Modern English seems to require this verb in several circumstances, where most other European languages don't seem to need it. (See? I just used it.)

For example, in questions: "Do you have a dog?" Whereas, "Have you a dog?" would be normal in other languages or in the English of days gone by.

Another example: "I do not know him." Again, "I know him not," could be used, but sounds stilted nowadays.

So, where did this 'do' come from?

  • 2
    Related, but theory-specific explanation for the use of do. But really, I think English works this way "because it does", just like many languages leave their question words in situ while others move them to the beginning of the sentence. Or how some languages have negative concord, while others don't.
    – Kosmonaut
    Commented Jan 3, 2011 at 21:59
  • 3
    "Our Magnificent Bastard Tongue" by John McWhorter goes over some of the more obscure aspects of English history, including the meaningless do. I don't have my copy available, but I remember there was something about early Celtic and Welsh influence that did it.
    – Jon Purdy
    Commented Jan 3, 2011 at 23:21
  • 3
    By the way, the reason do, of all verbs, would be chosen for this job is that it is a very common light verb. It doesn't have a lot of semantic content embedded into it. Many languages employ light verbs for various functions.
    – Kosmonaut
    Commented Jan 4, 2011 at 1:15
  • 2
    In Ireland you also hear "do" + verb for habitual actions. For example, "I do go there on a Friday" means "I usually go there on Fridays". Commented Jan 4, 2011 at 10:44
  • 2
    To elaborate on @Jon's comment, if I remember correctly McWhorter says that Welsh and other early Celtic languages had essentially the same type of do- construction, and he speculates that this is where it came from ... the Celtic inhabitants of England, who were conquered but not wiped out, adopted the AngloSaxon vocabulary, but some aspects of Celtic grammar remain in English. The alternative is to assume that English just by chance adapted the same relatively rare grammatical construction the dead Celtic language in the area had used. Commented Jun 6, 2012 at 13:48

3 Answers 3


Constructions using the equivalent of "do" as an auxiliary for reasons of emphasis or particular meanings are known in other languages.

An example that comes to mind is Middle Welsh: In the Mabinogion, for example the phrase "oruc ... a" occurs often before another verb, and means "did and ..." or "made and ...".

It is also theorised that the "-d/t" ending of weak Germanic past tenses (eg "walked") is a remnant of an ancestor of "did", so "walked" comes originally from the collocation "walk did" (not in English but in an ancestor language).

In modern standard English we use "do" in the affirmative only for emphasis or ("I do want it"), but in some dialects, and in older forms (eg Shakespeare), it is more common. In fact in Shakespeare, we find both affirmative and negative sentences with and without "do".

I have a theory about why "do" has spread to be compulsory in the negative in modern English: I think that Jespersen's Cycle operated to create an anomalous Head-modifier construction ("I go not", where the "not" follows the head word "go"), and the modern form has spread because it is more normal in English constructions for the modifier to precede the head ("I don't go"). I've not seen this idea suggested anywhere else, though the application of Jespersen's Cycle to Old English is well established.

  • Isn't the apparently arbitrary Shakespeare use of do attributed to poetic license, so he could follow theatrical stylistic guideline?
    – Eldroß
    Commented Jan 17, 2011 at 8:19
  • 2
    I don't know, is it? What "theatrical stylistic guideline" are you referring to? No doubt his choice of when to use "do" and when not to is partly determined by rhythm and metre, but it's clear that both options were available to him.
    – Colin Fine
    Commented Jan 17, 2011 at 10:21
  • Regarding collocation of verb + did for past tense: It would be a beatiful symmetry if this was the case, but alas; in Norwegian, a successor of old Norse, we have the verb 'vaske', meaning 'to wash', as an example of a regular verb with clear connection to the English counterpart. In the past tense this is conjugated 'vasket', again this seems to have the same origin as 'washed'. However, in Norwegian, there is no trace of any do-construct, so I think it is very unlikely that -ed stems from that, rather it seems to be a suffix with Germanic origin.
    – RolKau
    Commented Jul 29, 2018 at 8:09
  • @RolKau: I don't doubt that it is a suffix from Common Germanic. I am talking about the suggestion (I do not remember where I read it) that the origin of the suffix in Common Germanic might be a form of the verb underlying do (= German tun), suggesting that do-support was found in an unrecorded ancient language. I'm not aware that the verb do has a cognate in North Germanic.
    – Colin Fine
    Commented Jul 29, 2018 at 18:26

It comes from the Celtic languages spoken by the original inhabitants of the British Isles. (Irish Gaelic, Welsh, Cornish, etc). These languages were the original languages to have the odd use of the verb "do". In fact, there is not a single documented language on Earth outside of the Celtic languages and English to use "do" in this specific way. Here's how it ended up in English: When the Angles and Saxons (the original speakers of Old English) invaded from northern Germany, they took over most of the island of Great Britain. Because there were actually only a few thousand invaders, yet many more Celtic people there, it was not too long before most of the people speaking Old English were native speakers of those Celtic languages, and they spoke it the same way they spoke their native languages - with the odd use of "do". That's why a sentence such as "Have you a dog?" sounds old-fashioned today: it is old-fashioned. It is the original form of Old English.

Hope this helped! :)

  • 2
    Thanks for your answer and welcome to the site. It would greatly bolster your contribution to include references to this, because although there does exist documentation supporting the theory, it is not especially well known by the general populace.
    – tchrist
    Commented Sep 16, 2016 at 2:55
  • Also, if this explanation is not universally accepted among experts (and I have the impression that it is not) it would be appropriate for you to mention this.
    – herisson
    Commented Sep 16, 2016 at 4:31

This comment is correct that this question is answered by McWhorter, J. PhD Linguistics (Stanford), in ch. 1 of Our Magnificent Bastard Tongue (2009). The whole chapter's too long to quote; I quote one paragraph of the conclusion on p. 60.

 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.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.