Sea turtles rescued from Gulf spill released

« previous post | next post »

A most perfect garden path headline. It’s interesting that this one depends on our automatic processing of headlines with their own syntax.

The classic garden path sentence, “The horse raced past the barn fell”, is a full sentence and not a headline, and in that one the garden path is created by our preference for initially interpreting “raced” as a main verb; only when we hit “fell” do we backtrack and reprocess “raced” as a passive participle and “raced past the barn” as a modifier of “horse”.

In the sea turtles headline (Yahoo Science News, July 16, 2010), “rescued” is a passive participle in both the initial and final parsings – we don’t mistakenly interpret “rescued” as a main verb in the past tense, because we are not inclined to think that the sea turtles rescued anything, and the “from” phrase further makes it clear that the turtles were the rescued ones, not the rescuers. So what’s the garden path about?

It comes about because among other standard headline conventions, “are rescued” or “were rescued” will always be written just as “rescued”. So really the temporary ambiguity and its resolution is almost identical to that in the “horse raced” example: We initially take “[are] rescued” to be the main verb, and only when we hit “released” do we realize that the main verb must be “[are] released”, and that “rescued” is just a participle (with no understood “are”), and “rescued from [the] Gulf spill” is a modifier of “sea turtles”.

Every language will have its own constructions that help to set up garden path sentences, and probably every language used in newspapers will have its own conventions for headline syntax (and different newspapers or online news services may have their own as well – I recall some recent Language Log discussion of Reuters’ use of “says” + finite verb (“says finds” etc.), but I can’t find the entry). I would hazard a guess that English provides more sources for garden paths than, say, Russian, because Russian has such rich morphology, hence fewer grammatically ambiguous lexical forms.



16 Comments

  1. John Cowan said,

    July 16, 2010 @ 8:42 am

    From what I understand, Headlinese as a separate dialect with its own syntax is pretty much unique to English, though headlines in other languages certainly have their own lexicons, similarly driven by the need to save space.

    But I must take issue with your implicature that headlines are not, or need not be, sentences. They most certainly must be sentences in Headlinese, and indeed the development of the modern headline in the closing years of the 19th century was the transition from the use of mere NPs, like "The War", to actual sentences that told the story.

    Some sentence headlines, though, did appear much earlier, as in the entirely modern headline in the 1781 Boston Gazette after the Battle of Yorktown: "CORNWALLIS TAKEN!". This was apparently drawn directly from Washington's dispatch, as Dispatchese was much like Telegramese, which in turn is much like modern Headlinese.

  2. Mel Nicholson said,

    July 16, 2010 @ 8:57 am

    It took me a few rereadings to get the "wrong" parsing. I wonder if that is because I was reading it as a blog title rather than a headline.

    @Cowan Using a specialized syntax for headlines is not unique to English. When I was in Argentina, it took quite an effort to learn to parse headline-syntax as it differed from the syntax of the articles and spoken language.

  3. Morten Jonsson said,

    July 16, 2010 @ 9:15 am

    I'm having a hard time seeing any ambiguity. As Professor Partee points out, 'we don’t mistakenly interpret 'rescued' as a main verb in the past tense." But if we don't make that mistake–if we have to force ourselves to read the sentence the wrong way–then how can there be a garden path?

  4. Barbara Partee said,

    July 16, 2010 @ 9:45 am

    @John Cowan – Thanks, I stand corrected. I agree on reflection that headlines are indeed all sentences (in Headlinese) — I was unconsciously not distinguishing headlines from titles and headings.
    @ Mel Nicholson — maybe having as a blog headline, especially in this blog, makes a difference indeed. When I read it, it absolutely garden pathed me, so strongly that I imagined it would do so universally to everyone.
    @Morten Jonsson. Well, what if the headline had been "Sea turtles rescued from Gulf spill"? Wouldn't that be perfectly understandable? That's the garden path I was on: I expected the headline to end right after "spill", but then there came another word, "released", so I had to backtrack (unconsciously, automatically, and fast).
    In the first, shorter, headline, "rescued" is headlinese for "are rescued", and that's the main verb (more carefully, it's a tensed auxiliary plus a passive participle — in my entry I referred to whatever included the tensed element as the main verb). In the actual headline, "rescued" is still a passive participle, but now there's no understood "are", and "rescued from …" is a modifier. In the shorter headline I had expected, the subject of the sentence is "turtles". In the actual headline, the subject of the sentence is "turtles rescued from the Gulf spill".

  5. Ellen K. said,

    July 16, 2010 @ 9:59 am

    Checking actual newspaper headlines, I saw a few that had noun phrases rather than sentences, but, except one (which may be considered to have an implied "there is"), these were all features rather than news articles.

  6. Terry Collmann said,

    July 16, 2010 @ 10:34 am

    Ellen, you've spotted the convention. It's almost automatic in journalism that news heads are sentences (in headlinese, with verbs either explicit or inferrable) while feature heads are labels.

  7. John Roth said,

    July 16, 2010 @ 10:52 am

    I have to agree with Mel Nicholson. I can't see it as a garden path.

    I can certainly see the interpretation of 'rescued' as a main verb if the final verb wasn't there, but that's not the way I parsed it initially. Possibly my processing window is wider, so that I didn't commit to a parse for rescued before I saw the final verb?

  8. George said,

    July 16, 2010 @ 11:45 am

    @John Cowan

    "From what I understand, Headlinese as a separate dialect with its own syntax is pretty much unique to English . . ."

    In Standard Arabic, the unmarked declarative sentence form is VSO. However, headlines are typically SVO.

  9. Boris said,

    July 16, 2010 @ 1:05 pm

    For me, at least, the "released" does not force a reanalysis of the previous sentence. Although if that word were not there I would also have no problem parsing the headline with the implicit "were" as the main verb. I think the difference is that with the horse example, the word "raced" changes meaning depending on interpretation, while "rescued" does not.

    In other words, you cannot straightforwardly decompose the sentence into "The horse raced past the barn. And then it fell." This reading is not possible in English. You have to go back and say "The horse was raced past the barn (by someone)". Whereas "The turtles were rescued. And then they were released" is exactly what the headline means.

  10. nonpoptheorist said,

    July 16, 2010 @ 1:05 pm

    SVO aside, the fact it says "Gulf spill" rather than "Gulf-spill" or "Gulf Spill" directly caused an issue in my brain where I associate "spill" with "released" and it then led me to wonder how you "spill release" a box of turtles. Yes, a box, because I envisioned an image of rescue workers clutching boxes of turtles and performing a "spill release" action to hurl the turtles to safety. Although just how safe that maneuver would be to turtles then started to worry me.

    I then moved on from the headline to read the full text of the discussion, thought some more and maybe the issue here regards garden path headlines is that there are two words used for the object with the first being capitalized and that this would be an additional complication to acquiring comprehension.

  11. Ellen K. said,

    July 16, 2010 @ 1:37 pm

    I keep wanting to read it as "Gulf" being the full complement of "from", then reading "spill" as a verb (grammatical, if not sensical), and leaving "released" attempting to be a noun and failing.

  12. blahedo said,

    July 16, 2010 @ 1:57 pm

    I wonder if some of the people not seeing the garden path have induced a different grammar for headlinese, where the usual top-level constituent is an NP rather than an S. Even with "Cornwallis taken" one can take it as a noun with a postmodifier, lining up with a convention more often seen in book titles real and fictional like "Brideshead revisited", "Don Juan triumphant", and so on. (And that may suggest an etymological route by which the headlinese constructions, which I do normally read as abridged full sentences, would have originally come about.)

    If you have the S reading of headlinese, then the VPs are the heads of the S and you process as: "[Some] Sea turtles [were/have been/had been] rescued from [the] Gulf spill…", and then you hit "[were] released" and you're stuck. However, if you have the NP reading of headlinese, then a VP is a reduced relative modifier of the NP, of which you can have more than one, and you process: "[Some] Sea turtles [that were/that have been/that had been] rescued from [the] Gulf spill…", and then you have much less trouble adding "…[and were/and have been/and had been] released." (Not necessarily suggesting that we actually fill in the words as such, just using them to signal the Standard English sentences corresponding to the parses I'm attributing to the Headlinese.)

  13. Coby Lubliner said,

    July 16, 2010 @ 6:32 pm

    George said: "In Standard Arabic, the unmarked declarative sentence form is VSO. However, headlines are typically SVO."

    In Mexican Spanish its just the reverse, e.g. (from today's Excelsior) "Presenta el West Ham a Pablo Barrera".

  14. Michael Moncur said,

    July 17, 2010 @ 2:50 am

    I didn't walk down the garden path either, and I think the reason in my case is that I implied a comma that would have made a different headline:

    Sea turtles rescued from Gulf spill, released

    …which is a legitimate compound sentence in Headlinese. The turtles were rescued, then they were released.

    I re-parsed after I got to the end and found the comma missing, but the meaning is essentially the same.

  15. RW said,

    July 21, 2010 @ 5:09 am

    Sorry that this is not really on topic for this thread, but I just found a headline I could make no sense of at all. I've tried to understand what it means several times and now it's giving me a headache.

    Jail orbits moon man document thieves

    What am I missing?

  16. Gordon P. Hemsley said,

    August 13, 2010 @ 1:44 am

    @RW I think what you're missing about that crash blossom is that it also involves a pun, which, honestly, isn't all that funny or intuitive.

    Reading the article, it appears that thieves that stole documents about "moon man" (Neil Armstrong) were arrested and are likely to go to jail. Thus, this is the intended parsing:
    [ [ Jail ] [ orbits [ [ [ moon man ] document ] thieves ] ] ]

    However, I don't quite get the idea of jail orbiting thieves. Unless the thieves have some sort of gravity that is likely to eventually have the jail crash down on them. But still, it's a stretch, all at an attempt to have a humorous headline.

RSS feed for comments on this post