There’s been a small burst of recent postings on which vs. that as relativizers — a topic that seems to never die. Here’s an inventory of postings on the topic and on the distinction between restrictive and non-restrictive relative clauses, on Language Log and this blog, plus a small selection of other postings.

(Postings from Language Log unless otherwise noted.)

GP, 5/17/04: More timewasting garbage, another copy-editing moron (link) [one section]

GP, 9/17/04: Sidney Goldberg on NYT grammar: zero for three (link) [one section]

GP, 9/19/04: Which vs. that? I have numbers (link)

ML, 9/20/04: Which vs. that: a test of faith (link)

ML: 9/23/04: Which vs. that: integration gradation (link)

AZ, 5/3/05: Don’t do this at home, kiddies! (link)

AZ, 5/7/05: The people from the CCGW are here to see you (link) [one section]

AZ, 5/10/05: What I currently know about which and that (link)

AZ, 5/22/05: Five more thoughts on the That Rule (link)

AZ, 5/29/05: Smokin’ too much Fowler (link)

AZ, 7/4/05: That’s American (link)

AZ, 7/10/05: Still more Declaration of Independence (link) [one section]

ML, 10/7/05: Ann Coulter, grammarian (link)

ML, 10/8/05: Grammatical indoctrination at law reviews (link)

AZ, 10/31/06: If they do it too much, they should be told not to do it at all (link)

AZ, 3/4/07: Foolish hobgoblins (link)

AZ, 5/14/07: The unfab four (link)

GP, 3/4/08: For National Grammar Day: Copy editors, we do not hate your guts! (link)

AZ, 3/10/08: To Henry Fowler on the occasion of his 150th birthday (link)

AZ, 5/22/09: What’s wrong with this passage? (link)

GP, 8/24/09: Walking into a buzzsaw (link)

GP, 8/27/09: Where evidence counts for nothing and nobody will listen (link)

GP, 4/21/10: An HR bureaucrat, whom cannot write (link)

GP, 5/20/10: One comma too many (link)

GP, 5/23/10: Oddly enough, McArdle did not err (link)

GP, 7/3/10: That which doesn’t apply to English (link)

AZBlog, 10/19/11: that which won’t die (link): with a link to Stan Carey’s blog, and from there to John McIntyre  and Arrant Pedantry

GP, 10/22/11: Check all boxes (link)

5 Responses to “which/that”

  1. Joe Linker Says:

    Francis Christensen wrote a useful chapter in his original Notes Toward a New Rhetoric, “Restrictive and Nonrestrictive Modifiers Again,” now updated: http://assets.booklocker.com/pdfs/3213s.pdf

  2. arnold zwicky Says:

    The link doesn’t include the section on restrictives and nonrestrictives, so I can’t comment on Christensen’s treatment until I can get hold of the actual book (though I note that almost all usage manuals treat the question, so that any particular one is worth noting only if it has something novel to say — beyond, say, the coverage in MWDEU).

  3. That which is restrictive « Sentence first Says:

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

  4. Alex Segal Says:

    I wrote an essay on this topic: “That and Which: Muddle or Complexity?” The Vocabula Review, October 2010. To give you a flavour of my essay, here are a few paragraphs from the middle of the essay:

    Those who reject the notion that the two pronouns should differentiate two types of clause usually appeal to the disambiguating force of punctuation. Writing in 2004, Geoffrey Pullum contends that using which to introduce restrictive relative clauses does not “contribute ambiguity to properly punctuated prose.” Even some advocates of Fowler’s proposal hold that punctuation is sufficient. Judith Butcher argues that, strictly, “‘that’ should be used for defining clauses and ‘which’ for non-defining,” but also that the “punctuation distinction is the crucial one” (164).

    Sometimes the appeal to punctuation proposes that merely to follow current punctuation conventions involves always distinguishing two kinds of clause. Butcher writes: “Defining clauses have no punctuation, while non-defining must be between commas” (164). Yet current conventions do not mandate punctuation in front of all nondefining clauses, not even if the clauses are also supplementary:

    Gudrun was drawing upon a board which she held on her knee.

    In the greenhouse with the woman was a deaf elk which had got in somehow through a hole in the screen.

    A more plausible appeal to punctuation informs Wendalyn Nichols’s rejection of what she sees as Fowler’s too radical proposal:

    Why not simply teach and enforce the use of a comma before nonrestrictive which? After all, this is what the style manuals, the usage books, and the major modern grammars have in common: they all note that a comma is the marker of nonessential information, which means this punctuation practice comes far closer to being a universal rule.

    Nichols is in effect proposing that current punctuation conventions be changed so that they fit Butcher’s description of these conventions. Such a change would mean that the punctuation in the two sentences cited (that about Gudrun, that about the elk) would cease to be correct. But given that such punctuation is prevalent, rarely leads to ambiguity, and is arguably more elegant (in examples like those cited) than the alternative, her proposal seems to invite — albeit to a lesser degree — the kind of critique that she directs at Fowler: “it seems to me that making a guideline into a rule to avoid this sort of situation [restrictive–nonrestrictive ambiguity] is too heavy-handed, like imposing a curfew on an entire town because a handful of its teenage residents keep staying out too late.”

    Perhaps the most plausible appeal to punctuation makes just two claims: (1) in a correctly punctuated sentence that contains an integrated relative clause, it is a matter of indifference (at least from the point of view of clarity) whether the clause begins with that or which; and (2) a sentence that contains a supplementary relative clause with no punctuation surrounding the clause is incorrectly punctuated if this lack of punctuation leads to ambiguity. On this way of thinking, the following sentence is correctly punctuated only if the relative clause is integrated:

    He never circulated his research findings which were likely to cause embarrassment.

    If the absence of a comma before the relative clause guarantees that the clause is integrated, then, according to this approach, there is no reason to use that rather than which to introduce the clause.

    I agree that if the relative clause in the preceding example were intended to be supplementary, the sentence would be better punctuated if a comma had preceded the clause; and I think that if, in otherwise ambiguous cases, supplementary clauses were always surrounded by punctuation, then punctuation would be sufficient to mark the difference between integrated and supplementary clauses in most cases where there is a real chance of ambiguity. But I am not sure that failing to insert the comma would be incorrect. And even if it would be, given that there are correctly punctuated sentences with supplementary relative clauses not preceded by punctuation, and given that people do not punctuate as well as they should (a writer may not notice that context does not make clear that the clause is supplementary), I doubt whether we can in practice be very confident that the lack of a comma here ensures that the clause is integrated. Therefore, I think that where there is a real chance of integrated–supplementary ambiguity of the kind in which propositional content is at stake, it may be a good idea not only to surround a supplementary relative clause with punctuation but also to introduce an integrated relative clause with that. Differentiating the clauses in two ways may involve redundancy, but such redundancy does not seem to me to be a bad thing.

    We should note that punctuation alone is no more able than the pronouns alone to prevent a three-way ambiguity. Remember the sentence that involves both such an ambiguity (when the sentence is taken in isolation) and a nonrestrictive relative clause that cannot be preceded by a comma: “one cannot be an outlaw and escape the watchful eyes of Fricka, eager to protect the runes which secure her power too.” Remember also the Rousseau passage that Paul de Man discusses. He argues — convincingly I believe — that the relative clause is nonrestrictive; yet this being also an integrated clause, it cannot be preceded by a comma (at least not easily). And with respect to the ambiguous sentence about embarrassing research findings, there are contexts in which, even assuming that the relative clause is nonrestrictive, inserting a comma before the clause would be difficult:

    There is one clear instance of him seeking to avoid upsetting the scientific establishment: he never circulated his research findings which were likely to cause embarrassment.

    In order for what follows the colon to cohere with what precedes it, the relative clause, which we are assuming is nonrestrictive, seems best understood as an integrated clause — in which case there is considerable pressure not to insert a comma before it; and if we do not insert a comma, then there is considerable pressure to use which as a marker of its nonrestrictiveness. I think that the existence of nonrestrictive clauses that are best introduced by which and cannot be preceded by a comma might increase the chances of a nonrestrictive clause not being preceded by a comma even in cases where a comma can be inserted and is needed to show that the clause is nonrestrictive. If so, then this may increase the need in such cases for some way of differentiating the clauses other than the comma before supplementary clauses. So I doubt that the appeal to punctuation is as decisive a refutation of Fowler’s claim about lucidity as many think it is.

  5. Notes from school « Arnold Zwicky's Blog Says:

    […] school proscribes it. Readers of Language Log and this blog will remember that my linguistics crew deprecates this proscription. In fact, I have no quarrel with the which above — didn’t even notice it on my first […]

Leave a Reply

%d bloggers like this: