Arnold Zwicky's Blog

From a Comcast (cable tv) program description:

(Airdate January 9, 2007)  Stabler and Benson are at odds over whom to believe in a “he said, she said” rape case involving a husband and wife (Blair Underwood, Michael Michele) in the middle of an extremely bitter child-custody dispute.

I was struck by the whom of whom to believe. Not unacceptable, but very much not what I would say or write.

Meanwhile, Stan Carey has posted about a kerfluffle on Twitter, in which various tweeters have objected strongly to the name of the Twitter feature Who to follow. Carey finds this variant entirely acceptable (and the whom variant stilted), as do I. But Business Insider thinks it’s “bad English”; GalleyCat calls it “one of the most viewed and easily overlooked grammar mistakes on the Internet”, adding that it’s “reassuring to watch a major social network struggle” with grammatical rules; Jay Rosen, who teaches journalism at NYU, believes it’s a “grammatical error”; and other Twitter users are variously bothered, disappointed, or annoyed by the phrase. Carey provides lots of quotes, with links.

He maintains that all the critics are wrong and provides a long and detailed account of the who/whom issue, with many citations of sources. Well worth reading.

Here my concern is with the choice of pronouns in the specific construction in the Comcast and Twitter examples.

In both cases, we have marked-infinitival interrogatives: a wh-interrogative expression in combination with an infinitival VP marked by to:

who(m) [wh-interrogative phrase] + to believe/follow [to-infinitival VP]

Following CGEL (p. 873), I’ll call this the to-Infinitival Interrogative (or ToInt) construction. I’ll talk about its analysis in a moment, but first some account of why Carey and I are so wary of these ToInt examples.

The crucial observation is that ToInt is somewhat informal in style, while whom is decidedly formal in style. So, on top of a general preference for who over whom — except when the wh-pronoun is object in a PP or when it’s in extremely formal contexts — there is a specific preference for it in ToInt, so as to avoid stylistic discord (see Silva & Zwicky, “Discord” (1973), here). That is, I see ToInt as much like the construction Sluicing, which I posted about in connection with the awkward example

She’s been seeing someone else, but won’t say whom.

in “The siren song of whom” (here); Sluicing is also somewhat informal in tone.

One piece of evidence that ToInt is informal comes from its combination with fronted rather than stranded Ps. In general, P-fronting is decidedly formal; compare:

Where should I start from? / Which houses will you go into? [P-stranding; neutral in style]

From where should I start? / Into which houses will you go? [P-fronting; very formal in style]

P-fronting in combination with ToInt is remarkably awkward:

Where to start from? / Which houses to go into? [P-stranding]

?From where to start? / ?Into which houses to go? [P-fronting]

— indicating that ToInt is informal in style.

Digression on ToInt. CGEL notes (in chapter 11, written by Rodney Huddleston) that there are two types of ToInt, illustrated by

(1) What to do in the event of fire

(2) How to persuade her to forgive him?

Note the difference in punctuation; type (1) examples occur in titles or headings of material providing advice or instruction, while type (2) examples are (self-directed) questions. Type (1) examples tell, and they function like NPs, while type (2) examples ask, and function like clauses. I’ll call ToInt of type (1) Telling ToInt, and type (2) Asking ToInt. The Comcast example is of the Asking type, the Twitter example of the Telling type.

First observation: both types of ToInt can occur embedded as well as in main clauses (as in (1) and (2)):

(1′) He  told us / explained  what to do in the event of fire.

(2′) He  asked / wondered  how to persuade her to forgive him.

The Comcast example is embedded, the Twitter example is a main clause.

Second observation: both types of ToInt allow the full range of fronted wh-expressions that occur in finite clauses:

who(m) to see, what(ever) to do, when to go, where (else) to go, how to open the door, which/what road to take, how slowly to talk, …

Notes on who(m). Now back to the original concern, the choice of who vs. whom in object function. The earlier prescriptive standard (what I call System A in my “Object whom” posting) turns out to be difficult for people to master; people are inclined to use whom in more places than System A prescribes. Some of these — what I’ve called ISOC and ESOC here and here — are motivated by sentence syntax, but others are “unprovoked” instances of whom in subject function, with a range of symbolic associations (enumerated here). In the face of such examples, a reasonable response would be to simply advocate the use of System B, with whom used only for objects in PPs (To whom did you give the book?). That’s what John E. McIntyre says:

Colleagues, literate, educated, adult native speakers making a living with words, regularly come up to me to ask, “Should this be who or whom in this sentence?” They’re not barbarous. . . . They just don’t hear it. They don’t know whether to use the pronoun as subject or object unless they pause to parse the sentence. And that’s in ordinary, journalistic sentences, not rococo Nabokovian constructions.

Even when they do try to work it out, they commonly get it wrong.

Taking this tack also means that you won’t fall into discord in a construction where whom is prescribed in System A but the construction is informal in style: ToInt, Sluicing, and also echo or reclamatory questions like:

You saw who / ?whom?!

And that would be a good thing, to my mind.

This entry was posted on April 18, 2012 at 9:50 am and is filed under Constructions, Pronoun case, Style and register. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.