Linguistic commentary from a guy who takes things too literally

Archive for the ‘Sloppy and strict anaphora’ Category

You Can’t Go From Strict to Sloppy

Posted by Neal on March 4, 2011

A sign on a local electronics store says:

I can start my car from inside my house. Can you?

I’ve written about strict vs. sloppy anaphora (aka strict vs. sloppy identity) a couple of times before. The canonical example, at least for me, is an old joke that plays on it:

Wife: Jim kisses his wife goodbye before he leaves for work every morning. Why don’t you do that?
Husband: Because Jim wouldn’t like it!

Strict anaphora is the funny reading, where do that is understood as “kiss Jim’s wife”; sloppy is the wife’s intended reading, where do that is understood as just “kiss one’s wife,” resolving to “kiss your wife”.

In the store sign, though, we have two opportunities to choose between strict and sloppy, because there are two mys to deal with. Logically, there would seem to be four possible meanings. The intended one would be the Sloppy/Sloppy interpretation; i.e.

Can you start your car from inside your house?

Then there’s the Strict/Strict reading, implausible and funny, but as far as I can tell, available:

Can you start my car from inside my house?

Maybe I can even imagine a context in which this reading would be the intended one. Suppose I want you to house-sit for me and drive my car around while I’m away from home for an extended time. Furthermore, my car is temperamental, and requires ten minutes of gas-wasting idling time before it can reliably be driven on a cold day, so if you’re going to house-sit for me and drive my car, you need to be able to use my powerful remote-control ignition from inside my house. Knowing that you’re sometimes a bit of a technophobe, I ask you if you’ll be able to meet this prerequisite: “I can start my car from inside my house. Can you?”

Now for the mixed readings. There’s the Sloppy/Strict interpretation:

Can you start your car from inside my house?

Maybe I’m having a party, and the guests and I are bragging about the ranges of our remote-control ignitions. As I look out the window at my car and the guests’ cars parked on the street, I lay down a challenge: “I can start my car from inside my house. Can you?” We all whip out our tools and the contest begins.

OK, so how about the Strict/Sloppy interpretation?

Can you start my car from inside your house?

At this point, theoretical syntax has something to say. In reading a paper by Kyle Johnson on a different subject, I came across this line:

Dahl (1974) discovered that when an ellipsis has two pronouns in it, the first of them cannot get a strict interpretation if the second gets a sloppy interpretation. (p.4)

Johnson’s example:

James said he’d rob his constituents and Peter did too.

  1. James said, “I will rob my constituents” and Peter said “I will rob my constituents” too. [Sloppy/Sloppy]
  2. James said, “I will rob my constituents” and Peter said “James will rob his constituents” too. [Strict/Strict]
  3. James said, “I will rob my constituents” and Peter said “I will rob James’ constituents” too. [Sloppy/Strict]
  4. *James said, “I will rob my constituents” and Peter said “James will rob my constituents” too. [Strict/Sloppy]

I kind of agree with Johnson’s judgment, but I still wonder if the right context would make it OK. So how about with our RC car-ignition example? I don’t think I can get a Strict/Sloppy interpretation with that, either, but it might just be that I can’t imagine a suitable context where it might make sense. Maybe if you and I are trying to set up some kind of practical joke … yeah, let’s say that I’ve secretly removed the batteries from my remote ignition, and am going to drive my wife crazy by having her continually try and fail to start our car from inside the house. You are my confederate, sitting in your house a few blocks away, with my spare RC ignition, listening to me through a secret microphone transmitter we’ve put in my house. Now, every time I show my wife how easy it is to operate the remote ignition, you will hear me and at my signal, use the spare to start our car. But it will only work if the remote is powerful enough to unlock my car all the way from inside your house. How about that? With that scenario, can you get the Strict/Sloppy reading now?

Meanwhile, I’ve been looking into two categorial-grammar analyses of anaphora, with the aim of seeing if they can generate a Strict/Sloppy reading. If such a reading is truly ungrammatical regardless of context, and the CG analyses fail to generate that reading, woohoo! If you’re a CG fan and this question interests you, let’s talk.

Posted in Ellipsis, Sloppy and strict anaphora | 5 Comments »

Slippery Ellipsis

Posted by Neal on February 5, 2006

I remember sitting in my chemistry class one day during my junior year. The guy sitting in front of me was wearing his debate club T-shirt. On the back was our school mascot, a tiger, saying to another tiger, “Don’t panic.” The other tiger was replying, “I’m not!” (The debate club had some kind of inside joke involving the cathcphrase “Don’t panic” from The Hitchhiker’s Guide to the Galaxy.)

Weird, I thought. When the preacher asks, “Do you take this woman to be your lawful wedded wife?” and you say, “I do,” I thought, you have to mentally copy take this woman into the spot where something should be following do: “I do take this woman to be….” And when he asks, “Will you love, honor, and cherish…?” and you answer, “I will,” it’s understood that the love, honor, and cherish is carried over to complete the I will sentence: “I will love, honor, and cherish….” But here, when Tiger One says, “Don’t panic,” and Tiger Two says, “I’m not,” you can’t just take the verb from the first utterance and use it to complete the second one. If you do, you get *I’m not panic, which doesn’t work.

Shouldn’t Tiger Two be saying, “I don’t”? Well, he could, if he means to say he never panics. Or he could say, “I won’t,” if he and Tiger One are talking about some future operation. But what if Tiger One wants the non-panicking to happen right now, and Tiger Two is telling him that right now, he’s not panicking? It looks like I’m not is the only way to go, and you just have to mentally tweak panic and turn it into panicking before you use it to complete Tiger Two’s sentence.

I was reminded of this T-shirt at last month’s LSA conference during a session on verb-phrase ellipsis–that is, verb phrases that lack a verb phrase, such as the I’m not, I do, and I will examples above. The speaker whose talk particularly reminded me of the Don’t panic T-shirt was Ivan Sag, who talked about mismatches much bigger than trifling morphological details like panic vs. panicking. Those have been recognized for many years, and Sag summarized the standard view that as long as the present and missing VPs have the same meaning, the ellipsis is OK. So since Tiger One and Tiger Two were both talking about a situation in which Tiger Two is panicking, the mismatch is allowed. But Sag then went on to present cases where there is a mismatch in meaning…

  • Strict vs. sloppy identity
    John loves his wife, and so does Harry
    (under the reading in which Harry loves Harry’s wife)
  • Active-passive mismatch
    This needs to be done quickly, and I will.

  • Resolution of indexicals
    A: Do you think they’ll like me?
    B: Of course they will! [i.e., will like you, not will like me]
    (from Sag’s handout)
  • I’m gonna send them an email saying that Ling 1 is something they could take. I don’t think many of them will, though.
    (understood verb not the main verb, but take, buried in a relative clause)
    (from Sag’s handout)
  • Your parents aren’t home, but suddenly the police are.
    (at your house, that is, not theirs)
    (from Sag’s handout)

Some of the mismatches earlier in the list can be explained while maintaining the same-meaning story, but by page 6 of the handout, example after example of more elaborate mismatches made a good argument for Sag’s point: Trying to write rules for VP ellipsis based just on syntax or even semantics is hopeless. What really matters is what listeners can infer from some situation and use to fill in the missing VP.

Posted in Ambiguity, Ellipsis, LSA, Sloppy and strict anaphora | 2 Comments »


Posted by Neal on April 17, 2005

Doug’s been reading a few of Louis Sachar’s Marvin Redpost books, and the one he’s been reading now (Marvin Redpost: A Magic Crystal?) is full of fun linguistic lessons just in the first two chapters. It opens with a syntactic ambiguity, in this exchange between Marvin and his teacher:

“Excuse me, Mrs. North,” said Marvin. “When’s the book report due?”

“I told you Tuesday,” said Mrs. North.

Marvin nodded.

Mrs. North returned to her papers.

He still didn’t know when the report was due. Did Mrs. North mean that it was due Tuesday? Or did she mean that she told him on Tuesday when it was due?

In chapter 2, Sachar moves on to pragmatics. Marvin is invited to go home with his classmate Casey, whose house he has never visited before. On the way to her car, this dialogue ensues:

“I hope you like cats,” said Casey.

“Oh, sure,” said Marvin.

“You’re not allergic?” Casey asked.

“I don’t think so,” said Marvin.

“That’s good,” said Casey.

“Do you have a lot of cats?” Marvin asked.

“No, I’m allergic,” said Casey.

Casey expresses concern over whether Marvin likes cats seemingly apropos of nothing; Marvin assumes she is abiding by the Maxim of Relevance and has a reason for doing so, the most likely one being that there are cats in her house and she is looking out for the comfort and well-being of her guest. But no, she has no cats, and was apparently just violating the maxim. Casey, it seems, is a little bit off, an impression reinforced by the conversation two pages later:

“I’m going to have to call my mom when we get to your house,” he said. “She thinks I’m at Stuart’s.”

“Do you know your phone number?” asked Casey.

“Of course,” said Marvin. “Don’t you?”

“No,” said Casey.

That surprised Marvin. He’d known his phone number since kindergarten. “You should,” he said.

“Why should I?” asked Casey.

“I don’t know it either,” said Casey’s father from the front seat.

That really surprised Marvin. “Did you just move or something?” he asked.

So now Sachar is into the semantics lesson. Marvin intends his “Don’t you?” question to mean, “Don’t you know your phone number?” while Casey is taking it to mean, “Don’t you know my phone number?” Or as semanticists put it, Marvin intends the “sloppy identity” reading (where the you in know your phone number refers to either Marvin or Casey as appropriate), while Casey takes the “strict identity” reading (where it starts off referring to Marvin and continues to do so).

The resolution of the misunderstanding comes a few pages later, along with the revelation that Casey really was obeying the Maxim of Relevance when she brought up the subject of cats: The title of the book he’s been carrying for his report is A Thousand Cats.

A Magic Crystal? is clever and entertaining aside from the linguistically relevant parts, but I do have one minor complaint about it. When Marvin or Casey makes a wish with the magic crystal for something to happen in the future, they say, “I wish,” followed by a present-tense indicative clause–for example, “We wish nobody in Mrs. North’s class is sick tomorrow,” (p. 27) or “I wish I don’t get hurt” (p. 24). That’s not how wish is used in the English I speak. You could say, “I wish nobody would be sick tomorrow.” The trouble now is that it sounds as if it’s understood that someone is liable to be sick. If that’s not the case, you either substitute hope for wish, or go for an infinitival clause: “I wish for nobody to be sick tomorrow.” Sometimes Sachar uses wish in a more normal way, though. When the wish is a present-time counterfactual (i.e., for things to be other than they are right now), Sachar has the characters use a past subjunctive, the normal usage in my grammar: “I wish I had an ice cream sundae” (p. 20), or “I wish you’d shut up!” (p. 46). Does this pattern of usage for wish hold for any of you?

Posted in Ambiguity, Kids' entertainment, Lexical semantics, Quantity and Relevance, Reviews, Sloppy and strict anaphora, Syntax | 4 Comments »