Tuesday, 1 December 2009

Data, Singulars and Plurals

The data were gathered from weather stations around the world...

I nearly spilled my coffee all over the Sunday Times article. A typo? No.

Jones was not in charge of the CRU when the data were thrown away...

I know all about the notion that data is plural. I have seen the complaints; but I had never before seen data used as a plural outside of the pages of a style guide. It's like graffiti: any Italian will tell you it's plural and so will Fowler and his ilk, but nobody writes like that. I mean style guides are all very pleasant to read, but nobody, I thought, acted on them. Data is there to complain about, not actually pluralise.

Data is one of those words that have been picked out at random for special pleading. It is, in Latin, a second declension neuter plural nomintative. But the same argument is never made about agenda and stamina, both of which are also Latin plurals. Nobody ever says, 'My stamina are failing,' nor does any pedant or grammarian suggest that one should. There's a list of 55 Latin plurals here.

Gallows was originally a plural (because a gallows is made from several pieces of wood or galgi). But the word was already singular in Shakespeare's time (the gallows is built stronger than the church).

I even used to know an enthusiastic scholar of breakfasts who insisted that porridge was plural as it was a shortening of porridge oats: so 'How are your porridge?'

Etymology is no guide to number, and the grammar of a dead and foreign tongue cannot be applied to English (otherwise you would have to say "interpretations of the datorum vary" (it is strange that the rules tend to stretch to the little learning of the pedant)). Fjords would have to be fjorder, as in the Norwegian. When discussing Alans Hansen and Shearer you would have to go further and discover what the Hindi plural of pundit is.

We would all have to return to proto Indo-European grammar and there would be a great hush.

People write angry letters to the paper when data is singular, but the reader receives a strange jolt when the word is plural. One maniac correspondent wrote "I rebel at the phrase 'the data shows' which has become well-nigh universal", which means I suppose that he is reacting against the universe. If a usage is universal, it is correct. Fowler said that the word was 'plural only', but that was half a century ago. One might as well insist that the second person singular be thou.

The truth?

All of this comes under one simple rule that Bill Bryson and Kingsley Amis both agree on: number in English grammar is controlled by thought.

A gin and tonic is a drink, despite there being two nouns, because the drink is a single idea. One could conceivably say that gin and tonic are both ingredients in a gin and tonic, but it would be awkward. Fish and chips is a meal. Law and order (if considered as a single idea) is breaking down. The long and the short of it IS that:

Tomorrow and tomorrow and tomorrow
Creeps [singular] into this petty place

The Telegraph is therefore correct to write that "ministers have insisted that the Olympics is 'on budget'", because the budget would be for the games as a whole: a single idea.

Conversely, a singular noun can be treated as a plural. The couple are in love with each other. One could not reasonably say that the couple is in love with itself. The National Youth Orchestra are all in their teens, but the National Youth Orchestra is 61 years old. It depends how you're thinking about it/them.

Data is singular so long as I consider it so and plural when the whim strikes me. Teams, orchestras, armies, quintets, convocations, councils, countries, sets, groups, flocks, herds, prides and opera must bend  and bow to my ineffable will.

There is, of course, mistakes in number. Often a writer forgets, after a subordinate clause, how many subjects their verb had.  People forget that neither or nor nor give the plural. There are cases, occasional and rare - perhaps once in a century, that requires that two words distant in meaning cannot reasonably be considered a single idea. I even noticed in a previous post that I had made a country singular and then plural with no reason for the change, a shift that Fowler rightly objects to.

As a bit of trivia (another second declension nominative neuter plural), plus, being a preposition, does not change the number of a noun. One and one are two, but one plus one is two. "Tom, plus his friends, has arrived."

And then there are those confused serpents like Mr Fry who think that none has to be singular. It does not. Evidence and video are to be found in my previous post here.

How many legs?

P.S. There's a Polish word that's plural but used as a singular in English, but I can't for the life of me remember what it is. Anybody?


  1. Well said.

    It seems that Fry has recanted his crazy prescriptivist leanings:

  2. Is there still the Civil Service assumption that Her Majesty's Government is singular and foreign governments plural? It is a helpful distinction in writing.

  3. Panjomin, you may well be right and if you are that would make two words, as I had never in all my sheltered life heard of kielbasa (it seems to be an American term and I live in little thatched cottage in the middle of a cricket pitch).

    Anon, I had never heard that before, but I'm having supper with a senior civil servant tonight so I shall ask around and report any findings.

  4. Anon, the result of my dinner was that the British Government is alwasy referred to as HMG and is singular. The French Government would be the French government and might be pluralised. The civil servant I talked to had never heard of this as a formal rule. I did, though, find this over at Language Log: "When I started work in the Foreign Office some twenty-five years ago, the guidance was that the British government (Her Majesty's Government or more usually HMG) was always singular, but other countries' governments took the plural (perhaps because they were innately fissiparous, being composed of foreigners)."

  5. The Scifi channel is now called Syfy... "Syfy" is the plural of Polish "syf" meaning "grunge".

  6. Paczki is singular (and, I think, plural) in English, but it comes from pączki, which in Polish is the plural of pączek.

  7. Macbeth may have gone into some petty places in his day, but that's not what he was talking about in light of his wife's death and his imminent defeat. What he said was, "Tomorrow and tomorrow and tomorrow creeps in this petty pace..."

  8. I don't know if anyone is likely to read a comment on an older post, but I thought that agenda was not a plural, but a gerund (or gerundive, my Latin being somewhat rusty) like Amanda i.e. to be done, as Amanda is to be loved.
    Also, I hang out with a lot of scientists, and in scientific literature data as plural is still quite common as far as I can tell, though outnumbered by the singular (and datum occurs almost never, replaced by 'data point')

  9. Plenty of outrage still - and not only British! (Letter to The Economist 4/8/2012)
    Perfect plurals, please

    SIR – I recently read a headline in the New York Times in which the word data was used as a singular. I wrote to the newspaper about it, and in my letter I said that I also subscribe to The Economist and that it is extremely rare to see such a usage in that publication. Then, lo and behold, in your July 21st edition, I found at least two such usages—in the same article no less—that I found jarring (“Little peepers everywhere”).

    There are many good reasons to read The Economist. For me, one of them is the beauty and often near musicality of the language. Can you please post, clearly, a sign that data (and media) are plurals, for those of your writers and editors who are unaware.

    Ken Fradin
    Dedham, Massachusetts

  10. I can make a simpler argument for treating "data" as a singular: it's a collective noun, like "water" or "grain". It doesn't matter that "data" holds a form in line with a Latin plural. It's a cognate and so English rules are applicable, not Latin ones. Once imported into the language, a word is subject to that language's conventions, not its source language's.