It’s been a busy few weeks with research and the end of the semester. I promise I’ll get to writing substantive posts very soon. In the interim, I’d like to express my incredulity at John McPhee’s ignorance of the “irregular restrictive which” until it was pointed out to him by New Yorker editor William Shawn:
Mr. Shawn explained that under certain unusual and special circumstances the word “which” could be employed at the head of a restrictive clause. Ordinarily, the conjunction “that” would introduce a restrictive clause. Nonrestrictive: This is a baseball, which is spherical and white. Restrictive: This is the baseball that Babe Ruth hit out of the park after pointing at the fence in Chicago. The first ball is unspecific, and that sentence requires a comma if the writer wishes to digress into its shape and color. The second ball is very specific, and the sentence repels commas. There can be situations, though, wherein words or phrases lie between the specific object and the clause that proves its specificity, and would call for the irregular restrictive which.1
Perhaps I care more about restrictive and non-restrictive clauses than most do. After all, I’ve done research on the word that, and in the description for the freshman writing course I’m teaching this fall (more on that
soon here!), I spent a good 10% of my allocated words promising to show how set theory can tell you whether to put a comma before which.
But this should be common knowledge. There are three possibilities:
(1) This is the ball, which is red.
(2) This is the ball which is red.
(3) This is the ball that is red.
Sentence (1) has a non-restrictive clause. In this case, the clause works like a (non-restrictive) appositive: it adds more information about the already specified ball. Sentences (2) and (3) mean the same thing. Both have restrictive clauses, which means that they restrict from the set of all balls to the set of balls that are red. It’s a matter of Venn Diagrams. A non-restrictive clause doesn’t shrink the set. The restrictive clause does, so you have the set of balls that are red sitting inside of the set of balls.2
Many students struggle with when a clause should be non-restrictive—i.e., written with a comma and a which as in (1)—and when it should be restrictive—i.e., written without a comma using either a which or a that, as in (2) or (3). But it’s perfectly acceptable to use (2) and (3) interchangeably. Therefore, in particular, the question “should I use a which or a that?” misses the point: the relevant question is “should I use a restrictive or a non-restrictive clause?”.
It’s possible that the irregular non-restrictive which is historically recent. After all, linguistic norms change; punctuation a century ago was far different.3 But today it’s certainly well-established that the non-restrictive which is far from irregular. [Update, 6/19/13: See this post by Geoffrey Pullum for some historical background.]
McPhee goes on to discuss the New Yorker house style, albeit with an unfortunate attitude of naive and ignorant reverence of the grammarian copy editors. At some point, probably in the context of my freshman writing course, I’ll discuss further my views on grammar, prescriptivism, etc.4 For now, though, let me focus my remarks on the importance of the New Yorker’s house style.
Back in my student journalist days,5 I learned the Orwell quote “Good prose is like a windowpane.” You’re not supposed to notice the form; the point is the content.6 When I read the New Yorker, the window is clean, I can focus on the great content. (Indeed, I imagine the New Yorker’s stringent copy-editing played a role in helping to codify the grammatical conventions of educated writing today, just as dictionaries codified the scrambled orthography of a few hundred years ago.) Or at least this almost true: the sole exception that comes to mind is McPhee, whose style I often find impenetrable, his sentences unusual. Perhaps it has to do with his attitude towards grammar.7
In short: understanding the patterns of language matters, and it’s important that it be done in a sophisticated, non-dogmatic way. Our educational system has failed us if educated readers aren’t aware of these distinctions. The course I’m going to teach is going to be all about this. Stay tuned!
Let me end with a few stray remarks:
- There’s actually a fourth possibility, in addition to sentences (1)-(3), if the relative pronoun serves the role of an object instead of a subject, as in “This is the ball that I hit.” In that sentence, one can also omit the that: “This is the ball I hit.” This is the so-called optional that, the subject my research on relative clauses.
- There are some subtleties in the restrictiveness of relative clauses. For example, when the antecedent noun has an indirect article (“John is a friend of mine who…”) the relative clause is almost always restrictive. There are some interesting cases, though, where putting a comma before the who sounds more natural. These sorts of things have to do with having an ear for the subtlety of language. But one has to know the basic principles first.
- I was amused to see that McPhee’s example sentence had to do with a ball, since the canonical example I always used for my research on optional that was “This is the ball that I hit”.
- The existence of sentences (2) and (3) of seemingly identical meaning has puzzled some linguists, who have argued for a “principle of no synonymy”: any two sentences must have some subtle meaning difference. (My research about the word that had to do with the apparent synonymy between sentences like “This is the ball (that) I hit” with or without the that.8 My basic response to this is: Yes, there are probably in most cases subtle reasons why someone might choose which versus that, reasons which probably vary by speaker or writer and context. But certainly to a first approximation these sentences should be treated as completely synonymous.
- From the New Yorker, April 29, 2013, p. 35. Available behind paywall here. ↩
- To explain this better, I’d need to delve into the field of semantics, which studies the meaning of sentences. With a bit of explanation about natural language semantics, it really is just Venn diagrams. Alas, the way it’s taught in schools is hopelessly confusing; I couldn’t make head or tail out of what we learned in that rare opportunity for grammar instruction in 7th grade English. I hope to expand on this topic at some point in preparation for my writing course. ↩
- In particular, there are two competing forces governing punctuation: syntax and phonology/prosody. The syntactic view of punctuation puts commas—as well as em-dashes and parentheses—in generally recursive, parenthetic structures whose purpose is to help parse the sentence grammatically into its constituent parts. (There’s also a semantic role, in terms of restrictiveness, as in relative clauses or in noun phrases like “adjective(,) adjective noun”.) The phonological view of punctuation, by contrast, seeks to use punctuation to recreate the complicated prosodic and intonational structures of spoken speech. These two forces often align, but not always. Today, in American English—I’ve noticed British English is a bit different—the syntactic approach tends to win. As someone who enjoys recursive structures, I’m glad about this. But there are cases where the prosodic punctuation can win. For example, if I write “John decided that after this long and arduous journey, it would be best to rest”, I’m omitting the syntactically desired comma before after to follow spoken prosody better. (This is not a great example.) ↩
- In short, naive prescriptivism—foolish rules against split infinitives, ending sentences with prepositions, using passive voice—is idiotic. Of course, hypercorrecting against that prescriptivism by failing to acknowledge the standard conventions of prose, in favor of complete relativism, is no better. The appropriate attitude is a respect for the current conventions, combined with a chagrined acknowledgment that norms change over time, so people older than you are going to be too fussy and people younger than you are going to be too lax. For me, the main things I care about are comma placement, no dangling participles (of course there are many exceptions to that rule in great literature), and certainly no hypercorrected whoms. (Another thing I care a lot is appropriately conjoining different types of grammatical structure. To give an egregious—far worse than real life—example: “I played soccer, tennis and watched basketball” is bad but “I played soccer and tennis, and watched basketball” or “I played soccer, played tennis, and watched soccer” are acceptable.) But I don’t care at all about farther/further, since/because, like/such as, Oxford commas or not, subjunctive, hopefully, split infinitives, passive voice, prepositions at the end of sentences, conjunctions at the beginning of sentences, on the other hands without the initial first hands, etc. (Please don’t start a sentence with however. Although I probably do that from time to time.) And I acknowledge that, horrible as it sounds to my ears, the figurative use of literally might become acceptable in the future. ↩
- At the same Princeton High School that McPhee attended. ↩
This contrasts with poetry. An alternative theory about the difference between poetry and prose comes from the Irish writer Brendan Behan, a self-confessed “drinker with writing problems,” who is reported to have said:
There was a young fella named Rollocks
Who worked for Ferrier Pollocks.
As he walked on the Strand
With his girl by the hand
The tide came up to his knees.
Now that’s prose. If the tide had been in, it would have been poetry.
- To be fair to McPhee, I haven’t read much of his famous earlier stuff, and I do remember enjoying his book on Bill Bradley. ↩
- The omission of such an optional that in a recent quote from James Joyce has been the source of some controversy.) (Update: It was pointed out to me that there are multiple parsings of Joyce’s original sentence, only one of which corresponds to the optional that.) ↩