The wife and mother of two men killed in a fire

« previous post | next post »

Local radio station WFCR on Thursday, October 11 started a report with a sentence that gave me a big double-take:

“The wife and mother of two men killed in a fire in Northampton has filed suit …”

And the next morning, October 12, I saw almost the same words in the local paper, the Hampshire Gazette:

Photo caption:

Alleged arsonist Anthony Baye has been sued by Elaine Yeskie, the widow and mother of the two men killed in a Northampton house fire he allegedly set.

Beginning of story:

The widow and mother of men killed in a house fire in 2009 filed a wrongful death lawsuit Wednesday against alleged fire-starter Anthony P. Baye. Elaine Yeskie, 77, is seeking monetary and punitive damages against Baye, …

The version under the photo caption makes the description an appositive phrase, so we already know that it’s a description of one person. But the beginning of the radio story really took me by surprise and made me grab my pen. I feel subjectively sure, though I could of course be wrong, that I could never say that that way. All the ways I could express it take more words; about the shortest acceptable version I can find is “The wife of one and mother of the other of two men killed in a fire …”

I’m sure we can all interpret their versions once we get over our surprise. But I’m curious whether anyone thinks that English (or any other language you know) really allows “the wife and mother of two men killed in a fire” to mean what they meant. Has anyone seen other attested examples? And has any linguist ever discussed it, because I can’t see any way to get their intended meaning of the whole phrase from the meanings of its parts. (I’m worrying about what semanticists call ‘the principle of compositionality’: the meaning of a whole is a function of the meanings of the parts and the way they are syntactically put together.)

My own hunch: there’s nothing wrong with the syntax – it’s no different from “A friend and neighbor of the two men killed in the fire” – but the compositional meaning of that one is that the person spoken about is a friend and neighbor of both of the men, not a friend of one and a neighbor of the other. And another well-formed possibility (these types have both been studied) is “An aunt and uncle of the two men killed in the fire”, where now we’re talking about two people, but still one is an aunt of both men and the other is an uncle of both men. I don’t think it can have the meaning of an aunt of one and an uncle of the other. (Or can it? I could be wrong.)

I think they’re using a possible English sentence with a meaning it can’t really have, but getting away with it (perhaps also in their own minds) because we can figure out what they must mean.

I’d love to have a theory of sentence production that could explain how they came up with it, if indeed I’m right that it’s not really a combination of form and meaning that our unconscious grammar really legitimates.



48 Comments

  1. B.T.Carolus said,

    October 16, 2012 @ 9:41 pm

    I have definitely seen women and men described as "the wife and mother of two" or "the husband and father of one" in news articles (here's one http://www.krem.com/news/northwest-news/Vehicle-strikes-kills-pregnant-wife-and-mother-of-two-122784869.html). I think that this might have come out of an attempt to over-explain that phrase because either the writer or editor was trying to clear up the fact that the two were actually her adult children. Then, satisfied that they'd cleared up that ambiguity, they didn't notice that it actually left the larger sentence a great deal more muddled.

  2. Brian said,

    October 16, 2012 @ 10:19 pm

    I feel subjectively sure, though I could of course be wrong, that I could never say that that way.

    Really? I feel like this is one of those sorts of sentences that I could easily produce, even though I had same reaction you did upon hearing it. This highlights the difference between what is acceptable English when you're a reader/listener, and what is acceptable when you're a writer/speaker. The latter tends to be a much larger set.

    I think the best way to rewrite it is to turn it around: “The woman whose husband and son were killed in a fire in Northampton has filed suit …”

  3. Aanel said,

    October 16, 2012 @ 10:37 pm

    I think we're dealing with "headline English" in all three of those contexts: radio news report lead-in, photo caption, and news-burb lead-in. All three, like a headline, often demand the most succinct wording possible, even if it strains comprehension. Headlines telescope and abbreviate and strain comprehension all the time.

    I actually think "The widow and mother of men killed in a house fire" is fine, although it would have been better if it had started with her name: "Elaine Yeskie, widow and mother of men killed in a fire" (an appositive, as you say), but journalistic standards require putting particulars separately and secondarily, where they can be spelled out in detail in together one place, for ease of comprehension in a short piece.

    In sum, we're dealing with standard newspaper journalistic methodology here. Not that big of a deal in my opinion. Unless your mind is skewed and prone to imagine Oedipal or incestuous implications where none normally exist.

  4. uebergeek said,

    October 16, 2012 @ 10:38 pm

    Yes, why didn't they just use the construction Brian mentioned: "The woman whose husband and son were killed… ?" Not only is it cleaner syntactically, it also seems more appropriate conceptually. The woman is now the focus of the story, so it makes sense to start with her. I think the other sentences get tangled because they try to use the deceased men as the starting point even though they're discussing the woman's actions.

  5. Paul Clapham said,

    October 16, 2012 @ 10:45 pm

    Consider "the sisters of the two men killed in the fire". In this phrase there are two (or more) women who are variously sisters of one or other of the men. There isn't any implication that the men are brothers — at least, not to me there isn't. Others may feel differently, since the construction is way out in the far fringes of English grammar.

  6. Ran Ari-Gur said,

    October 16, 2012 @ 10:56 pm

    @Brian:

    > This highlights the difference between what is acceptable English when you're a reader/listener, and what is acceptable when you're a writer/speaker. The latter tends to be a much larger set.

    Really? I would think that the exact opposite is true: most people would accept many utterances that they themselves would never produce. No?

  7. RW said,

    October 16, 2012 @ 11:09 pm

    I find it very interesting that this strange construction was used. The wording that Brian suggested is much clearer and to me it seems a more natural and obvious choice for a writer. The original seems like implicitly sexist language, defining the woman, who is the focus of the story, only in terms of her relationships to the men.

  8. Martha said,

    October 16, 2012 @ 11:11 pm

    I've read that sentence several times, and while I understand the situation now that it's been explained, I have trouble having that phrase describe something other than a woman who is the mother of the two men killed, who happens also to be a wife (other than, of course, a woman who is married to her two sons), although I suppose that would have to be "A wife and THE mother of two men killed in a fire …”

  9. Barbara Partee said,

    October 17, 2012 @ 12:11 am

    @B.T. Carolus — but "the wife and mother of two" is something entirely different; she's a wife, and she's a mother of two (children). In this example "of the two men killed …" is the complement of "wife and mother". But the perfect acceptability of your example, which is superficially similar to the one under discussion, may have had something to do with the odd one happening to occur to the writers.
    @Aanel — Although I wouldn't quite call it "fine", I do agree that ""The widow and mother of men killed in a house fire" is better than the original, which had "the two men" instead of "men". The 'bare plural' "men" seems to contribute a "category", or "kind", rather than particular reference. Yes, I'll bet we could find examples like "The mother and daughter of Marines killed in battle was honored at the ceremony." But I'll bet it would be much harder or impossible to find anything like "The mother and daughter of those two marines", meaning just ONE woman.
    @Paul Clapham — your example, "the sisters of the two men killed in the fire" — is perfectly fine and semantically compositional (though working out exactly how the semantics works with plurals is challenging, and there's not a single standard account of this case, I don't think). "Sisters" is plural, so we know we're dealing with more than one person. "The wife and mother of … " can be either one person or two, and the problem example was when it was one person — standard semantics (as for the "friend and neighbor" case) would predict that she should be the wife of the two men and the mother of the two men, but we know that can't be so, so we recompute however we have to. (When grammar and common sense conflict, even a linguist lets common sense win – but then posts on Language Log ;-) )

  10. Barbara Partee said,

    October 17, 2012 @ 12:15 am

    P.S. – about why they made the men primary and described the wife-and-mother in terms of them: this event was very big news in 2009 – it was part of an even bigger arson spree all in one night, in normally quiet Northampton – and everyone knew about the deaths of those two men. The men and the accused arsonist could be considered well-known; the widow/mother hadn't been in the news before.

  11. uebergeek said,

    October 17, 2012 @ 1:07 am

    Probably even more straightforward: "Elaine Yeskie, whose husband and son were killed.." or "Elaine Yeskie, whose husband and son were the two men killed…" (Latter option could address Barbara Partee's point that readers would already be familiar with the two men.)

  12. John Lawler said,

    October 17, 2012 @ 1:29 am

    Like Mark, I'd love to have a theory of sentence production that could explain how this happened, and I think part of it would have to point out the custom in our culture of describing female active human subjects of sentences, wherever possible, by their relation to males, even dead males.

    On another note, however, I did notice in the news recently a description of a man as someone's widower, which I had not seen before. Robin Lakoff remarked on this possessive asymmetry in Language and Women's Place.

  13. boynamedsue said,

    October 17, 2012 @ 1:37 am

    My initial reaction was that the writer had made a mistake conjugating the verb and it should have read:

    'The wife and mother of two men killed in a fire in Northampton HAVE filed suit'

    Meaning that there were two women, with different relationships to the different men. I think that sentence would be acceptable. It took me a while to realise there was actually only one woman who had different relationships to the two men.

    Does everyone agree that in the case of there being two women the sentence loses its double-take value? And if this is the case, why is the original sentence be confusing (which it clearly is)?

  14. boynamedsue said,

    October 17, 2012 @ 1:39 am

    "why is the original sentence be confusing (which it clearly is)?"

    That is not a grammatical mistake, I am a speaker of London Multicultural English (cerca 1998).

  15. boynamedsue said,

    October 17, 2012 @ 1:42 am

    "I think part of it would have to point out the custom in our culture of describing female active human subjects of sentences, wherever possible, by their relation to males, even dead males."

    I suspect it has more to do with news values: "Men die in fire" is more newsworthy than "woman files suit" so the journalist subconsciously wrote a sentence in which the woman is defined be her relationship to her more newsworthy relations.

    By the way, does anyone else feel like a bit of ghoul for picking over the syntax of this woman's tragedy?

  16. Peter Taylor said,

    October 17, 2012 @ 1:51 am

    The natural interpretation is easily rejected on the basis of extralinguistic context because of the extreme rarity of polyandry.

    I'm not sure whether the word "respectively" would suffice to clarify the situation, but I'm sure many dissertations have been written on its use.

  17. Dan M. said,

    October 17, 2012 @ 2:33 am

    Could the problem also be extra-linguistic? What if the headline writer misunderstood their notes or their quick skim of the article, shrugged at what newfangled marriage equality movement has made possible, and written this headline thinking the situation was really as they described it?

  18. pj said,

    October 17, 2012 @ 5:12 am

    @Peter Taylor, 'respectively' would definitely lead me to expect two women, one the wife of one of the men, and the other the mother of the other.

    What this whole problem clearly reveals is that English has a terrible gap in its vocabulary for a term meaning something along the lines of 'very-close-family-woman'. That word could be used in the first sentence (where the known 'two men' apparently need to be central), and then the exact relationships clarified. Something like:

    'The [____] of two men killed in a fire in Northampton has filed a suit against alleged arsonist Anthony P. Baye. Elaine Yeskie, 77, whose husband, [name], and son, [name], died in a house fire in 2009, is seeking monetary and punitive damages…'

    Does any other language have a word that allows this solution?

  19. maidhc said,

    October 17, 2012 @ 5:14 am

    "Kinswoman"?

  20. pj said,

    October 17, 2012 @ 5:58 am

    maidhc,
    Yeah, I thought of that, but besides being archaic-sounding (obsolete?), 1) I'm not sure it really covers 'wife' as well as blood relations; and 2) it suggests (in the singular) a more distant relationship doesn't it? If the men's 'kinswoman' was pursuing an action, I'd be imagining, I dunno, a second cousin or something, that might never even have met them. It doesn't hit the spot for 'very-close-family-(even-if-not-blood)-woman' for me.
    A stepmother, or a daughter-in-law, could be a [____] to you, but she isn't a kinswoman, I don't think.

  21. Ginger Yellow said,

    October 17, 2012 @ 6:34 am

    All the ways I could express it take more words; about the shortest acceptable version I can find is “The wife of one and mother of the other of two men killed in a fire …”

    Why not "A woman whose husband and son were killed in a fire…"?

  22. David Scrimshaw said,

    October 17, 2012 @ 8:13 am

    I think a factor in the writer using this construction is that we have an implicit understanding that a woman cannot be both the mother and wife of a man.

    If she had been the sister of one man and the employee of the other, the writer might not have written "The sister and employee of two men killed…"

  23. MonkeyBoy said,

    October 17, 2012 @ 8:16 am

    While this example could have been written more clearly it falls under the case of a "distributive" interpretation, such as for:

    "John and Bill love their mothers"

    where the normal interpretation doesn't involve John loving Bill's mother.

  24. Q. Pheevr said,

    October 17, 2012 @ 8:22 am

    @John Lawler, with apologies for the thread drift:

    In 2008, I had a similar response to the New York Times' description of Asif Ali Zardari as "[Benazir] Bhutto's widower." Some people pointed out (including Laurence Horn on ADS-L) that it's possible to find several earlier examples (such as Ted Hughes, widely described as Sylvia Plath's widower).

  25. Faldone said,

    October 17, 2012 @ 8:44 am

    I don't think my first interpretation has been addressed here, that the wife and mother (who I took to be one person*) was the one killed in the fire.

    *We can pass over the problem of the one woman being both wife and mother to two individuals. That issue has been sufficiently covered.

  26. Q. Pheevr said,

    October 17, 2012 @ 8:56 am

    And, on the actual topic of the post:

    The intended (non-Jocastan) reading is certainly possible in a context such as "the wife and mother of A and B," where it can be made explicit by the adverb respectively. So the question is, can we distribute out the roles of "wife" and "mother," one per man, even when the two men are merely quantified over and not separately identified?

    I wouldn't be entirely comfortable* saying something like "the wife and mother (respectively) of two men," which ought to force the intended reading if it is indeed semantically available; for me, respectively really wants to have two explicit lists to collate. But this kind of use of respectively is certainly attested; the OED (s.v. respectively) gives the example "Of the three defendants.., two were respectively president and secretary of the..Society." So on the whole, I'm inclined to suspect that for at least some speakers, the intended reading is grammatically possible, and not just an inferential leap from the literal meaning of the sentence motivated by a pragmatic presumption against incest and bigamy.

    *I'm not sure here whether my discomfort is a grammatical one, or just a matter of stylistic preference.

  27. Ellen K. said,

    October 17, 2012 @ 9:54 am

    Ginger Yellow: Why not "A woman whose husband and son were killed in a fire…"?

    Because that does not specify that these are particular men who have previously been reported about in the news.

  28. Darryl Shpak said,

    October 17, 2012 @ 10:00 am

    I understood it right away, and tripped more over whether the men were killed in the fire, or the woman was. But I wasn't particularly confused about that either; that's clear from the context.

    It actually took me a moment to even figure out what it was that confused you. But of course, now that I've seen it, I can't un-see it, and definitely agree that looks awkward. Nevertheless, it didn't cause any problems for me on first reading.

  29. Ellen K. said,

    October 17, 2012 @ 10:17 am

    I'm pretty sure the ambiguity that allows the reading of the woman being the one killed would not be there in the original, since it was spoken. Even before continuing on to "has filed suit" (which the original does, though the LL headline does not), there are differences in how each meaning would be said. I think if we heard the audio it would be clear the two men are the ones who died even before we get to "filed suit".

    I agree — BHP

  30. boris said,

    October 17, 2012 @ 11:40 am

    I wonder if the fact that "wife and mother" is such a common phrase contributed to this construction, at least in speech. (I don't think "widow and mother" is that common, though).

  31. Barbara Partee said,

    October 17, 2012 @ 12:01 pm

    @John – I’m not Mark ;-). It’s a fair default assumption, though. (Not that I’m Mark, but that the author is. ;-) )
    @boynamedsue: I agree that if it referred to two women, no problem. And that if one woman were the wife of both men and the mother of both men, no problem. The problem is just with getting their intended interpretation out of that very phrase. – And yes, I did feel mildly uncomfortable doing a linguistic post about a tragic event, and I sat on it for a couple of days, but the construction was just too interesting, and the arson event was three years ago, and the lawsuit event isn’t a tragedy, so I decided it wasn’t wrong to post about it.

  32. Barbara Partee said,

    October 17, 2012 @ 12:01 pm

    @pj and @maidhc: Interesting. It’s clear that words like “kinswoman” are TOO general for this case, and pragmatic principles explain why if you use a word like that people will think it’s not any relation so close as wife or mother. And since ‘wife’ and ‘mother’ are so important, and at the same time so different, I doubt that any language would have a word that lumps them together without including a lot more relations. So I don’t think it’s a terrible gap in the language, and I don’t think the potential need for it arises very often. (We do have words that cover just two important relations, like spouse and parent, but those are relations that have all their properties in common except for gender.)

  33. Barbara Partee said,

    October 17, 2012 @ 12:02 pm

    @MonkeyBoy and @QPheever: Yes, they’re intending a distributive interpretation – wife of one and mother of the other. But distributive interpretations could also involve two women, e.g. “The wife and mother respectively of Mr. A and Mr. B have/has filed suit”. (If the phrase is the subject, then agreement on the verb may tell us whether it’s one woman or two. I suspect I would do a double-take even here if I hit a singular verb.) I agree with QPheever. I don’t think I can get a distributive interpretation for that phrase when the complement is just “of the two men …” rather than “of Mr. A and Mr. B” or some such conjunction. But it’s interesting that they do occur. Sometimes the context will set up an explicit pair of referents for a plural noun phrase, in a specific order, and that could do it. But there was no context in my example, and there’s no context in QPheever’s example either, so it’s a little surprising. But “respectively” is at least helpful in eliminating the understanding that I’m otherwise forced to, that she (or they) bear(s) both relationships to both men.

  34. Gene Callahan said,

    October 17, 2012 @ 12:14 pm

    @John Lawler: "I think part of it would have to point out the custom in our culture of describing female active human subjects of sentences, wherever possible, by their relation to males, even dead males."

    Since this is ALWAYS possible, I guess we should never in speech or writing find females except so described, hey?

  35. Ethan said,

    October 17, 2012 @ 12:21 pm

    Re: "kinswoman" not implying a strong enough relationship. "Family member" works in one direction: "A woman who lost two family members in a fire…" but sounds a bit odd in the other direction " A family member of two men who were killed…"

  36. Dave K said,

    October 17, 2012 @ 12:24 pm

    I think Martha's right. Phrasing it "The wife and the mother" makes it clear (at least to me) that two different people are involved. I don't think it's actually a rule but it does provide an extra set-off that clarifies things.

  37. Mark F. said,

    October 17, 2012 @ 12:45 pm

    I think Q. Pheever has answered the question of how the meaning of the sentence can be seen as compositional based on rules that might plausibly belong to many idiolects of English. But isn't compositionality a matter of degree? In any sentence where there is ambiguity to be resolved, there is obviously some input beyond the meanings of the parts and how they are joined. And I have the sense that the rules relating syntax to semantics are fuzzy, at least at the edges. Following Pheever's analysis, for me as well, "respectively" does feel like it should have an explicit list to quantize over. But there is also the sense that a list of names is like a definite description of the set. In this case, replacing the list with the description (while dropping the "respectively") may just increase the badness score for the construction, with the exigencies of newspaper writing changing the reporter's threshold for acceptability.

  38. Mark Johnson said,

    October 17, 2012 @ 3:13 pm

    What about "the wife or mother of two men .."?

    There's no chance that Partee and Rooth style type-lifting will save they day?

  39. boynamedsue said,

    October 17, 2012 @ 3:23 pm

    The obvious answer as to how to phrase this sentence in the simplest way possible is:

    A (The?) woman whose son and husband were killed in a fire in Northampton has filed a suit…

    @Barbara, the time interval does seem respectful enough in this case, I'd assumed it was a more recent incident.

  40. boynamedsue said,

    October 17, 2012 @ 3:25 pm

    I could render it into British headline English, but I think that probably is inappropriat.

  41. boynamedsue said,

    October 17, 2012 @ 3:26 pm

    damn my spelling.

  42. Grover Jones said,

    October 17, 2012 @ 5:14 pm

    What our language DOES need is a collective term for nieces and nephews. Can't tell you how many times in conversation I've searched my brain for the term, before realizing that it doesn't exist!

  43. William Steed said,

    October 17, 2012 @ 6:04 pm

    The less comfortable (and contextually humorous) version occurs in Hot Fuzz with dialogue something like this:
    Butterman: That's Lurch. Real name Michael Amstrong. Dad says he's got a child's mind.
    Angel: ok.
    Butterman: He lives up Summer street with his mother and sister.
    Angel: And are they as big as he is?
    Butterman: Who?
    Angel: The mum and the sister?
    Butterman: Same person.

  44. Barbara Partee said,

    October 17, 2012 @ 7:58 pm

    @ Mark F. Oh boy, now things are getting serious. I think all I can say in this space is 'yes and no', those are all great points, and they get into deep issues about "what's in our unconscious grammar(s)" and "how we use it". And about fuzziness on all levels. The only part I'll try to answer explicitly is about how compositionality and ambiguity can co-exist: ideally, whenever there's ambiguity, we can track down two different "inputs" to the semantic interpretation rules — some syntactic ambiguity or some lexical ambiguity or both. But even that is not always enough, and there are interesting debates about how to best theorize about further context-dependent effects. Is there "invisible material" in the syntactic input to the semantics, or maybe some kinds of "pragmatic enrichment", or maybe places where the semantics fills in something like a "variable", leaving it to pragmatic principles to figure out the most likely intended value for that variable. (We have something like that with many uses of simple pronouns like "he", or even "you".) Big questions, all interesting. This is a fun field.

  45. Steve Morrison said,

    October 17, 2012 @ 7:58 pm

    Of course, the woman in question could be described as the "survivor" of the two men killed in the fire.

    As for a gender-neutral term for niece/nephew, I've seen "niebling" used to mean this.

  46. Barbara Partee said,

    October 17, 2012 @ 8:16 pm

    @ Mark Johnson, that's getting in too deep for this forum, and besides, I don't know the answer. I don't think any kind of Partee & Rooth shiftings would get that relation to be one for one of the men and the other for the other. Maybe (I'm really not sure and am not going to stop and compute) our treatment could generate some broader meaning where it meant a woman who stands in at least one of those two relations to each of those two men — I think so — but that would also include the case where she was mother of both, for instance, so it's too weak a reading.

  47. J.W. Brewer said,

    October 17, 2012 @ 9:30 pm

    I am surprised by the surprise at seeing instances of actual use of the perfectly common English word "widower." For what it's worth, the google n-gram viewer shows a substantial decrease in the widow:widower ratio over the course of the 20th century, although the former remains more frequently used than the latter. That's talking about instances of the words in the relevant corpus. But it's worth remembering that there's also an asymmetry in referents – unlike e.g. sister/brother or niece/nephew, there are dramatically more actual widows than actual widowers in existence (about four times as many in the U.S. as of about ten years ago, according to the first semi-reputable-looking source I googled up) as a result of sex differences in average lifespan combined with the fact that older-husband/younger-wife couples are much more common than the reverse (at least in the U.S. and I expect in Anglophone societies more generally). The word-usage ratio in the n-gram is more skewed than 4:1, but you still shouldn't be comparing it to an unrealistic 1:1 baseline.

    I take it the Lawler and/or Lakoff observation/complaint is that women are *gratuitously* referred to in terms of their familial status/title/etc more often than men are, which might well be the case, but is not a hypothesis advanced by commenting on a non-gratuitous example. Here, not only is Mrs. Yeskie's relationship to the two decedents journalistically salient as a human-interest aspect of the story, it's crucial because this is really just a story about the filing of a lawsuit, and this is a statutory wrongful death suit that can only be filed by someone in a statutorily-specified familial relationship with the decedent. (There's usually a hierarchy of who the proper plaintiff is – unless the law in Massachusetts is weird, she's probably suing for the deceased son only because he was unmarried, otherwise the right to do so would have been held by her daughter-in-law.) I have heard lawyers with more experience in this area than I have use the phrase "statutory survivor(s)" as jargony shorthand for "those relatives by blood or marriage who have a right to sue under the relevant jurisdiction's wrongful death statute," but that's probably too insiderish for journalism. I suppose "next of kin" is simultaneously folksy and lawyerly and might compress both of the relationships at issue here into a single NP, although idiomatically "next of kin" is perhaps more commonly used when the individual occupying that role is more remotely related to the decedent than a spouse/parent/child.

  48. Dan M. said,

    October 18, 2012 @ 1:20 am

    J.W.Brewer has nailed it, I think.

RSS feed for comments on this post