You Say Day-ta, I Say Daa-ta

YOU SAY DAY-TA, I SAY DAA-TA….Enough about healthcare, let’s take on a real issue: is the word “data” plural or singular? Ditto for “media.”

Right-thinking people already know that the answer to this is “singluar,” but one of my readers takes exception to this. “It’s hard to imagine any reasons that I would find remotely plausible,” he says, “but it might be interesting to see you take a shot at it.”

Well, my first shot at this came when I wrote my original blog back in 1997 ? which, by the way, probably makes me one of the oldest bloggers around, even if my hiatus from 1998-2002 was a little longer than is usually acceptable ? and it went like this:

David wanted to know today if I thought it was OK to use the word mediums (as in, “TV and radio are two mediums worth advertising on”). My opinion: the word media was long ago transformed into a singular noun in common usage, so it is no longer very useful as a plural of medium. Therefore, mediums is acceptable usage when you are talking about two or more specific cases, whereas media commonly refers to the whole enchilada (as in “media conspiracy”). Just another case of language evolving.

This is the standard descriptivist argument: language evolves, and rules should evolve to take account of common usage. “Media” and “data” are now singular simply because most people use them that way.

But there’s more: I think “data” is legitimately singular anyway because it belongs in the class of collective nouns, words like “hair” that obviously refer to a collection of individual things but that nonetheless take the singular. It’s like the old joke: “I’m going to get my hair cut today.” “Oh yeah, which one?”

But just as the singular “hair” is usually taken to refer to a full head of individual hairs, I think that “data” is most often used to refer to a mass of information that, like hair, is an identifiable single thing in its own right, with its own properties and its own attributes. “The data is inconclusive” indicates an opinion about the entire set of data, not of any individual datum.

So that’s it. Is everyone convinced?

POSTSCRIPT: Those of you who don’t buy this will be even more outraged by an obvious corollary, which I touched on above: if you are truly referring to several specific pieces of data, it’s all right to use the word “datums.” Ouch! Maybe “pieces of data” is better after all.