here, it has traditionally been concluded that Celtic languages could not have had any impact on English for the simple reason that no Celtic speakers survived the genocide to influence the language.

But the truth is that the genocide of an entire society inhabiting vast expanses of territory is possible only with modern technology. The Angles, Saxons, and Jutes did not possess anything we would consider modern technology. How, precisely, were they to kill practically every Celt outside of Wales and Cornwall—that is, in an area about the size of New England? With swords? How many people can you get at? Remember, there weren’t even guns yet. And even when Dutch and English colonists in South Africa had guns, African peoples there, like the Xhosa and the Zulu, gave them enough of a run for their money with spears that even though the whites ended up subjugating the blacks under apartheid policies, the blacks still today vastly outnumber the whites. There was no way to kill everybody.

In that light, whatever havoc the Germanic invaders wrought, there were not, apparently, very many of them. Early Anglo-Saxon chroniclers like the Venerable Bede had it that the invaders “overran” Britain. But writers of their era did not have access to substantial and regular news from all over the land, satellite photography, or our conceptions of demography or even scholarship. Bede was even writing three centuries after said “overrunning,” which as Bill Bryson notes “is rather like us writing a history of Elizabethan England based on hearsay.” Bede could easily document as “overrunning” what was actually a compact number of violent, destructive encounters.

Comparative genetics has recently confirmed that this was the case. By tracing mutations in mitochondrial DNA in women and on the Y chromosome in men, we can reconstruct the migrations of human populations since the emergence of Homo sapiens. It turns out that only about 4 percent of British men’s genetic material is traceable to a migration from across the North Sea. Moreover, essentially none of British women’s genetic material traces back to such a migration, meaning that the invaders were not couples with children, such that women and young’uns would bulk up the total. Rather, the invaders were just a bunch of guys. In fact, evidently the famous Germanic invaders numbered about 250,000, about as many people as live in a modest-sized burg like Jersey City.

We will never be able to bring the Celts of this era back to life to ask them whether they felt terribly “exterminated,” nor do official records survive that would allow us to check for ourselves. However, there have always been clues that are problematic for the genocide account. A burial site with graves both in the style of Germanics across the North Sea and in the style of Celts (with the body buried crouching and facing north or northeast) suggests not genocide, but Celts living alongside Germanics. The very fact that after the invasion, archaeologists find no abrupt transformation in material culture suggests that Celts survived in numbers robust enough to pass on their cultural traditions permanently.

A valuable snapshot comes in the laws established by Ine, a seventh-century king of Wessex (in an era before any individual considered himself the king of England as a whole). Two centuries after the Angles and company supposedly exterminated the Celts, the stipulations of Ine’s laws indicate a Britain where Celts are numerous and well integrated into society. The wealhs (Welshmen in modern parlance) Ine repeatedly refers to and legislates for include lowly slaves, respectable landowners, and even horsemen serving the king. The main lesson, as Ine devotes one law after another to establishing precisely how much compensation a Welshman’s family or owner gets if he is killed, is that subjugated though they usually were, the Celts were there, in numbers.

The scenario Ine’s laws depict brings to mind, in fact, the situation of American blacks before Emancipation, right down to the fact that wealh, while coming down to us as Welsh, was not the name the people had for themselves (which was Cymry), and in Old English meant “foreigner,” with a goodly tacit implication as well of “slave.” In southern America before the end of the Civil War, Africans and their descendants were subjugated, but were still part of the warp and woof of existence for whites, outnumbered them, and included in their number a class of free farmers and artisans.

The genocide story, then, has fallen apart. Genes, archaeology, documentary evidence, and sheer common sense leave it dead in the water. Typical assumptions such as magisterial popular chronicler David Crystal’s that the Celts hung around for a brief spell as slaves and brides but their “identity would after a few generations have been lost within Anglo-Saxon society” can no longer be accepted.

This leaves us with a simple fact about what happens when languages come together: they mix. There is no recorded case in human history in which languages were spoken side by side and did not spice one another with not only words, but grammar. This means that even without recordings of seventh-century Celts speaking “Englisc” and peppering it with phrasings copying Celtic grammar, we can assume that this was the case, because it quite simply must have been.

Taking a cue from the slavery analogy, we see that Jamaicans today speak a hybrid language, popularly called patois, that was born when African slaves learned English and filtered it through the languages they had been born to. Here is a sentence in patois:

Unu main mi tingz, no tiek non gi im. 

You all mind my things, don’t give him any.

The word unu is unfamiliar in English itself, but it is the word for you in the plural in the Igbo language of Nigeria, which many early slaves in Jamaica had grown up speaking. Patois phrases it Don’t take any give him because many West African languages string verbs together in just this way, such as the Twi language also spoken by many African slaves in Jamaica:

O de sekan no ma me. 

he took knife the gave me 

“He gave me the knife.”

The patois case is an example of what happens when there are so many people speaking a language in a non-native way that new generations speak it that way instead of the original way.

There are similar cases around the world. Another one involves the famous click sounds in a family of languages, Khoisan, spoken in southern Africa by hunter-gatherers. Those clicks are extremely rare worldwide. Outside of the southern half of Africa, the only language in the entire world with clicks is a way of talking, now extinct, that speakers of one tiny aboriginal Australian language made up for use in male initiation rites! (The people’s everyday language is Lardil; the click talk was called Damin.) Among the languages with clicks in southern Africa, as it happens, are several that are not in the Khoisan family but are spoken nearby, such as Zulu and Xhosa (the native language of Nelson Mandela). Given that clicks essentially do not exist anywhere else, it is obvious that earlier forms of Zulu and Xhosa mixed with click languages.

The way that English uses do and -ing just like Celtic, then, is predictable. Celts, less exterminated than grievously inconvenienced, had to learn the language of the new rulers. The Celts’ English was full of mistakes—that is, ways of putting words together that worked in Celtic but were new to Old English. However, over time, this Celtic-inflected English was so common—after all, there had only ever been 250,000 Germanic invaders—that even Anglo kids and Saxon kids started learning it from the cradle. After a while, this was Englisc—just as in Jamaica, after a while, Don’t take none give him was the way one spoke English there, and in what is now South Africa, using click sounds like the hunter-gatherers became the way one spoke Xhosa.

Right? Well, for History of English scholars, still not. The genocide is, to them, only one reason to see the Celtic-English mirrorings as accidental.

Assumption Number Two: Shitte Happens

The body of scholarship on The History of English is replete with detailed descriptions of meaningless do and the verb-noun present tense just “happening.” We are to assume that chance alone could have nudged English into coming up with meaningless do and a verb-noun present.

To the scholars working in this vein, meaningless do and the verb-noun present are just same-old same-old as languages go. To them, saying that English got these features from Celtic just because Celtic has them is like proposing that anteaters, because they have long tongues and eat insects, must have evolved from chameleons. In that case, we easily see that plenty of animals have long tongues and even more eat bugs, such that there is no scientific reason to assume that anteaters evolved from chameleons.

The problem is that these scholars have usually had little occasion to look hard at languages outside of the Germanic family. They are unaware that, as it were, in actuality very few languages have long tongues, and that even fewer both have long tongues and also eat bugs. They do not realize how very special English is—or that it is inescapable that Celtic languages made it that way.

In the do case, the distraction is that in many colloquial dialects of Germanic languages, one can use do in a way kind of like English’s meaningless do. But only kind of.

The issue is sentences like this colloquial German one:

Er tut das schreiben. 

he does that write

This means “He writes that,” and in terms of word order, it certainly looks like Ye Olde do in the Hamlet passages—that is, it looks like German’s version of He doth write. However, in fact, it isn’t—it is something quite different.

For one, German’s version is optional. One might say Er tut das schreiben, but the simple Er schreibt das is also alive and well, and in fact, much more usual.

Then, most importantly, meaningless do is meaningless, but German’s do is meaningful. It is used when you want to emphasize some part of the sentence. When you put stress on what you want to emphasize, you might also toss in a do. So: imagine if now and then you fall into moods where you enjoy taking a knife and stabbing pillows open. Suppose you run out of pillows but you still have that nagging urge, and then you see a laundry bag bulging full of clothes. A thought balloon pops up over your head: Maybe I’ll cut the bag open!

Well, in German the thought balloon would read:

Ich tue vielleicht den Sack aufschneiden. 

I do maybe the bag cut-open

So, German do is an optional trick, used only in the present tense, as one factor in the way you emphasize something. Obviously this is nothing like the way we use do in English, in which to negate a sentence or to make it a question, you have to stick in a do no matter what, and the do has no meaning of its own. No dialect in any Germanic language other than English uses do in this way; at best, there are dialects that use it in variations on how German does.

History of English specialists seem to suppose that it’s just that English merely drifted one step beyond German’s do—making it required instead of optional. But if that were so natural, so same-old same-old, then surely it would have happened in some other Germanic language sometime. Also keep in mind that each Germanic language comes in a bunch of dialects, many of which are quite different from the standard. If meaningless do is so unremarkable and could have “just happened,” then surely some small dialect of something somewhere—some villagers in the northern reaches of Sweden, some farmers down in some Dutch dell, some Yiddish speakers in a shtetl—somebody, somewhere would have come up with their own meaningless do just by virtue of shitte happening. But they haven’t.

Nor do, apparently, any other human beings beyond Europe. Meaningless do is not a long tongue—it’s a tongue used as a leg. Some readers will think perhaps of a language like Japanese, where quite often a verbal concept is expressed as “doing” a noun, such as travel being rendered as to do travel; here is Taroo travels:

Taroo ga ryokoo o suru. 

Taroo travel does

Persian is like this, too, so much that it has only a few hundred verbs per se—to speak Persian is to be accustomed to “doing a waking up” instead of awakening someone, and so on. But in both of these cases, do has literal meaning: one is “doing,” performing, the noun. And in neither language is do used with all nouns as meaningless do is used with all verbs. Japanese and Persian’s do is a meaningful word; meaningless do is a little cog of grammar that happens to have the shape of the actual word do.

I make no claim to have checked all six thousand languages in the world for a meaningless do, but I am aware of precisely two approximations of it anywhere but in Great Britain, and then, only approximations.

One is in a small language called Nanai spoken in Siberia, where it is be rather than do that is used meaning-lessly. To say They died, you can say “They died was.”

Hjoanči buikiči bičin. 

they died was

This does not mean “They were dead” or “They were dying”: the was is the third-person singular form, not the third-person plural form that would agree with they. The was word is just tacked on, with no meaning: “They died—uh, was.” A more graceful translation would be something like “What it was is that they died.” But even this is optional, like the Germanic dos.

Then there is the Monnese dialect of Italian and a few other close relations, out of dozens of Italian dialects, where the word for do is used in questions:

Ngo fa-l ndà? 

where do-he go 

“Where is he going?”

But in these Italian varieties, do is not used in negative sentences, whereas in English, do is used in both negative and question sentences—just as in Welsh and Cornish. And then, in Welsh and Cornish, do is also used in “default” affirmative sentences and was as well in earlier English—Gertrude in Hamlet’s “That you do bend your eye on vacancy.” So, in some Italophone hamlets, so to speak, do has been yoked into service in a meaningless fashion—but not in the particular way that it was in English, which mirrored precisely how do was used in the Celtic languages spoken by the people whom Anglo-Saxon speakers joined in invading Britain.

The only languages in the world that are known at present to have meaningless do as English does are (drumroll, please) none other than the Celtic languages. Can we really believe that the Celts had nothing to do with English’s meaningless do, which parallels it so closely and once did so even more? In fact, do we have any reason to consider that the Celts were anything less than the crucial factor, without whom English would have no meaningless do?

This question looms ominously over all of the specialists’ “shitte happens” versions of how English got meaningless do. All of them are brilliant in themselves, but also seem to ignore that meaningless do as it exists in English is about as weird as finding an AMC Gremlin on the moon.

One specialist tries that it all started with do being used to indicate that something is done on a regular basis—Cats do eat fish would mean “Cats are in the habit of eating fish”—and that something odd happened in negative versions of sentences like that. At first, Cats do not eat fish meant “Cats are in the habit of not eating fish,” as a kind of description of something specific about cat’s gustatory disinclinations. It was a description of a habit of cats, with do as the habitual marker. But obviously that sounds like a rather labored way of saying “Cats don’t eat fish,” and that’s exactly how people started processing it. Instead of “What cats do is not eat fish,” people heard “Cats eating fish is a ‘no’ ” That is, they heard the negation as the most prominent feature in the sentence and thought of the “habit” part as background. Thus a sentence that was first about do-ness became one about not-ness. Do started to seem like just some bit of stuff hanging around. It lost its “juice” and stopped meaning “regularly,” and eventually meant nothing at all, functionless like a hallowed old politician given a sinecure in acknowledgment of services rendered back in the day. Voilà, meaningless do.

Now, if you didn’t quite get that or had to read it again in order to do so, it’s not surprising. To be sure, it follows more gracefully when expressed in terminology that academic linguists are trained in, and the article in question is one of the most elegantly written pieces of scholarship I have ever read; it has always been, to me, almost pleasure reading—I’d take it to the beach. Yet the explanation is still a distinctly queer, Rube Goldberg turn of events. Nothing like that is documented to have happened to the word do in any other language on earth, and besides, the author even admits that in almost half of the sentences in the Early Middle English documents he refers to, do does not, in fact, indicate that something happens regularly.

“Future research” will figure out why, he has it—but how about if future research shows that what created meaningless do was not that English speakers for some reason drifted into the peculiar hairsplitting reinterpretation of Cats are not in the habit of eating fish as meaning “It is not that cats eat fish,” but the fact that the people who lived in Britain long before English got there had meaningless do already?

Then there are those who claim that meaningless do was a natural development in response to various ways that English’s grammar changed from Old to Middle English. For example, in Old English, verbs could sit in various places in a sentence—at the end, at the beginning, and so on, depending on what was next to it. In Middle English and beyond, verbs started sitting in the middle, after the subject and before the object, the way they do now (The boy kicked the ball). But there was an intermediate point when the general pull was toward the verb’s being in the middle, but there were still sentences like Wherefore lighteth me the sonne? (“Why does the sun light me?”) where the verb lighteth is before the subject sun. One way of thinking has it that meaningless do came in because when you use it, the verb ends up in the middle the way sentences by Middle English were supposed to go:

Wherefore lighteth me the sonne? 

verb object subject (Oh, no no!)

Why does the sun light me? 

subject verb object (That’s the ticket!)

Yeah—but the question is where English even got a meaningless do to use in this way. In languages all over the world since the dawn of human speech, at first the verb can hang around at the ends or beginnings of sentences and then at some point the verb is restricted to sitting in the middle; it happens all the time. And nowhere—nowhere—else on earth have such languages taken the word do and turned it into a meaningless little helper in order to nudge this along.

Biblical Hebrew put verbs first; Modern Hebrew puts them in the middle. Yet no one in Israel today is using the Hebrew verb for do in a meaningless way. The Arabic of the Koran puts the verb first; it has since morphed into the vast array of today’s actual spoken Arabics, stretching from Morocco across northern Africa into the Middle East, as well as down into Nigeria, Chad, and Sudan. Not one—not one—of these modern Arabic dialects has a meaningless do.

So, where oh where might English have gotten that meaningless do in order to whip its verbs into line? According to the History of English folks, in trying to figure that out, we are to ignore that the languages already spoken in Britain . . . I don’t even need to finish the sentence. It may well be that corralling verbs into the middle of the sentence made a meaningless do useful. But only English had a meaningless do available in the first place—as used by Welsh and Cornish speakers. Welsh and Cornish, then, were together the reason English has a meaningless do today.

English’s verb-noun present also looks, to traditional specialists, as if it is just one step past something in other Germanic languages. Again, there is a big picture they are missing.

Now, the sheer presence in a language of a progressive construction using a verb-noun is not all that extraordinary. It happens here and there that people say that they are “in” an action to say that they are in the process of accomplishing it at this very instant. This includes the Germanic languages. In German you can say, as I mentioned before, Ich bin am schreiben for “I am writing.” Similarly, in Dutch it would be “I am on the writing,” Ik ben aan het schrijven, and Norwegian has something similar with Jeg er åt å skrive.

English is peculiar, however, in taking the ball and running with it, to the point that the bare verb is nosed out completely. That is something much rarer in languages, popping up only in obscure corners here and there. Basque is one, a language related to no other one on earth. Or there is one little dialect of Greek (Tsakonian, for the record) that has booted bare verbs in the present and uses a progressive, for no reason anyone can see.

But then, another of the obscure corners in question is good old Celtic. English is the only Germanic language that developed in a context where Celts were the original inhabitants—and English is also the only Germanic language that turned its verb-noun progressive into its only present tense.

However, History of English specialists adhere to a just-so story in which the verb-noun present “just happened” by itself. Old English had two progressive constructions. One was the verb-noun one, I was on hunting, but another one used a participle form of the verb, marked with -ende. “I was following” was Ic wæs fylgende.

Now, today we do not say “I was huntende.” To the experts, I was hunting rather than I was on hunting happened because for some reason speakers started having a hard time telling -ende and -ing apart and settled on using -ing in the -ende construction.

But that -ende was an inheritance from Proto-Germanic. As such, other Germanic languages have their versions of -ende today, and no one confuses them with anything. German’s -ende, for instance, is doing just fine. Once in Germany I told a waitress not to put onions in my salad or I would become der kotzende Fremder (“the vomiting foreigner”). In earlier German, just as in earlier English, there was an -ende progressive; here is an Old High German sentence:

Ist er ouh fon jugendi filu fastenti. 

is he indeed from youth much fasting 

“From his youth on he has been fasting much.”

But then meanwhile, German drips with -en suffixes on both nouns and verbs, including their version of I was on huntingIch bin am schreiben (“I am writing”). Yet in a thousand years, this -en suffix has not been “confused” with -ende. Was it really just a shrug of the shoulders that supposedly led English speakers to confuse their -ende with -ing, which is much less similar to -ende than is German’s -en? Nowhere else in the Germanic family is the -ende ending so prone to “collapsing” into other ones. Why was -ende so uniquely subject to the vapors in only English? Historians of English are producing a description based on what they see in the documents over time rather than explaining it.

And the traditional version is hopeless in explaining why the verb-noun progressive, once established when -ende was “confused” with -ing, metastasized and became the verb-noun present, i.e., the only way to express the present short of sounding like a Martian by answering “I write” when someone asks what you’re doing. Why couldn’t I am hunting have stayed as meaning “I am in the process of hunting at this very instant”? That is, as it has in all of the other Germanic languages? Scholars charting the triumph of the verb-noun progressive over the bare verb diligently note how the progressive becomes ever more common—14 percent progressive! 45 percent! 67 percent! 92 percent!!!!!—rather than wondering why nothing remotely similar ever happened to any of the other languages that Proto-Germanic morphed into. They are describing, but not explaining.

The Celtic account, then, is more useful in the sheer scientific sense than the old one. It provides an answer to what specialists have been shrugging their shoulders about for eons: just why the verb-noun (-ing) ejected the participle (-ende) and just why the progressive became the only way to express the present tense. Welsh and Cornish express their progressive with a verb-noun and not with a participle. Welsh and Cornish use their progressive instead of bare verbs in the present tense. Period.

Yet among the specialists, to propose that the English progressive construction is a copy of Celtic’s is considered a renegade point. Certain wobbly speculations continue to be reproduced in sources both scholarly and popular. Maybe, we learn, Anglo-French scribes were hazy on the difference between -ende and -ing—which leaves unexplained how scribal errors on codexes read in candlelight by a tiny literate elite would have affected the way millions of mostly illiterate people out on the land spoke.

The likeness between English and Celtic is so close in this case that the only thing that would seem to save a traditionalist approach is the old assumption that there were barely any Celts around. And as we have seen, that assumption can’t stand.


1. The Angles, Saxons, and Jutes encountered Celtic speakers.

2. Meaningless do in the affirmative, negative, and internegative is found nowhere on earth except in Celtic and English.

3. English is the only Germanic language that uses its verb-noun progressive as the only way to express present tense; Welsh and Cornish do the same.

Asked what those three facts signify as to why English has meaningless do and and verb-noun present, all human beings would draw the same conclusion—except those who know more about the history of the English language than anyone else in the world. What could the blockage possibly be?

Assumption Number Three: Writing Is How People Talked

It’s a timing issue. On its face, the specialists seem to have a point, especially with meaningless do. Old English speakers met Celts starting in A.D. 449. This would be when Celts started learning and transforming the English language. Yet there is not a hint of meaningless do in any English document until the 1300s, in Middle English. Century after century of Old English writings and no meaningless do at all. Why does it show up so late?

The verb-noun progressive pops up in Old English now and then, in sentences like Ic wæs on huntunge for “I was hunting.” But today other Germanic languages also have their “on hunting” constructions when they want to stress “progressiveness,” and so on its face this does not make Old English look at all strange as its family goes. And in any case, more often, the progressive is the -ende one as in Ic wæs fylgende for (“I was following.”) In fact it is the -ende progressive that becomes a little more common in later Old English.

So Old English does not look terribly Celtic, and to traditional scholars this proves that the Celts cannot have been the source of meaningless do or the verb-noun present. What historians of English see is that English with meaningless do and a verb-noun present ruling the roost does not show up in documents until past the halfway point of English’s entire documented life span. If Celts mashed their mix-ins into English, then why did they take almost a thousand years to do it?

But that question proceeds upon a fundamental misconception about what ancient written evidence tells us—or doesn’t—about how the language was spoken every day.

Writing and talking are very different things. This is clear to us when it plays out in our own times with our own languages: in the entire eighty-five-year run of Time one could miss that in casual speech people say “whole nother.” But when we are dealing with languages of antiquity, whose casual renditions we cannot experience, the gulf between writing and speaking cannot help but be less apparent. We will never know how Old English was spoken by illiterate farmers; the written version that survives for us to peruse is the only rendition of the language we will ever know. However, people way back when were no more given to gliding around talking like books than we are, and in fact, writing and talking were much more different for them than for us.

In ancient times, few societies had achieved widespread literacy. Writing was primarily for high literary, liturgical, and commercial purposes. Spoken language changed always, but the written form rested unchanging on the page. There was not felt to be a need to keep the written form in step with the way people were changing the language with each generation.

For one, each language was actually spoken as a group of dialects very different from one another, such that there was no single spoken variety to keep up with. As long as the written form was relatively accessible to the general population, however they actually spoke, then the job was done. Old English, for example, came in four flavors: Northumbrian, Mercian, Kentish, and West Saxon. Most Old English documents are in the West Saxon dialect, because Wessex happened to become politically dominant early on. But this means that what we know as Old English is mostly in what is properly one dialect of Old English, and the speakers of the other dialects just had to suck it up. They did, and there is no evidence that anyone much minded.

In addition, there was always a natural tendency, which lives on today, to view the written language as the “legitimate” or “true” version, with the spoken forms of the language as degraded or, at best, quaint—certainly not something you would take the trouble of etching onto the page for posterity with quill and ink. As such, the sense we moderns have that language on the page is supposed to more or less reflect the way the language is spoken would have seemed peculiar to a person living a thousand years ago, or even five hundred.

In Europe, for example, it was the technology of the printing press and the democratic impulses in the wake of the Reformation that led to calls for written material in local languages. Until then, people in France, Spain, Italy, and Portugal readily accepted Latin—a different language entirely from what was spoken “in the street”—on the page. Similarly, for many centuries, Slavic language speakers were used to ancient Old Church Slavonic as a written lingua franca, although no one spoke it. Sanskrit was long the written language par excellence in an India where its offshoots, such as Hindi, had long since emerged and thrived as spoken languages.

Until well into the twentieth century, for Indonesians, the written language of books and newspapers and the one taught in school was Classical Malay, about as different from how people actually spoke as Shakespeare’s English is from how we do. Even today, although standard Indonesian is more in step with everyday speech than Classical Malay, it is still not the way anyone would actually converse casually who wanted to date or have friends. Indonesians giggle to see the spoken rendition of their language put on a blackboard, just as English speakers would find it inappropriate to see Time written in the dialect of rappers.

Among people worldwide before five hundred years ago, this kind of gulf between written and spoken language was, in a word, a norm. In the late 1200s, Dante considered himself to be venturing a special gesture in writing La Vita Nuova in Italian instead of Latin, only doing so because he wanted the work to be accessible to a woman who did not know much Latin. In Dante’s world, Italian as we know it had long existed as an everyday language—Dante himself spoke it 24/7—but the ordinary thing was to write in Latin. Latin was what Italian (and French, Spanish, and the other Romance languages) had been a thousand years before. But since Latin hit the page first—Latin “called it!” as kids say in grabbing a seat—a millennium later, there still reigned a sense that Latin was fit for the page, while Italian was just “the vulgar tongue,” as even Dante put it, “in which even housewives can converse.”

There is a similar situation today in Arabic-speaking countries. In, say, Morocco, the Arabic used in writing is one thing, preserved on the page and kept as close to the language of the Koran as possible. The Arabic actually spoken is another thing, morphed fifteen hundred years away from the written variety and now a different language entirely. A Moroccan will recall learning “Arabic” in school, so different is the standard variety from the “Arabic” they learned at home.

Moreover, in each Arab nation, the standard has drifted into a spoken dialect in different ways such that the spoken Arabic in each place is a different language from the others. So, good in the standard is jayyad, but look at what it is in different countries in the Arab world:


To Arabic speakers, the preservation of Standard Arabic on the page, and only occasionally or even never seeing the language they actually speak in print, feels normal. It is even seen as an advantage in forging ties between different Arab nations.

It is in this light that we return to the fact that the Celtic impact shows up in English documents only long after the Celts and the Old English speakers had first come into contact. The personnel in question did not live in anything like our world. They lived lives in which there were no potatoes, tomatoes, coffee, tea, chocolate, spinach, broccoli, or sugar. They didn’t need last names because most of them spent their lives in small villages where everybody knew one another. Dismemberment and murder were so common that adjudicating their outcomes was the main focus of suites of laws penned by kings like His Highness Mr. Ine.

And, by and large, speakers of Old English did not read. Writing was better described as scripture—a formal, ritualized, elite pursuit, preserved via scribes copying old texts century after century, sequestered in thick-walled edifices from the hurly-burly of actual everyday speech. In this rigidly classist world, casual English—especially the kind associated with wealhs, who were usually slaves—was no more likely to wind up engraved in ink than the charming babblings of toddlers.

As such, the only reason that Celtified English started coming through in writing even in the 1300s was a historical accident.

The year 1100 is when, largely, Old English stopped and Middle English, an almost curiously different thing, began. Middle English was, indeed, a profound transformation of Old English. Partly, yes, in terms of words—a bunch of French ones started pouring in—but also in terms of grammar. When the Norman French conquered England in 1066 and established French as the written language of the land, for the next century-and-a-half there is almost no written English that has survived. Then after relations with France began to sour in the early 1200s and English started to be used as a written language again, we see a brand-new, slimmed-down English, as if it were in an “after” picture in a diet ad.

Old English had been jangling with case markers, and nouns had three genders as in Latin, Greek, and Russian. In Middle English, waking up like Rip Van Winkle around 1200, case and gender were largely as they are now: vestigial and absent, respectively.

In Old English, words like this one for stonestān, took different suffixes according to case and number. Suddenly in Middle English, the only case suffix left is genitive (possessive), and the plural one is the same in all cases. Suddenly, in Middle English we are, in other words, home.


The question, though, is whether this all really happened in 150 years, and certainly it did not. Languages do not suddenly chuck away their case markers, almost as if people speaking languages with lots of them find them, deep down, as burdensome as we Modern English speakers do when exposed to them in classrooms in Latin and Russian. Greek, like Russian, has been chugging along for millennia with enough case markers to sink a ship.

Sure, languages will slough off a suffix here and a suffix there, just like we do our eyelashes. Just as sure, some languages will slough off more suffixes than others, just as many men go bald on their heads but retain hair, well, elsewhere. But even that takes time. No language “goes bald” in just a century-and-a-half.

Rather, in languages’ documentation we watch these things happen gradually—century by century, as with those few more hairs you keep spotting on the shower drain as the years pass. Old High German had a Latin-style fecundity of case and gender inflection. Modern German has much less (although still a lot by English standards). We can watch that happening in a majestic procession of documentation over almost a thousand years—not just 150, roughly the time since the end of the Civil War! And even then, the end result was not a language as denuded of case and gender as English, but the German that so frustrated Mark Twain.

For this reason, on the stark difference between the English of Beowulf and the English of The Canterbury Tales, even specialists agree that there is some discrepancy between what we see in the documents and what was the living reality of everyday Old English. Namely, Middle English is what had been gradually happening to spoken Old English for centuries before it showed up in the written record.

The Old English in writing, then, is the language as it was when the Germanic invaders brought it across the North Sea, preserved as a formal language, a standard code required on the page, kept largely unchanging by generation after generation of scribes and writers imitating the language of the last. The language used every day was quite different, not policed and preserved the way the written language was, free to change naturally as all spoken language does, such as by losing suffixes one by one.

The only thing that led writers to start actually putting this “real” Old English on the page was the 150-year blackout period. When people started writing English once more, the written Old English standard could not exert the pull that it once had. These were now documents of another time. One hundred and fifty years was a vaster amount of time to a Dark Ages Englishman than it is to us—he had no photos or newspapers as we do of the Civil War, and no audio recordings as we do of the 1890s onward—and the continuity between generations of scribes preserving the old language had been broken. It was as the French had taken over Dante’s Italy for 150 years, imposing French as the language of writing. Imagine if after the French left, writers were no longer competent in Latin and felt more comfortable writing in the language people actually spoke, Italian.

In this light, our timing problem with the Celtic features is solved.

Traditional specialists understand that Old English was losing its case markers gradually even though writers wrote as if this wasn’t happening. As such, they should be able to accept that “Celty” English would also have been spoken out on the ground even though no one would have deigned to transcribe it amid the formality of the written word. This is not a studied argument designed to get around something about Old English, but a call to bring scholarship on The History of English in line with the realities of how different writing was from casual speech in the ancient, semiliterate world.

We can assume that Celts were speaking Celtified English starting with the first generation who grew up bilingual, as far back as the fifth century, and throughout the Old English period. However, this was not the English from across the North Sea—Celtified English was likely thought of as “mixed” or at least funny-sounding English for a long time. As such, it would never have been committed to print—and in a world without audio recording technology, this means that this kind of English as spoken during the entire reign of Old English is hopelessly lost to us.

However, starting in the Middle English period, when it became acceptable to write English more like it was actually spoken, this would have included not only virtually case-free nouns, but also our Celticisms. Therefore, it is not that Celticisms only entered English almost a thousand years after Germanic speakers met Celts in Britain. It is merely that Celticisms did not reach the page until then, which is quite a different thing.

People writing the way they actually talked was quite rare anywhere in the world until rather recently, and even today it is by no means universal.

The truth, then, is that if meaningless do and the verb-noun present did pop up in the first Old English documents, or even in Old English documents at all beyond the occasional peep, it would be very, very strange. We would expect that the constructions would show up only after a historical catastrophe such as the Norman occupation, after which, in many ways, England learned to write again. If the Battle of Hastings had not put a 150-year kibosh on written English, then “real” English might not have been committed to print until as late as after the Reformation, in the 1500s.

In the obituary of someone who started some famous chain of stores, often the date that the first establishment opened seems much earlier than you would have expected. The first McDonald’s, for example, opened in 1955. That doesn’t “feel” right: McDonald’s was an entrenched part of American life only ten years after that or more. For example, there is an I Love Lucy episode from 1956 where Lucy and Ethel are making a long road trip and running low on food, as fast-food restaurants alongside interstate highways were not yet ubiquitous. For a long time after 1955, McDonald’s restaurants were in business, but because they had yet to proliferate widely, to most people they were barely known. The first Wendy’s was opened in 1969—my intuition would have put it in about 1978.

Likewise, the Celtic imprint on English would have thrived below the radar long before it appeared regularly in print, even when meaningless do and the verb-noun present had long been well established as ordinary speech. They just weren’t being publicized in commercials yet, so to speak. Since there was no recording technology, we can’t hear Old English speakers using them. But they did. We know that because English was the only Germanic language spoken by people whose native languages had the selfsame traits.

One Last Assumption: Where Are the Celtic Words?

There is one last thing that misleads linguists into thinking the Celts could not have had any significant impact on English: the fact that there are, essentially, no words in English that trace to Celtic.

One might expect there to be some, after all. The Vikings left a whole mess of their words, as did the Normans. One would presume that when large numbers of people start using a language imposed on them and start speaking it in their own way, that they will sprinkle their version of the language with a lot of their own words. The Vikings left behind their get and skirt and even their their; the Normans left behind seemingly every word we use to step beyond humility. So where are the long lists of Welsh and Cornish words?

Instead, there are only a dozen-odd words that have been traditionally traced to Celtic, and most of them are arcane, obsolete ones introduced by Christian missionaries from Ireland. Naturally, then, experts assume that the Celts must have just learned English the way they encountered it and added nothing to it. This assumption is reasonable. It is also mistaken.

The fact is that people scattering their own words into their new language is not a universal. It might happen—the Vikings did it; the Slavonic-speaking people who picked up Latin in what is now Romania did it to Romanian. But it might not.

Russian, for example, has some quirky features that show that at some point way back, it was learned by so many speakers of another language that it was never the same again in terms of grammar. The culprit was a language of the family called Uralic. Its most famous members are Finnish and Hungarian, but other ones have long been spoken across a vast expanse of what is now northern Russia. In Russian, it seems odd that in negative sentences an object has to be rendered as “of” itself: “I see a girl” is Ja vižu devočku, but “I do not see a girl” is Ja ne vižu devočkiwhere the -iending connotes “of-ness” (“I do not see of a girl”). Odd, that is, until you notice that Finnish and its relatives do that same thing. In Russian, unlike in a “card-carrying” Indo-European language, you do not “have” something: rather, something “is to” you: “I have a book” is U menja kniga (“to me is a book”). Again, something similar is par for the course in Finnish.

No one interested in the Russian-Uralic encounter denies that Russian picked up these and other things from speakers of Finnish-related languages. It’s as if your child comes back from summer camp with some downloaded music they never listened to before, from some friends they met who were into that kind of music.

Yet there are at very best about a couple of dozen Uralic words in Russian, most of them obscure. The Vikings left about a thousand in English, and the Normans left ten thousand. Yet the Uralic speakers left just a handful in Russian. We will never know just why; certainly it was due to specific cultural factors lost to us because the people had no writing.

It is the same in India: in the southern part, there is a smallish family of languages, Dravidian, completely unrelated to the other ones, which are of the Indo-Aryan subfamily of Indo-European. When you hear that a person from India speaks Tamil, for instance, that is a Dravidian language, as unlike Indo-Aryan Hindi, Bengali, Gujarati, and the others as Finnish is unlike English. In any case, along the barrier between the Dravidian area and the Indo-Aryan area, people have often been bilingual in Dravidian and Indo-Aryan languages—but over the past thousand years, almost no Dravidian words have seeped into Indo-Aryan languages. Yet Indo-Aryan words are fairly dripping with features in their grammars which, again, no linguist denies are the result of Dravidian speakers learning Indo-Aryan ones.

It also bears mentioning that, really, etymology is not the most rigorously policed of fields. Much of the basic work was done long ago under different standards of evidence than linguists would admit today; there are a great many holes (“etymology unknown”), and legions of etymologies that, if linguists were moved to seriously examine them today, would fall apart. In that light, there is some work suggesting that there are at least a few more Celtic words in Modern English than once thought. Candidates include bragbrat, curse, and baby.

In any case, the paucity of Celtic words in English is no argument at all against meaningless do and present-tense -ing being due to Celtic influence. It’s interesting—the work that argued that Dravidian languages decisively shaped Indo-Aryan grammar is today cherished as sage, classic, and incontrovertible. Yet a very similar argument about Celtic and English is received as quirky, marginal, and eternally tentative.

Celtic Underground Even Today

To show how ordinary it would have been for a “Celtified” expression to almost never make it onto the page over centuries’ time, here is a living example. There is a queer little wrinkle in regional dialects in the north of England. Standard English verb conjugation in the present tense involves one thing: tacking on -s in the third person singular:


In the northern dialects in question, instead the rule is that you tack on the -s in all persons and numbers:


Except for one thing: in the third person plural, when you use the pronoun they instead of nouns like Betty and Shirley, or children, or McDonald’s outlets, you drop the -s:


So Betty and Shirley walks, but they walk.

Weird, isn’t it? There is nothing like it in any Germanic language but English. But there is something just like it in—need I even finish the sentence?

With the Welsh verb, in the third person plural, when nouns like Betty and Shirley are involved, the conjugational ending is the same as for the third person singular one. Again, verbs are first, and so Welsh has learned she for she learnedlearned Betty and Shirley for Betty and Shirley learned:

dysgodd hi (“she learned”) 

dysgodd Betty a Shirley (“Betty and Shirley learned”)

But if you use the pronoun they, the verb takes a third person plural ending:

Dysgon nhw (“they learned”)

Cornish has the same thing.

Thus in Welsh, Cornish, and these dialects of English, how you conjugate the verb in the third person plural varies according to whether the subject is a noun or a pronoun. In itself, that seems an arcane and, to anyone but a linguist, dull thing. But for our purposes, the crucial fact is that no Germanic language other than English knows anything like this.

And overall, in terms of English or any European language beyond Celt land, this quirk, which linguists call the Northern Subject Rule, is one of those “Who’d a’ thunk it?” things. Even History of English specialists see it as an oddity: it is not a run-of-the-mill development that happens in this and that language randomly like, say, conjugational endings.

In fact, it is something that happens occasionally in one specific type of language: the roughly one in ten worldwide that put the verb first. Like Tagalog in the Philippines. Or like . . . hmm. Thus we can form a good idea as to why these English dialects have taken on such a bizarre trait.

Yet the reader, especially if American, is unlikely ever to have known of the Northern Subject Rule, because it happens to have taken hold only in northern British dialects. Standard English developed from dialects far southward, and so the Northern Subject Rule has remained a strictly spoken feature, uttered countless times daily and evaporating into the air, recorded on the page only by occasional diligent dialecticians. It is unassailably Celtic, and yet unknown in the pages of The Economist, and always will be.

Crucially, there is no reason that meaningless do and the verb-noun present could not have thrived in obscurity in the exact same way until “real” English got to come out of the closet in the 1200s. The Northern Subject Rule is living the same closeted life today, showing us clearly that what is written can often be strikingly different from what is said.

What Is Proof?

As to whether English has a goodly dose of Celtic in it, at this point there is little that The History of English orthodoxy has left to deny it.

The scholars working in the traditional vein seem unable to arouse genuine interest in changes in the language that they cannot trace step-by-step in the documents starting as soon as they emerged. Hence the judgment on the issue in a benchmark study of Middle English: “There might be something to say for Keller’s and Miss Dal’s assertions that the ancient Britons were not exterminated but became amalgamated with the Germanic invaders and assumed their language while retaining some syntactical peculiarities of their ancient native tongue, but such statements remain necessarily hypothetical for lack of documentary evidence.”

Even though that was written in 1960 (hence the “Miss Dal”), mainstream sentiments have not changed since. Developments that cannot be followed from when they started are, to the experts, not worth extended engagement.

But following changes in English starting from when they hit the ground in casual speech is a luxury available only from documents dating from when English was written more or less as it was spoken. Old English was almost never written that way. The Celtic impact must be embraced in the frame of mind of, say, a paleontologist who reconstructs the behavior of dinosaurs from fragmentary but indicative clues.

There are pathways of footprints left by herds of sauropod dinosaurs, the Brontosaurus (okay, Apatosaurus) type, in which smaller footprints run in the middle while the bigger ones run along the sides. Paleontologists have inferred from this that younger sauropods were protected by being flanked by the big older ones, as among some animals today. We will never have film to prove this, and most likely will never resurrect sauropods with DNA and watch them do it. Yet it is accepted that the paleontologists’ reconstruction is a valid approach to the evidence available, and almost certainly correct.

The likenesses between Celtic languages and English are a similar case. Realities of the history of writing among human beings in ancient semiliterate societies make it impossible that we would find meaningless do in Old English documents like Beowulf, the Lindisfarne Gospels, Aelfric’s Colloquy, or Cædmon’s Hymn, even if meaningless do was being used casually every day all over England. Yet the presence of the same feature in Welsh and Cornish, and its absence used this way anywhere else in the known world, make treating it as something that just happened all by itself in English seem almost strange.

Overall, scholars of English’s history are less resistant to than uninterested in the impact of Celtic. The reason, one senses, is that charting how Celtic languages shaped English does not involve using the tool kit they are accustomed to. These scholars are trained to examine aspects of English grammar that really did emerge by themselves and were never thought of as “bad” or “peculiar,” and thus were committed to the page not long after they got going.

Going is, in fact, a good example, in the going to future marker, English’s alternate to good old will. This is the kind of thing English specialists love to sink their teeth into. In Old English, there was no such thing as using the word for go to put a verb in the future as in I’m going to think about thatGo was about going somewhere and that was that. Even as late as Shakespeare, at the end of the 1500s, go still meant go. In Two Gentlemen of Verona, the Duke asks Valentine, “Sir Valentine, whither away so fast?” and he answers, “Please it your Grace, there is a messenger that stays to bear my letters to my friends, and I am going to deliver them (III, I, 54-57).” Valentine means that he is literally going in order to deliver the letters.

However, if you are going in order to do something, then automatically what you are going in order to do will actually occur in the future. As such, Valentine’s statement could be taken as meaning that his delivery of the letters will occur in the future—that is, that he will deliver the letters. Because of that ever-looming implication of futurity whenever one said going to, after a while going to started to actually mean the future rather than actual going.

It is about fifty years after Two Gentlemen that Charles I, amid the crisis that would soon cost him his head, rallied the gentry of Yorkshire saying, “You see that My Magazine is going to be taken from Me.” (Poor Charles, for the record, was not complaining that he was to be deprived of his Sports Illustrated; by magazine he meant “arms depot,” more pertinent to his situation.) This was a usage of going to that was not literal—the arms depot could go nowhere. Going to here had become a future marker like will, and wouldn’t you know, around the same time in 1646, a grammarian popped in specifying that now “ ‘going to’ is the signe of the Participle of the future.”

There is ample scholarly work on how going to went from referring to locomotion to becoming a future tense marker, complete with statistical analysis, tables, and so on. It’s great stuff, and it’s what a scholar of language change is trained in.

However, charting how Celtic languages impacted English involves different strategies. It requires being a different brand of linguist. Often, that brand is language contact specialist. That person has an eye on what sorts of features are common around the world and what sorts are not, is obsessed with not just one language family but with several, and has a native taste for history as well as linguistics. Such linguists are less tickled by things that sprung up in a language by themselves than by things that languages did to one another.

As such, it’s as if scholars of The History of English are engaged in a lusty game of Monopoly when adherents of the Celtic idea bust into the room asking who wants to play a game of Clue. Or, some people are building things with an Erector set and someone pops in with a little car made of Legos. To the traditional specialist on how English got from Beowulf to The Economist, drawing parallels between English and some other language is just Not What They Do, especially not at any length. That feeling is understandable, but the problem is that the language contact specialist’s analysis, in this case, squares with logic in a way that the same-old same-old analysis simply does not.

Frankly, another likely factor is that Irish, Welsh, and Cornish are not languages anyone is apt to become familiar with who is not of Celtic ancestry. Andrew Dalby, working outside of the academy, has a way of getting such things tartly right. On the Celtic question, he gets in that “few English linguists know Welsh, so the similarities tend to be overlooked or played down.” Yep—I highly suspect that if Welsh were, say, for some reason regularly taught in schools across Western Europe and in America, as French and Spanish are, then to linguists, raised with “schoolboy” Welsh, the parallels between Celtic and English would seem glaringly obvious and would long ago have been accepted as having a causal rather than correlative relationship.

However, here in real life, even to seasoned linguists, Celtic languages are, as often as not, remote oddities, bristling queerly on the page à la the likes of Sut rydych chi? meaning “How are you?” in Welsh. Rydych??? How do you even pronounce that?? To someone whose foreign language competence is in French and German, there is nowhere to grab on to here. One moves on.

All that understood, the facts tell a story even if we will never have the “documentary evidence” of the kind the scholar quoted above was accustomed to working with. Swords and grimaces could not have exterminated a race of millions of Celts and left a few huddling in Wales and Cornwall. Rather, Celts, albeit subjugated, lived on throughout Britain in vast numbers. The Germanic invaders, like dominant classes worldwide at the time, enshrined a version of their language on the page that reflected what it was like before it came to be spoken and reshaped by the people who, albeit subjugated, continued to vastly outnumber them, and who passed their rendition of the language on to future generations both Germanic and Celtic. After the Norman French conquered the country, English was rarely written for a century-and-a-half, and when English was reawakened on the page thereafter, it suddenly had a grammatical flavoring that paralleled no languages on earth but Celtic ones, while English’s relatives over on the Continent developed nothing similar.

Those facts lend themselves to an analogy about people we will call the Robinsons and the Joneses.

In 1870, Mr. Robinson and his family move to a small town in Illinois called Summerfield. Thirty years later, in 1900, the town’s newspaper does a story about how Mr. and Mrs. Robinson and their three offspring have developed an unusual deftness in playing the piano with their feet. They play only with their feet, never with their hands, and can manage pleasant renditions of classical sonatas. The story also notes that the Robinsons’ elderly next-door neighbors, the Joneses, have the same skill, as do their kids.

The news story does not tell us whether the Robinsons learned to play the piano with their feet from the Joneses. However, it does note that the Robinsons were close friends with the Joneses and that the Joneses’ son Thaddeus even married the Robinsons’ daughter Minerva.

Researching the issue in 2008, we find two other things. First, in 1880, researchers in the new field of sociology did an extensive study of the town the Robinsons moved to Summerfield from, Wistful Vista. And even in their chapter on the arts in Wistful Vista, which includes a detailed description of the town’s musical scene, there is nary a mention of anyone playing the piano with their feet. Nor in the annals of descriptions of, or reports from, any other mid-nineteenth-century Illinois towns is there any record of people playing piano with their feet, just as today the practice is unheard-of in Ohio or anywhere else.

Second, Mr. Jones, having made his way into serving as Summerfield’s water commissioner, left his papers to the local museum, and among them is a daguerreotype of him playing the piano with his feet way back in 1850, long before the Robinsons moved into the house next door.

Obviously, this evidence makes it rather plain that the Robinsons picked up their quirky approach to piano playing from the Joneses. However, imagine modern historians instead insisting that the Robinsons learned to play the piano with their feet on their own, despite that the Joneses right next door, their close friends, had been doing just that long before the Robinsons moved to Summerfield.

Our historians craft elaborate webs of motivation that would lead the Robinsons to take off their shoes and socks, hoist up their legs, and attempt “Chopsticks” with their toes. Mr. Robinson was a banker—maybe he developed repeat stress syndrome in his hands from using the telegraph machine while communicating with banks out of town and found that the only way he could play the piano was with his pedal digits. Maybe Thaddeus, whom the article described as a spirited fellow “full of the dickens,” was as a tyke given to athletic stunts like putting his bare feet on the keyboard.

Yes, maybe. But all of this leaves the outside observer wondering what the use is of concocting scenarios like this. The scenarios would seem, ultimately, to be for some reason turning a blind eye to an obvious explanation. What purpose does it serve, we ask, to deny that the Joneses taught the Robinsons how to tickle the ivories with their feet? And what is the use of pointing out that the Robinsons don’t dress much like the Joneses? (That’s the part about Celtic words, in case the analogy is slipping!)

Or even: why conclude that the Joneses may have been “just one influence” on the Robinsons? “Acknowledging” both sides is of no use in this case. The Robinsons learned how to play the piano with their feet from the Joneses. Period. If the Joneses had not already been playing the piano with their feet, the Robinsons would not be, either.

The judgment must be the same on Celtic’s impact on English. The facts in this particular case do not lend themselves to mere parenthetical civil surmises that Welsh and Cornish “may have influenced ” English grammar, with the treatment otherwise proceeding as usual, describing meaningless do and the verb-noun present drifting into existence by themselves for no reason. The facts do not indicate that the Welsh and Cornish features merely pitched in on a process that would have happened by itself anyway. If Old English had been brought to an uninhabited island—or, say, Cyprus, Greenland, or Fiji—rather than an island where Celtic languages were spoken, then there would be no such thing as a Modern English sentence like Did you see what he’s doing? That sentence would be rendered as See you what he does?, as it is in any normal Germanic—or European—language.

English is not normal. It is a mixed language not only in its words, but in its grammar. Every time we say something like Did you see what he’s doing?, we are structuring our utterance the way a Welsh or Cornish person would in their own native tongue. When well-intentioned chroniclers take in from scholarship on The History of English that “the English language has been indifferent to the Celts and their influence” (Robert McCrum, William Cran, and Robert MacNeil’s The Story of English) or that “the Celtic language of Roman Britain influenced Old English hardly at all” (David Crystal’s The Cambridge Encyclopedia of the English Language), they have been misled, despite the brilliance of their books overall (both of these are among my favorite books of all time).

English is not, then, solely an offshoot of Proto-Germanic that inhaled a whole bunch of foreign words. It is an offshoot of Proto-Germanic that traded grammar with offshoots of Proto-Celtic. The result was a structurally hybrid tongue, whose speakers today use Celtic-derived constructions almost every time they open their mouths for longer than a couple of seconds. Do you want to leave now? What’s he doing? Did he even know? What are you thinking? I don’t care. She’s talking to the manager.

Celtic grammar is underneath all of those utterly ordinary utterances in Modern English. Our language is a magnificent bastard.



If you find an error or have any questions, please email us at Thank you!