Which and that are used nearly interchangeably in everyday speech. Yet since which seems more formal, many people conclude it must be better in written material. Why is there so much confusion surrounding the use of these two words?
H.W. Fowler—the granddaddy of authoritative English style, usage, and grammar—explains a crucial part of the muddle in his 1926 classic, A Dictionary of Modern English Usage.
A supposed, & misleading, distinction is that that is the colloquial & which the literary relative. That is a false inference from an actual but misinterpreted fact; it is a fact that the proportion of thats to whichs is far higher in speech than in writing; but the reason is not that the spoken thats are properly converted into written whichs, but that the kind of clause properly begun with which is rare in speech with its short detached sentences, but very common in the more complex & continuous structure of writing, while the kind properly begun with that is equally necessary in both. (“that, rel.,” p. 635)
As it turns out, which and that have very different grammatical functions—at least according to grammarians—and careful writers take care to distinguish between them in their work.
To understand when to use which or that, you’ll need a brief—and, I promise, painless—introduction to relative pronouns and relative clauses. That and which, both relative pronouns, are joined by a third relative pronoun, who, in introducing relative clauses. (That plays a remarkable number of grammatical roles in the English language—conjunction, adjective, adverb, demonstrative pronoun—but when it’s being confused with which, it is generally being used as a relative pronoun.)
Here are a few examples of relative pronouns, bracketed along with the clauses they modify:
The bookstore did not have the book [that she wanted].
The thrift store, [which she stumbled upon one afternoon during a walk], happened to have the book she’d been searching for.
The teacher, [who had suffered through a repressive childhood], seemed to have no sensitivity to the moods of teenagers.
Now, relative clauses can be either restrictive or nonrestrictive. For clarity and as a memory aid, the Associated Press Stylebook calls these “essential” (for restrictive) and “nonessential” (for nonrestrictive) clauses. I think these terms are helpful, so I’ll keep the same terminology here.
Essential clauses, as the term suggests, are essential to the meaning of a sentence. This is where grammarians insist that is appropriate. An example from Mark Twain:
“A classic is something that everybody wants to have read and nobody wants to read.”
In Twain’s quote, the first relative clause is “that everyone wants to have read.” His second relative clause is prefaced by an understood that: “[that] nobody wants to read.” It would be impossible to remove either essential clause from this sentence without changing its meaning.
On the other hand, nonessential clauses add information to the sentence in which they appear but, if deleted, would not change the basic meaning of the sentence. This type of clause calls for which:
The books, which she found at the thrift store, sat waiting to be read on the desk.
The Senate, which agreed to reconvene in a week, adjourned at midnight.
In both of these sentences, the which clauses are nonessential; they provide descriptive information that can be removed without changing the essential meaning or structure of the sentence.
What do grammarians mean by essential and nonessential clauses? Consider Gore Vidal’s tongue in cheek etymology:
“‘Politics’ is made up of two words. ‘Poli,’ which is Greek for ‘many,’ and ‘tics,’ which are bloodsucking insects.”
In this sentence, both whichs are being used correctly to introduce nonessential clauses. But could the two clauses really be deleted without removing information “essential” to the statement? In this example, I think that’s questionable—without the which clauses, the quote would have virtually no meaning. As a result, it might be best to consider the “essential” aspect of a restrictive clause to refer to the grammatical structure of the sentence rather than its actual content (even if content is often used to explain the distinction). If we do remove the two which clauses from Vidal’s quote, it reads:
“‘Politics’ is made up of two words. ‘Poli’ and ‘tics.’”
The grammatical structure of the sentence is still intact, and the deleted information can be added, though perhaps less wittily, in a separate sentence. Therefore, when the sentences are rewritten, it’s easy to see that nonessential, or nonrestrictive, clauses aren’t essential to the grammatical structure of the sentence.
Grammatical explanations aside, there’s one very easy way to remember the difference between which and that: watch your commas. Just remember that a which should always have a comma before it, and a that, almost never.
To put it another way, as a general rule, upon encountering a which in your writing, try first replacing it with that. If the result doesn’t make sense, it will be clear you’re working with a nonessential clause. In that case, keep which and surround the clause it introduces by commas.
Try to apply the rule while reading the following examples.
Wrong: The books which she had bought at the thrift store seemed to be eyeing her from afar, longing to be read.
Right: The books, which she had bought at the thrift store, seemed to be eyeing her from afar, longing to be read.
Right: The books that she had bought at the thrift store seemed to be eyeing her from afar, longing to be read. (Note that this sentence could just as easily, and perhaps more elegantly, be written with an understood that: The books she had bought at the thrift store seemed to be eyeing her from afar, longing to be read.)
There’s a tiny exception to this rule: once in a while, a comma may correctly precede that. For example:
She picked up the books, paperback and frayed, that she’d bought at the thrift store.
In this example, two adjectives intercede between the relative clause (“that she’d bought”) and the word it modifies (“books”), and are rightly set off by commas. So to remember the which-that distinction, pay attention to commas, but don’t get tripped up by the rare that in legitimate need of a comma.
Now that you’ve mastered the easiest way to impress a copyeditor, I’d understand if you’re asking: Is this really worth all the effort? Some writers—for example, business or legal writers, who tend to like gussying up their text—will dismiss the which-that distinction as fussy and unnecessary. But is it?
Part of the trouble with relative pronouns is that British English does not respect distinctions between them in the least. The British sprinkle their speech and written works liberally with whichs and thats, with nary a concern for grammar. Americans take their example to mean which must somehow be more literary or formal, and believe it must dress up their language in admirable ways. Meanwhile, American copyeditors valiantly struggle to hold back the tide of British English invading from across the sea, insisting on grammatical exactitude and consistency.
While they may only partially succeed, in the United States, following the distinction between which and that is the mark of an educated writer. As Bryan Garner somewhat archly remarks in his Modern American Usage (widely considered the most authoritative reference on these matters since Fowler): “The distinction between that and which makes good sense. It enhances clarity. And the best American editors follow it.” (“That. A. And which,” p. 783)
Meanwhile, “elegant variation,” Fowler insists, is not sufficient reason to use one or the other. But it is perhaps the most pressing argument for writers, who—I suppose—may wince at the overuse of the narrow, abrasive-sounding that and wish, just once, to replace it with the much smoother-reading, not to mention seemingly more “literary,” which.
Is there a “right” answer here? Without a doubt, becoming more aware of the many unnecessary thats in your writing, as well as the necessity of translating spoken ambiguities into written clarities, will make you a better, stronger writer.
In addition, this is an area in which the grammarians have steadily won ground over the past century, at least in the United States—to the point that they may as well declare victory. It’s a fact that well-edited publications, whether in print or online, scrupulously observe the distinction, as Garner so emphatically points out. And every style guide on my shelves recommends the restrictive use of that and nonrestrictive use of which. As a result, though I’m not often one to counsel following the crowd, following the crowd in this case can only accrue benefits.
In a final, ironic twist, readers—certainly your most important audience—may not even notice your hard-won efforts to distinguish whichs from thats. Still, they will, even if unconsciously, perceive your prose to be clear, well-written, and easy to follow if your relative clauses are in good order and preceded by the appropriate relative pronouns.
And that may be the best argument for knowing your whichs from your thats of all.