Paradigm gaps
Here's another linguistically oriented cartoon, this time confronting the serious problem of paradigm gaps and the complicated mental operations that go into deciding if something 'sounds right' or not. A famous case in English is a coordinated possessive NP: if John and I write a paper together, is it my and John's paper? John's and my paper? Mine and John's paper? Me and John's paper? Paradigm gaps are rather odd, in that they represent indeterminacy in a corner of the grammar where indeterminacy is not the norm. Speakers choose different variants on different occasions, usually find that none are particularly satisfactory, and will sometimes deliberately rephrase to avoid using the construction at all, as the Soup 2 Nutz kid is doing below:
On a similar but syntactic, rather than morphological, note, while car camping this weekend a friend noticed the following instruction on the 'warning' tag of a fold-out camp chair:
(1) Do not sit on the arm of this chair to avoid danger.
We all agreed that if a mountain lion should jump out of the bushes, it would be dramatically ineffective to try to avoid it by sitting on the arm of the chair. Of course, what the label intended is this:
(2) To avoid danger, do not sit on the arm of this chair.
What's interesting is how the right-adjoined position for the purpose clause 'to avoid danger' in (1) does not allow matrix scope -- that is, it can't be construed with the imperative 'Do not!" part, as intended, but has to be construed with the main verb ('sit"). The thing is, right-adjoined material normally can allow multiple scopes if there's multiple modifiable constituents, e.g.
(3) I told her to go to the party to annoy Bill.
...where 'to annoy Bill' can be construed with 'go to the party' (embedded scope) or 'told' (matrix scope); in the first case, the speaker is personally trying to annoy Bill (by doing the telling to go to the party); in the latter, the object of 'tell' ('her') is the one presumptively doing the annoying (by going to the party). That is, the rightmost position for the purpose-clause allows both scopes. Compare that to (4):
(4) To annoy Bill, I told her to go to the party.
Here, only matrix scope is available. So normally, rightwards attachment allows multiple scopes, leftwards only matrix scope.
But with the imperative in (1), rightwards attachment doesn't allow the highest (imperative) scope, only the lower (verb-oriented) scope. It's a kind of a paradigm gap.
P.S. The camp chair purpose-clause, we all agreed, was likely the result of a bad translation. But I'm curious about the source language and its behavior with purpose clauses. If it's a head-final language, like Japanese or Korean, it could just be that the translator didn't know how to properly adjust the word order to get the English scope s/he wanted. On the other hand, if it's a head-initital language like Mandarin or Cantonese, perhaps that language has interestingly different purpose-clause properties than English, and the clause-placement here reflects an L1 effect. Going to check the 'Made In:' tag now...yep! Made in China. So maybe there's something about Chinese purpose clauses, or Chinese imperatives, that means the matrix scope is available in the equivalent Chinese sentence. Free paper topic! :)
posted by hh @ 9:33 PM
8 Comments:
Hm, cool post, hh. The cartoon is excellent, too.
Fascinating!
Do you see a possible exception to the matrix scope of the sentence:
To get to Carnegie Hall, I told her to practice.
I see it as ambiguous -- but would you say that's only because of a strong semantic force?
That's a nice example, Michael! I don't find it so bad, either. But I tend to think that it's really just because of the semantic context, as you suggest, that allows the embedded scope to emerge...kind of a post-hoc reanalysis/reinterpretation, not really licensed by the syntactic structure. Though that might be too strong... maybe a preposed adjunct clause can get embedded scope for 'real', with just a strong preference for matrix. Hard to sort out the possibilities! Need more data.
Of course, if John and me write a paper together, everybody knows that the correct phrase is "our paper" ;)
But now, here's a conversation I had recently with a Californian friend.
ME: so, are you going to let your mum mess up with the organization of your wedding?
MICHELE: absolutely not! This is Bruce and my wedding!
A wild thought about the camp chair example: could negation have something to do with the lack of matrix scope? Something like a weak island effect? What do you think about
1) I didn't tell her to go to the party to annoy Bill
It seems to me like the matrix reading is more difficult to get here. But then, I can't really speak English :P
Hi Luis! Interesting thought about the negation. In your example, I still find I can get matrix scope a-ok ("I didn't tell her to go the party to annoy Bill" where you're denying that you did the telling on purpose to annoy Bill). BUT, in the imperative example, I think it makes a difference:
(1) Don't sit on the arm of the chair to avoid danger!
--negation necessarily scopes over the purpose clause
(2) Do sit on the arm of the chair to avoid danger
--no negation, therefore no scope issue at all! the imperative is suggesting that to avoid danger, you should sit on the arm of the chair.
So in (2) either the matrix imperative force can be modified by the purpose clause, or else it doesn't matter that it can't be so modified since the interpretation ends up the same no matter what. It's only in (1) that negation creates a clear difference between imperative scope and embedded scope, and makes it clear that embedded scope is the only available one. Good eye! maybe it is a negative island of some kind...
I looked in the Cambridge grammar - they acknowledge the *awkward* and indicate that several options are all used - John and my, John's and my, and my and John's. They also say that with plural objects (those are John and my papers; those are John's and my papers) the first reading is unambiguous and preferred when that's the meaning wanted, and "my and John's papers" is less preferred.
I'm a little puzzled by the idea of a "paradigm gap." Of course people are always trying to say things like "my and John's paper," and finding it unsatisfactory, along with the alternatives. In the days of prescriptivism (up until last Tuesday), I was taught that the correct form is "John's paper and mine," and although you rarely hear this in everyday speech, I don't think it violates anyone's intuitions about correctness. In Spanish, too, which has no genitive, it is impossible to say *"mi y su papel," but natural to say "mi papel y el de Juan," ("my paper and that of Juan,")or "su papel y el mio," (his paper and mine.) Would you say that, because it is not just impermissible, but impossible to modify one noun by two possessive pronouns, Spanish has a "paradigm gap"? And if not, why would you say so about English?
This is a real, not a rhetorical, question. Anyone?
Post a Comment
<< Home