May 24, 2007

Journalists' quotations: unsafe in any mood

Following up on Barbara Partee's note about "Baseball conditionals", I've gotten a lot of mail full of great examples and insightful analyses, and I'll summarize it all for you as soon as I have some time. What worries me, though, is that most of the examples come from quotations in the newspapers, and experience suggests that quotations in the newspapers are almost as fictional as the dialogue that I cited from Elmore Leonard's novels.

A little searching in Google News confirmed my suspicion in this case: here's David Ginsburg's version from the AP, compare to Allan Ryan's version from the Toronto Star:

David Ginsburg, AP "He could have been a little rusty early on, and then the inning he gave up four runs I think he kind of lost his composure a little bit," Orioles manager Sam Perlozzo said. "He just did a little damage control in that situation, we're OK."
Allan Ryan, Toronto Star "He could've been a little rusty early on because it's, like, his seventh day," said Baltimore manager Sam Perlozzo. "He gave up four runs and I thought he lost his composure a little bit and wasn't able to do damage control for us."

"Wait", you're probably saying to yourself, "Perlozzo probably used similar but not identical phrasing in two different interviews. Surely journalists aren't so cavalier with direct quotations, even in sports stories."

But if you said that, you'd be wrong. As I understand it, quotes like Perlozzo's usually come from group post-game interviews, staged in a press-conference-like format. And the press reports of direct quotes from these interviews are remarkably divergent, from one another and from a careful transcription of the recorded interview (which is sometimes available on line). For an example with some discussion, see "What did Rasheed say", 6/23/2005; "Ipsissima vox Rasheedi", 6/24/2005; "'Quotations' with a word error rate of 40-60% and more", 7/30/2005.

Does this affect the linguistic point that Barbara and all my other correspondents are interested in? Well, sort of. Making a quick scan of today's baseball news, I found a counterfactual conditional with the first clause in past indicative (Phils beat Marlins in wild finish; closer Myers hurt, Sports Network, 5/24/2007):

Then, on Myers' second pitch to Miguel Olivo, the right-hander uncorked a wild pitch and immediately grabbed his right shoulder before leaving the game with a strained right shoulder.

"Something wasn't right when I threw the first pitch to Olivo," Myers said. "Then the second (pitch), I couldn't stop it. I didn't even know (where my control) was going. I don't think (my shoulder) popped or tore because if it did I wouldn't have any strength in it right now, but I have strength in it."

But instead of "if it did", the version quoted by Todd Zolecki in the Philadelphia Inquirer ("Phils Win a Disaster", 5/24/2007) has "if it would have":

After the Phillies had blown a four-run lead in strikingly improbable fashion, Myers uncorked a wild pitch to Miguel Olivo and immediately grabbed his upper right arm. He left the game.

"I felt it after the first pitch to Olivo," Myers said, seated in a chair in the visitors' clubhouse with his shoulder wrapped in ice. "I didn't even know where the second one was going. I don't think [the shoulder] popped or tore. If it would have, I wouldn't have any strength in it right now. I have strength in it. I'll just find out how it feels tomorrow. A couple days, maybe? I don't know."

That doesn't detract at all from the linguistic interest of the examples and their analysis. But it means that the examples represent one dialect and register -- the players and managers, speaking in interview style -- filtered through the perception, memory, and note-taking abilities of speakers of different dialects using a different register -- the journalists, writing in newspaper style. And maybe, in some cases, a copy editor has intervened as well.

Whatever the source and the chain of transmission that brings them to us, these examples definitely fall into consistent patterns. In this mornings's email from Barbara Partee:

Oh no, look what they have followed up with today! This has not only an indicative counterfactual, but a haplologous "if". And the third sentence in the quote shows that this pitcher does have "would" in his grammar and can use it when he wants to (David Ginsburg, AP Recap of May 23 Baltimore-Toronto game):

"McGowan (0-2) gave up three runs and eight hits in six innings. The right-hander pitched well enough to win, but his poor toss during the rundown proved costly.

"I'm not sure if I make a good throw I get him," McGowan said. "I tried to hurry it up, dropped my elbow and it just sailed on me. It'd have to be a perfect throw."

(That was the Toronto pitcher made the bad throw and Baltimore won this one!)

As far as I can tell from a quick scan of Google News, no one else reported this particular quotation from McGowan. However, a Google search for {"if I make a good throw"} turns up 1,470 examples like these:

"I knew that we should have at least gotten two," Wilson said. "A play like that, I don't think you ever expect to get a triple play. You never know. If I make a good throw, there's probably a good chance that Ty makes that play and you end up with three outs right there. But I threw low, and we only ended up with one. It really cost us."

"I jumped for the ball, I should have made the play, I have to take the heat and I'm man enough to take the heat," Perez said. "If I make a good throw, the game is tied. I didn't make a throw and that's it."

"The throw to second, that was the whole game right there," admitted Rivera. "I didn't have a good grip on the ball. If I make a good throw, he's out by far. It just got away from me."

“That whole inning was tough,” Lidle said. “Anderson looked like he was out at first on the replay, and if I make a good throw to second it’s an easy double play. It was just a frustrating inning.”

Interestingly, leaving out the "if" only turns up one paratactic example:

"I make a good throw, we should be playing right now," Guillen said. "I make a bad throw. That's it."

And Jose Guillen, from the Dominican Republic, is probably not a native speaker of English.

Does this mean that baseball players and coaches don't use the if-less form much, in this particular context? Or does it mean that sportswriters and editors usually add "if" to paratactic conditionals in such cases? I'm not sure.

[Update -- Language Hat writes:

You write "That doesn't detract at all from the linguistic interest of the examples and their analysis." But I think it does. It's not necessarily the case that sportswriters are simply rewriting utterances in their own dialect and register; they may well be rewriting them according to their ideas of how baseball people "should" talk. This phenomenon is inescapable when reading pre-civil-rights writing by white people trying to represent black speech, and to the extent that it occurs it invalidates any linguistic interest aside from a peripheral curiosity about imaginary dialects.

He's basically right, as usual. But in my opinion, it's a bit more complicated. The syntactic patterns in journalistic quotations can still have several sorts of interest: they may suggest hypotheses to test in conversational transcripts from more reliable sources; and when we share the patterns in question -- in this case, most varieties of paratactic conditionals are well within my own norms for informal conversation -- our reactions to the journalistic examples can lead us to see new things.

At least in the sports context, the cases where I've checked journalists' quotes against a recording show mainly (a) extensive ellipsis and reordering, (b) random substitutions and paraphrases apparently caused by memory errors or careless note-taking, and (c) partial editing towards standard-English norms. Unlike the case of "imaginary dialect creation", errors of these kinds will leave some interesting cases of paratactic (and other non-standard) conditionals more or less intact, or at least will not create too many imaginary examples. ]

[Update #2 -- another reader writes:

When I worked as a managing editor at a small paper with well-known sports coverage, it was routine for the sports reporters to clean up the grammar of athletes and coaches they interviewed, particularly coaches. (As long as the grammar-clean didn't change the meaning of the quote.) They got better repeat access, particularly with the coaches, when they did so. Even athletes who TALKED like no-necked neanderthals didn't like to be REPORTED as talking like no-necked neanderthals. My understanding is that this is reasonably common in the sports reporting world. In fact, one way sports reporters took minor vengeance on figures they didn't like was to report their bad grammar exactly as spoken.

On a related note, re: Jose Guillen:
Multi-lingual reporting is a tricky issue with a lot of ethical implications; international reporters and city beat reporters on immigrant-heavy beats could talk really interestingly about that, since there's a lot of debate about when you should say, "'He shoved me,' she said, in Spanish." vs. "'He shoved me,' she said." with no indication of language used. It's a particularly hot topic in cities where immigration is a tense issue and choosing to identify the language spoken when reporting a crime story could even induce jury pool bias! This is not the kind of beat I ever covered, so my knowledge of the topic is fairly general, but I did listen to colleagues debating it extensively.

My understanding is that different papers have different official policies about how much "cleaning up" of what type is possible and/or mandatory, but that the rules are not very uniformly enforced, even at places like the New York Times.

One crucial point here is that things like evaluation of someone's accent, and even of their general perceived social status, may have a lot to do with how much 'cleaning up" a writer feels is appropriate. And in some cases, there is even a sort of "dirtying down", which can take the form of "eye dialect" mis-spellings of standard pronunciations like "sez" or "an'". I say this as someone whose informal conversation could very easily be presented as that of a "no-necked neantherthal". Of course, I'd probably be amused or even pleased if someone decided to go that way in writing about me.]

[Update #3 -- Jim Lewis wrote:

There's another reason, I think, why reporters regularly clean up quotes, and that's simply that people rarely speak in complete sentences, and often not in coherent ones. If you look at a word for word transcription of normal conversation (as I would imagine you have), it's full of sentence fragments, stops and restarts, ums and uhs, interruptions, switching tenses midstream, and so on -- all sorts of noise.

Jim gave more details about the incoherence of spontaneous speech rendered as text, but editing out filled pauses and false starts is not what we're talking about here. In one of the posts that I cited at the start of this article ("Ipsissima vox Rasheedi", 6/24/2005), I gave a specific example of how to clean up such disfluencies honestly -- which none of the reporters whose stories I quoted had bothered to do. I don't mean that they intended to communicate something that is false, in those cases, just that they didn't exert the minimal effort required to create accurate (if appropriately edited) direct quotations. Of course, sometimes the effect of such "approximate" quotations is to communicate a falsehood, and perhaps that's the intent as well, I don't know -- for an example, see "'Approximate' quotations can undermine readers' trust in the Times", 8/27/2005.

Chris Phipps writes:

Reading your post about Journalists' quotations on Language Log reminded me of a dissertation one of my colleagues wrote for the U. Buffalo linguistics department: Krainz, Stacy - Professionalism vs. Audience Appeal in News Analysis Discussions: A Journalist's Dilemma. (9/02)

One of the sub-topics Stacy covered was the differences between "direct" quotes and what she called "constructed dialogue" or "the rendering of what another speaker has said from a first person (direct speech) perspective" which she found prevalent in political roundtable discussion (which I suspect, is close in style to sports reporting-speak).

I think her basic point was that reporters try to create a sense of "interpersonal appeal" so they unknowingly modify the quotes of others.

If you're interested in following up, here's Stacy's GURT paper on the topic:

Prof. Krainz's paper focuses on the methods that broadcast journalists use to generate and maintain audience interest in round-table discussions. One of those methods is what she calls "constructed dialogues" -- but these are more like puppet-theater personifications of the positions of others, as in an example she quotes of a reporter re-imagining Ken Starr's request to a judge as "So he goes to the judge in the Paula Jones case, and he says to her 'Look, stop those depositions. Shut em down. My case is more important than that case. ...'" Her discussion is interesting, but I don't think it's relevant to the examples under discussion, where the reporters are not supposed to be acting the part of someone in the news, but rather claim to be supplying a direct quote. Of course, you could argue that this is the whole problem -- the journalists involved actually do think that their job is to make the whole story up. But I don't believe this -- they're not dishonest, they're just lazy. Or rather, their culture doesn't give them any incentive to take the trouble to be accurate, so they take the easy route of passing half-remembered paraphrases off as direct quotations.]

Posted by Mark Liberman at May 24, 2007 08:19 AM