by

A Certain Closeness

Do you see any grammatical mistake in the sentence “He had developed a closeness to his recent suffering”? A classics teacher married to an author wrote to me a while ago to ask me this:

I am doing some editing on my wife’s new book (really, it’s just an excuse for me to get to read it a few times!), and she has a fairly consistent usage that Word (and the Internet) find to be completely unacceptable.

His wife was using phrases like a closeness, and Word was reporting that the first of those words should be dropped.

“Not one single example sounds better without the indefinite article,” my correspondent wrote. “Can you shed light on this? Are they just examples of noncount nouns shifting roles and acting as count nouns, or is there something else going on here?”

Word is clearly trying to implement the usual story about count nouns and noncount (or “mass”) nouns: that count nouns like chair take determiners such as a(n), one, and several, while noncount nouns like furniture don’t. You can talk about a chair or several chairs, whereas phrases like *a furniture or *several furnitures are ungrammatical.

And Word is trying (inconsistently, as we shall see) to impose a general rule that a noun can take the indefinite article if and only if it is a count noun.

That is not a correct rule for English. Almost every noncount noun allows at least some count-noun behavior. Beer is noncount, but you can say I’ll have a beer, (coercing beer into meaning “bottle or glass of beer”). Cement is noncount, but you can say Let’s try a different cement, forcing cement to be understood as “type of cement.”

So, is closeness grammatically like cement or like furniture? That question is empirical: To find the answer we should look at our practice as competent users of English, rather than assume that the grammar hackers at Microsoft have codified our practices accurately and can be taken as the relevant authority.

What I’m saying deserves boldface: It is vastly more likely that you know how to use your native language than that Microsoft programmers have achieved fully accurate automated grammar analysis and advice.

Computational linguists regularly test systems on a corpus of about 44 million words from 1987–89 issues of The Wall Street Journal which Mark Liberman persuaded the Journal to make available to the Association for Computational Linguistics about 25 years ago. There are occasional typographical errors in the files, yes, but for many purposes the WSJ corpus is as good a place to look for evidence about English as anywhere else.

A few seconds of computerized searching in WSJ yielded these examples (the first of which has the second occurrence of closeness in the corpus):

  1. I believe there is an increasing closeness among IOC members, as opposed to the political barriers that have existed in the past.
  2. We have garbage, I have to admit, but there’s a closeness.
  3. When you work with someone in that staff-member relationship, you develop a closeness that continues off of the Hill.

Any English speaker can see that these are not typos. There is nothing wrong with letting closeness take the indefinite article. Yet Word flags one of these examples, the second, as a possible error–despite treating the other two as fine. It is clearly inconsistent.

Word also (as my correspondent noted) identifies He had developed a closeness to his recent suffering as a possible grammar error, recommending replacement by He had developed closeness to his recent suffering, which has a different meaning and sounds wrong.

Worse still, when I tried typing They had a certain closeness, Word fingered that as an error, and recommended changing it to *They had certain closeness, which is definitely ungrammatical.

Finally there is the truly mysterious fact that Word identifies a frenzy as erroneous use of English. Frenzy is clearly and obviously a count noun. Dictionaries say so, and show the plural as frenzies. Yet Word impugns a frenzy in the same way as genuinely ungrammatical phrases like *a crockery and *a furniture.

The lesson for my correspondent is simple. Taking advice on usage from Word is like taking advice on investments from Bernie Madoff. The grammar-checking tool is a chaotic, unreliable, inconsistent, brain-dead piece of junkware. It can’t tell what’s grammatical and what isn’t, yet still it presumes to query every passive or split infinitive, and advise you falsely on when to use an indefinite article and hundreds of other points. Go to the Preferences pane and switch it off. Read your drafts carefully and revise assiduously, letting your own intuition be your guide. Don’t even think of getting your word processor to do it for you.

Return to Top