by

Grammatical Shades of Grey

shutterstock_138339137I responded cautiously when my nonlinguist partner, staring at the wording on a supermarket yogurt container, cried “That’s wrong!”

The words on the label promised that the yogurt was made of cow’s milk. “That’s wrong!” she said; “This yogurt didn’t come from just one cow!” It ought to be spelled cows’ milk, with the genitive plural, she insisted.

We happened to have a carton of a different brand of yogurt in the refrigerator, so we could double-check. Sure enough, down in the small-print list of ingredients on the second carton it said cows’ milk.

“There you are!” she said triumphantly, as if that had settled the issue.

I conceded that indeed the milk probably did not come from a single cow, and left it at that, holding in check my desire to protest that the contents of the rules of English grammar cannot be read off either informal logical intuitions or snippets of raw data. I didn’t want to act the stereotypical academic with his inevitable arcane reasons why things are actually much more complicated. I always hate being told that things are actually much more complicated (though I know they usually are); it only seems to underline my subordination to someone else’s expertise while denying me the clear-cut answer I sought.

Later, however, I did a little quiet empirical investigation. Sure enough, things are more complex. It’s not just that one yogurt manufacturer has the wording right and another has it wrong.

You can’t tackle a problem like this by simply going to the places on the web where people have asked “Which is correct, cow’s milk or cows’ milk?” The answers given there tend to be off-the-cuff assertions by soi-disant grammar know-it-alls, and don’t settle anything.

Yet settling a question by reference to actual usage calls for huge amounts of data. The 44 million words of the computational linguists’ favorite Wall Street Journal corpus is nowhere near enough (there is one article with three or four occurrences of cow’s milk, and no occurrences of cows’ milk, which establishes that the singular version has occurred in newspaper text, but not how often, or whether it was a slip).

The trillion words on the web are not that easy to use: Google’s hit counts are estimates that can be off by orders of magnitude, and you only get page counts (not item counts). And worst of all, Google plays fast and loose with apostrophes: searching for X’s yields results for Xs’ as well, and vice versa. In fact searching for woman’s magazine elicits a suggestion that you might want women’s magazine instead.

More useful figures can be obtained with the Google Ngram Viewer, which allows searching a gigantic list of letter strings found in millions of books (and treats apostrophes with respect). Results obtained there do not support my partner: cow's milk has always been significantly more frequent than cows' milk throughout the 19th and 20th centuries.

Of course, that still doesn’t tell us which (if either) is correct: We can’t conclude that an expression is an error simply because it has a frequently encountered rival alternative.

Don’t worry, I am not going to leave you guessing. I will tell you what I think the right answer is. But—as the giant computer in The Hitchhiker’s Guide to the Galaxy said before revealing that millions of years of computation concerning the answer to the ultimate question of life, the universe, and everything had led to the result “42”—you’re not going to like it.

From various kinds of grammatical evidence, I conclude that both phrases are grammatical. That is, either a genitive singular or a genitive plural noun can be used as a prenominal modifier.

Moreover, contrary to my partner’s intuition, it is clear that cow’s milk doesn’t entail that the milk is from only one cow. The name of Professional Woman’s Magazine doesn’t entail that it is for only one professional woman.

You probably won’t welcome this information. Everybody seems to want the answers to grammar dilemmas to be resolved in black and white, not shades of grey. They want a single correct form for each meaning, not choices for users to make according to taste. But I see no more warrant for affirming that Professional Woman’s Magazine should have been called Professional Women’s Magazine than for insisting that Look the number up is wrong and should be replaced by Look up the number. These things just aren’t true.

Still, if you are anything like most of the public you will feel cheated: I haven’t sternly ruled that that one is the right form and the other is wrong. You want discipline. So all right, just this once, I’ll play Christian Grey to your Ana Steele, and satisfy your repressed desire to be grammatically dominated.

You know the form that you normally use? Well, it’s the other one that’s grammatically correct. So shape up. Or prepare to feel the kiss of the lash.

Return to Top