Linguistic commentary from a guy who takes things too literally

Those Three Little Words, in 16 Languages

Posted by Neal on May 12, 2007

Faithful readers of this blog (and sporadic ones who happened to read just the right posts) know that I’ve been interested in questions that contain more than one wh-word. In the course of doing some research on them, I’ve been having a set of 24 questions translated into as many languages as I can find willing native speakers for. As a result, I can now say “Who read what?” in 16 languages. In fact, I think I will:

  1. Chinese (Mandarin): shei du le shenme?
  2. Czech: kdo co cetl?
  3. English: who read what?
  4. Estonian: kes luges mida?
  5. German: wer hat was gelesen?
  6. Greek: pjos djavase ti
  7. Hebrew: mi kara ma
  8. Hindi: kisne kyaa paRhaa?
  9. Hungarian: ki mit olvasott?
  10. Japanese: dare ga nani o yomimasita ka?
  11. Korean: nwuka mwuesul īlkessnī?
  12. Macedonian: koj čto čital?
  13. Russian: kto čto čital?
  14. Spanish: quién leyó qué?
  15. Tagalog: sino ang nagbása ng anó
  16. Vietnamese: ai (đã) đọc gì

I bring this up today because I just got back from a one-day conference in Athens (Ohio, that is) where I presented some of this research. It concerned a strange property of multiple-wh questions that many linguists have remarked on over the years. Consider this multiple-wh question:

Who saw Elvis where?

To answer this question, it’s not good enough just to say something like:

Glen saw him at the gas station.

It has to be a list (of length greater than 1) of seer/location pairs, e.g.

Glen saw him at the gas station, Doug saw him in the school lunchroom, and Adam saw him on the swings at the playground.

If you don’t believe this is true, look what happens when you replace saw with a verb that dictates that your answer will be a single actor/location pair:

#Who killed Elvis where?

The hashmark indicates that this sentence sounds inappropriate; it seems to imply that Elvis was killed by multiple people in multiple locations. At least, that’s how it sounds to me, and how it has sounded to other researchers. (If you disagree, read the final paragraph of this posting.) For this question to get a single-pair answer, it would have to be rephrased as something like

Who killed Elvis, and when?

That’s about the only alternative we have in English for this particular question, though in some languages, you can coordinate the who and the where to get something like this (which is ungrammatical in English, as indicated by the asterisk):

*Who and where killed Elvis?

You can see a similar effect in this next pair, where English allows both a multiple-wh version and a coordinated-wh version:

#When were you born where?
When and where were you born?

For me, When were you born where? is grammatical, but is appropriate only in a conversation among believers in reincarnation.

Does this pattern of pair-list answers for multiple-wh questions and single-pair answers for coordinated-wh ones hold up for other languages? This question piqued my interest after I read papers by Anikó Lipták and Konstantin Kazenin noting the effect for Hungarian and Russian, respectively. So for the past few years, I’ve been trying to find out by having people translate those 24 questions and tell me what kind of answers they can or must have.

The answer so far? No; the pattern does not hold up for other languages. (A little but for Hebrew, but nothing else so far.)

But you know, there are a lot of languages out there, and I’ve only looked at 16. Furthermore, the data I do have could be enough to allow some interesting generalizations to be made that don’t concern SP and PL answers. Therefore, I am now making this data available to the public, to look at with other perspectives than my own, and to add data from other languages if they are so inspired. I hope the Coordinated-Wh Project will become a useful resource for linguists interested in syntactic, semantic, and typological issues concerning multiple-wh and coordinated-wh questions. Right now the page has all the background information, and the data from a few of the languages I’ve investigated. Check back during the coming weeks, as I will gradually upload the data from the other languages. If you speak a language that isn’t represented there yet, you and I can correct that situation. And if you speak a language that is already represented, it’s still good to get judgments from additional speakers. Just send me an email, and we can decide when and how to do it, or when to do it and how, or maybe even how to do it when.

8 Responses to “Those Three Little Words, in 16 Languages”

  1. flounder said

    Finnish would be “kuka luki mitä?”. As for the Japanese version, I would rather keep the particles separate from the question words when romanizing, like this: “dare ga nani o yomimasita ka?” (誰が何を読みましたか?)

  2. Fire Next Time said

    I saw an anti-smoking advert in Spanish that piqued my curiosity as a non-native Spanish speaker. It said, “Quien se fuma a quien?”. Who is smoking whom? Which is a good question in my opinion, if you view a cigarette as having that ability. But, I’d like to know if it should have been: Who is being smoked by what? My Spanish isn’t good enough to translate that correctly.

  3. Glen said

    Those inappropriate-sounding questions don’t sound inappropriate to me, at least in the right context. For instance, “Who killed Elvis where?” sounds just fine to me, especially if the speaker is asking someone to repeat information already given. Like so:

    “Colonel Tom Parker killed Elvis in the bathroom.”

    “Wait… Who killed Elvis where?”

  4. Neal said

    Flounder: Japanese romanization has been corrected as suggested. As for the Finnish, thanks! How’d you like to do another 23 for me? Warning: Some of the original English isn’t grammatical, and some of the Finnish translations I’m looking for may not be grammatical.

    FNT: From my limited knowledge of Spanish, I know that the a does not mean “by,” but is the accusative marker for animate direct objects. But as for the se fuma part, I don’t know. I’d’ve expected just plain fuma. Maybe a Spanish-speaking reader will elucidate.

    Glen: Darn it, I always forget to mention an important point that you’ve brought up. The context you describe is what we call an echo question. It is not so much asking for an answer to a question as for a repeat of what was just said, because you didn’t hear it, or don’t think you heard it right, or don’t believe it. Notice that if you’d said, “Who killed Elvis [mumble]?” someone could ask “Who killed Elvis WHERE?” and this time, the response they’d expect would not be a declarative “Parker killed him in the bathroom,” but a repeat of the original question (but hopefully enunciated better).

    While I’m at it, I should also mention what are known as Ref questions, asked when you don’t know who or what a pronoun refers to. So if I said, “Who killed Elvis there?” and you don’t know what the heck I’m talking about when I say there, you might ask, “Who killed Elvis where?” — again, not asking for a declarative answer, but for a repeat of the question, with there replaced with an explicit location: “Who killed Elvis in the bathroom?”

    So, taking these questions not as echo or Ref questions, but just as ordinary questions that happen to be asking about multiple things, I believe you’ll find my claim holds up.

  5. Ingeborg S. Nordén said

    Swedish: Vem läste vad? for which I would also expect multiple answers.

  6. Petroglyph said

    Dutch: “Wie las wat?” Or less literary and more informal: “Wie heeft wat gelezen? / Wie heeft er wat gelezen?” (lit.: “Who has read what?”). I’d expect a list of paired answers to this question, too. If this were an ordinary question in a conversation, it’d be about a group of people who read (friends, students, …). For a single answer pair “Wie heeft (er) gelezen, en wat?” (“Who read, and what?”) would be appropriate.

    “Waar en wanneer ben jij geboren?” (“When and were were you born?”) would be a formal but fully natural way of getting some basic information on someone — a single-pair answer. Even “Waar en wanneer ben jij geslagen?” (“When and where were you hit?” – something that can occur more often than being born) definitely invites a single-pair answer. Replying with a complete index of fights and where they took place would be an invalid answer. Unexpected, funny.

    “?Wanneer ben jij waar geboren?” (When were you born where?”) sounds nonsensical, and “*Wanneer ben jij geboren waar?” is ungrammatical (the participle “geboren” requires final position here). The first one sounds barely acceptible as a ref guestion, but I suspect this kind of construction is English influence on Dutch anyway. Perhaps that’s why it sounds off.

    As to killing Elvis… “Wie heeft Elvis waar vermoord?” (“Who killed Elvis where?”, again, participle comes last) sounds a little off to me, but perhaps that’s because it doesn’t make sense at all (like questions about the king of France). Let’s try “Wie heeft Kennedy waar vermoord?”. That sounds a little strange, too. Perhaps your theory works for Dutch as well, and the question is unacceptable because it has a single-pair answer.

    In fact, the only contexts I can think of using “Wie heeft Kennedy waar vermoord?” are quite specific. For instance, it’s perfectly acceptable when you assume the Hearer(s) already know the answer. This question would then elicit a single-pair answer because it seems like two single-answer questions clumsily conflated into one. Something you’d quickly ask a group of students to refresh their memory before moving on. “Wie heeft Kennedy vermoord, en waar?” (“Who killed Kennedy, and where?”) is much the same: two brief reminders of familiar material, only less clumsily formulated.

    What if you assume the Hearers don’t know the answer? Asking “Wie heeft Kennedy waar vermoord?”, and stressing the question words, would imply “There’s so many theories; who is to say?”. The question would definitely elicit a list of paired answers. The focus, however, would not be on the murderer/location pairs, but you’d be making a more general point about conspiracy theories.

    “Wie heeft Kennedy waar vermoord?” apparently doesn’t work in Dutch (possibly because it does _not_ have a list of paired answers) except in highly specific contexts.

  7. flounder said

    Neal: Oh, I didn’t notice there were more of them. 🙂 I can’t guarantee anything, but if you drop me an e-mail, I’ll take a look at them when I have time. (I hope you can see the email address I entered in this comment form.)

  8. Deolici said

    FNT: Your almost literal translation of the anti-smoking advert is right: “¿Quién se fuma a quién?” = ‘Who is smoking whom?’, although your wanting for a ‘better’ translation is perhaps not that pertinent. I explain: if you say in Spanish “¿Quién se fuma qué?” the advert makes sense no more, as the idea was actually to view the cigarette as having the ability to smoke -even more, the ability to ‘eat up’ a human being. ‘Who is being smoked by what?’ -“¿Quién es fumado por qué?”- sounds rather bad in Spanish and does not have the impact of the original, besides of becomimg ambiguous: “¿…por qué?” also means ‘why?’ so you get ‘Who is somked why?’. Nevertheless, leaving metaphores and advertising strange use of language to one side, you and Neal make remarkably accurate observations: no cigarette on earth can smoke a human being, and the verb ‘se fuma’ must be only ‘fuma’.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: