The leader of the IMF and a possible candidate for president
« previous post | next post »
The first sentence of this news report is perfectly fine, but it presents a linguistic puzzle:
The leader of the International Monetary Fund and a possible candidate for president of France was arrested Sunday in connection with the violent sexual assault of a hotel maid after being yanked from an airplane moments before it was to depart for Paris, police said.
The puzzle is how such a conjunction can denote a single person, as it clearly does in this sentence. It could even more easily denote two, but then we’d see “were arrested”, not “was arrested”.
First a descriptive query: do all languages allow such a conjunction of a definite and an indefinite singular noun phrase in subject position, interpreted as referring to a single person? And does English allow it quite generally, or is this a special newspaper style?
If the conjunction occurred in the predicate, “Dominique Strauss-Kahn, 62, is the leader of the IMF and a possible candidate …”, I would know how to analyze it – I’ve written about interpreting indefinite and definite noun phrases in different ways, with different “semantic types”, including a “predicate type” for such predicate positions. Then the “and” gets a straightforward interpretation as a conjunction of properties – So-and-so is both the leader of the IMF and a candidate for president, fine.
But the subject of a sentence isn’t normally interpreted predicatively. I know how to explain a subject noun phrase containing two descriptions of a single person when it has a conjunction of common nouns under a single article, as in “The president and commander-in-chief”, or “A prominent poet and essayist”, since the common nouns themselves have “predicate type”, and the interpretation of “the” or “a” then gives the whole noun phrase its referential or specific interpretation as a single individual.
But in the sentence above, if we start from the usual referential or specific interpretation of each of the two conjoined noun phrases, everything I know about the interpretation of “and” would lead me to conclude that the conjunction would have to come out plural. (In fact that’s so even if they have the same denotation: Superman and Clark Kent are (not is) the same person. Though come to think of it, you can't very well put two co-referential proper names in an example like the sentence we started from, neither with singular nor with plural — it sounds anomalous to say "Superman and Clark Kent was/were arrested ..", doesn't it?)
This is mainly a puzzle for analysis; maybe I should be searching the linguistic literature and if I don’t find a solution there I should start working on one. Intuitively, it seems to call for something similar to an analysis Emmon Bach long ago suggested for all noun phrases (not something he held onto), where there is some ‘covert’ head noun with those two noun phrases connected to it in a predicative manner, e.g. “a person who is the head of the IMF and a candidate for president.” That would get the right semantics just fine. (And it would probably even explain why it's hard to do the same thing with two proper names, since proper names don't so easily get predicative meanings.) But I don’t like to posit otherwise unmotivated invisible “deep structure” (or extra invisible stuff at some level of “logical form”) just to get the semantics to come out right – that feels like cheating.
Yerushalmi said,
May 15, 2011 @ 7:15 am
I had the same reaction to the sentence.
Why didn't they just write this?
"The leader of the International Monetary Fund, a possible candidate for president of France, was arrested…"
Barbara Partee said,
May 15, 2011 @ 7:19 am
Ah, but that's a different reaction! Mine was, wow, English can do that?! I hadn't noticed, and I don't think any of the theories of the semantics of noun phrases and of conjunction that I know of can handle this!
If they had just written it your way (which would indeed probably be an improvement, avoiding the temporary ambiguity), I wouldn't have gotten the "wow" moment that gave me the impetus to write this post! Less fun!
MattF said,
May 15, 2011 @ 7:31 am
I'd leave out the 'a' for a minimal edit:
The leader of the International Monetary Fund and possible candidate for president of France was arrested…
Still inelegant, but without that change, it's ambiguous until you hit the was/were branching point.
army1987 said,
May 15, 2011 @ 7:47 am
I find it mildly confusing (after seeing was I took a couple hundred milliseconds to realize what was going on) but not ungrammatical.
I agree with MattF, but what about similar constructions of the foo bar and a baz quux in which there's no possible or candidate or any other clue that the foo bar and baz quux would mean that the foo bar is one of several baz quuxes rather than the only one? (the foo bar, a baz quux would work).
army1987 said,
May 15, 2011 @ 7:49 am
BTW, it is grammatical in Italian too, where the fact that all verbs agree in person and number with their subjects makes the construction unambiguous with all verbs.
James said,
May 15, 2011 @ 7:49 am
Hm, interesting.
It does sound very newspaperish, or anyway very newsy. I can imagine a nightly news anchor using your conjoined article construction.
This doesn't work — at least not for me:
So the articles are important.
Mark Liberman said,
May 15, 2011 @ 7:50 am
Searching the New York Times Corpus for similar examples, all the cases that I've found so far seem to be incompatible with the "two descriptions of one entity" reading, e.g.
The Marriott Corporation and a private Japanese company will spend about $400 million to develop up to nine Marriott hotels in Europe …
The ozone hole over Antarctica and a smaller hole over the Arctic have become seasonal occurrences …
The President of Botswana and a grassroots organizer from Burkina Faso have received the Hunger Project's annual leadership prize
The private National Radio Network chain and a police spokesman reported the bomb was left in a sports duffle bag under a table …
So I agree with Barbara's reaction: "wow, English can do that?"
As Barbara observes, it's always interesting that so many commenters see such examples as an editing problem rather than an analysis opportunity!
Gaston said,
May 15, 2011 @ 7:51 am
But that's not English, it's English encoded into writing and has therefore lost a crucial aspect of the language, the intonation, which is what makes it clear (or at least an available interpretation) that the conjunction refers to a single entity. I imagine the writer had the appropriate intonation in mind so wasn't alert to the unclarity that results from the lossy encoding (I'm sure we've all had this problem in eg email when trying to convey irony, etc. and being misunderstood). It could of course have been made clearer with commas.
Bruce Rusk said,
May 15, 2011 @ 7:52 am
One reason this form is relatively rare might be its ambiguity in many cases, such as when both terms are plural and with certain conjugations. In the sentence, "Several members of the Legislative Assembly and opponents of the ban on cheese doodles spoke at a rally on snacking freedom yesterday," it's unclear whether the "members" and "opponents" are identical sets of people. Even two singular terms could be ambiguous here, since unlike "was/were," "spoke" remains the same with a singular or plural subject.
It would also be possible to differentiate the two in some cases by the placement of articles: "A professor of linguistics and accomplished puppeteer addressed the conference / was the keynote speaker." vs "A professor of linguistics and an accomplished puppeteer addressed the conference / were the keynote speakers." The combination of definite and indefinite articles in the example perhaps adds to the initial confusion.
army1987 said,
May 15, 2011 @ 7:52 am
no, wait. Now that I translate it literally in Italian, keeping both articles, it doesn't sound right.
Marc L said,
May 15, 2011 @ 7:58 am
As a retired copy editor, I can tell you that the problem resolves itself in a lack of commas. The editor, who probably had three seconds to read the copy, should have made the phrase appositional. It happens all the time on the copy desk. It's built into the system.
The Ridger said,
May 15, 2011 @ 7:59 am
But if you delete the "an" (as suggested by MattF) it sounds as though he's the only "possible candidate".
I have to say I didn't really register the oddness, because "was" so clearly told me it was one man. Had it been, say, "went", I might easily have thought it was two people and been startled at a following "he".
Jacqueline Dubois said,
May 15, 2011 @ 7:59 am
I understand that ' the leader of the IMF and a possible candidate…' is an entity, a description of one person, whatever you put in it which obeys to other grammatical rules, and true it seems specific to English as in French I would say 'le patron du FMI et candidat potentiel…'
John said,
May 15, 2011 @ 8:12 am
Then there's your "he held onto" which I would have written "he held on to"
Ø said,
May 15, 2011 @ 8:15 am
What if he hadn't been caught?
"The leader of the International Monetary Fund and a possible candidate for president of France narrowly escaped arrested Sunday in connection with the violent sexual assault of a hotel maid …"
Martin J Ball said,
May 15, 2011 @ 8:16 am
Surely the real ambiguities are:
1. 'sexual assault of' rather than 'sexual assault on'. To me it's at least possible that the 'of' denotes the actor
2. 'the violent sexual assault of a hotel maid after being yanked from an airplane …' Who was 'yanked from the airplane'? The maid?
Ø said,
May 15, 2011 @ 8:25 am
oops: escaped arrest
edI was just trying to make the same point made more clearly by Bruce Rusk with his cheese doodle example.
James said,
May 15, 2011 @ 8:35 am
But, again, the post isn't asking for editorial advice. It's about the correct analysis of the rather distinctive conjunction.
How is the given construction related to the following (which needs no articles):
Barbara Partee said,
May 15, 2011 @ 8:45 am
So far only @army 1987 has ventured an answer about whether this kind of conjunction "The A and a B" in subject position can refer to a single individual in other languages. Interestingly he first said 'yes' for Italian and then changed it to 'no', apparently having first thought about examples like "The A and B" or "An A and B", which are unproblematic.
That raises another descriptive question: In languages without definite and indefinite articles (probably the majority of the world's languages), what interpretations are there for a conjunction like "leader of IMF and possible candidate for president", used in subject position? If it's possible at all, is it ambiguous between one person and two?
Mark Liberman said,
May 15, 2011 @ 8:49 am
Turning the definiteness around, we find things like the following, where the two NPs are also distinct:
A Soviet diplomat and the Vietnamese Ambassador here, Ngo Dien, said there had been a few cases of violence against recent Vietnamese settlers …
A copy of that statement and the videotape were given to The New York Times by the Salvadoran authorities this week.
A lack of computers and the need for paperwork now use up as much as 60 percent of a caseworker's time.
A smiling Prime Minister Li Peng and the Communist Party General Secretary, Jiang Zemin, appeared before what was described as "an enlarged meeting " of the party committee …
A Federal grand jury and the agency's inspector general have been investigating the matter for more than a year.
A Postal Service official and the designer of the stamp vigorously deny all three suggestions.
A Federal District Court in North Carolina and the Court of Appeals for the Fourth Circuit dismissed her suit.
But I did turn up these examples, where the conjoined noun phrases are an initial appositive offering two descriptions of a singular subject:
A writer for The New Yorker and the author of "Moonshine: A Life in Pursuit of White Liquor," he started visiting the Florida cane fields in 1984, and over the next four years kept going back.
A senior at Newtown High School and the principal cellist with the Greater Bridgeport Youth Symphony, he will play Faure's "Elegie " in a program that will also include Mozart, Schubert and some pops.
A member of the royal caste and the firstborn of a well-off family, Mrs. Rai was honored with a splendidly decorated cremation tower and noble bull sarcophagus.
A lecturer in English and American Literature at the University of Kent and the author of "Dickens on England and the English," Mr. Andrews has that pleasant kind of wit that lurks in the nooks and crannies of scholarship.
This structure is fairly common in journalistic writing, I think. The example cited at the beginning of Barbara's post has a similar feeling, which is no doubt why she thought of the "covert head" idea. But such examples are rare, though apparently possible.
VMartin said,
May 15, 2011 @ 9:22 am
"The leader of the International Monetary Fund and a possible candidate for president of France was arrested"
I would translate it in Slovak this way:"Hlava MMF a jeden z kandidátov na prezidenta Francúzska BOL zatknutý…"
BOL means there was one person involved, BOLI means there were two different people involved. The other part of sentence remains untouched (except arrested – zatknutý changes in plural to zatknutí.)
I think that in Latin (no articles as well) the first part could be the same.
The sentence is perfectly understandable in both ways.
Zythophile said,
May 15, 2011 @ 10:11 am
What's the betting the sentence originally started with his name, and read: "Dominique Strauss-Kahn, 62, the leader of the International Monetary Fund and a possible candidate for president of France, was arrested Sunday …" before someone at AP decided to take the name out?
Incidentally, is in only me that has problems with the second half of the sentence? "… the violent sexual assault of a hotel maid after being yanked from an airplane moments before it was to depart for Paris …" What, the maid was assaulted after being yanked from an airplane?
Barbara Partee said,
May 15, 2011 @ 11:07 am
@Mark – Thanks! Question: how do you do those searches? I had imagined that trying to do something with "a" and "the" and "and" would be hopeless, so I didn't even try. And are you searching within the NYTimes site? I imagine not, because I think they care only about keywords. Are you doing it in some linguistic corpus database?
@ VMartin — thanks, that's interesting about Slovak. I notice that you didn't just drop the article from the indefinite member of the pair, but added the explicit "one of". Would it be possible to do it without that, i.e. just "Hlava MMF a kandidát na prezidenta Francúzska BOL zatknutý…"
If it's possible at all, could you still interpret the second one as indefinite?
@Mark, Zythophile, and some others: Yes, I think it's relevant that that conjunction would naturally be interpreted predicatively if it were in apposition to a referential term like a name (See Chris Potts's work on non-restrictive modifiers). And that may indeed be what makes it possible — but I still don't know any synchronic analysis that would predict that.
In fact I noticed after my initial post that there's a related phrase at the head of the recent LL post about Teutonic feminists: "Former Berkeley County delegate and current Republican gubernatorial candidate Larry Faircloth said Thursday …". Those are both indefinites, but it's interesting that in that construction we use them anarthrously, and they could just as well have been one definite and one indefinite.
Luis said,
May 15, 2011 @ 11:09 am
Spanish behaves like Italian in this respect:
(1) El líder del Fondo Monetario Internacional y posible futuro presidente de Francia fue arrestado / * fueron arrestados. –> one person.
(2) El líder del Fondo Monetario Internacional y un posible futuro presidente de Francia fueron arrestados / *fue arrestado –> two people.
John Cowan said,
May 15, 2011 @ 11:10 am
In my opinion, the reason people are talking about what edits to make is that the sentence is unacceptable English. Indeed, the more I look at it, the less English it seems to me. Zythophile's theory accounts for it very neatly: the sentence started out fine and was altered by an editor who didn't notice that the result was no longer grammatical.
Last week I wrote the phrase from taps to reveille meaning the whole of the daylight hours, but that doesn't mean that from … to … in a temporal context in English can specify the later time before the earlier. My attention was on remembering how to spell reveille and I simply didn't notice what I had written.
Ran Ari-Gur said,
May 15, 2011 @ 11:25 am
This sentence reminds me of Dr. Liberman's many examples of misnegation over the years: it sounds right, with the right meaning, until you try to pick it apart. Shortening and simplifying it — "The leader of the IMF and a presidential candidate was arrested Sunday" — makes it sound quite wrong, at least to me, so I'm going with the "poor monkey brains" hypothesis; it only sounds O.K. because it's too much effort to notice that it's not.
(Of course, I can't say for sure that it's ungrammatical English: there could well be speakers who accept even the shortened/simplified version. But I think it's only my poor monkey brain that makes the original quotation work for me.)
Coby Lubliner said,
May 15, 2011 @ 11:27 am
What I gather from Mark's examples is that "a Y and the X" is OK as specifying a single subject, but "the X and a Y" is not, unless "and a Y" is clearly made parenthetical, whether with punctuation (in writing) or intonation (in speech).
Mark Liberman said,
May 15, 2011 @ 11:31 am
@Barbara: Question: how do you do those searches? I had imagined that trying to do something with "a" and "the" and "and" would be hopeless, so I didn't even try. And are you searching within the NYTimes site? I imagine not, because I think they care only about keywords. Are you doing it in some linguistic corpus database?
I started with a digital copy of The New York Times Annotated Corpus
http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2008T19
which
I wrote a little script to extract the text and divide it into sentences, one per line, and then ran egrep with a command like
egrep '^(The|An*) [a-zA-Z ]* and (the|an*) [a-zA-Z ]* (has|is|was|will) '
Most of the hits are of course not relevant, and so far I just picked out the relevant few by eye.
A somewhat more ambitious effort would run a parser like this one over the output, and use the results to find relevant examples automatically. Today's parsers have a hard time with coordination, so the accuracy would leave something to be desired.
The NYT site search does allow you to search for quoted strings, but (as far as i know) these can't contain wildcards.
Justin said,
May 15, 2011 @ 11:49 am
I was not surprised or confused, but probably because I already had heard the news elsewhere and therefore knew and expected that both descriptions refer to the same person. Actually, I only noticed the ambiguity after having my attention directed at it.
As to other languages: In German "Der Chef des Weltwährungsfonds und ein möglicher französischer Präsidentschaftskandidat […]" seems to exhibit the same ambiguity and sounds a little more "off" than in English, albeit still valid.
I was also told that in Hungarian (which has definite and indefinite articles) the construction might be completely invalid.
Sunny said,
May 15, 2011 @ 12:04 pm
This may be a case where pragmatics does some heavy lifting; if I hadn't already known who Dominique Strauss-Kahn was, I would have found the sentence ungrammatical. Real world knowledge probably hides a lot of grammatical sins.
ProudToBeAMammal said,
May 15, 2011 @ 12:06 pm
Will "Mr. X", a gentleman and a scholar went to Paris" fall into the same puzzling category?
Tadeusz said,
May 15, 2011 @ 12:08 pm
Here's an almost verbatim translation into an article-less language, Polish, from a Polish daily:
"Dominique Strauss-Kahn, przewodniczący Międzynarodowego Funduszu Walutowego i [=and] prawdopodobny kandydat na prezydenta Francji, usłyszał zarzuty napaści seksualnej,"
As you, hopefully, can notice, the construction is reversed, the proper name firmly in the subject position, and the descriptions (modifiers) used appositively; if, instead of a proper name, it were a definite noun referring to a unique object, then it would be natural to use this construction, too. I believe that other Slavonic languages would use this construction, too (I am not quite certain about Bulgarian, which does have the post-nominal definite article).
The conjunction "i" (=and) in the translation does not seem to me used idiomatically, my impression is that a comma would be better.
Tadeusz said,
May 15, 2011 @ 12:10 pm
Oh, and the verb is singular.
Coby Lubliner said,
May 15, 2011 @ 12:13 pm
By the way, the latest version of the news story (from AP) reads like this:
T said,
May 15, 2011 @ 12:13 pm
I think Ran Ari-Gur and Sunny are right here. I believe it's one of those cases where your brain is tricked into interpreting it as some kind of apposition with invisible head subject (my second guess would indeed be some invisible deep structure). And the more I look at it, the more ungrammatical it seems, at least to me.
And German/Spanish/French make it sound even worse …
adriano said,
May 15, 2011 @ 12:15 pm
I've tried to translate the first sentence into Italian keeping both articles, but I think it doesn't work.
You can get something similar to that construction, but only getting rid of them both:
"Presidente del FMI e possibile candidato alle presidenziali francesi, DSK è stato arrestato domenica…".
Not quite the same thing.
Barbara Partee said,
May 15, 2011 @ 12:47 pm
Wow, @Mark, it definitely was much fancier than a simple google search. (does the middle term just mean any string of any letters at all? No constraint on length?). Thanks for doing it!
And it's interesting that now one of the leading hypotheses is that it's actually ungrammatical. As you can see, I started the original post by saying the sentence was perfectly fine, but a puzzle to analyze. But I could have been wrong; I could well have been misled by the combination of helpful context and closeness to other things that ARE perfectly fine. But I still don't know. If it's really not fine, that's good for everything I've ever written about noun phrase type-shifting and about conjunction. But my 'wow' balloon is popped.
And Mark, isn't it interesting that some of the editorial advice that's been prevalent among these comments is really quite relevant to the linguistic issues. I'm grateful to all our readers.
VMartin said,
May 15, 2011 @ 1:16 pm
@Barbara Partee
I notice that you didn't just drop the article from the indefinite member of the pair, but added the explicit "one of". Would it be possible to do it without that, i.e. just "Hlava MMF a kandidát na prezidenta Francúzska BOL zatknutý…"
If it's possible at all, could you still interpret the second one as indefinite?
It sounds even better now and I would say it is the same person from the beginning. But there must be the IMF abbreviation! Once I read in some treatise by Anton Marty (where he criticised Husserl) that before the sentence is ended we expect how it would end and if it doesn't, the sentence is percevied as written in bad style.
So if the first name (I understand by "name" in scholastic way the whole compound) "The leader of the International Monetary Fund" is not short and somehow logically connected with the second name " a possible candidate for president of France" or exclude it – like: "The president and a possible candidate for spokesman was arrested" ´- we might expect they are different people, at least in my language where there are no articles.
Der Manager und der Einkäufer des Hotels waren festgenommen worden, ….
Tom Recht said,
May 15, 2011 @ 1:29 pm
For another cross-linguistic data point, my intuition as a native Hebrew speaker is that such a definite+indefinite coordination in subject position would be completely ungrammatical in Hebrew, though it would be fine as a predicate. (Though it wouldn't be too surprising if it enters Hebrew journalese as a calque from English at some point in the future, as lots of other constructions have done.)
I don't know how reliable such intuitions are, though, given that native English speakers can't agree on whether the construction is grammatical in English.
John Walden said,
May 15, 2011 @ 2:09 pm
I immediately thought "Dan Brown!" and hunted down
http://itre.cis.upenn.edu/~myl/languagelog/archives/001628.html
Perhaps a bit unfair because journalism is not novel-writing, even when both are clunkily written.
John Roth said,
May 15, 2011 @ 2:17 pm
I seem to be late to the party. However, the noun phrase, as written and without information from earlier in the sentence, seems to refer to two people. Then the singular verb renders the sentence ungrammatical.
If we relax the constraint that the NP, in the absence of lexically prior information, must refer to two people, then it's a garden path sentence.
Hope I'm not totally confused.
John Roth
Jerry Friedman said,
May 15, 2011 @ 2:29 pm
An appostive example in French:
De son côté, le ministre du Commerce et un des leaders de la formation centriste, Vince Cable, a été plus critique face à ses alliés gouvernementaux…
Would that be grammatical without the name?
GeorgeW said,
May 15, 2011 @ 2:32 pm
Ran's monkey-brain explanation wins first prize. It is clearly ungrammatical when shortened.
Mark Liberman said,
May 15, 2011 @ 3:12 pm
@Ran Ari-Gur: Shortening and simplifying it — "The leader of the IMF and a presidential candidate was arrested Sunday" — makes it sound quite wrong, at least to me, so I'm going with the "poor monkey brains" hypothesis; it only sounds O.K. because it's too much effort to notice that it's not.
I agree, the shortened version is clearly bad. But I'm not so sure about the original. "Possible" seems to play a role — consider "The Suns got a capable point guard and a possible star", describing a trade where the capable point guard and the possible star are the same person. Or this triple play: "When the Baltimore Orioles acquired Eddie Murray in July, they knew they were adding a true professional, a link to their past and a possible gate attraction in September."
It's suspicious that these seem to have to be in object rather than subject position. But still…
Barbara Partee said,
May 15, 2011 @ 3:37 pm
I forgot to reply to @ProudToBeAMammal, who asked, "Will "Mr. X", a gentleman and a scholar, went to Paris" fall into the same puzzling category?". And this is also to @adriano, and maybe some others.
No, when there's a proper name (or other referential noun phrase) that is modified by a non-restrictive noun phrase or conjunction of noun phrases, the noun phrases in the modifier are interpreted as predicates (I'm following Chris Potts's analysis of non-restrictive modifiers.) (Same goes for those Dan Brown constructions.) So those are not puzzling at all, and as some have suggested, they may form the basis for how we interpret the sentence in question if we accept it at all.
Interesting to know that the sentence in question is definitely ungrammatical in a number of languages. But we have a report that it's fine in Czech. Interesting.
James said,
May 15, 2011 @ 3:39 pm
MYL, I don't think they do have to be in object position.
"A capable point guard and a possible star joins the Suns, as Phoenix completes the trade…"
No? My confidence is a bit shaken, but this seems okay to me.
[(myl) I agree, this does seem OK. None the real-world examples that I've found so far have been in subject position, though.
We do seem to understand such things as quasi-predicative, e.g. "[someone who is] a capable point guard and a possible star".]
David said,
May 15, 2011 @ 4:12 pm
I would translate the sentence into Swedish perhaps as: "Chefen för Internationella valutafonden, tillika (en) möjlig fransk presidentkandidat, anhölls på söndagen…" "Tillika" here means roughly 'and also', which makes it clear that only one person is involved. I judge that the indefinite article ("en") is optional. Including it somewhat suggests a sense like "one of the candidates" rather than "a candidate": Swedish often prefers dropping the indefinite articles when using titles as complements.
If the middle clause is translated as "och en möjlig fransk presidentkandidat" ("och en" = 'and a'), it would strongly suggest, in my opinion, that two persons were involved. This suggestion would be removed if the indefinite article was dropped: "och möjlig fransk presidentkandidat", since the article-less form must function as a complement, but then the sentence sounds, if not ungrammatical, then certainly rather clumsy, perhaps because we actually aren't talking about a complement here, but about a subject. (I hope I get the terminology right.)
ShadowFox said,
May 15, 2011 @ 4:24 pm
I'm not convinced the problem has anything to do with definite/indefinite mix. Another publication referred to him as "the leading Socialist candidate for president". So that would make the sentence, "The leader of the IMF and the leading Socialist candidate for president of France was arrested …". Best I can tell, there is no difference. Dropping the second article sound better prosodically, but gives me pause otherwise…
Hjordis said,
May 15, 2011 @ 4:35 pm
I agree that it's probably ungrammatical, but I didn't notice because I was too busy trying to figure out who got pulled from an airplane and when. It took me a few reads.
@James I don't think that's grammatical either. At least, my brain is protesting to it, which is a good indicator.
ShadowFox said,
May 15, 2011 @ 5:07 pm
Consider some variants:
(1Aa) "The chairman of the Cheese-Doodle Party and the leader of the opposition was arrested…"
(1Ab) "The chairman of the Cheese-Doodle Party and a leader of the opposition was arrested…"
(1aA) "The opposition leader and the chairman of the Cheese-Doodle Party was arrested…"
(1bA) "An opposition leader and the chairman of the Cheese-Doodle Party was arrested…"
(2Ba) "A member of the Cheese-Doodle Party and the leader of the opposition was arrested…"
(2Bb) "A member of the Cheese-Doodle Party and a leader of the opposition was arrested…"
(2aB) "The leader of the opposition of the Cheese-Doodle Party and a member of the opposition was arrested…"
(2bB) "A leader of the opposition and a member of the Cheese-Doodle Party was arrested…"
These will certainly elicit different responses from different people. But, it seems, the relationship between the dual subject is important and may affect sensitivity to the construct. IMO, the ones that are right out as impossible are the two double-indefinite constructions–2Bb and 2bB–which is contrary to what was claimed in some comments above. The rest require some more thinking, but they don't look obviously wrong to me (which is not to say that they are perfectly acceptable). This is in contrast with other examples, where a double-indefinite sounds more plausible than other versions. So it's not just a question of syntax.
ShadowFox said,
May 15, 2011 @ 5:22 pm
Here's what I mean:
(3Aa) "The leader of the Socialist Party and the leading candidate for president…"
(3Ab) "The leader of the Socialist Party and a likely candidate for president…"
(3Ba) "A member of the Socialist Party and the leading candidate for president …"
(3Bb) "A member of the Socialist Party and a likely candidate for president…"
In this case, I have no objection to any of the constructs, because of a plausible direct relationship between A/B and a/b. In this case, membership in the Socialist Party is conducive to becoming a candidate (even if the actual selection mechanism is completely obscure). A similar but weaker relationship exists in my other examples, but it's lacking entirely in the original (no relationship between the IMF and the French government).
In my view, this means that we focus on the syntactic structure and ignore semantics at our peril. Compare "The King of the Jungle and a water buffalo…" with "The King of the Jungle and a proud father of a three-cub litter…". Ignoring for the moment the fact that "the King of the Jungle" being a lion is a cultural artifact (as is the reference to lion offspring as "cubs"), the former makes no sense, but the latter is just fine, despite the fact that they are structurally quite similar.
Ran Ari-Gur said,
May 15, 2011 @ 6:13 pm
@Dr. Liberman: I don't think it's the "possible" that makes those acceptable. Your second example, "[…] they were adding a true professional, a link to their past and a possible gate attraction in September", seems just fine to me when we drop the third coordinand, leaving "[…] they were adding a true professional and a link to their past". (Incidentally, it stops seeming fine to me if we change the first coordinand to be definite, as in "[…] they were adding the oldest true professional and a link to their past". I'm not sure if the relevant difference really is just indefinite+indefinite vs. definite+indefinite, or if making it definite triggered something else that's causing the problem.)
Even so, a sentence like James' suggestion ("A capable point guard and a possible star joins the Suns, as Phoenix completes the trade…") is interesting for the reasons that Dr. Partee gives in the main entry, since (so far as I can see) none of them is specific to the case where the first coordinand is definite.
Ian Preston said,
May 15, 2011 @ 6:35 pm
Is there any reason why the word "now" should make a difference to the grammaticality of examples like these?
V said,
May 15, 2011 @ 7:21 pm
Tadeusz: This sentence works fine in Bulgarian also, the articles don't change anything. The equivalent of przewodniczący would get the definite suffix and that's all that's different. It's interesting to note a minor false friend, in that that the cognate of "prawdopodobny" in Bulgarian means plausible, not probable. But it's still understandable, I don't know if you can call it a flase friend. The equivalent of prawdopodobny getting a definite article would make it sound like it's almost certain he'll be the candidate, rather than merely probable.
Also, what Sunny said: "This may be a case where pragmatics does some heavy lifting; if I hadn't already known who Dominique Strauss-Kahn was, I would have found the sentence ungrammatical. Real world knowledge probably hides a lot of grammatical sins." Yay pragmatics.
V said,
May 15, 2011 @ 7:27 pm
If I didn't know who he is, I probably would have thought the indefinite article before "possible" breaks the grammatical structure, but I didn't becuase I do.
Mark F. said,
May 15, 2011 @ 7:36 pm
Hmm. If I change the verb I can add ambiguity, but it doesn't feel any worse to me:
The leader of the International Monetary Fund and a possible candidate for president of France fell into the hands of the police in connection with the violent sexual assault of a hotel maid . . . .
Ran Ari-Gur said,
May 15, 2011 @ 7:40 pm
@Ian Preston: Those seem a bit different to me, not because of the "now", but because they're referring to, and adding more information about, an already-mentioned referent. I'm not sure why that's an important difference, but I'm pretty sure that it is.
Victor Mair said,
May 15, 2011 @ 10:29 pm
"The first sentence of this news report is perfectly fine, but it presents a linguistic puzzle…."
Right now, the linked article has this as the first sentence: "Dominique Strauss-Kahn's reputation with women earned him the nickname 'the great seducer,' and not even an affair with a subordinate could knock the International Monetary Fund leader off a political path pointed in the direction of the French presidency." We are not told of DSK's being yanked off the plane until the second paragraph. It looks as though the editors have thought better of the original first sentence.
Peter G. Howland said,
May 15, 2011 @ 10:49 pm
After 50-some thoughtful, scholarly, linguistically sophisticated comments (the consensus of which seem to come down to “we get it, but it’s kinda clunky”) I thought it might be safe now to insert an aside @John Cowan’s 11:10 am.
How is it possible for “…from taps to reveille…” to mean “…the whole of the daylight hours…” when taps=evening and reveille=morning? That sounds like the whole of the *nighttime* hours to me.
Might you have used the old field-worker’s saw of “we’s workin’ ever day from can to cain’t” (from when one can see to when one cannot) for the whole of daylight?
J. Goard said,
May 15, 2011 @ 11:30 pm
I find it interesting that, although the definite article can be used in an apposition construction when the second nominal itself doesn't provide sufficient accessibility (uniqueness, if you prefer):
(1) Barbara Partee, the influential semanticist and all-around great person
…this possibility seems to be blocked by an article (definite or indefinite) in the second half of the conjunct, i.e. when the conjunction is of DPs/full nominals rather than NPs/ungrounded nominals. That is to say, (2) and (3) force me into the strange reading where Partee is the only influential semanticist:
(2) Barbara Partee, the influential semanticist and an all-around great person
(3) Barbara Partee, the influential semanticist and the author of "Compositionality in Formal Semantics"
So, apposition can apparently screw around with the criteria for definiteness, but only for a single determiner over the entire apposed nominal. I'm not sure that this effect holds in every case, though. I'll see what insight I can get from Google…
@boris_tweets said,
May 16, 2011 @ 12:46 am
@ Jerry Friedman:
I say this is ungrammatical both with and without "Vincent Cable." In my opinion, this is nothing more than a big giant anglicism (not a rare occurrence in French media…).
Also, I'd feel much better if a "l'" were added to the sentence: "De son côté, le ministre du Commerce et l'un des leaders de la formation centriste, Vince Cable, a été plus critique face à ses alliés gouvernementaux…"
Barbara Partee said,
May 16, 2011 @ 1:13 am
@Ian Preston — YES! Thanks! That added "now" makes it unambiguous that at least the second conjunct is definitely a predicate, since you can't use "now" as part of a regular noun phrase. #Now a partner in a hedge fund added that … — you can't parse that with "now" as part of the NP. So either the whole conjoined subject NP in these cases is being interpreted as if it's an appositive to an implicit head term, or the second conjunct is being interpreted as appositive on the first.
@Shadow Fox — nice research, and your conclusions look plausible to me.
@J Goard — smileys, and the content is interesting too!
@Peter G. Howland — After 50-some thoughtful, scholarly, linguistically sophisticated comments (the consensus of which seem to come down to “we get it, but it’s kinda clunky”) — I really appreciate the seminar this has turned into, and I take the gradually emerging conclusions differently: the line between grammatical and ungrammatical is not entirely clear; the semantics of these things is clearly predicative; how exactly they should be "analyzed" is still not clear; pragmatics is playing a big role; many other languages don't allow it at all, so pragmatics can't overrule grammar altogether.
And to everyone — while I was sleeping (I'm in Moscow), this got more and more interesting, thank you all!!
Glenn Bingham said,
May 16, 2011 @ 2:37 am
First of all, I read this article a couple of hours before reading this post, and the sentence worked just fine. Not knowing the person and his history, I parsed the sentence as referring to a single person and figured him to be the one yanked from the plane. Uncharacteristically, I didn’t even notice the opportunity for ambiguity until reading Barbara’s post. One more vote for grammatical.
(100) The Chief of the Chicago Police and a rookie lawman appeared in court.
(102) The Sheriff of Tombstone and a rookie lawman appeared in court.
Sentence (100) indicates 2 people—in most scenarios. (102) is ambiguous: 1 or 2? Probably 1.
(103) The former President of the United States and a now (or “current”) MMA superstar sauntered into the room.
(104) The former governor of California and a reborn actor got terminated by Maria today.
There is no one in our experience for which both descriptions fit in (103), whereas the combination of the two descriptions in (104) is a definite description, picking out a unique individual. [I admit that (104) sounds better without the article, but is impossible (or maybe marginal for some) with the definite article.] This illustrates that the “former/now” addition does not eliminate the ambiguity. The original sentence (maybe after truncating the yanking and airplane stuff), these sentences, and the sports sentences contributed in the thread all seem OK, so the syntax does not want to eliminate these possibilities. Putting aside pragmatic biases toward one reading or other, they are ambiguous, so we need at least two ways to map them to the semantics. [It was Jimmy Carter who was secretly working out in his basement and became an overnight celebrity, graduating from command of the Pentagon to command of the hexagon.]
(105) I saw the former governor of California and a reborn actor.
(106) The former governor of California and a reborn actor saw me.
(107) She is a friend of the former governor of California and a reborn actor.
(108) He is the former governor of California and a reborn actor.
(105-107) most naturally take the 2-people reading, but allow the 1-person reading with the right intonation. On the other hand, (108) takes an unambiguous 1-person reading. In (108), the NPs are obviously predicative. They clearly are predicated of a common individual. The variation from this is not just with subjects, but with objects of verbs and prepositions, etc.
So if you are less easy than I am with Bach’s suggestion that there is an underlying logical predication in these cases, then I suggest you might look to explanations of restrictives vs. unrestrictives for some guidance. Somewhere in the difference between “My brother who is 6’8” (as opposed to the taller ones) will help you” and “My brother, who (incidentally) is 6’8”, will help you” is the same spirit that we find in this set of possibilities. Do we have necessary information for identifying the person we are speaking of or just additional information about the person already assumed? The sentence pair immediately above demonstrates a difference in pronunciation, and it is indicated by the commas. The sentence under scrutiny (correctly?) lacks any commas that would indicate that “The (now) leader of the International Monetary Fund” is a definite description and “a possible candidate for president of France” is unrestricted added information. The intuition of the folks who thought the sentence should be re-written with an appositive indicates that this construction must be related in some close way to that construction.
Barbara Partee said,
May 16, 2011 @ 3:37 am
@Glenn Bingham: Thanks, interesting. (i) But you wrote "a now X", and the disambiguating example had "now an X". Cf "a/the former X" vs "formerly a/the X". The variants with "a now" vs "now a" etc can be used almost interchangeably as predicates, but referring NPs in subject position can't begin with 'now' or 'formerly', only with "a/the now" or "a/the "former". When "now an X" is a predicate, it is probably two constituents, not just one (I haven't worked on that, so I'm not positive. It may be rather that 'now' can indeed be an initial modifier inside a predicative NP though not inside a referential NP. That would make semantic sense.)
(ii) The relevance of appositives has indeed become clear in this discussion. What I still don't know, if I want to analyze this construction in the language of those for whom it's fine (you, and maybe me, and several others), is whether (a) to posit some invisible noun phrase to which the whole conjunction is a modifier, or (b) to posit that the second NP is really an appositive modifier on the first, but with its commas replaced by an uncomma'd "and", for some unknown reason. (I suppose the rules shouldn't be TOO simple or every language would do it.) In the case of (a), since deletion of a semantically empty NP like "a person" is much more precedented than deletion of a term referring to a specific individual not previously mentioned, the empty NP should probably be assumed to be INdefinite and non-specific rather than definite or specific; but then, as in Bach's analysis, the modifier needs to be restrictive, not appositive, unfortunately. So maybe that's an argument for (b), that it's the second conjunct only that's appositive.
If only the second conjunct can ever start with "now" or "formerly", that might be an argument in support of hypothesis (b).
Andrew said,
May 16, 2011 @ 4:09 am
This reminds me of the old joke where a headstone reading "Here lies a Member of Parliament and a man of impeccable honour" prompts a passer-by to remark that he didn't realise two people could be buried in the same grave.
Ross said,
May 16, 2011 @ 5:02 am
@Mark: that egrep expression
egrep '^(The|An*) [a-zA-Z ]* and (the|an*) [a-zA-Z ]* (has|is|was|will) '
would also match "Ann Landers and annnnn oh no will my typewriter please stop stuttering"
Not exactly likely to be in the corpus, of course. But if you want to match specifically "The", "A" or "An", you should use alternation for all of them — (The|A|An) — or else at least limit it to one "n" — (The|An?)
[(myl) Well, the NYT is well enough editing that such things don't occur often. A larger problem is that about 90% of the patterns matched are not even close to being instances of what we're looking for, because the "and" doesn't conjoin two sentence-initial noun phrases (thus "The IND subway was closed and the openings and gratings were covered with plastic" or "The scene between Becket and the Fourth Tempter is strong drama" or "A subdural hematoma heals when a membrane forms around it and the blood is resorbed").
We can easily improve precision at the expense of recall in various ways; and we can improve both less easily, say by running a part-of-speech tagger over egrep's output and selecting from the results. But if all we want to do is to sample a few examples, linguistic ore at 10% purity is good enough to get started with, since five minutes of scanning the output will yield about 50 valid cases.]
iching said,
May 16, 2011 @ 5:57 am
I wonder what the possible use of "but" instead of "and" says about this question. e.g. "The leader of the International Monetary Fund but a hopeless manager of his own household financial affairs was arrested Sunday …", or "Here lies a Member of Parliament but a man of impeccable honour", or "Fire, a good servant but a bad master, is…"
The "a" after the "but" seems expendable here as well as after "and".
Does the lack of ambiguity about number make this use of "but" less problematic than the analogous use of "and"…?
Barbara Partee said,
May 16, 2011 @ 7:14 am
@iching — That's an interesting observation. It's further evidence that these are predicative and not referential, I think, because you can never say "John but Mary went to the party" (though you can say "John but not Mary" and "John but also Mary".) For predicates, 'but' is fine: He is a violinist but a poor one.
So far, we're getting consistent evidence that these are understood predicatively, and the puzzles concern syntax and the syntax-semantics relation – where can they occur, how are they derived, and is the syntactic account such that the predicative semantics is explained.
Mr Fnortner said,
May 16, 2011 @ 9:23 am
I was searching via Google for expressions of the subject problem, when I came across this: http://tinyurl.com/3aresos, whose title is, I Wish You, a Japanese Monk and a Friend of Mine a Smooth Journey Home. I have no idea whether these are one, two, or three people. Even my inspection of the web page produced no real clues. This is very interesting, not even mentioning the possible "I wish you a Japanese monk…" crash blossom, barely averted with a well-placed comma.
Aatu Koskensilta said,
May 16, 2011 @ 6:29 pm
The construction certainly exists in Finnish, and is sometimes genuinely ambiguous. The sentence "Tunnettu loogikko ja filosofi pidätettiin." could be read as reporting either the arrest of a single individual who is a well-known logician and philosophers or the arrest of two distinct people, a well-known logician and a philosopher. Further, the choice of verb form does not resolve the ambiguity — "pidätettiin" would be used regardless of the number of people arrested.
Alexander said,
May 17, 2011 @ 8:53 pm
FWIW: While it is (marginally?) possible, (1), for a quantifier in a left conjunct to bind something in a right conjunct, I think it seems clear, (2), that the quasi-predicative second conjunct in the construction under discussion cannot be interpreted as predicating of a variable bound from the first conjunct. This might be more evidence that the second conjunct is more like an appositive than just predicate, due to the apparent dependence on reference.
(1) Every new faculty hire and his Chair attended the orientation.
(2) *Every senior professor and a chair of his department attended the senate meeting. [can't mean: Every x [ SeniorProf(x) & Chair(x) ] [ Attend.Meeting(x) ]
(3) The senior professor and a chair of his department attended the senate meeting.
Bathrobe said,
May 19, 2011 @ 5:23 am
I've heard that ballet dancers develop quite distorted feet from the practice of their craft. I would say that newspaper writing does the same to English prose.
Modern journalistic style appears to have developed a formula whereby the really important information is pushed to the start of the sentence, and various techniques are then used to add in background information. Such techniques include pre-head modifiers (or kennings, if you want to be poetic about it) and later explanatory expansions of the main information. The result is extremely distorted prose that sometimes makes a mockery of traditional English theme-rheme, given-new information structures, etc. (however you want to characterise them).
In this case the important information is that a really important person (The leader of the International Monetary Fund and a possible candidate for president of France) was arrested. The juicy bits are also included as early as possible (in connection with the violent sexual assault of a hotel maid after being yanked from an airplane moments before it was to depart for Paris).
I have to translate English news stories into Chinese as part of my work, and believe me, some of this twisted journalistic style doesn't translate well at all. In fact it is downright confusing. And when non-native speakers get hold of it, this can be especially hard to follow, particularly when time-sequences come into it. One problem with is that this style doesn't respect either identities or time sequences as they are normally used in English. The result is that unless you know what they are talking about, you can get awfully confused.
Graham said,
May 19, 2011 @ 11:51 am
This construction is discussed extensively (and in a cross-linguistic setting) in: C. Heycock and R. Zamparelli, "Friends and colleagues: plurality, coordination, and the structure of DP" in NATURAL LANGUAGE SEMANTICS, v. 13, (2005), p. 201-270
http://www.springerlink.com/content/ug10756653402256/
Bathrobe said,
May 19, 2011 @ 12:47 pm
I still prefer my explanation :)
I've seen lots of explanations of the minutiae of linguistic structure. Without having read Heycock and Zamparelli's paper (which is not available to me), I tend towards a purposive view of linguistic structure: within the limits of larger syntactic restraints, language is basically expression-oriented. The syntax is the raw material on which a particular expressive style is built. It's how the journalistic style manipulates syntax that counts. Journalistic style is unnatural from the start, but having become innured to it we don't notice how unnatural it is until it goes completely overboard with structures like this one.
(Sorry if I sound fuzzy, but syntax doesn't tell us how people write; it just analyses linguistic structures in a static way. Without understanding the dynamic of journalistic prose it's impossible to understand how this kind of sentence could come about).
Roberto said,
May 20, 2011 @ 3:27 am
Thanks Graham for pointing out my paper with Caroline Heycock. It is accessible from
http://semanticsarchive.net/Archive/jdmMTdkN/
The theory proposed in that paper is that conjunction is set-product, i.e. the union of each element in the denotation of the first conjunct with each element in the denotation of the second (note that both singular and plural individuals are treated as sets in that paper, so predicates are sets of sets). So for instance:
tall = {{a} {b}}, fat = {{b} {c}}
tall and fat = {Union({a},{b}),Union({a},{c}),Union({b},{b}),Union({b},{c})}
= {{ab} {ac} {b} {bc}}. Note that if an individual appears in both sets the union operation will yield a singleton set (here {b}). This was a feature that played an important role in the paper.
In this story a proper name, e.g. Dominique Strauss-Kahn (or a referential definite description) would be a singleton: {s}. To conjoin two proper names you then need to lift them to a predicative type ({s} => {{s}}), conjoin them with set product, then lower them ({{s}} =>{s}). This would normally lead to a plural individual, but if you really want to conjoin "DSK and DSK" you can do it: ({s} => {{s}}; {union({s},{s})} = {{s}} => {s}). So this theory predicts that as long as two names or descriptions are coreferential they will be collapsed in a single individual, no problem. This is obviously too strong (see the Clark Kent and Superman example cited above). My suspect is that the problem is detected by pragmatics (why? you went through all that lowering and lifting trouble to get a meager {s}?), and that it might be enforced by adding constraints on the lifting and/or lowering part. As many have pointed out, these cases are possible insofar they are predicative (there are various decent cases with superlatives: "the stongest man in the world and the only person who can survive a fall from a skyscreaper is here"), so maybe you can end up with a singular individual only when the lifting bit is independently licensed by the fact of being a "good predicate". But what is a predicate, pardon, two predicates, doing up there? Open question.