The irregular restrictive what ?

John McPhee’s “Draft No. 4” (in the April 29 New Yorker) has a passage that puzzled me, about “the irregular restrictive which ” (italics mine), a term that McPhee learned from the New Yorker editor William Shawn, who “explained that under certain unusual and special circumstances the word ‘which’ could be employed at the head of a restrictive clause.” In other words, which can sometimes substitute for that .

Which in fact often substitutes for that ; there is no absolute rule that that divides their use. Many writers, however, prefer to reserve that for restrictive clauses, which for nonrestrictive. The distinction brings a small (very small) element of consistency to writing:

There are three magazines on the table. I want to read the magazine that arrived today.

[That introduces a restrictive clause: “that arrived today” identifies magazine .]

There are two pieces of junk mail and a magazine on the table. I want to read the magazine, which arrived today.

[Which introduces a nonrestrictive clause: “which arrived today” is not needed to identify magazine .]

McPhee explains that “the irregular restrictive which ” is reserved for sentences in which “words or phrases lie between the specific object and the clause that proves its specificity.” The term “irregular restrictive which ” seems to be a Shawn (or Harold Ross?) creation: the only evidence for it that I can find is McPhee’s essay.¹ This use of which , however, goes back at least as far as H. W. Fowler’s Modern English Usage (1926), where it is explained in the entry “That, relative pronoun.” Fowler’s defining and non-defining are the equivalents of restrictive and nonrestrictive :

Each that -clause is, or at least may be meant as, defining; but between each & the actual noun of the antecedent . . . intervenes a clause or phrase that would suffice by itself for identification. In such circumstances a that-clause, though correct, is often felt to be queer, & it is usually possible, though by no means necessary, to regarded it as non-defining & change that to which .

McPhee gives three examples from his recent writing of “the irregular restrictive which .” Here is one:

In 1822, the Belgian stratigrapher J. J. d’Omalius d’Halloy, working for the French government, put a name on the chalk of Europe which would come to represent an ungainly share of geologic time.

Try it the other way:

In 1822, the Belgian stratigrapher J. J. d’Omalius d’Halloy, working for the French government, put a name on the chalk of Europe that would come to represent an ungainly share of geologic time.

I find it hard to see any difference: name seems the obvious antedecent each time. McPhee’s other examples leave me just as confused.

But I found a way out of my muddle by consulting Garner’s Modern American Usage (2009). In the entry "Remote relatives," Bryan Garner addresses “the exceptional which”, his name for what Shawn called “the irregular restrictive which.” Garner presents this use of which as an attempt to avoid ambiguity:

Garner’s final four sentences are a model of clear reasoning about usage. Now I know that I need not spend another second thinking about “the irregular restrictive which.” Clarity!

*

11:21 a.m.: Not done yet. The New Yorker ’s Eleanor Gould credited William Shawn:

[Quoted in Barbara Wallraff’s Word Court (2000). The sentence in question referred to “a dispute about language which they would like this column to resolve.”]

¹ Harold Ross, who founded the New Yorker, was a Modern English Usage devotee. From a 1949 letter to Kay Boyle:

We think ourselves into knots over style things around here, although we’ve long since cracked most problems. We’re having one now on when to use which and when to use that that is a little gem. Fowler, in Modern English Usage, differentiates between them and, somehow or other — I don’t know how, so help me — we got to following him in the editing of all house (or unsigned) stuff and then in practically all fact stuff (the writers are around the office and can be talked to from hour to hour), and then in more or less all the fiction, most of the writers falling into line.

Notice that Ross doesn’t use “the irregular restrictive which ”:

We’re having one now on when to use which and when to use that that is a little gem.

The logic of “the irregular restrictive which ” would have the sentence read:

We’re having one now on when to use which and when to use that which is a little gem.

I found this letter quoted in John Updike’s More Matter: Essays and Criticism (2009).

[I wonder: does “exceptionally well-edited” mean persnickety ?]