Matrix verbs as "ghostly adverbials"?
« previous post | next post »
Last week, fev at Headsup: The Blog featured an unusual referential tangle ("March of the pronouns", 4/24/2009):
A 30-year-old Pontiac man is in the Oakland County Jail and facing felony charges after authorities said he rammed a man’s car after finding his wife in the backseat with him.
Beyond the amusing combinatorics of reference (two possible antecedents each for his and him, plus the question of whose backseat it was), fev notes that
"Authorities said" isn't really part of the subordinate clause. It's acting like the British "claim quotes" — sort of a ghostly sentence adverb, reminding us that there's some sort of privileged source to which (or whom) the otherwise potentially libelous assertion can be pinned.
In this particular example, it's possible that the authorities issued a statement before the guy went to jail, and it's likely that "facing felony charges" involves something that could be construed as an authoritative statement early in the process. But in general, I think that fev is right: there are matrix sentence frames that indicate how you know something, or how you feel about it, while being otherwise transparent to the semantic relationships expressed by the structures around them.
Courtesy of a search of newspaper archives, here are a couple of examples of "authorities said" in after-clauses where it's almost certain that the authorities told the story long after the events whose chronological order is being explained:
A man was tased at SeaWorld on Tuesday afternoon after authorities said he failed to pay admission and fought with security personnel.
Three children were drowned tonight in the Scudog River after authorities said their mother, identified as a Mrs. Wallace, tossed them into the swirling waters and then leaped in after them.
But these examples might also be analyzed as parentheticals without the punctuation, since there is no that: "A man was tased at SeaWorld … after (authorities said) he failed to pay admission …"
And it's somewhat suspicious, or at least disappointing, that a search of the Proquest Historical Newspaper archives turns up no instances of the string "after authorities said that", although there are 58 instances of "after authorities said".
It's easy to find examples where a matrix sentence frame is used as if it were a modal or epistemological adverbial of some sort. But some of these are cases where the matrix might actually be a parenthetical (as in the case of the "authorities said" examples above), and some of them are cases where logic can be restored by treating a subordinate clause as within the scope of the matrix verb, as if by shifting the order of clauses:
Unless it's shipped soon, I'm afraid that I'll have to cancel my order. (= I'm afraid that [I'll have to cancel my order unless it's shipped soon].)
If there's a swine flu pandemic, she says that hundreds of millions will die. (= She says that [hundreds of millions will die if there's a swine flu pandemic].)
Can anyone find a genuine example where neither of these dodges will work?
This might be a sentence of the form "X after police said that Y", where it's clear that X and Y both took place long before police said anything about the matter. Or it might be "X, because I'm afraid that Y", where the relationship between X and Y is entirely independent of my emotional states.
Chris Straughn said,
May 1, 2009 @ 8:56 am
Yet another reason English needs evidentials. Leave the police statement out all together and encapsulate all that meaning in a tiny morpheme.
Ryan Denzer-King said,
May 1, 2009 @ 9:50 am
I'll second that, and add that English needs obviation too.
I got lots of ghits for "X after police said Y" in situations where the order of events is clearly X, Y, police statement.
"Man arrested after police said they found drugs in his purse"
"A 23-year-old Blauvelt woman was arrested early yesterday morning after police said she kicked officers and damaged a cruiser"
Dierk said,
May 1, 2009 @ 9:56 am
Chris, why not just do the appropriate and enclose 'authorities said' in commata. Never understood why punctuation is used so … sparingly in many English texts, particularly commata:
Once upon a time every journalist, even the lowliest police report copier, had a novel in his drawer; unfortunately today they have trouble using punctuation they way it has been used for ages by novelists.
[(myl) Wow. Not many of those novels, in or out of the drawers, have used the Greek plural of comma. The OED mentions it in the entry for comma, calling it the plural form "as L. or Gr.", but gives no examples, perhaps because its readers found none; and Literature Online finds only one instance in its index of more than 350,000 works — in a Latin passage attributed to Anonymus, Siue Pessimus omnium Poëta:
Quistamen hic Mystes tragico qui Fulmina abore
Torquet? Num doctus? Cortè. Nam Metra Catonis
Quattuor edidicit, totidem quoq; Commata Tullî.
And the only examples of commata that I can find on the web are in phrases like "cola and commata", referring not to marks of punctuation, but to phrasal divisions ("There are no marks of punctuation, but the skilled reader was guided into the sense by stichometric, or verse-like, arrangement into cola and commata, which correspond roughly to the principal and dependent clauses of a sentence.")
So I think you get the prize for Pretentious Comment of the Week. ]
Ellen said,
May 1, 2009 @ 10:52 am
What are evidentials? And what is obviation as used here (as something a language can have)? (Commata I understood. :))
[(myl) For evidentials, try here. As for obviation, the Wikipedia entry for Ojibwe Grammar explains that "Like most Algonquian languages, Ojibwe distinguishes two different kinds of third person, a proximate and an obviative. The proximate is a traditional third person, while the obviative (also frequently called "fourth person") marks a less important third person if more than one third person is taking part in an action." ]
Richard said,
May 1, 2009 @ 11:52 am
I think the chronology in the following will serve to illustrate what Mark wants, in that the whole event took place long before the authorities had a chance to say anything:
(My emphasis.)
In any case, the loss of parenthetical status (and before that disagreement between speakers about whether something of this kind is or should be parenthetical) is quite typical of locutions becoming discourse markers, i.e. pragmaticalisation; this regularly involves a process of weakening and integration (and is in many ways like grammaticalisation).
There are other examples, even just in English ('say' and 'please' spring to mind), of independent units originally with the syntactic status of clauses that have 'become unparenthetical' (and thus apparently illogical, because they aren't or can't be interpreted semantically where they appear in the syntactic structure) in this way. (Some of them are still written either often or occasionally with commas, but I suspect there is often no phonetic realisation of the original parentheticality of, say, 'please'.)
There is much interesting work still to be done on the mechanisms by which changes like these happen – and indeed on the phonetics and phonology of 'comma intonation' (I'd be delighted to see any bibliographic references people can suggest on this!). But they are, I think, definitely changes, and as such there will be variation: I suspect many people now will not pronounce these kind of phrases as if comma'd off from the rest of the sentence, and so I'm not sure we can expect they will write as if they do.
ray said,
May 1, 2009 @ 12:17 pm
I think this is just a relic of the CYA style of writing, in which journalists are instructed never to write that a person actually committed some illegal act, lest their newspaper be sued for defamation. Instead, they are taught to insert "allegedly," "police said," or some other weasel phrase in order to deflect the accusation of, well, accusation. Then, in a fit of wanting to smash all the info, including weasel-speak, into a single sentence, they write the sentences on display here.
Bryn LaFollette said,
May 1, 2009 @ 12:43 pm
My thinking was along similar lines to that of Ryan Denzer-King, but from the direction of wondering if there were and newspaper publications in languages with Switch Reference that could be searched for similar constructions to this. I figure this would be a great way to determine which clause was in fact being treated as the matrix! Obviation, which often seems to go hand-in-hand discourse-structure-related stuff like Switch Reference would be pretty useful in these cases too, it seems, since in text of the length of newspaper articles, you'd usually expect to see the "main character" marked as proximate, and so it would be interesting to see whether the statement-making-authority figure in the above examples was marked as obviative.
Just for the record, I'm ok with English not having Obviation or Switch Reference, though I have certainly had situations where they would have been useful in relating a story.
Bill Walderman said,
May 1, 2009 @ 1:31 pm
"So I think you get the prize for Pretentious Comment of the Week."
Maybe Dierk happens to be a native speaker of a language other than English in which the plural of "comma" is regularly and unpretentiously "commata" (or "komma"/"kommata") and he simply made a mistake.
Bill Walderman said,
May 1, 2009 @ 1:45 pm
If you had tried searching for "kommata" you would have found about 2,750,000 hits.
Here's one that comes up early in the search, in which a person who maintains a website on writing correct German tells us that "kommas" is the more common form but, she says, the "foreign" form "kommata" is also permitted. A commenter notes that "kommas" is no less a foreign plural form in German than "kommata."
http://erfolgreich-schreiben.typepad.com/schreiben_im_beruf_/2005/11/kommas_oder_kom.html
Dierk said,
May 1, 2009 @ 1:56 pm
Mark, nothing to do with pretension but with me being German. Not only by birth but also by education. And we learned the Latin plural in school, later used it at university. Comes quite naturally to me – admittedly I also chose Latin and Ancient Greek as subjects in school.
Dierk said,
May 1, 2009 @ 2:00 pm
Forgot [and can't edit]:
I am over 40. Younger Germans may be more comfortable with 'commas', the pluralisation following German rules, since it has become a standard form after I left school.
[(myl) OK, the award is withdrawn. (Insert joke about urine sample indicating recent use of illegal punctuation terminology…)]
Charles Smith said,
May 1, 2009 @ 2:36 pm
In regards to obviation, a classmate and Ojibwe scholar once told me that, in Ojibwe, if obviation becomes too difficult at some point (e.g. if pluralization is necessary or with terms for family members), then the obviation is just switched around (so proximate becomes obviative and vice versa). This is mainly done by storytellers according to what he told me. So, it's not as much of who's the most important 3rd person to the discourse (although that's the main reason), it's also distinguishing between two 3rd persons and how hard it is to form the obviative if a lot of other stuff needs to be done, too.
Nathan Myers said,
May 1, 2009 @ 3:06 pm
Dierk's closing paragraph citing a mythical golden age may have biased Mark against him; it chafed me, too. But Dierk's first sentence seems to me to deserve an answer, particularly because I had wondered it myself.
Rubrick said,
May 1, 2009 @ 4:20 pm
It's a pity 'commata' isn't in common use in English, as it would form the basis of a very solid knock-knock joke. (Constructing said joke is left as an exercise to the reader.)
Bill Walderman said,
May 1, 2009 @ 4:28 pm
"A 30-year-old Pontiac man is in the Oakland County Jail and facing felony charges after authorities said he rammed a man’s car after finding his wife in the backseat with him."
Dierk's point about commas is interesting. The reason why it's not felt to be necessary to set off "authorities said" by commas is that when the sentence is spoken, "authorities said" is not set off by pauses. Doesn't the intonation pattern support the theory that "authorities said" is no longer a parenthetical but rather has been transformed into a "ghostly adverbial," a "matrix sentence frame[] that indicates how you know something, or how you feel about it, while being otherwise transparent to the semantic relationships expressed by the structures around [it]?"
John Lawler said,
May 1, 2009 @ 5:09 pm
Probably it's both, to different people, or to the same person under different circumstances.
Depends to a certain extent, for instance, on how formulaic the particular expression is. The more an individual automates their use of a given phrase, the more likely it is to have several category memberships, as the speaker finds new uses for it or repurposes old ones, or simply misinterprets its original structure. Nothing new here, really; this is ordinary language change (fascinating, granted — even ironic in places, like some instances of "I regret to inform you that S"). But normal.
And commas, no matter how much we love'em, are phenomena of written language, not spoken. Even though I adhere to the theory that effective use of commas in writing does represent intonation contours in one sense or another, it's still the case that either the majority or a large minority of English speakers are either outright illiterate or, let's say, don't write enough, or well enough, to make use of commaticity theory. Spoken intonation usually does make things less ambiguous; written commas maybe not so much.
Mary Bull said,
May 1, 2009 @ 5:41 pm
Bill Walderman's argument from spoken speech is interesting. However, in my experience, pitch (meaning, the frequency of sound waves) can also stand in for commas. I've just now tried reading the sentence under discussion aloud, and I find that the pitch of my voice drops on the words "authorities said." I find, also, that I speak those two words more softly than the rest.
So for that reason, as well as for clarity, I think setting off "authorities said" as a parenthetical insertion is the best way to go.
As soon as I finished reading Dierk's comment this morning, I thought he had zeroed in on the heart of the matter. I wasn't distracted by "commata" for "commas." And I found his ironic closing paragraph quite telling. But then, I'm a couple of years past 80, and I remember how well even small-town newspapers in the U.S. used to be written and edited.
Mary Bull said,
May 1, 2009 @ 5:54 pm
I wrote and submitted my first comment before reading John Lawler's. And I see that he and Bill Walderman are using the word "intonation" to mean "pitch," whereas I took it to mean the general rising or falling direction of a phrase or clause.
So we're all three talking about the same phenomenon.
However, I stand by my judgment that, in the particular case under discussion, a pair of commas setting off "authorities said" resolves the ambiguity.
Russell said,
May 1, 2009 @ 7:38 pm
Not sure why, but I think putting commas or parens around "X said" seems a bit strange. I know it happens, but somehow the commaless version seems better.
Also, if this were parenthetical, would we expect it to take scope only over the second part, and never over both? (probably Haj Ross's paper on slifting [sentence-lifting] has something to say about this, but it's not in hand ATM)
And do the ghostly adverbials behave the same way? The closest I could find to wide scope is this:
Of course the police are still saying that so-and-so was arrested and faces charges in the original sentence, but the part that a journalist needs to qualify is the part in the after-phrase.
Finally: you can also get present tense: "…after police claim/say/…" Are there other contexts for present tense in after-phrases when the main clause is past tense?
Russell said,
May 1, 2009 @ 7:46 pm
Whoops, I missed the drowning and tasering examples, which are basically like the Jason Lee Clough one. Still, one wonders if "…was arrested after police say" and "…were drowned after police say" are interestingly different.
Ellen said,
May 1, 2009 @ 8:36 pm
I'm thinking that this structure is something particular to newswriting. Setting it off by commas would make it sound like a different kind of writing. If I read something like that, I'd read it like there were commas there. But I can't imagine a news anchor reading it that way.
Ellen said,
May 1, 2009 @ 8:38 pm
Read it aloud, that is.
Bill Walderman said,
May 1, 2009 @ 9:04 pm
'Nam Metra Catonis
Quattuor edidicit, totidem quoq; Commata Tullî.
'And the only examples of commata that I can find on the web are in phrases like "cola and commata", referring not to marks of punctuation, but to phrasal divisions'
Interestingly, and irrelevantly, the neo-Latin example cited, which is apparently from an obscure Jacobean work by Thomas Heywood entitled "An Apology for Actors" (London 1612) uses "commata" in the sense of phrasal divisions ("He's learned four lines of Cato and as many Ciceronian periods"), not punctuation marks. Isn't it amazing what you can find on the internet with Google?
http://books.google.com/books?id=nqCX5t-CiksC&pg=RA1-PA7&lpg=RA1-PA7&dq=%22metra+catonis%22&source=bl&ots=Recw2ign4r&sig=j-q93LcJo-7fxlAeQC61FRSlHVM&hl=en&ei=aZj7SZjTJZPkNcbEgesH&sa=X&oi=book_result&ct=result&resnum=1#PRA1-PA7,M1
"Pessimus omnium poeta" is how Catullus describes himself with self-deprecation in a thank-you poem to Cicero (No. 49). But the excerpt quoted above isn't from Catullus.
Spectre-7 said,
May 1, 2009 @ 9:30 pm
*knock knock*
Who's there?
Commata.
Commata who?
Commata dat house so I can break yous face.
…
Thank you. I'll be here all week.
pfc said,
May 2, 2009 @ 1:06 am
@Charles Smith
At least in Ojibwe, the verb form is switched from direct to passive (or vice versa) at the same time. According to this site, verb form switching is preferred:
It is considered better to use passive voice if needed, than make changes between chosen proximate and obviative persons in a narrative depending on who is performing the action.
I think this is a really cool feature to help resolve ambiguity that I never would have thought of before studying that language. But still, nothing beats indexed pronouns in ASL for avoiding ambiguity!
Vasha said,
May 2, 2009 @ 1:53 am
Prompted me to spend a few minutes rewriting the tangled sentence, which is not as interesting as everyone else's comments, but I like my result well enough to submit it here: "…facing felony charges, having rammed a car after finding his wife in the back seat with the owner, authorities said."
Faldone said,
May 2, 2009 @ 1:53 pm
Say what you will about the sentence about the Pontiac man, I see no temporal problems. Presumably he was in jail facing felony charges at the time the sentence was published, which was after the authorities said what they believed he had done. Likewise, the statement by the authorities undoubtedly came after whatever actions by the Pontiac man provoked the authorities to make their claim.
Private Zydeco said,
May 3, 2009 @ 4:46 am
the order of events denoted in the textual example in question
seems in fact to be thus: first, said Pontiac man committed (or
didn't committ) the acts described in the allegations and legal
charges he now faces and, after arrest, was booked and brought
to the county jail. The latter events being dependent on jotting of
police reports, field-questioning, and/or preliminary hearings, it
stands to reason to assert that a policeperson or group thereof
"saying", or averring, that those allegations are true acted some-
what performatively upon the processes of booking and jailing.
peter said,
May 3, 2009 @ 8:36 am
In your post of 29 April, Mark, about the death of Jack Good, you cite a transcript of a spoken interview with Good published in Statistical Science in 1996. In that interview, Good uses "lemmata" as the plural for "lemma".
Aaron Davies said,
May 3, 2009 @ 11:45 am
you'll also see "schemata" fairly frequently. i wonder how many people who know what "stigmata" are know that they're plural?
Bill Walderman said,
May 3, 2009 @ 11:59 am
Let me try to restate more clearly a comment I posted previously. The intonational contours of matrix sentence frames such as "authorities said," when read aloud or spoken, are different from the intonational contours of parentheticals, and this difference is perhaps a clue that there is a difference in the underlying syntax.
Private Zydeco said,
May 4, 2009 @ 1:03 am
Backseat driving had invariably proved itself hazard enough
to all motorists – not only to those with passengers bent on
appropriating command of what vehicles they should chance to
hitch rides in – before car operators themselves, spouses in
tow, no less, began undertaking the practice.
Matt B in LA said,
May 5, 2009 @ 9:39 pm
@Bill Walderman
"Let me try to restate more clearly a comment I posted previously. The intonational contours of matrix sentence frames such as "authorities said," when read aloud or spoken, are different from the intonational contours of parentheticals, and this difference is perhaps a clue that there is a difference in the underlying syntax."
Ah, you said what I was thinking. It's more attributive than parenthetical, but I don't know if that means it doesn't need commas. It's not qualifying one aspect of the sentence; it's qualifying the entire sentence. The sequencing would seem to make more sense with the attributive at the end of the sentence, though:
A 30-year-old Pontiac man is in the Oakland County Jail and facing felony charges related to ramming a man’s car after finding his wife in the backseat with him, authorities said.
Much more work is required to make the pronouns "sensible."
May Links « Literal-Minded said,
May 16, 2009 @ 12:53 am
[…] finish, a few from Language Log. Mark Liberman expands on a brief post on Headsup, regarding sentences like, "A 30-year-old Pontiac man is in the […]
Karen said,
May 16, 2009 @ 8:54 am
The problem with putting it at the end, as in "A 30-year-old Pontiac man is in the Oakland County Jail and facing felony charges related to ramming a man’s car after finding his wife in the backseat with him, authorities said." is that now the authorities are saying he's in jail and facing charges, when all that is really being modified is the "ramming a man's car" part of the sentence; the "in jail on charges" part is being justified by the "authorities said" part, it's not up for doubt.