A Rule Which Will Live in Infamy

“Yesterday, December 7, 1941—a date which will live in infamy—the United States of America was suddenly and deliberately attacked by naval and air forces of the Empire of Japan.” That was how President Franklin D. Roosevelt opened his famous infamy speech, 71 years ago. Ignoring the writing handbooks, he opened with a passive construction, which of course is just right for the rhetorical context (America as innocent victim). And he also ignored another bogeyman rule: He introduced a restrictive relative clause with which.

The false belief that restrictive which is an error stems from a quixotic reform effort of the early 20th century. That attempt to change English failed, but American teachers and editors took note of it, and misinterpreted it as legislation, blithely ignoring the evidence of FDR’s sentence and thousands of others in all kinds of literature.

Grammarians in the 19th century had been worrying for some time about the options for opening a relative clause: who or that or nothing for the ones modifying human-denoting nouns; which or that or nothing for the rest; so much dangerous freedom of choice. In a book called The King’s English (1906), Henry W. Fowler and his brother Francis proposed to reduce that freedom by altering the language. Since that was only rarely used to introduce nonrestrictive relative clauses (the parenthetical kind, with the commas), the Fowlers decreed that all nonrestrictives with that were mistakes, and although which very commonly introduced restrictive relative clauses (the kind without the commas), they opined that English would be neater if this didn’t happen.

Any rule disallowing restrictive which would have to have an ugly clutch of exceptions. Three examples:

  • The putative ban can’t apply when a preposition precedes the relative pronoun: the town in which she lived is grammatical but *the town in that she lived isn’t.
  • The supposed rule should be ignored when modifying demonstrative that, because that which you prefer is clearly preferable to ?that that you prefer.
  • The rule can’t apply to a conjoined which: We must trust the unknown entity who or which created us is grammatical but *We must trust the unknown entity who or that created us isn’t.

The Fowlers noticed the first two of these, and cited some others. They acknowledged that the reader might think “a rule with so many exceptions to it is not worth observing.” They also admitted that “it is not easy to draw any distinction that is at all consistently supported by usage”: Literature testified against them. The Bible, for example. “Render therefore unto Caesar the things which are Caesar’s,” said Jesus according to Matthew 22:21. Did King James’s team of learned translators not know English grammar? Or is it that it was OK to put restrictive which in Jesus’ mouth in 1611, but today we know better?

(I give a little more evidence of just how grossly the invented rule fails to accord with literary usage in this post.)

William Strunk knew nothing of the Fowlers’ reform proposal when he wrote the first version of The Elements of Style more than a decade after the Fowlers’ book appeared. Strunk wrote that we should keep related words together and “keep apart those which are not so related.” But as Jan Freeman noticed, when E.B. White revised Strunk’s little ragbag of ukases in 1959, he added the Fowler rule, and then altered Strunk’s original prose, changing which to that throughout, to conceal the facts of his mentor’s actual usage! He also ignored the rule in his own prose: a departure which the community marks solemnly on its calendar was perfectly grammatical for White, and it occurs in “The Death of a Pig”.

Never believe that the Fowler rule prevents ambiguity. Punctuation suffices to distinguish restrictive from nonrestrictive relatives. In “E. B. White, who covertly altered Strunk’s text,” the red part is a nonrestrictive relative; in “the pedant who bowdlerized Strunk’s text” the underlined part is a restrictive one. The commas tell all. Ambiguity doesn’t afflict relatives with which any more than those beginning with who or in which.

Rather than acting as a bulwark against ambiguity, the rule is used mainly for nitpicking and sniping. Ann Coulter dinged Supreme Court nominee Harriet Miers for a “grammatically incorrect ‘which’ instead of ‘that’” and called it “the sort of error that results from trying to sound ‘Ivy League’ rather than being clear” (see Mark Liberman’s dissection on Language Log). Noreen Malone of New York Magazine got English literature assistant professor Matthew Hart to state that Barack Obama had a B-minus intellect on the grounds of a letter that he wrote to his girlfriend when he was an undergraduate (!), citing the phrase “an ecstatic vision which runs from Münzer to Yeats” for “confusing that and which” (again, see Mark Liberman’s analysis).

Grammar snobs trying to show off their linguistic rectitude by playing gotcha with an invented rule that never matched educated usage; copy editors slaving away trying to enforce it; Microsoft Word blindly putting wavy green underlining under every relative which not preceded by a comma. What a senseless waste of time and energy.

Follow the Fowler rule if you want to; it’s up to you. But don’t tell me that it’s crucial or that the best writers respect it. It’s a time-wasting early-20th-century fetish, a bogeyman rule undeserving of the attention of intelligent grownups.

Return to Top