That which is restrictive

This is quite a long post about a distinction some people make between that and which as relative pronouns — an oft-disputed point of English usage. Feel free to skip ahead if you’re familiar with the territory.

Restrictive clauses (aka defining or integrated relative clauses) provide information that’s essential to a sentence. Take this one:

The bike that I keep in the garage is ideal for short trips.

The underlined clause is integral to the sentence, for reasons context would normally make clear. For example, there may be an implication that I have access to other bikes, so the restrictive clause defines or restricts what bike I’m talking about.

Non-restrictive clauses (also non-defining or supplementary relative clauses) are bound less tightly to the sentence: they can be removed without changing its essential point. Thus:

The bike, which I keep in the garage, is ideal for short trips.

Here, there’s only one bike I could be referring to, and the information about where I keep it is supplementary, non-defining, dispensable.

In speech, non-restrictive clauses are intoned separately; in writing, this separation is marked by punctuation: normally commas, as above, sometimes dashes or parentheses.

There’s a good case for calling non-restrictive clauses supplementary relative clauses, and restrictive ones integrated relative clauses. But these terms are quite new, and in this post I use the more familiar names.

So far so uncontroversial. Then there are sentences like this:

The bike which I keep in the garage is ideal for short trips.

And shazam!, opinion is divided. To me, this sentence is essentially equivalent to the first example — the one with that — though if I spoke it I’d probably use that or neither. (What is a non-standard alternative.) But some would call this which ambiguous and improper, saying it should not be used with a restrictive clause.

The issue cropped up in a recent post about whom. A reader objected to restrictive which; I said restrictive that was not a grammatical requirement — it never has been — and I quoted the Chicago Manual of Style (“which can be substituted for that in a restrictive clause”) and other sound authorities in support of this position. See John E. McIntyre for a clear and common-sense summary.

Restrictive which is avoided in AmE more than in BrE. Garner’s Modern American Usage (1998) describes two schools of thought, neither of which I fit into. He says British writers “have utterly bollixed the distinction between restrictive and nonrestrictive relative pronouns”. But it’s not about distinguishing between pronouns: it’s about different types of clause; and commas, used correctly, signal that distinction.

* * *

Yesterday, The Guardian’s Mind your language blog firmly advocated the that/which pseudo-rule. It began with a reader’s niggle and ended by claiming that the line “All molecules which are drugs bind to receptors” fails to be “clear, unambiguous and factually correct”. But it looks fine to me. I responded with a couple of comments, which I’ll cannibalize here rather than rewrite. First:

I find nothing unclear about “All molecules which are drugs bind to receptors.” So long as the sentence is punctuated correctly (i.e., without commas), its meaning is unambiguous. No one on earth thinks all molecules are drugs.

Which can be used with restrictive clauses. Jane Austen did so, as did Dickens, Melville, Stoker, and countless good writers. It’s in the King James Bible [“Render unto Caesar the things which are Caesar’s, and unto God the things that are God’s”]. Non-restrictive that appears in Macbeth. Arnold Zwicky has described some of the problems with this “rule”, and there’s a useful history in MWDEU.

The continued popularity of which-hunting probably owes at least in part to the dubious influence of Strunk and White, but not even White himself held to this unhelpful proscription. From Death of a Pig: “…no one took the event lightly and the premature expiration of a pig is, I soon discovered, a departure which the community marks solemnly on its calendar”

In a comment, the blog author David Marsh said that he didn’t use the word rule; that he accepted Austen, Shakespeare and others “broke many of the so-called grammatical rules”; and that it seemed irrational not to follow a “‘useful distinction’ that aids clear and elegant writing”. My second comment, at #52, addresses these points:

It isn’t a question of Shakespeare, Austen et al. breaking so-called rules. Those so-called rules were not there to be broken. The “That Rule” was invented and disseminated later, by prescriptivists who either didn’t know or didn’t care that restrictive which had graced impeccable standard English for centuries. It still does. . . .

Geoffrey Pullum, co-author of the Cambridge Grammar of the English Language, says the restriction “has no basis”. He has collected some Language Log posts on the subject here; this one is particularly thorough.

Both that and which are commonly and correctly used as restrictive relativizers, though there are subtle differences in how people use them. David, I enjoy your articles (and your tweets), but I don’t think the arbitrary, awkward, and anachronistic “That Rule” does credit to your style guide. You might not have used the word rule, but you might as well have done, given the severity of your last paragraph.

If the distinction “aids clear and elegant writing”, why do great writers persist in ignoring it?

Calling it “arbitrary, awkward, and anachronistic” might have been a bit harsh. (I’m a slave to alliteration.) But the concluding question stands.

There’s also the problem of fronting prepositions: I can’t say *”This is something of that I’ve written before.” I use which, or I strand the preposition — “something I’ve written of” — not that there’s anything wrong with that.

* * *

The Fowler brothers, though they did not invent the restriction, popularized it in The King’s English. H. W. Fowler returned to it in his influential usage dictionary, seeking “clearer differentiation” between that and which and arguing his case at eloquent and extravagant length. Yet he acknowledged that “it would be idle to pretend that it is the practice either of most or of the best writers”.

Consider now the following line from Simple and Direct by Jacques Barzun:

Next is a typical situation which a practiced writer corrects “for style” virtually by reflex action.

What’s funny and instructive is that on the previous page, Barzun had recommended “using that with defining clauses except when stylistic reasons interpose”. This example of immediately self-contradictory which-hunting appears in Joseph M. Williams’s excellent essay The Phenomenology of Error. Williams continues:

after Barzun stated the rule, and almost immediately violated it, no one noticed — not Barzun himself who must certainly have read the manuscript several times, not a colleague to whom he probably gave the manuscript before he sent it to the publisher, not the copy editor who worked over the manuscript, not the proof reader who read the galleys, not Barzun who probably read the galleys after them, apparently not even any-one in the reading public, since that which hasn’t been corrected in any of the subsequent printings.

Another possibility is that people noticed but left it alone. Some years ago, I began collecting literary examples of restrictive which for a blog post showing how common and standard it is — much as I’ve done with comma splices. I soon gave up because the usage is ubiquitous: it would be like collecting examples of semicolons.

A sensible entry in The Columbia Guide to Standard American English says:

most of us use which almost interchangeably with that in restrictive modifiers and rarely but sometimes use that to introduce nonrestrictive modifiers. . . . Best advice: use that or which or nothing, depending on what your ear tells you. Then, when writing for certain publications, know that you may have to replace a good many whiches with thats, and perhaps a that or two with a which, to conform to the “rule” almost no one follows perfectly in other than Edited English and few can follow perfectly even there.

The almost is significant: even those of us who use that and which to some degree interchangeably tend to prefer one or the other in certain ways or with certain kinds of antecedent. There are contextual subtleties, and they have been explored in some depth at Language Log and elsewhere.

Another caveat: heeding your ear is not always a good guide in these cases, because we can habituate ourselves to hear something as “wrong” when it clearly is not. I know I have. Conversely, if we learn that a usage is grammatically and stylistically kosher, our decision to accept it — perhaps despite an instinctual, long-cherished peeve — gradually improves how it sounds to our ears.

And lo: we have more freedom than we had before, and at no significant cost I’m aware of.

If your dialect is an American English one, there’s a fair chance you abide by the “rule”. You’re fully entitled to, just as style guides are entitled to stipulate it for reasons of simplicity or supposed clarity. (Some don’t: The Economist‘s says “Americans tend to be fussy” about making the distinction, but that “good writers of British English are less fastidious”. Hat tip to Lane Greene for that.)

But I don’t think it’s useful or beneficial to outlaw which from restrictive clauses, and I like having a choice of relative pronouns. Let punctuation do the work of clarifying. Used with skill, it does it well.

Postscript: My comments at The Guardian helped convert at least one editor. This morning, I received confirmation of a second. One more, and I’ll call it a trend.

.

Updates:

A follow-up post, “A comma, which muddles meaning“, quotes a Guardian editorial with a comma muddle that shows how obeying the rule can undermine communication. The Guardian style guide editors conceded that I may be right (i.e., that they may be wrong). There’s hope yet.

The funny thing about 'that' and 'which' is that the pseudo-rule causes precisely the confusion its defenders think it will resolve.

— Stan Carey (@StanCarey) February 10, 2012

Arnold Zwicky, following up on this post, laments what may be “a hopeless battle”. He supplies an example of the pseudo-rule being taught to school kids as fact and seemingly being “mangled in transmission”.

He has also put together a most convenient collection of posts on which vs. that.

Robert Lane Greene, in a comment on Google+, advises that you “delete this ‘rule’ from your memory and free up some space for something useful.”

Arrant Pedantry explores in greater detail the distinctions between the relative pronouns, and finds that “we do have rules—just not the ones that are proposed and promoted”.

Tom Freeman wonders whether in abandoning the That Rule we would lose a useful distinction. After examining the evidence, he finds there are “any number of ways to rephrase for clarity”, and honourably recants.

Robert Coren, in a comment on Arnold Zwicky’s Google+ page, tells an illuminating tale: “Several decades ago, my workplace had a tech writer who was a serious which-hunter, and she had so terrorized at least of [sic] the developers that he produced a draft that avoided “which” entirely, using “that” even in non-restrictive clauses, resulting in pretty much unreadable, or at least incomprehensible, prose.”

Copyediting.com kindly includes this post in its October 21 news roundup and, interestingly given what I wrote in this comment, italicises which in the title.

Geoffrey Pullum, at Language Log, says that far from helping American users of English, the pseudo-rule “ruins their lives”. He presents and analyses an example that shows how insane one’s prose can become because of confusion over the rule, and describes the Fowler brothers’ proposal as “unmotivated, unimplementable, and stupid”.

Peter at 9 Months with the Chicago Manual of Style agrees with me, noting that the that/which distinction “was never a rule to begin with”, and elaborating on the CMOS‘s “lukewarm” endorsement of our right to choose.

I recently got a copy of the new fifth edition of the American Heritage Dictionary; here are a few lines from its usage note on the matter:

this use of which with restrictive clauses is very common, even in edited prose. Moreover, in some situations which is preferable to that. Which can be especially useful where two or more relative clauses are joined by and or or: It is a philosophy in which ordinary people may find solace and which many have found reason to praise. Which may also be preferable when introducing a restrictive clause modifying a preceding phrase that contains that: We want to assign only that material which will be most helpful.

Arrant Pedantry, in a post examining what corpus data reveal, finds that the distinction is not useful and that “its usefulness is taken to be self-evident, but the evidence of its usefulness is less than compelling”.

In a subsequent post, Arrant Pedantry unpacks the strange behaviour of relative clauses, especially that, and justifiably describes it as “far more complex than most editors or pop grammarians realize”.

Peter Harvey neatly summarises the commas-and-relative-clauses situation and describes the pseudorule as “nonsense”.

Geoffrey Pullum, at Language Log, calls the rule “a time-wasting early-20th-century fetish, a bogeyman rule undeserving of the attention of intelligent grownups”. And he blames me for infecting him with “false optimism about changing people’s minds”. [*evil cackle*]

Random Idea English offers a useful overview of the usages and says the rule “has nothing to do with grammar”. A later post on which-hunting points out that “the use (or not) of commas is the main distinguishing feature (apart from context) between defining and non-defining clauses in writing, rather than the use of that or which“.

Mark Liberman at Language Log laments a link at Reddit perpetuating the “phony” rule, and provides a brisk summary of the history of the contentious usage.

Revisiting the topic with a terrific post at Arrant Pedantry, Jonathon Owen shows how the that/which rule, “rather than regularizing the language and making it a little more consistent, actually introduces a rather significant irregularity and inconsistency.”

Advertisement